2019 AnalyzerPro

Technical Manual
After the work of Dr. W. Gratzer and Dr. H. Burg

Revised by M. Schmidt

ATTENTION: The technical manual is currently

under review. Slight changes may occur.

Matthias SCHMIDT
ANALYZER PRO
INTRODUCTION

Dear accident experts!

The following handbook is based on the great scientific preparatory work done by
Dr. Werner Gratzer and Dr. Heinz Burg in the field of accident analysis. The summary,
which by no means claims to be complete, is intended to serve as a quick reference
book for the most important formulas.

The manual is made available free of charge to all accident experts. As a token of
your appreciation, I would be delighted to welcome you personally to one of our
seminars.

I wish you exciting work and lots of fun with the manual!

Matthias Schmidt
CONTENT

Introduction ..................................................................................... 1
Kinematics ...................................................................................... 4
Introduction .................................................................................. 4
Basics.......................................................................................... 4
Sizes and units ............................................................................ 4
Uniform movement ....................................................................... 4
Uniformly accelerated movement...................................................... 4
Uniform change of acceleration ........................................................ 5
Distance-time data .......................................................................... 5
Distance-time calculation (Module) ...................................................... 6
Turning-in crash and rear-end collision (Modules) ..................................... 7
Forward calculation ...................................................................... 7
Backward calculation .................................................................... 9
Avoidability consideration .............................................................. 10
Avoidability (Module) ...................................................................... 11
Spatial avoidability ...................................................................... 11
Temporal avoidability ................................................................... 11
Drive off-brake (Module) .................................................................. 12
Pedestrian – Barrier method (Module) ................................................... 14
Kinematics and dynamics of a pedestrian accident ................................ 16
Casting distance of pedestrians ....................................................... 17
Sight obstruction by oncoming traffic (Module) ....................................... 22
Overtaking (Module) ....................................................................... 24
Change of viewing angle (Module) ....................................................... 24
Attention (concentrative - distributive) .............................................. 26
Visual system ............................................................................. 26
Visual information recording ........................................................... 28
Lane change procedure (Module) ........................................................ 33
Casting distance (Module) ................................................................. 35
Kinetics ......................................................................................... 36
Introduction ................................................................................. 36
Basics......................................................................................... 36
Impact hypothesis ....................................................................... 36
 The impact coefficient k ............................................................... 37
The friction coefficient µ ............................................................... 39
Collision analysis ........................................................................... 39
Backward analysis ....................................................................... 39
forward analysis ......................................................................... 42
Wall impact ............................................................................... 45
Coasting analysis ........................................................................... 46
Tracking analysis (Coasting analyis backwards) ..................................... 46
Calculation of driving dynamics (Coasting analysis forward) ...................... 50
Calculation of forces between tyres and road surface ............................... 59
Introduction .............................................................................. 59
Metrological recording of tyre properties ............................................ 60
Mathematical replacement models for tyres ........................................ 60
Modelling .................................................................................. 61
Tyre model .................................................................................. 62
Theoretical principles ................................................................... 62
Replication of tyre characteristics .................................................... 63
Approximation in the sliding area ..................................................... 64
Equivalence approach for slip and slip angles ....................................... 65
Wheel load dependence ................................................................ 67
Horizontal dynamic module ............................................................ 69
„IPG-Tire“ Tyre Approximation ........................................................ 73
Further Literature ....................................................................... 77
Dimension Data .......................................................................... 78
Serial collision (Module) ................................................................... 80
Change of vehicle velocity ............................................................. 80
Calculation of passenger acceleration due to collisions ........................... 85
Relationship between EES, permanent deformation and structural stiffness,
calculation of collision duration ....................................................... 90
KINEMATICS

INTRODUCTION

Kinematics is a subfield of mechanics that describes the movement of bodies purely

geometrically. All kinematic processes can be described by the quantities time,
place, velocity and acceleration. Forces, masses, impulses and energies are not
considered in kinematics.

BASICS

SIZES AND UNITS

a Acceleration or deceleration [m/s²]
v Velocity [m/s] or [km/h]
t Time [s]
s Distance [m]

Index A, f.e. vA Initial, e.g. initial speed

Index E, f.e. vE Final, e.g. final velocity
Index S, f.e. tS Swelling, e.g. swelling time
Index R, f.e. tR Reaction, e.g. reaction time
Index B, f.e. tB Braking, e.g. braking time

In general, the following applies:

 a > 0: Braking
 a < 0: Acceleration

UNIFORM MOVEMENT
„Drive with constant velocity“

𝑎 = 0

𝑣 = 𝑐𝑜𝑛𝑠𝑡. = 𝑣𝐴 = 𝑣𝐸

𝑠 = 𝑣𝐴 𝑡 + 𝑐𝑜𝑛𝑠𝑡.

UNIFORMLY ACCELERATED MOVEMENT

„Braking or accelerating“

𝑎 = 𝑐𝑜𝑛𝑠𝑡.

𝑣 = 𝑎 𝑡 + 𝑣𝐴

𝑠 = ½ 𝑎 𝑡² + 𝑣𝐴 𝑡 + 𝑐𝑜𝑛𝑠𝑡.
UNIFORM CHANGE OF ACCELERATION
"Swelling phase" - This type of movement is characterized by a constant change in
acceleration over time and occurs, for example, when a brake is applied.

Mathematically, this means that the acceleration increases linearly from a = a1 to

the maximum value a = a2.
𝑎2 − 𝑎1
𝑡 + 𝑎1 𝑓𝑜𝑟 0 < 𝑡 < 𝑡𝑠
𝑎 = { 𝑡𝑆
𝑎2 𝑓𝑜𝑟 𝑡 > 𝑡𝑆

For 0 < t < tS the following applies:

𝑎2 − 𝑎1 2
𝑣 = ∫ 𝑎 𝑑𝑡 = 𝑡 + 𝑎1 𝑡 + 𝑣𝐴
2 𝑡𝑆

1 𝑎2 − 𝑎1 3 1
𝑠 = ∫ 𝑣 𝑑𝑡 = 𝑡 + 𝑎1 𝑡 2 + 𝑣𝐴 𝑡 + 𝑠𝐴
6 𝑡𝑆 2

At the end of the swelling phase, the following speed results:

1
𝑣(𝑡𝑆 ) = 𝑣1 = 𝑣𝐴 + (𝑎2 + 𝑎1 ) 𝑡𝑆
2
For the swelling distance, the following results:

1
𝑠(𝑡𝑆 ) = 𝑠𝑆 = 𝑣𝐴 𝑡𝑆 + (𝑎2 + 2 𝑎1 ) 𝑡𝑆2
6
The initial velocity vA is obtained by solving the quadratic equation:

(2𝑎1 + 4𝑎2 )𝑣𝐴2 + 𝑣1 (2𝑎1 − 2𝑎2 )𝑣𝐴 − 𝑣1 ²(4𝑎1 + 2𝑎2 ) + 3(𝑎1 + 𝑎2 )2 𝑠𝑆 = 0

DISTANCE-TIME DATA

In each individual movement section, a certain form of movement is assumed -

namely either a uniform or a uniformly accelerated movement, or a uniform change
in acceleration.

Of 5 possible values, 3 must be given. The phases are seamlessly connected to each
other, therefore the initial and final velocity of two successive phases must always
be identical.
Example: Phase 1: Uniformly accelerated movement - Phase 2: Uniform change of
acceleration - Phase 3: Uniform movement.

DISTANCE-TIME CALCULATION (MODULE)

In this module, the phases Brake, Buildup and Reaction are calculated as a uniform
process. No acceleration or deceleration is assumed during the reaction phase. The
column requires 5 input values for calculation.
1
Final velocity: 𝑣𝐸 = 𝑣𝐴 − 2 𝑎 𝑡𝑆 − 𝑎 𝑡𝐵

1
Velocity at end of buildup phase: 𝑣1 = 𝑣𝐴 − 2 𝑎 𝑡𝑆

1 1
Total distance: 𝑠𝑔𝑒𝑠 = 𝑣𝐴 𝑡𝑅 + 𝑣𝐴 𝑡𝑆 − 6 ∗ 𝑎 ∗ 𝑡𝑆2 + 𝑣1 ∗ 𝑡𝐵 − 2 ∗ 𝑎 ∗ 𝑡𝐵 ²

Reaction distance: 𝑠𝑅 = 𝑣𝐴 𝑡𝑅
1
Buildup distance: 𝑠𝑆 = 𝑣𝐴 𝑡𝑆 − 6 𝑎 𝑡𝑆 ²

1 1
Braking distance: 𝑠𝐵 = 𝑣𝐴 𝑡𝐵 − 2 𝑎 𝑡𝑆 𝑡𝐵 − 2 𝑎 𝑡𝐵 ²

2
𝑣12 −𝑣𝐸
Braking time: 𝑡𝐵 = 𝑎
Total time: 𝑡𝑔𝑒𝑠 = 𝑡𝑅 + 𝑡𝑆 + 𝑡𝐵

Depending on the problem, the equations are

linked together and solved accordingly. In this
way, 41 different calculation specifications are
possible.

The following values can be roughly assumed as

guide values for the threshold duration:

Car tS = 0,2 – 0,4 s

Motorbike tS = 0,3 – 0,5 s
Truck tS = 0,3 – 0,4 s

TURNING-IN CRASH AND REAR-END COLLISION (MODULES)

These modules deal with the situation of a preceding vehicle which is caught by a
descendant. Either the front vehicle accelerates and turns in (turning-in crash) or it
brakes and drives into a column (rear-end collision).

The consequences of a certain initial situation (forward calculation) or vice versa

(backward calculation) can be calculated.

FORWARD CALCULATION
Both initial velocities, final distance and arc correction must be given. Either the
difference in velocity, the initial distance or the reaction time of the rear accident
participant are calculated.

CALCULATION OF DIFFERENCE IN VELOCITY

sge1 Total distance of the turn-in or front vehicle (vehicle centre)
sge2 Total distance of the vehicle driving into the back of the other vehicle
dist Final distance
so Initial distance
korr Arc correction, which takes into account any arc travel, vehicle length
and collision angle

The total time results from the equation: 𝑠𝑔𝑒2 + 𝑑𝑖𝑠𝑡 = 𝑠𝑔𝑒1 + 𝑠𝑜 − 𝑘𝑜𝑟𝑟

Korr = Distance of the centre of the vehicle of the bender minus difference of the
projections of the positions:

𝑘𝑜𝑟𝑟 = 𝑅𝑀 (𝛼 − 𝛽) + 𝑅𝑉 sin(𝛽 − 𝜇) + 𝑅𝐻 sin(𝛽 + 𝛾)

 𝐿−Ü
𝐵 2 𝐿 2
𝜇 = arctan( )
𝑅𝑀 = √(𝑅 + ) + ( − Ü) 𝑅+𝐵
2 2
Ü
𝛾 = arctan( )
𝑅𝑉 = √(𝑅 + 𝐵)2 + (𝐿 − Ü)2 𝑅+𝐵

𝑅𝐻 = √(𝑅 + 𝐵)2 + Ü2

L … Vehicle length R … Inner curve radius

B … Vehicle width RM … Radius of the centre of the vehicle
Ü … Rear vehicle overhang RV … Radius of the front corner
α … Curve angle (rotation angle) RH … Radius of the rear corner
β … Collision angle

The following applies to the collision angle (without upper velocity limit)):
𝑠𝑔𝑒1 1 2
𝛽=𝛼− = 𝛼 − (𝑣𝐴1 𝑡𝑔𝑒 + 2 𝑎1 𝑡𝑔𝑒 )
𝑅𝑀

The resulting equation for tge cannot be solved conclusively if korr is not assumed to
be constant. Therefore, the total time tge is first calculated with ß = 0 and the
resulting value of korr, then a new value of ß and korr is calculated from t ge and with
this value tge is calculated again, so that the correct value is reached iteratively.

tge (total time) and the final velocities are calculated. A possible prematurely
reached speed limit of the front vehicle and the possibility that tge < tR +tS can be
taken into account. In this case, the formulas are modified accordingly.

CALCULATION OF DEPTH DISTANCE

There are always 2 solutions for a certain initial situation, one with the collision
during the reaction or buildup phase and one with the collision during the braking
phase or after reaching the velocity limit of the front vehicle. For each value of the
reaction time there is a maximum value of the differential velocity.

If the given value of the differential velocity is greater than the maximum value, the
reaction time is calculated which was at least required for the given initial situation
to produce the given differential velocity. The largest possible value of the depth
distance is calculated (at the beginning of the braking phase if the deceleration
values are different, at the end of the braking phase if the deceleration values are
equal).

If the deceleration values are equal, it must be taken into account that if the collision
occurs during the braking phase of the front vehicle, the differential velocity is
independent of the depth distance and depends only on the reaction time.
For a1 + a2 ≠ 0 is valid:

1
𝑣𝐴2 − 2 𝑎2 𝑡𝑆2 − 𝑑𝑢 − 𝑣𝐴1 + 𝑎2 (𝑡𝑅2 + 𝑡𝑆2 )
𝑡𝑔𝑒 =
𝑎1 + 𝑎2

du … Differential velocity

If the requested vehicle comes to a standstill before the collision or reaches the final
speed before the collision, the following applies:

1
𝑣𝐴2 − 2 𝑎2 𝑡𝑆2 − 𝑣𝑜𝑏1 − 𝑑𝑢 + 𝑎2 (𝑡𝑅2 + 𝑡𝑆2 )
𝑡𝑔𝑒 =
𝑎2

This formula is also used if a1 + a2 = 0 applies under the condition:

1
𝑑𝑢 < 𝑣𝐴2 − 𝑎 𝑡 + 𝑣𝐴1 − 𝑎1 𝑡𝑠2
2 2 𝑠2
If the differential speed is too high, the reaction time can be calculated.:

1
𝑑𝑢 − 𝑣𝐴2 + 2 𝑎2 𝑡𝑆2 + 𝑣𝐴1 − 𝑎1 𝑡𝑠1
𝑡𝑅 =
𝑎1

The final speed of vehicle 2 is calculated from the final speed of vehicle 1 and the
differential velocity. This allows the braking time of vehicle 2 to be calculated and
the total time to be calculated.

BACKWARD CALCULATION
The path of vehicle 1 must be given. The following values can be calculated:

1. the initial speed, if the reaction time of the rear vehicle and the differential
velocity are given at the time of collision.
2. the reaction time, if the initial speed of the rear vehicle and the differential
velocity are given.
3. the differential velocity, if the initial speed of the rear vehicle and the
reaction time are given.
4. the reaction time and brake deceleration.

First, tge is calculated from the data of the front vehicle:

- From the collision angle, if this was entered > 0: 𝑠𝑔𝑒1 = 𝑅𝑀 (𝛼 − 𝛽)

2 2
(𝑣𝐸 −𝑣𝐴 )
- Or from vE: 𝑠𝑔𝑒1 = 2 𝑎1

𝑣𝐸 −𝑣𝐴
And further: 𝑡𝑔𝑒 = + 𝑡𝑘𝑜𝑛𝑠𝑡
𝑎1
tkonst is the time with which the vehicle continues to travel after the speed limit – if
existent - has been reached, vE is then equal to the speed limit vgr.
2
(𝑣𝑔𝑟 − 𝑣𝐴2 )
𝑡𝑘𝑜𝑛𝑠𝑡 = 𝑡𝑔𝑒 −
2 𝑎1

1. Calculation of the initial speed of the approaching vehicle:

𝑣12 = 𝑣𝐸2 + 𝑎2 𝑡𝐵2

𝑡𝐵2 = 𝑡𝑔𝑒 − 𝑡𝑅2 − 𝑡𝑆2

𝑣𝐴 = 𝑣12 + 𝑎2 𝑡𝑆2

2. Calculation of the reaction time of the approaching vehicle:

𝑎2
𝑣12 = 𝑣𝐴2 − 𝑡
2 𝑆2
𝑣𝐸2 = 𝑣𝐸1 + 𝑑𝑢

3. Calculation of the differential velocity of the approaching vehicle:

It is calculated analogously v12 and tB2 and further vE2. The differential velocity
then results from the difference to vE1.

4. Calculation of the reaction time and deceleration:

The total distance of vehicle 2 can be calculated from the distance of vehicle
1 and the initial distance. Thus the total distance, the total time, the initial
velocity and the final velocity are given for vehicle 2. From this the other
values can be calculated.

AVOIDABILITY CONSIDERATION
The different possibilities to avoid collisions are calculated. The avoidability
consideration can be done with or without safety distance at the end (if the reaction
time is given).

The collision is avoided if the final velocities are equal in due time. This can be
achieved by:

Increase of acceleration a1, deceleration a2 or initial distance so.

Or

Reduction of the brake initial speed vA2.

The size that is to be calculated is changed until

𝑣𝐸1 cos(𝐶𝑜𝑙𝑙𝑖𝑠𝑖𝑜𝑛 𝑎𝑛𝑔𝑙𝑒) = 𝑣𝐸2

is reached, or until the equation can no longer be solved.

This is again done by iterative adaptation to the new arc correction that occurs
during the avoidability calculation.

AVOIDABILITY (MODULE)

With regard to avoidability, there are 3 different questions:

 How high should the initial speed have been...

 What should the brake deceleration have been…
 Where would have to be reacted (with the original deceleration and initial
velocity)…

… so that the collision could have been avoided.

The avoidability can be calculated spatially and temporally.

SPATIAL AVOIDABILITY
The opposing vehicle (2) must cover the missing braking distance. The collision-free
end point is shifted by this value.
2
𝑣𝐾2
Missing braking distance: 𝑠𝑓𝑒ℎ𝑙 =
2 𝑎2

Avoiding initial velocity:

2
(𝑎1 𝑡𝑆1 )2
𝑣𝐴1 + 𝑎1 (2 𝑡𝑅1 + 𝑡𝑆1 )𝑣𝐴1 − − 2 𝑎1 (𝑠𝑔𝑒 − 𝑠𝑓𝑒ℎ𝑙 ) = 0
12
Avoiding deceleration:
2
𝑎12 𝑡𝑆1 2
+ (2 (𝑠𝑔𝑒 − 𝑠𝑓𝑒ℎ𝑙 ) − 𝑣𝐴1 (2 𝑡𝑅1 + 𝑡𝑆1 )) 𝑎1 − 𝑣𝐴1 =0
12
Avoiding reaction point: 𝑠𝑔𝑒 => 𝑠𝑔𝑒 + 𝑠𝑓𝑒ℎ𝑙

TEMPORAL AVOIDABILITY
The clearance time of the opposing vehicle (2) must be available for this purpose.

Clearance time: 𝑡𝑔𝑒𝑧 = 𝑡𝑔𝑒 + 𝑡𝑟ä𝑢𝑚

Avoiding initial velocity:
2
𝑎 𝑡𝑆1
(𝑠𝑔𝑒 + 21 (𝑡𝐵1
2
+ 𝑡𝐵1 𝑡𝑆1 + 3 ))
𝑣𝐴1 =
𝑡𝑔𝑒𝑧

Avoiding deceleration:
𝑣𝐴1 𝑡𝑔𝑒 − 𝑠𝑔𝑒
𝑎1 = 6 2
𝑡𝑠1 + 3 𝑡𝐵1 (𝑡𝑠1 + 𝑡𝐵1 )

Avoiding reaction point:

𝑡𝐵1 = 𝑡𝑔𝑒𝑧 − 𝑡𝑅1 − 𝑡𝑆1

𝑣𝐸1 = 𝑣𝐴1 − 𝑎1 𝑡𝐵1

𝑠𝑔𝑒 = 𝑣𝐴1 𝑡𝑅1 + 𝑠𝐵1 + 𝑠𝑆1

DRIVE OFF-BRAKE (MODULE)

A vehicle is accelerated from an initial speed v A at acceleration a1 and decelerated
with a2 to its final speed vE after a period tU of constant speed.

For the time t, until the maximum speed is reached, the following applies:

𝑎1 (𝑎1 + 𝑎2 )𝑡 2 + 2 (𝑎2 𝑣𝐴 + 𝑎1 𝑎2 (𝑡𝑈 + 𝑡𝑆 ) + 𝑎1 (2 𝑣𝑎 − 𝑎2 𝑡𝑆 ))𝑡

𝑡𝑆2 𝑡𝑆 2
+ 2 𝑎2 𝑣𝐴 (𝑡𝑈 + 𝑡𝑆 ) − 𝑎22 + (𝑣𝐴 − 𝑎2 ) − 𝑣𝐸2 − 2 𝑎2 𝑠𝑔𝑒 = 0
3 2
Maximum speed reached: 𝑣𝑚𝑎𝑥 = 𝑣𝐴 + 𝑎1 𝑡
𝑣1 −𝑣𝐸
Braking time: 𝑡𝐵 = 𝑎2

Total time: 𝑡𝑔𝑒 = 𝑡 + 𝑡𝑈 + 𝑡𝑆 + 𝑡𝐵

Reaction time: 𝑡𝑅𝑒𝑎𝑘𝑡 = 𝑡 + 𝑡𝑈 − 𝑡𝑅 (after drive-off)

If tU > tR applies, then the reaction location is calculated as follows:

𝑎1 2
𝑠𝑅𝑒𝑎𝑘𝑡 = 𝑣𝐴 𝑡1 + 𝑡 + 𝑣𝑚𝑎𝑥 (𝑡𝑢 − 𝑡𝑅 )
2 1
𝑎1 2
Otherwise as follows: 𝑠𝑅𝑒𝑎𝑘𝑡 = 𝑣𝐴 𝑡𝑅𝑒𝑎𝑘𝑡 + 𝑡𝑅𝑒𝑎𝑘𝑡
2
In the same module it is possible to calculate the last possible (necessary) reaction
point at which the vehicle comes to a standstill on the specified total distance.
Alternatively, a possible reaction point can be entered, whereby the required
distance to a standstill is calculated.

sAR … Distance until the beginning of the reaction

tAR … Time until the beginning of the reaction

Relationship between sAR and tAR:

𝑣𝐴 𝑣𝐴 2 𝑠𝐴𝑅
𝑡𝐴𝑅 √
= − + ( ) + 2
𝑎1 𝑎1 𝑎1

2
𝑡𝐴𝑅
𝑠𝐴𝑅 = 𝑣𝐴 𝑡𝐴𝑅 + 𝑎1
2
For the velocity at the reaction time, the following applies:

𝑣𝑅𝑒𝑎𝑘𝑡 = √𝑣𝐴2 + 2 𝑎1 𝑠𝐴𝑅

For the maximum velocity reached, the following applies:

𝑣𝑚𝑎𝑥 = 𝑣𝑅𝑒𝑎𝑘𝑡 + 𝑎1 𝑡𝑅

𝑎1 𝑡𝑅2
𝑠𝑅 = 𝑣𝑅𝑒𝑎𝑘𝑡 𝑡𝑅 +
2
Using the general formulas follows:

𝑠𝑔𝑒 = 𝑠𝐴𝑅 + 𝑠𝑅 + 𝑠𝑆 + 𝑠𝑅

𝑡𝑔𝑒 = 𝑡𝐴𝑅 + 𝑡𝑅 + 𝑡𝑆 + 𝑡𝐵

PEDESTRIAN – BARRIER METHOD (MODULE)

Accidents involving pedestrians can be divided into several groups on the basis of the
different types of contact between pedestrian and motor vehicle. The three main
groups are: hitting at the front, running over and brushing laterally. These groups,
determined by the contact geometry, can be further subdivided by geometric and
kinematic quantities.

The correct determination of the type is essential, as different reconstruction

methods have to be used for each subgroup. The transitions between most groups
are fluent, which makes clear assignment even more difficult.

1. Hitting at the front (head-on collision):

The pedestrian is located with his whole body in front of and within the outline of
the vehicle. The pedestrian's body is accelerated to the speed of the vehicle.

1a. Decelerated collision:

The pedestrian disengages from the vehicle and, after a flight phase, hits the road,
where he comes to rest after a certain phase of slipping or rolling. The pedestrian's
end position is in front of the vehicle's end position. Very good test results are
available for this case and therefore precise statements can be made about the
collision speed and the collision point.

1b. Unbraked collision:

Different impulse geometries result in different end positions here:

 The pedestrian remains on the vehicle until braking is initiated. The

pedestrian then disengages and falls onto the road. Depending on the position
of the pedestrian on the vehicle and the intensity of the braking, the
pedestrian can remain on the vehicle until its standstill.
 The pedestrian falls down laterally after touching and comes to lie behind the
final position of the vehicle.
  If the pedestrian is located behind the vehicle after the accident, the accident
tracks indicate that the pedestrian has been thrown over the roof of the
vehicle. The throw distance is then almost identical to that of the braked
impact.

2. Hitting at the front (partial head-on collision):

In contrast to a full frontal collision, the pedestrian is not completely within the
outline of the vehicle at the moment of impact. This impact geometry does not
necessarily cause the pedestrian to slide to the side of the vehicle. The partial frontal
collision is divided into two categories, one going in and one going out, depending
on the walking direction of the pedestrian.

2a. Going in:

The pedestrian only comes into contact with the edge areas of the vehicle front.
Most often only with the leg with which he has just taken a step. After the primary
contact, the pedestrian turns around his longitudinal axis. The energy transferred
during the impact is almost completely converted into rotational energy of the
pedestrian, who then glides along the side of the vehicle. This results in damages at
the side of the vehicle and further injuries to the pedestrian..

2b. Going out:

The procedure is similar to that of a pedestrian walking in. The pedestrian only comes
into contact with the edges of the vehicle front. Most often only with the leg with
which the pedestrian is about to take the next step. After the primary contact, the
pedestrian turns around his longitudinal axis. The energy transferred during the
impact is also almost completely converted into rotational energy. Since the
pedestrian moves away from the vehicle, there is no contact with the side of the
vehicle.

3. Lateral brush:

One speaks of lateral stripes if the pedestrian only comes into contact with the side
of the vehicle. If it is a pedestrian going in, he or she hits the vehicle side, is thrown
away from the vehicle and comes to lie behind the line of the front side of the
vehicle.

If the pedestrian moves parallel to the vehicle, an atypical lateral stripe can form,
in which the pedestrian is touched by parts protruding from the side of the vehicle.
The pedestrian is then immediately thrown back by these parts and leaves no further
traces on the vehicle.

4. Running over:

When driving over a vehicle, at least one wheel drives over the pedestrian's body.
Without prior contact between the pedestrian and the vehicle driving over him, we
speak of a simple crossing, i.e. the pedestrian has been hit by another vehicle and
thrown onto the road or has already been lying on the road for other reasons.

If the pedestrian is caught by a vehicle and driven over by it, this is a complicated
crossing. This case is very rare. Mainly for box-shaped vehicles with low deceleration.

KINEMATICS AND DYNAMICS OF A PEDESTRIAN ACCIDENT

The collision between pedestrian and motor vehicle can be divided into three phases:
Contact, flight and slip phase.

The contact phase is initiated by the primary impact and ended by the detachment
of the pedestrian from the vehicle. The energy transferred to the pedestrian during
the primary impact is now dissipated during the secondary impact, during the slip
phase and, if necessary, during the tertiary impact.

In order to be able to explain the sequence of events during an accident more

precisely, the vehicle geometry must also be included. Depending on the profile of
the chassis, this is divided into four basic types.
Type A:

In these wedge-shaped vehicles, the primary impact is usually caused by the bumper,
which hits the pedestrian below the centre of gravity and triggers rotation towards
the vehicle. In addition, the pedestrian moves away from the vehicle and impacts at
different distances from the front of the vehicle, depending on the length of the
bonnet and the collision speed. At low speeds, the angular momentum is cancelled
during head and fuselage impact.

At collision speeds above about 30 km/h, the angular momentum can become so
great that the head and fuselage impact is not sufficient to compensate for the
angular momentum. The pedestrian's centre of gravity is raised, which in extreme
cases can even lead to a throw over the vehicle.

Type B:

In contrast to type A, such high angular momentum is not achieved here because the
pedestrian is not hit so far below the centre of gravity. Only at higher speeds can
the pedestrian's head reach the windscreen.

Type C:

The pedestrian is gripped in full length. The main kick-off point is near the centre of
gravity. Therefore there is almost no rotation or elevation.

Type D:

The raised position of the bumper in contrast to type C leads to a rotation away from
the vehicle. The danger of overrunning is particularly high here, as the negative
angular momentum results in very small casting distances. The pedestrian is thrown
onto the ground directly in front of the vehicle.

CASTING DISTANCE OF PEDESTRIANS

Under the term casting distance a number of values is summarized, from which very
good statements about the collision speed can be derived.
Casting distance: distance between the point of collision and the end position of the
pedestrian in the direction of the longitudinal axis of the vehicle..

Lateral casting distance: distance between the pedestrian's hip impulse point on the
vehicle in collision position and the body's end position in the direction of the
vehicle's lateral axis.

Slipping distance: distance between the first point of contact of the pedestrian with
the road after separation from the vehicle and the final position of the body in the
direction of the longitudinal axis of the vehicle.

Transverse slipping distance: distance between the first point of contact of the
pedestrian with the road after separation from the vehicle and the final position of
the body in the direction of the transverse axis of the vehicle.

PEDESTRIAN CASTING DISTANCE AFTER KÜHNEL (GLOBAL)

Tests by Kühnel have led to the following relationship between the throw distance s
and the collision velocity vk as a function of the deceleration a:

𝑣𝑘2
𝑠 = 0,0178 𝑎 𝑣𝑘 + 0,0271
𝑎
If one resolves the equation after the collision velocity vk, then the result is:

𝑣𝑘 = √0,107855 𝑎4 + 36,90037 𝑎 𝑠 − 0,328414 𝑎³

The formulas are only applicable if the following conditions are met:

 It must be a full frontal collision.

 The collision is triggered by a car.
 The kick-off takes place in a braked state.
  The deceleration must be at least 3 m/s².
 They must not be very small persons (e.g. children), as the high impact point
compared to the centre of gravity results in shorter casting distances.

PEDESTRIAN CASTING DISTANCE AFTER KÜHNEL (DECELERATION)

This test series from Kühnel shows the relationship between the throw distance s and
the collision speed vk without taking the brake deceleration a into account.

Upper barrier: 𝑠 = 0,00375 𝑣𝑘2 + 0,175 𝑣𝑘

Lower barrier: 𝑠 = 6,81818 ∙ 10−6 𝑣𝑘3 + 2,55682 ∙ 10−3 𝑣𝑘2 − 0,827273 𝑣𝑘

PEDESTRIAN CASTING DISTANCE AFTER RAU, OTTE AND SCHULZ

After the real experiments and evaluations of Rau, Otte and Schulz, the following
relationship applies:

𝑠 = 0,0052 𝑣𝑘2 + 0,0783 𝑣𝑘

The following conditions apply:

 The pedestrian must be hit entirely. The kick-off takes place at the front.
 Impact by car or van.
 Deceleration > 4.5 m/s² continuously up to the end position and starting at
the latest immediately after the start of the impact.
 End position of the pedestrian in front of the vehicle front.
 For very small children there are shorter throwing distances..

PEDESTRIAN CASTING DISTANCE AFTER DETTINGER

The Dettinger (Dekra) model is to be used for an impact where braking was only
applied during the collision or up to 0.6 s after the collision. If the vehicle is not
braked during the collision, the impact point is slightly higher and the impact factor
greater due to the missing brake pitch. This means that greater longitudinal casting
distances can be expected than with a braked impact.

𝑠 = 0.0000164 𝑣𝑘3 + 0,00452 𝑣𝑘2 + 0,071 𝑣𝑘 + 2,5

Dettinger specifies the limits for vK with ± 4 km/h.

CYCLIST CASTING DISTANCE AFTER KRIEG, RITTER ET. AL.
Tests resulted in an overall diagram which is valid for both intersection collisions and
rear-end collisions.

Minimum tolerance: 𝑠 = 0,0028313 𝑣 2 + 0,3 𝑣

Total curve: 𝑠 = 0,0028313 𝑣 2 + 0,1635798 𝑣

Maximum tolerance: 𝑠 = 0,0028313 𝑣 2 + 0,0271596 𝑣

CASTING DISTANCE OF GLASS SPLINTERS

From the experimental investigations by Braun and Kühnel on the casting distances
of headlight and windscreen glass splinters, conclusions can be drawn about the
collision point and the collision velocity.
If the total splitter field is selected, the casting distance is calculated as follows
after attempts of Dekra:

First glass splinter: 𝑠 = 0.0018 𝑣𝑘2 − 0,0544 𝑣𝑘

Last glass splinter: 𝑠 = 0,5 𝑣𝑘

THE SETTLEMENT
From the settlement of the pedestrian, approximate statements can be made about
the collision speed. The settlement is defined as the length between the roadway
and the head impact point along the vehicle surface.

According to Kühnel, the following applies to the settlement:

𝐿𝐾 = 0,7 𝐻𝐹𝑔 + 𝑐 𝑣𝑘

LK … Settlement c … Constant of chassis (1 – 2,5 cm/km/h)

HFg … Height of pedestrian

Kühnel gives c with about 1 cm/km/h, however measured at older vehicles with
steep front. The flatter the bonnet, the larger c can be assumed.

It should be noted that at collision speeds of up to approx. 20 km/h the pedestrian's

head tends to hit the vehicle at the same point. Only at higher speeds does the
development increase by about 1 to 2.5 cm per 1 km/h according to the constant c.
The development depends very much on the chassis constant c, which in turn
depends on the shape of the vehicle front. If this shape constant is not well known,
the settlement should not be used or at least a correspondingly large tolerance range
should be indicated. However, the settlement serves as a good control value.

MOVEMENT VELOCITY OF PEDESTRIANS

According to investigations by Eberhardt and Himbert (speed in m/s):

Type of Age of pedestrian

movement 6-7 14 - 15 20 - 30 30 - 50 50 - 60 70 - 80
m f m f m f m f m f m f
Walk 1,5 1,5 1,7 1,6 1,2 1,4 1,5 1,3 1,4 1,4 1,0 1,1
Fast walk 2,0 2,0 2,2 1,9 2,2 2,2 2,0 2,0 2,0 2,0 1,4 1,3
Run1) a 3,4 2.8 4,0 3,0 3,0 4,0 3,6 3,6 3,5 3,3 2,0 1,7
b 3,1 2,8 3,4 3,0 3,0 3,2 3,2 3,2 3,0 3,0 2,0 1,7
Run2) a 4,2 4,0 5,4 4,8 7,4 6,1 6,5 5,5 5,3 4,6 3,0 2,3
b 3,6 3,4 4,2 3,9 4,9 5,0 5,0 4,7 4,0 4,1 2,5 2,1
1 … normal endurance run a … flying start
2 … fastest possible type of movement b … standing start
m … male f … female

SIGHT OBSTRUCTION BY ONCOMING TRAFFIC (MODULE)

The upper green vehicle represents the visible oncoming traffic and the blue lower
vehicle the accident participants. The red dots mark the pedestrian in different
positions. The top one shows the pedestrian in the position at the time of entering
the road, the bottom one the pedestrian in the position at the time of collision.
The figure shows the positions of the vehicles at the time of the first view of the
pedestrian who is in the middle position at this time.

Based on the intercept theorems, the following relationship between the distances
is formed:

𝐴𝐵 𝐵𝐷
=
𝐴𝐶 𝐶𝐸
The following applies to the AC line: 𝐴𝐶 = (𝑠2 − 𝑠1 ) sin(𝛼) − 𝑑

𝑠1 = 𝑣𝐹 𝑡

s2 Pedestrian's distance to collision

s1 Pedestrian's path to first sight
d Collision position referred to the left side of the vehicle
minus the seat position of the driver also referred to the left side of the
vehicle.
α Direction angle of the pedestrian
vF Velocity of the pedestrian on the distance s1
t Time from the beginning to the first sight. Variable used to solve the
equation
BD Distance of oncoming traffic from the beginning to t
CE Distance of the person involved in the accident up to t

BD is derived from the speed and, where appropriate, from the acceleration and
speed limit of oncoming traffic and t minus the distance rear - crossing position of
the pedestrian.

CE results from collision velocity, initial velocity and acceleration as well as from
the total time minus t. The distance of the front from the seating position and s2
cos() is considered as well.

Depending on the situation, an equation of 2nd or 3rd degree in t results, which can
be solved. It is then checked whether the speed limit of oncoming traffic has already
been reached at the time of t, or whether the pedestrian has already reached the
position where his speed is different. If necessary, the calculation is carried out again
using the corrected formulas.

The reaction time taken up by the driver involved in the accident is the total time
less braking time, buildup time and t.

If the collision velocity or the initial velocity is to be calculated from the permissible
reaction time, a closed solution is not possible, therefore an iterative method is
used. The same formulas are used and the required quantity is changed until the
calculated reaction time corresponds to the permissible time.
OVERTAKING (MODULE)

The basic formula is simply the equation between the distances of the vehicle being
overhauled and the vehicle being overhauled.

𝑠𝑔𝑒2 − 𝑘𝑜𝑟𝑟 = 𝑑1 + 𝑠𝑔𝑒1 + 𝑙1 + 𝑑2 + 𝑙2

𝑠𝑔𝑒2 Distance of the overtaking vehicle

𝑠𝑔𝑒1 Distance of the overhauled vehicle
𝑘𝑜𝑟𝑟 Correction size due to lane change
𝑑1 Initial distance
𝑑2 Final distance
𝑙1 Length of the overhauled vehicle
𝑙2 Length of the overtaking vehicle
The distance of the overtaking vehicle consists of a lane change during pulling out,
a straight section and a lane change during shearing. The lane change is calculated
along an inclined sine line. Therefore - instead of the distance during the lane change
- the length space requirement (i.e. the projection onto the driving line of the
overhauled vehicle) must be used. This is achieved by the correction element "korr".

The input possibility to consider a velocity limit and a final velocity for each vehicle
or not, results in a total of 16 different possibilities with correspondingly many
formulas.

CHANGE OF VIEWING ANGLE (MODULE)

Explanation of terms:

 Defensive distance (defensive time): distance (duration) from the point of

hazard identification (hazard identification point) to the collision position
(first contact).
 Accommodation: Focusing of the eye (eye lens) on an object.
 Accommodation time: Required time for accommodation.
 Conspicuity value: Measure of perception and recognizability.
 Conspicuity point: Point (location) at which an object becomes conspicuous.
 Attention: attention of the activity, interests and desires of the person to the
subject of the activity.
 Resolution ability: Ability to still perceive 2 separate points separately.
 Field of vision: The entirety of the points that can be fixed with the eyes
moved and the head immobile: Approx. 60° to the left and right and approx.
40° up and down. (cf. with visual field).
 Eye-jump (saccade): jerky eye movement caused by the narrow limitation of
the central visual area.
 Eye-turning: Time between the start of the eye jump and the end of the
correction acccade.
 Decision-making time: Time from the recording of the content of a perception
(recognition) to the making of a decision.
 Fovea centralis: Area of the retina in which sharp vision is possible, comprising
approx. 1° to 1.5°.
 Hazard detection: Detection of information content as danger.
 Hazard detection position (point): Current position at the time of hazard
detection.
 Field of vision: A section of the environment that can appear on the retina of
an unmoving eye.
 Used field of vision: A section of the field of vision that is determined by the
situation. (Includes objects that impair vision such as glasses, helmet, car A-
pillar, etc.).
 Reaction: A change in behaviour following the perception of information.
 Spontaneous reaction: reaction without conscious decision-making processes.
 Choice reaction: Reaction after evaluation of alternative possibilities.
 Reaction cause: Information to which a potential reaction can be made.
 Reaction request: Information that must be reacted to.
 Reaction time: Duration from hazard detection to the start of braking /
steering (start of the respective threshold time). This definition makes the
most sense for forensic practice. In other scientific disciplines other and more
differentiated definitions are used.
 Visual ray: Straight from the eye to a point of an object.
 Visual angle: The angle between the rays of vision and the boundary points of
an object.
 Visibility point: Point at which the visual beam passes the obstacle for the first
time.
 Visual distance: A distance within the visual range.
 Visible dead space: A non-visible area around a vehicle that depends on the
seating position.
 Distance of seconds: Path travelled in one second.
 Blind spot: A section of road not shown in interior and exterior mirrors and
not visible to the driver.
 Visus (Visual acuity): reciprocal value of the smallest viewing angle (in arc
minutes) that 2 points can include to be just visible separately.
 Perception: Capturing information that can potentially be reacted to or that
is a reason to turn one's view to it.
ATTENTION (CONCENTRATIVE - DISTRIBUTIVE)
A distinction must be made between concentrative and distributive attention:

Attention
concentrative distributive
Recognizability Details of a few objects Rough overview of many
objects
Orientation Performance Limited Good
Information Scope Small Large
Information Accuracy Large Small

Concentrative and distributive attention are used according to subjective

requirements and can only be present successively, not simultaneously.

In the alternation between distributive and concentrative attention, there is usually

a leap in the gaze (gaze turn), unless a "coincidentally" fixed object is the reason for
a sudden concentrative attention turn.

From the appearance of the stimulus in the peripheral visual area to recognition
(reaction request), 0.4 to 0.7 s elapse.

VISUAL SYSTEM
Accommodation time:

The accommodation time depends on the age and the size of the accommodation
jump.

Example: From remote setting up to 50 cm, the accommodation jump lasts approx.
0.5 s for a 28-year-old and approx. 0.75 s for a 40-year-old.

Distribution of sensory cells on the retina:

Two different types of sensory cells are distributed on the retina: Cones for color
vision and rods for light-dark vision.

The cones are located in a narrow area around the central fovea. Towards the
periphery the cone density decreases strongly. The distribution is colour-related.

In the area of the fovea the density of the rods is 0, towards the periphery it first
increases strongly up to approx. 20° and then decreases again by half.

Visual field:

The visual field in binocular vision under optimal conditions extends vertically to a
maximum of 130° (60° upwards and 70° downwards) and horizontally to about 180°.
The resolution is very limited in the peripheral visual field and therefore also the
perception of stationary objects away from the line of sight. On the other hand, the
perception of movement in the periphery is good.

Static visual acuity:

The average visual acuity of the normal-sighted population is 1, which means that
one angular minute can still be resolved. An object with a diameter of almost 3 m
can still be seen at a distance of 100 m in the area of the fovea under good visual
conditions.

Dynamic visual acuity:

Dynamic visual acuity is the ability to recognize details in moving objects. Dynamic
visual acuity is greater than static visual acuity when an object moves slowly across
the line of sight and less when the object moves quickly.

Depth perception:

Depth perception in binocular vision is mainly achieved by the different viewing

direction (convergence angle) from the two eyes to the object. The images of an
object created in the eyes show a lateral shift. This enables spatial vision. Sometimes
it is also possible to estimate the depth distance based on the size of the image of a
known object.

Depth perception and thus distance estimation is influenced by many factors. For
example, clear vision provides shorter distances and blurred vision (dust, fog) greater
distances. For example, the feeling is created that the mountains are close when
there is a foehn - i.e. clear vision.

The perceptible distance difference depends on the distance of the object:

Distance of object Perceptible distance difference

1 m 0,4 mm
3 m 1,3 mm
10 m 4,0 cm
50 m 1,0 m
100 m 3,5 m
1000 m 275 m
Motion perception:

The perception of the movement of an object can be achieved simply by fixing the
object in place, i.e. by constantly imaging the object in the retina in the fovea, and
thus requiring movement of the eyes or head or accommodation. While the fixed
object is always displayed at the same position, the image of the environment shifts.
If, on the other hand, the object is not fixed, the image is taken in the peripheral
area of the retina. If the object now moves, the pixel moves over the retina, while
the image of the environment remains the same.

VISUAL INFORMATION RECORDING

Perception of static objects:

When a vehicle moves along its route, direct visual contact occurs at a certain point,
i.e. at this point (visibility point) the visual beam passes the obstacle concealing the
view.

Depending on the size of the conspicuousness of the other psychological parameters,

the object is conspicuous (conspicuous point) and perceived more or less late (or
never!).

After further approximation and comparison with memory contents, the object is
recognized in its meaning (recognition point). Depending on whether the object is
important for further driving behaviour, the driver will initiate an action.

Perception of one's own movement (velocity):

In the eye of a resting observer, the environment forms a constant image on the
retina. If the observer moves forward, the image in the eye changes. The fixed object
point is still imaged in the fovea and does not change its position. The larger the
angle to another point, the smaller the distance and the greater the own speed, the
faster the image point travels over the retina. This results in a kind of flow pattern
of points that move from the fovea to the outside, the faster the further outside
they already are.

Since many parameters influence at the same time, a clear interpretation of the flow
pattern and assignment to a certain velocity is not possible, whereby the estimation
of the own velocity is only possible inaccurately.

The estimation of one's own speed is also influenced or coined by other sensory
perceptions (auditory and mechanosomatic). For example, in a quiet and calmly
gliding vehicle, one's own driving speed tends to be underestimated. Not to be
underestimated is also the habituation effect.

Perception of the movement (speed) of an object

The movement of an object generally causes a displacement of the image on the

retina. If the object moves away, the image becomes smaller. When approaching, it
is the other way round. A lateral movement causes a pure displacement of the image,
or a head or eye movement is necessary.
The physiological limit value for the perceptibility of a movement (velocity)
correlates with the required minimum angular velocity with which the image on the
retina changes. If this perception threshold is exceeded, it is possible to perceive a
movement without estimating a speed. Head and eye movements are calculated and
do not influence the perceived measure of angular velocity.

The required minimum angular velocity depends on contrast, luminance, object size,
reference point, direction of movement and observation time.

PERCEPTION OF RELATIVE MOTION (RELATIVE VELOCITY)

Relative movement in longitudinal direction:

The perception of a relative movement in longitudinal

direction is proportional to the change of the depth
distance. If two vehicles drive in the same direction, the
problem can be reduced to the movement of the object in
front. The relative velocity causes the angle of vision to
change on vehicle contours or to the brake lights.

Trigonometric considerations lead to the following results

for the calculation:

𝑑𝛼 4 𝐵 (𝑎 𝑡 + 𝑣)
𝜔= =
𝑑𝑡 (𝑡 (𝑎 𝑡 + 2 𝑣) − 2 𝐷)2 + 𝐵 2

Investigations with test persons have provided the following limit values for
conspicuous angular velocities under laboratory conditions:
The figure shows the determined values depending on the different brake
acceleration values:

It can be seen very clearly that the velocity level had no influence on the reaction
and that only the angle change as a function of the braking acceleration was
relevant. It is also noticeable that the angular velocity is indirectly proportional to
the distance between the vehicles: double the distance requires only about half the
angular velocity.

Relative movement in transverse direction - inner field of vision:

The investigation of crossing obstacles in the inner field of vision is made more
difficult by the fact that different effects overlap here, so that it cannot be said
whether and to what extent the change in viewing angle was the attention-grabbing
moment.

Assuming the situation of an intersection with stop lines, the question arises as to
when an approaching road user can notice that another road user is entering the
intersection from a standing position. The points on the retina form a flow pattern.
If an object unexpectedly starts to move, this represents an abnormality in the flow
pattern, which can lead to a reaction request. If one takes into account the aspect
that points further out in the field of vision move outwards more quickly, a right-
angled approach of the obstacles results in three parameters that are decisive for
such a situation:

 The speed of approach - the faster the observer approaches, the faster the
flow pattern flows outwards.
 The angle of observation - the further outside the corresponding point is, the
faster it already flows outwards.
  The acceleration of the obstacle - this is relevant for the conspicuousness of
the anomaly.

The change in viewing angle in the inner area becomes particularly relevant if the
view of a potential obstacle is obscured by an unfortunate relationship of spatial
conditions (e.g. A-pillar).

Relative movement in transverse direction - external field of view:

Constant observation angle:

Particularly dangerous are situations in which the

viewing angle of a transverse vehicle practically does not
change. This occurs when the following applies:

|𝑣
⃗⃗⃗⃗𝐻 | = |𝑣
⃗⃗⃗⃗𝐹 | sin 𝛽 csc(𝛼 + 𝛽)

Measured values of the detection time at a constant

observation angle β:

The measured values determined depend on the approach angle:

It can be seen from the figures that a greater distance to the observed object
requires a lower angular velocity to complete an approximation process. to
recognize. This corresponds to intuition, since small angular velocities of distant
objects produce a relatively much larger image on the retina than near objects that
appear large from the outset. The "relative size" of the observed object thus changes
faster with distant objects than with near objects.
Variable observation angle:

Due to the practicability of the application, the angular change that occurs when
the obstacle deviates at a certain velocity from the velocity at which the
observation angle β is constant was investigated.
In the experiments, different combinations of approach angle α and observation
angle β were investigated, with the collision partners driving at different
velocities, deviating from the constant observation angle.

The graph clearly shows that the perceived change in angle

rises linearly with Δv. The detection time is indirectly proportional to the angular
velocity.
LANE CHANGE PROCEDURE (MODULE)

The lane change procedure (e.g. to the left) consists of 4 sections. First, the vehicle
is steered to the left, then the steering angle is increased from 0 (straight ahead
travel) to a certain maximum value (first section) and then steered back again to 0
(second section). At this moment, the vehicle again reaches straight-ahead travel,
at the same time the maximum yaw angle is reached. Afterwards, the vehicle is
guided analogously to the right and then back again.

The curve traversed is characterized by the fact that its radius of curvature, which
corresponds to the current radius of the curve, decreases from infinity to a minimum
value and then becomes infinitely large again. In the right-hand drive phase, the
same process takes place once again.

The model used in the accident reconstruction program AnalyzerPro is an oblique

sine line, i.e. the vehicle travels with its center point on a sine curve which is rotated
so that the tangents at the beginning and end are horizontal and have a distance
equal to the entered side distance. The centre of the vehicle is defined as the centre
of the wheelbase and the width.

With a sinusoidal curve, the radius of curvature is, exactly as desired, at the
beginning in the middle and at the end infinitely large (steering angle is zero) and
reaches after approximately one quarter and three quarters of the path in each case
the smallest radius. The maximum lateral acceleration is obtained approximately at
one of these two points, namely at the point where the velocity is greater.

The equation of the oblique sine line is as follows:

𝑥 1 2𝜋𝑥
𝑦 = 𝑆𝑣 sin ( )
𝐿2𝜋 𝐿
y Momentary offset
x Path along the straight road, x goes from 0 to L
Sv Side offset (width for evasion)
l Alternative route (actually the space required in the longitudinal
direction of the road)
For the calculation of the smallest radius of curvature, the oblique sine line is rotated
in the x-axis:

𝑆𝑣 2𝜋𝑥
𝑦= √𝑆𝑣2 + 𝐿2 sin ( )
2𝜋𝐿 √𝑆𝑉2 + 𝐿2

The smallest radius of curvature is achieved if the following applies:

2𝜋𝑥 𝜋 2𝜋𝑥 3𝜋
= 𝑜𝑑𝑒𝑟 =
𝐿 2 𝐿 2
Also if:

𝐿 3𝐿
𝑥= 𝑢𝑛𝑑 𝑥 =
4 4
The smallest radius of curvature is calculated as follows:

𝐿 √𝑆𝑉2 + 𝐿²
𝑅=
2 𝜋 𝑆𝑉

The lateral or normal acceleration depends on the velocity and the curve radius:

𝑣2
𝑎𝑞 =
𝑅
From the last two equations, L can be calculated as a function of velocity v and
lateral acceleration. This means that the maximum lateral acceleration that is
reached at the point with the smallest curve radius must be entered. If the lane
change process is accelerated, then L must be calculated iteratively. The distance
of the vehicle centre point is calculated by numerical integration of the sinusoidal
curve.
CASTING DISTANCE (MODULE)

The gravity of the earth acts on a dropped body in a vertical direction and the friction
against the current flight direction.

The following applies to air friction:

𝑅 = 0,5 𝑐𝑤 𝐴𝑞 𝜌 𝑣 2

The following applies to the inflow velocity:

𝑣𝑥 𝑣
𝑣𝑤 = (𝑣 ) + ( 𝑤 )
⃗⃗⃗⃗
𝑦 0

The calculation is carried out by means of numerical integration:

𝑅𝑥 = 0,5 𝑐𝑤 𝐴𝑞 𝜌 𝑣𝑥2 𝑢𝑛𝑑 𝑅𝑦 = 0,5 𝑐𝑤 𝐴𝑞 𝜌 𝑣𝑦2

𝑅𝑥
−
𝑎=( 𝑚 )
𝑅𝑦
− −𝑔
𝑚
𝑣 = 𝑣0 + 𝑎 𝑡

𝑟 = 𝑟0 + 𝑣 𝑑𝑡

(vx, vy) Velocity vector (spatial)

(vW, 0) Headwind velocity
cw Coefficient of air resistance (cw value)
Aq Cross-sectional area
ρ Air density
v Momentary inflow velocity
a Acceleration vector
v Velocity vector
r Position vector
KINETICS

INTRODUCTION

When calculating collisions, in contrast to kinematics, all forces must be directly

taken into account and their effects calculated. A collision analysis can be conducted
in different variants, whereby depending upon situation certain methods offer
themselves. It is to be pointed out that the determination of a collision speed is not
the Ultima Ratio with the production of an accident reort, but rather only an
important intermediate step.

BASICS

IMPACT HYPOTHESIS
An impact is by definition an exchange of forces between two bodies. In a
macroscopic view, such as a car accident, it is assumed that the distances taken by
the bodies during the impact time can be neglected. Instead, a single impact point
is assumed to be representative. Furthermore, all individual forces occurring in an
impact are combined to form a total resulting force.

The simplest example here are two ideal, non-deformable bodies:

t… Contact tangent n… Contact normal

F1 … Force acting on body 1, F2 analogue

At the collision point, the movement of two rigid bodies results in forces which point
in opposite directions with the same magnitude.
In reality one works with deformable bodies, which is why the impact hypothesis is
only approximate. For example, the permanent deformation of a vehicle after a
collision must be taken into account.

Characteristic values for the impact are in the contact area:

 The elasticity of the material (represented by the k-factor = impact

coefficient)
 The friction occurring (represented by the friction µ)

THE IMPACT COEFFICIENT K

After the impact of a real body, a deformation remains. Depending on the elasticity
of the material, this can be very large or practically non-existent. In addition, the
elasticity of all the materials involved is not normally the same. The impact
coefficient k is a concept that makes it possible to look at the phenomenon in
summary.

The straight impact:

In this case, the resulting forces are in the same plane as the velocity vectors, so
rotation does not occur. The impact is divided into 2 sections:

1. At the point of contact, opposite, equally large impact forces occur (F 1, F2).
The impacting body 1 is decelerated, the impacted body 2 is accelerated, both
bodies are deformed. At the end of the first section the bodies have reached
their maximum deformation, the speeds are the same.
2. The deformations decrease depending on their material properties. Body 2 is
further accelerated, body 1 is further decelerated - until the bodies separate.

The two extreme cases are here:

...the completely plastic impact, in which the two balls do not deform back, remain
together and move at a common speed.

...the completely elastic impact, in which a complete traceability takes place, ball
1 stands after the impact and ball 2 has taken over the entire impulse.
The description of the elasticity of an impact is given by the k-factor:

∆𝒗′ 𝒗′𝟐 − 𝒗′𝟏

𝒌= =
∆𝒗 𝒗 𝟏 − 𝒗𝟐

Or in words: The impact coefficient is the velocity difference after the impact
divided by the velocity difference before the impact..

In general, the following must be fulfilled: 0 ≤ k ≤ 1. It means:

k = 0 … plastic impact 0 < k < 1 … real impact k = 1 … elastic impact

A special case is the sliding collision, where v1‘ > v2‘ is possible and the k-factor can
become negative.

The eccentric impact:

In reality, it is not 2 centres of mass that collide, but e.g. 2 parts of a vehicle chassis.
The impact coefficient k must then of course be determined from the velocities of
the points of contact. Only the velocity components in the direction of the contact
normal are of interest here (vBn and vBn‘).

Normal component of the contact point Normal component of the contact point
velocity before the impact velocity after the impact

The following results for the k-factor of the eccentric impact:

∆𝒗′𝑩𝒏 𝒗′𝑩𝒏𝟐 − 𝒗′𝑩𝒏𝟏

𝒌= =
∆𝒗𝑩𝒏 𝒗𝑩𝒏𝟏 − 𝒗𝑩𝒏𝟐

As control parameter in AnalyzerPro you can use ∆𝑣𝐵𝑛 ′ ′

= 𝑣𝐵𝑛2 ′
− 𝑣𝐵𝑛1 , because it is
usually between 2 and 7 km/h for car - car collisions.
THE FRICTION COEFFICIENT µ
The friction coefficient µ is used to describe the friction in the contact zone. The
definition is:

𝑇𝑎𝑛𝑔𝑒𝑛𝑡𝑖𝑎𝑙 𝑝𝑎𝑟𝑡 𝑜𝑓 𝑡ℎ𝑒 𝑑𝑒𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑓𝑜𝑟𝑐𝑒 𝐹𝑡

𝜇= =
𝑁𝑜𝑟𝑚𝑎𝑙 𝑝𝑎𝑟𝑡 𝑜𝑓 𝑑𝑒𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛 𝑓𝑜𝑟𝑐𝑒 𝐹𝑛

The size of the friction coefficient is influenced by the surface quality of the contact
zone and the type of impact.

COLLISION ANALYSIS

BACKWARD ANALYSIS
Momentum-backwards analysis:

From the conditions after the collision (translational and rotational movements of
the vehicles) the conditions before the collision are concluded. The principle of
linear momentum and the principle of angular momentum are used for the
calculation. The coasting impulses are known by amount and direction (from
tracking) as well as the direction of the run-in impulses.

The solution to the principle of linear momentum results from the fact that the shock
drive vector must be the same size for both vehicles and act in the opposite
direction. Then it is checked whether the principle of angular momentum is fulfilled
by the shock drive calculated from the principle of linear momentum. The position
of the touch tangent or the impact coefficient need not necessarily be considered in
this procedure.

EES-backwards analysis:

In the case of oncoming traffic collisions or rear-end collisions, the application of the
drive-balance method, i.e. the sole application of the principles of linear
momentum, is problematic, since small angular changes in the momentum vectors
produce large changes in the magnitudes of these vectors.

In order to overcome this problem, the EES procedure was developed. In addition to
the principles of linear and angular momentum, the law of conservation of energy is
also taken into account. The difference between the energy of the system before
the collision and the energy of the system after the collision corresponds
approximately to the deformation work that becomes visible when the vehicles are
damaged. This deformation work can be calculated by suitable methods. This way a
further equation of determination is obtained, which replaces the direction of the
input momentum. This facilitates the calculation of such accident types, in particular
accidents with slipping. The position of the contact tangent or the impact coefficient
are not necessarily important for the EES method as well.

The influencing variables in the backward analysis are summarized and discussed in
the following table. The term "momentum backward" means that the solution is
found primarily by applying the principle of linear momentum. EES backward" means
that the direction of an input pulse is replaced by the law of conservation of energy
equation. In the table, an "E" means that this is an input value. The values without
"E" are calculated. Values that are marked with "SV" are obtained from tracking
analysis.
 Momentum EES
backwards backwards
Vehicle 1
Direction momentum after collision SV SV
Amount momentum after collision SV SV
Direction momentum before collision E
Amount momentum before collision
Angular momentum after collision SV SV
Angular momentum before collision E E
EES value E*

Vehicle 2
Direction momentum after collision SV SV
Amount momentum after collision SV SV
Direction momentum before collision E E
Amount momentum before collision
Angular momentum after collision SV SV**
Angular momentum before collision E E
EES value E*

Orientation of contact tangent and normal

Touch point velocities after collision in normal
direction (k)
Touch point velocities after collision in tangent
direction (kt)
Friction coefficient in the contact zone (µ)
* Can also be calculated from the structural formulas
** This value can be calculated iteratively from the inlet data

In the column "EES backwards" you can alternatively enter the direction of the
momentum for vehicle 1 and leave the direction of the momentum for vehicle 2
blank.

The information about the contact point velocities and the friction value in the
contact zone are not required as input variables, but they can be calculated if the
orientation of the contact tangent and normal is defined and serve as control
parameters.

TIPS AND TRICKS FOR BACKWARD ANALYSIS

The following problems can occur during a backward analysis:

High friction value µ:

This happens when the calculation provides a impact drive that is too strongly
twisted in relation to the normal direction. The consequence of this is that the
friction in the contact zone becomes greater than is necessary for an adhesive
impact. Therefore, the friction would have an accelerating effect.

High DvBn‘ / High k-value:

The differential velocity of the points of contact after the impact should be roughly
2 - 7 km/h. Sometimes, however, the calculation result can be significantly higher.
The same applies to the k-factor, which is proportional to DvBn'.

If these problems occur, the input parameters must be corrected. The correction can
only be made indirectly either by shifting the contact point, twisting the tangent, or
changing the coasting data (in the tracking analysis). In most cases it will be
necessary to change the course angle. This can often be changed by correcting the
second position after the collision position.

FORWARD ANALYSIS
With the forward calculation, less is known about the collision sequence with this
approach than with the backward analysis. First of all, there is nothing that can be
regarded as certain. Only it is assumed that two vehicles drove in given directions at
given speeds. One wants to calculate from this which conditions existed after the
collision.
So only amounts and directions of momentums before collision are assumed. Nothing
is known about momentums after collision. Also the collision location related to the
roadway does not necessarily have to be known. By varying different input
parameters, the collision speed and other parameters are determined using the "trial
and error" principle.

Assumptions must therefore be made about the conditions in the contact zone, i.e.
assumptions about the components of the contact point velocities after the collision.
For this purpose, the orientation of contact tangent and normal must be determined.
The difference of the components of the contact point velocities in normal direction
must be determined via the impact coefficient or directly as input. The coefficient
of friction in the contact zone determines the degree of slip. The collision without
slipping can be regarded as a special case and the impact with hooking at the contact
points can be regarded as a special case of it again.

The input values for the forward calculation differ from the backward calculation as
shown in the following table. It should be noted that the orientation of the touch
tangent and normal can have a very large influence on the result, which is why
extreme care should be taken when determining them. The difference of the
components of the contact point velocities in normal direction is determined by the
impact coefficient, result is a difference in km/h, which is on average about 5 km/h.

The difference between the components of the contact point velocities in the
tangent direction depends on how large the coefficient of friction is applied in the
contact zone, whether this coefficient of friction is used at all and whether there is
an impact with or without slippage. In the case of an impact without slippage, the
components have the same length in the tangent direction (static friction exists); in
the case of an impact with slippage, they may differ considerably from one another
(sliding friction exists).

The two available methods work in principle according to the same formulas and
differ only in the interpretation of the input of the friction coefficient:

Momentum forward (friction value as limit): The friction coefficient entered is a limit
value which has the maximum friction forces occurring in the contact zone. If the
friction cone specified by the friction coefficient is not left, there is an adhesive
impact and the components of the contact point velocities parallel to the tangent
are the same. If there is an adhesive impact, the coefficient of friction occurring
during the collision is smaller than the input (maximum equal).

Momentum forward (R fixed): With this method, the impact is calculated exactly
with the specified coefficient of friction.

The size of the coefficient of friction is checked next. The friction force acts in the
opposite direction to the relative movement of the contact surfaces. This means that
the force must be opposite to the faster parallel component of the contact point
velocities. If the coefficient of friction is increased, the difference between the
parallel components is reduced. If the coefficient of friction is set above the value
required for an adhesive impact, the previously slower vehicle would become faster
and the friction force would act in the direction of the now faster component, which
is of course physically wrong. If this occurs, the error message "Coefficient of friction
too high" is displayed.

Momentum Momentum
backwards forwards
Vehicle 1
Direction momentum after collision SV
Amount momentum after collision SV
Direction momentum before collision E E
Amount momentum before collision E
Angular momentum after collision SV
Angular momentum before collision E E
EES value

Vehicle 2
Direction momentum after collision SV
Amount momentum after collision SV
Direction momentum before collision E E
Amount momentum before collision E
Angular momentum after collision SV
Angular momentum before collision E E
EES value

Orientation of contact tangent and normal E** E

Touch point velocities after collision in normal E
direction (k)
Touch point velocities after collision in tangent
direction (kt)
Friction coefficient in the contact zone (µ) E*
* Depending on the variant as default or limit
** Only required for structural data and distribution of EES values

In addition, a check is carried out to determine whether the vehicles would continue
to penetrate each other after the collision. If the vector of the contact point velocity
of vehicle 1 shows between that of vehicle 2 and its centre of gravity or vice versa,
the warning is given that the input is to be checked.

Manual input will be used in cases where there is a slight slippage. It should be noted
that sliding is always associated with friction; the greater the relative movement,
the greater the friction to be expected up to a certain limit.
TIPS AND TRICKS FOR FORWARD ANALYSIS
If a collision analysis in forward calculation is to be performed, the user must
consider, on the basis of the available clues, how the collision probably occurred and
what type of collision it was. The general procedure is clearly different from
backward calculation and requires very good knowledge of collision mechanics on
the part of the user.

At the beginning of the calculations, an initial hypothesis is made about the probable
collision sequence and the data in the following table is used to check whether the
hypothesis was correct. Normally, several changes to the initial hypothesis are
necessary before a correctly descriptive result can be obtained for the current case.

The initial hypothesis includes the following data:

Initial data Evaluation bases

Collision velocities Vehicle damage and leakage paths

Yaw and course angle Driving behaviour before collision
Location of the contact point Damage to both vehicles
Orientation of the contact tangent Damage patterns, penetration behaviour
Friction in the contact zone Type of collision, case with or without
slipping
Elasticity of the impact partners Differential speed of the contact points
after the collision, structural stiffness of
the affected vehicle parts
Penetration depth Permanent deformation, dynamic
deformation, creep tendency of plastic
parts
Stiffness ratios in the contact zone Position of the collision zone on the
vehicle, evaluation of the affected
structures
Deformation energy, EES values Damage symptoms
Contact point velocity directions Type of collision, case with or without
after collision slipping
Road-related collision location Lanes on the road, movements before the
collision
End positions of the vehicles Accident sketch, information from the
parties involved
The initial hypothesis can be tested by a dynamic calculation of the casting
movement.

WALL IMPACT
Since the "coasting movement of an obstacle" cannot be determined, the backward
analysis is not applicable in the conventional way. However, if the coasting
movement of the impacting vehicle is known, the backward analysis would be
desirable. In the method developed here, the backward analysis was combined with
the forward analysis. From the impacting vehicle the coasting is known, from the
fixed obstacle it is known that the collision speed is 0, i.e. the inlet is known.

The data for the impacting vehicle must be entered in the vehicle data on the left;
the right column is provided for the wall or obstacle.

You must assign a vehicle number to the wall or obstacle and enter a correspondingly
large mass in the vehicle data. In the input mask of the geometrical data you can
select "Obstacle" as vehicle, then predefined data are taken over.

In contrast to the normal backward analysis, the coefficient of friction must be

entered for the wall impact. This method can always be used if the impact is against
a standing object and the direction of impact is in the direction of the centre of
gravity.

COASTING ANALYSIS

TRACKING ANALYSIS (COASTING ANALYIS BACKWARDS)

Tracking is used above all when you can have a relatively good idea of the coasting
of a vehicle due to traces.

Intermediate positions are defined starting from the end position (0). In the previous
figure, two further positions have been defined from the beginning of tracking
analysis (start of coasting, position 3). The course must be determined between the
points. The accuracy is increased if this is done by a "spline calculation" instead of
seeing. The course angle then results from the direction of the tangent of the spline
at the relevant point.
The side slip angle results from the relation: 𝛽𝑖 = Ψ𝑖 − 𝑣𝑖

βi … side slip angle Ψi … yaw angle vi … course angle

The velocity calculation is based on a unicycle model.

The deceleration at the defined positions are calculated. The calculation of the
speed change in the section takes place as a buildup phase, which means that a linear
change of the deceleration is assumed in each interval. The deceleration at the
defined positions is calculated from the side slip angle and the partial braking factor.

The standard formula for this is as follows:

𝑎𝑖 = 𝑎𝑚𝑎𝑥 (𝑇𝐵𝑖 + (1 − 𝑇𝐵𝑖 )| sin(𝛽𝑖 ) |)

This formula is based on a linear relationship between the lateral guiding force and
the circumferential force. The partial braking factor is included in the calculation
here irrespective of the direction. The maximum lateral guiding force is assumed to
be the same as the maximum circumferential force.

This formula can now be modified by assuming a different maximum deceleration for
the longitudinal and transverse directions.

Modified Linear Model:

𝑎𝑖 = 𝑎𝑚𝑎𝑥 𝑇𝐵𝑖 + 𝑎𝑆 𝑚𝑎𝑥 (1 − |𝑇𝐵𝑖 |)| sin(𝛽𝑖 ) |)

𝑎𝑚𝑎𝑥 = 𝑎𝑈 𝑚𝑎𝑥 √cos(𝛽𝑖 )2 + (𝑅𝑉 sin(𝛽𝑖 ))²

𝑎𝑆𝑚𝑎𝑥 = 𝑎𝑈𝑚𝑎𝑥 𝑅𝑉

𝑎𝑈𝑚𝑎𝑥 = 9,81 𝜇𝐵

µB … Coefficient of friction between tyre and road surface

TB … Partial braking factor

RV … Ratio of coefficient of friction: longitudinal to transverse = 1 : RV

u… longitudinal

s… transverse

If the partial braking factor is entered negatively, this means acceleration.

If the linear model is used, the partial braking factor occurs as a direction-
independent summand.
Elliptical model:

In AnalyzerPro, you can also use a relationship according to an ellipse equation:

𝑎𝑖 = 𝑎𝑚𝑎𝑥 √(𝑇𝐵𝑖 cos(𝛽𝑖 ))2 + (𝑅𝑉 sin(𝛽𝑖 ))²

In order that an acceleration in longitudinal direction can also be taken into account,
this formula must be modified:

𝑠𝑔𝑛(𝑇𝐵𝑖 )(𝑇𝐵𝑖 cos(𝛽𝑖 ))2 + (𝑅𝑉 sin(𝛽𝑖 ))²

𝑎𝑖 = 𝑎𝑈𝑚𝑎𝑥
√(𝑇𝐵𝑖 cos(𝛽𝑖 ))² + (𝑅𝑉 sin(𝛽𝑖 ))²

In this formula, the partial braking factor is related to the longitudinal direction
(direction of the circumferential force). With small side slip angles and small partial
braking factors, the resulting deceleration tends to be smaller than according to the
standard formula.

In the ellipse model, the partial braking factor occurs as a direction-dependent

variable. This means that the deceleration (acceleration) in the longitudinal
direction caused by the partial braking factor becomes smaller with increasing side
slip angle, while the deceleration generated by the lateral force becomes larger.

COMPARISON LINEAR AND ELLIPTICAL MODEL

A comparison of the two models shows that with sides slip angles the deceleration
increases somewhat more slowly with increasing partial braking factors.

Assumed values for the comparison:

Friction value: 0,8
Friction coefficient ratio longitudinal to transverse: 1 : 0,9

Partial braking Phi (°) a (m/s²) elliptical model a (m/s²) linear model
factor
0.1 0 0,8 0,8
1 0 7,8 7,8

0 10 1,2 1,2
0,1 10 1,4 1,9
0,2 10 2,0 2,5
0,3 10 2,6 3,2
0,4 10 3,3 3,9
0,5 10 4,1 4,5
0,6 10 4,8 5,1
0,7 10 5,5 5,8
0,8 10 6,3 6,5
0,9 10 7,1 7,2
1 10 7,8 7,8
0 20 2,4 2,4
0,1 20 2,5 3,0
0,2 20 2,8 3,5
0,3 20 3,3 4,0
0,4 20 3,8 4,6
0,5 20 4,4 5,1
0,6 20 5,0 5,6
0,7 20 5,7 6,2
0,8 20 6,4 6,7
0,9 20 7,1 7,2
1 20 7,8 7,8

0 30 3,5 3,5
0,1 30 3,6 3,9
0,2 30 3,8 4,4
0,3 30 4,1 4,8
0,4 30 4,5 5,2
0,5 30 4,9 5,6
0,6 30 5,4 6,0
0,7 30 5,9 6,4
0,8 30 6,5 6,8
0,9 30 7,1 7,2
1 30 7,7 7,7

0 60 6,1 6,1
0,1 60 6,1 6,2
0,2 60 6,2 6,3
0,3 60 6,2 6,5
0,4 60 6,3 6,6
0,5 60 6,4 6,7
0,6 60 6,6 6,8
0,7 60 6,7 6,9
0,8 60 6,9 7,0
0,9 60 7,1 7,2
1 60 7,3 7,3

0 90 7,1 7,1
0,1 90
0,2 90
0,3 90
0,4 90
0,5 90 7,1 7,1
0,6 90
0,7 90
0,8 90
0,9 90
1 90 7,1 7,1
CALCULATION OF DRIVING DYNAMICS (COASTING ANALYSIS FORWARD)
The complete simulation of a driving process requires a large number of input
parameters. In order to make the process more estimable, parameters with only
minor effects on the motion sequence are more or less extensively neglected. It
depends on these simplifications for which purposes such a model can be used and
for which ones it cannot be used.

General input parameters are:

 Environmental data including road inclination

 Traction coefficient
 Separation curve (various friction coefficient surfaces)
 Criteria for trace marking
 Minimum slip limit
 Factor for the slip angle
 Vehicle data / vehicle geometry
o Length, width, wheelbase, overhang, track width
o Centre of gravity reserve
o Empty weight, permissible total weight
o Loading situation
o Spring stiffness, roll stiffness, structural stiffness
o Force distribution
o Drive power
o Type of drive
o Brake system type, brake force distribution
o Tyres
o Maximum load capacity
o Position of the maximum slip angle
o Quotient between sliding coefficient and static friction coefficient
o Axle geometry
o Steering ratio

Time-dependent input parameters are:

 Steering wheel angle setting

 Brake pedal forces
 Accelerator pedal positions
 Time-dependent coefficients of friction, tyre condition factors and partial
brake factors

Initial conditions are:

 Initial speed of the vehicle centre of gravity

 Initial position
 Course and yaw angle
  Yaw rate

Termination conditions are:

 Maximum simulation time

Program properties in driving dynamics:

 A three-dimensional vehicle model is used in which, however, only the three

degrees of freedom transverse motion, longitudinal motion and yaw motion
are calculated as independent quantities and displayed in one plane.
 Dynamic wheel loads are calculated from lateral and longitudinal
acceleration. The over-determined system is solved by introducing roll angles
and axle spring stiffnesses (rolling moment distribution).
 External forces acting on the model are: Weight force, dynamic wheel loads,
circumferential forces, lateral forces, inertial forces.
 The temporal specification of any braking forces is possible. They can be
entered in the form of value pairs (brake pedal force as a function of time).
In the same way, drive forces can be assigned to a specific point in time by
specifying accelerator pedal positions and steering wheel angles. The required
values of the intermediate time steps are determined by linear interpolation.
By providing information on the installed brake force distribution, the brake
forces can be distributed to the wheels in the correct ratio. The drive force is
distributed by specifying the drive concept. With the aid of the steering ratio,
statements can be made about the speed of rotation of the steering wheel.
The model has front wheel steering. The consideration of toe-in or wheel
position errors due to deformation is possible.
 The program has a simple driver model. Driving manoeuvres can be carried
out by specifying certain steering angles, accelerator pedal or brake pedal
positions.
 Tyre lateral forces are determined from the slip angle, wheel load,
circumferential force and coefficient of friction. Kinematic variables for this
are yaw speed, center of gravity speed, instantaneous pole coordinates, wheel
contact point coordinates and steering angle. The lateral forces are calculated
using the (static) tyre map "IPG-Tire".
 It is possible to consider an area with different coefficients of friction in the
calculation. A different coefficient of friction is then defined within this
traction coefficient. For example, it can be simulated that there was an oil
slick on the roadway or that the vehicle left the roadway and continued its
movement on the green strip.
 The current values of distance, speed, total acceleration, lateral
acceleration, yaw rate and course angle can be displayed in a coordinate
window as an aid for evaluating the calculation.
TIPS AND TRICKS IN DRIVING DYNAMICS
It should be borne in mind that the driving dynamics calculations are based on a
model. Therefore, deviations from reality are possible. However, these can be
minimized by a suitable definition of the parameters. It is recommendable to
calculate a test that comes close to the case to be investigated and to check the
applicability of the model there.

In general:

The term skidding is used when a vehicle performs movements that can no longer be
controlled by the driver. Skidding can be caused by driver actions without external
forces or by external forces, for example by collision with another body.

Most often the expert has to examine the skidding coasting after a collision. To a
lesser extent, e.g. in the assessment of single accidents, skidding processes caused
by driver actions are to be assessed.

When investigating skidding processes, various characteristics must be taken into

account that allow classification and interpretation:

Skidding by driver action:

 Drift tracks, from which one can see from the course of the stripes in the tire
tracks whether the driver has additionally steered or braked
 Blocking traces of empty tyres or rim flanges, which may be an indication of
technical defects.
 Deviation procedures, even partially from the roadway. Differences in
coefficients of friction causing yaw movements due to braking.

Skidding after a collision:

 Traces of mostly all wheels and a largely straight centre of gravity path
indicate that the wheels of the vehicle were blocked.
 A curved course of the centre of gravity path indicates that none or not all
wheels were blocked.
 In the case of unblocked wheels, it is important for the correct dimensioning
of the decelerations whether translation and rotation end at the same time or
at different times.

The following pictures show examples from mathematical simulation. They are
intended to show some basic differences and influences.
 Simulation A: Skidding coasting with blocked wheels, rotation v = 40 km/h
to left (positive) s = 11,8 m
µ = 0,8
a = 5,22 m/s²
t = 2,14 s

Simulation B: Skidding coasting with blocked wheels, rotation v = 40 km/h

to the right (negative) s = 11,8 m
µ = 0,8
a = 5,22 m/s²
t = 2,14 s

Simulation C: Full braking with four blocked wheels v = 40 km/h

s = 9,63 m
µ = 0,8
a = 6,39 m/s²
t = 1,73 s

You can see from the graphics that even with blocked wheels a completely straight
line of the centre of gravity is not to be expected with a skidding coasting path.

Furthermore, the actual decelerations in relation to the maximum traction (here µ

= 0.8) are not insignificantly reduced both during emergency braking and even more
so during skidding. This decrease results from the tire characteristics in connection
with the wheel load fluctuations. The comparison of the deceleration during
emergency braking with that during skidding coasting with blocked wheels shows a
significant difference. The deceleration during skidding coasting is about 18 % less
than during emergency braking. In general, it is advisable to expect 80 % of the
deceleration to be expected during emergency braking if the wheels are blocked
during a skidding coasting. The following diagrams show the progressions of
important parameters for simulation A. Noteworthy are the almost linear reduction
of center of gravity and yaw speed (here e.g. 5.22 m/s² and 140 degrees/s²), the
phase shift of the maxima for circumferential and lateral forces and the fact that
the roll angles are always larger than the pitch angles.

Course of centre of gravity and yaw velocity:

Circumferential
force course
 Lateral force
course

Roll and
pitch angle

If the wheels are not blocked, the movements of the skidding vehicles look different:
Tyre tracks are only fragmentary, the path of the centre of gravity is often clearly
curved (wheels deformed, steering turned), translation and rotation usually do not
end at the same time. The following three simulation calculations show basic facts.
In particular, the very different degradation of translation and rotation must be
taken into account when applying approximation formulas.

Simulation D: Skidding with free-rolling wheels, slightly braked

 Simulation E: like D, but less rotation and braking

Simulation F: like E, but less rotation and steering angle at t = 3 s

In the simulations D to F, the initial speed was 40 km/h, the same as in the previous
simulations. The yaw rate for simulation D was 300 degrees/second.

The reduction of the speeds is shown on the following diagrams:

D In contrast to the
blocked wheels,
there is now a stair-
like reduction in
speed. The rotation
is after 1.8 s or 13.5
m at 0. The average
deceleration is
calculated with 167
degrees/s². At t =
1.8 s the linear
speed is still 15.6
km/h. The
E remaining coasting
distance is 5.5 m,
the corresponding
deceleration is 2.75
m/s². The
deceleration for the
speed reduction
from 40 to 15.6
km/h took place
over 13.5 m and
thus averaged only
1.7 m/s².
 F
In simulation E, the
initial yaw rate was
only 200
degrees/sec. The
braking effect on
the rear wheels was
reduced from a
partial braking
factor TB = 0.1 to
0.04.

By chance, translation and rotation now end almost simultaneously. The mean
decelerations are calculated at 3.8 m/s² and 67 degrees/s². The diagrams of the
velocity curve during the skidding process show that neither the linear nor the
rotational velocity reduction are constant. The rotational speed is reduced to less
than half during a distance of almost 4 m, then even an acceleration takes place and
then the decay to 0 takes place within a distance of about 1 m. The yaw deceleration
reaches maximum values of 330 degrees/s². This raises the question at which angle
the largest yaw deceleration occurs and when it is largely constant or even
accelerated. This problem will be discussed later.

In simulation F, the initial yaw rate was further reduced to 100 degrees/second. 3 s
after the start of skidding the steering wheel was turned 90 degrees to the left. On
the corresponding diagram you can see that the rotation is reduced to almost 2 m
distance, the corresponding yaw deceleration has a maximum value of 220
degrees/s² and an average value of 60 degrees/s², related to the total distance
during which the reduction took place. The linear deceleration was 2 m/s² on
average until the complete degradation of the rotation, then about 0.9 m/s² until
standstill.

If you take a closer look at simulation E, where translation and rotation end at about
the same time, you can also make statements about when rotational speed is reduced
and when it is not. For this purpose, the skidding process, the velocity reduction and
the associated decelerations were drawn on a comparable scale. From this and from
further calculations, as well as from the evaluation of measurement records or film
evaluations of accident tests, it can be deduced that the rotation is mainly reduced
at side slip angles between 0 and +/- 30 degrees. The yaw deceleration also depends
on the ratio of translation to rotation.
CALCULATION OF FORCES BETWEEN TYRES AND ROAD SURFACE

INTRODUCTION
In mathematical simulation, the simulation of the effect of the forces in the tyre
contact area is of the greatest importance. The automotive industry is making
enormous efforts to meet the ever-increasing demands on the reliability of its
products. This means that it is necessary to obtain further detailed knowledge about
the safety-relevant components of a vehicle. The vehicle characteristics of the
vehicle are optimally adapted to the driver, the car should always remain easily
controllable and its behaviour predictable in all situations. These requirements can
be summarised under the term "dynamic driving safety". To achieve this goal, driving
stability, steering and braking behaviour must be coordinated. Design measures on
the chassis and on the tyre design are necessary for this purpose.

The tyre plays a decisive role here as the connecting link between the vehicle and
the road. Its behaviour ultimately determines how safely braking and steering forces
can be transmitted.

In order to better control and predict the influence of tyres on vehicle behaviour,
extensive driving behaviour studies and measurements have been carried out over
decades and simulation programs have been developed. An important step in such
investigations is the development of a suitable simulation model for the tyre
properties.

Especially the properties of the tyres are very difficult to describe mathematically.
An important reason is that the friction laws of classical mechanics (Coulomb
friction) cannot be applied to tyres as used in automobiles. Rubber is not a rigid
body, but has viscoelastic properties of high complexity. There is therefore no single
discrete coefficient of traction between the tyre and the road surface. According to
the theory of rubber friction, the frictional force currently transmitted is composed
of the following four components:

 Adhesion component (shear of molecular bonds)

 Hysteresis component (deformation of the rubber)
 Viscose component (shear of a liquid layer in the contact surface)
 Cohesion component (energy losses due to abrasion or cracks)

The adhesion component of the tyres is of primary importance for the driving
movement of the vehicle. Their maximum occurs at very low sliding velocities. These
are present in the area of drive and brake slip during normal driving. It is important
to note that no force is transmitted without slippage.

The properties of tyres are usually determined on test benches. The transmittable
forces are measured as a function of slip or slip angle. They also have an influence:
Road condition, road state, vehicle and tyre condition and driving condition.
When considering the condition of the road, the materials used in the road surface,
the age of the road surface, the traffic load, the season and the micro- and macro-
roughness should be taken into consideration. The road builders have developed
various measurement methods to monitor road grip and use them to check when road
surfaces need to be replaced.

The road condition describes the concrete type of surface, i.e. dry, wet, snowy, icy
or dirty road.

In terms of vehicle and tyre condition, the design, axle suspension, tyre dimensions,
tread design, tyre inflation pressure and other factors must be taken into account.

Driving conditions depend on driving speed, longitudinal and lateral acceleration,

vehicle vibrations and all driver activities. They have a transient effect on wheel
load, slip and skew.

METROLOGICAL RECORDING OF TYRE PROPERTIES

The use of testing machines such as internal or external drum test benches and flat
belt test benches is common for the measurement of tyre properties and tyre
characteristics. With these machines, a wheel (tyre with rim) can be pressed onto
the rotating drum or a running belt. The wheel can be driven, braked, steered or
tilted. The wheel load can be varied.

This results in the typical characteristic curves as can be seen from the systems.
These characteristic curves must now be available in analytical or numerical form so
that they can be used in simulation programs. There are many different approaches
to the mathematical representation of tyre properties in the technical literature. A
distinction must be made between closed solutions, which can contain numerous
simplifications, and iterative methods, in which measurement data is entered and
intermediate values are determined point by point.

MATHEMATICAL REPLACEMENT MODELS FOR TYRES

The replacement models for describing circumferential force via slip and lateral
force via skew can be of any complexity. For a sufficiently precise description of the
tire properties, the representation of the dependency of the transferred
circumferential force on slip and the tire side force on the slip angle is necessary.
Both depend considerably on the wheel load. Therefore, there are tyres with a
certain load carrying capacity.

The tyre load capacity is determined by the tyre designation. There you will also find
information about the change in the tyre load capacity depending on the internal
tyre pressure. As an approximation, it can be said that every 0.1 bar change in
pressure results in a change in the tyre load capacity of 100 N.

The profile of the tyre lateral force over the skew is strongly dependent on the tyre
construction, but also on the wheel load, the tyre inner pressure and the camber.
The difficulty lies in the mathematical simulation of this characteristic curve.

MODELLING
Before a model is created, it must be decided which influences and factors are to be
taken into account and which are not expected to have a lasting influence on the
model with regard to the issue at hand. In their simplified form, models of this kind
can help to narrow down problem areas and then concentrate on the relevant
questions.

The tyre properties can be displayed in different ways and also processed in
programs:

 Representation by tables
 Representation by graphs
 Representation by formulas

The first two possibilities are quite difficult to use. The evaluation may require a lot
of effort or may not be accurate, especially with graphs. However, the third
possibility offers the best handling, namely the use of a closed formula.

In the case of a decision in favour of the formula, there is a choice between:

 Formulas containing series, e.g. Fourier approaches or polynomials of n th

degree
 Formulas that contain special model functions.

The use of rows has some disadvantages:

 Relatively many coefficients are needed to adjust a closed curve to a set of

data.
 Extrapolating beyond the range to be adjusted can result in very large
deviations.
 The coefficients generally do not describe the measures in a recognisable way
that would allow the values to be changed in a controlled and targeted
manner.

The best way to avoid these disadvantages is to use a model function tailored to the
problem. Due to its special structure, it should be able to describe the measured
data with great accuracy. In addition, it should have parameters related to the
typical sizes of the measured data.
In order to be able to process the recorded data more easily and to enable a fast and
simple simulation of the measured values on the computer, the tyre model should
be able to describe the following in particular:

 The circumferential force Fx as a function of slip Χ

 The lateral force Fy as a function of the slip angle α
 The self-aligning torque Mz as a function of the slip angle α

The model should meet the following requirements:

 It must be able to describe stationary and dynamic behaviour.

 The coefficients should be easy to obtain from measurements.
 It must be physically meaningful, i.e. the parameters should describe
characteristic parameters of the tyre in order to draw conclusions about the
stability behaviour by changing them.
 The number of parameters should not be too large for the formula to be
compact and easy to use.
 The model must accurately reproduce the measured values.

The creation of a mathematical tyre model is in any case a very difficult task, which
is why AnalyzerPro uses the "IPG-Tire" model from IPG in Karlsruhe. This company
has an intensive cooperation with the University of Karlsruhe and supplies tyre
models to many research and development departments in the automotive industry
all over the world. The tyre model has been adapted so that only those values are
required as inputs which are either available to the experts or can be easily
estimated. These values are: Load capacity, lateral stiffness and ratio of sliding
friction to static friction. The tyre model IPG-Tire is described in the following in its
complete properties.

TYRE MODEL

THEORETICAL PRINCIPLES
The quality of a tyre model depends decisively on how exactly, with what effort and
in which operating range tyre reactions can be simulated and tyre characteristics
approximated. The approximation of tyre characteristics and the model of the
physical equivalence of slip and slip angle are important theoretical foundations of
IPG-Tire.

In addition to the static tyre reactions, IPG-Tire also determines the horizontal
dynamic behavior. For this purpose, the low-pass character of the tyre is modelled.
REPLICATION OF TYRE CHARACTERISTICS
The tyre model calculates the slip angle, slip and wheel load from the kinematic
condition of the rim and returns the reaction vector associated with this operating
point from lateral force, circumferential force during acceleration or braking and
restoring torque.

The corresponding dependencies

 Lateral guide force as a function of slip angle, wheel load and optional
additional parameter z,
 Circumferential forces as a function of slip, wheel load and z,
 Self-aligning torque as a function of slip angle, wheel load and z

are determined in measurement series. This means that individual discretized

measurements are available in a limited measuring range, while the simulation
program requires continuous information in any load range. Therefore, a simulation
is only possible when the series of single data are available in analytical form, i.e.
as continuous functions. Due to the relative ignorance of the physical causes and the
complexity of the occurrence of the tyre reactions, a mathematical procedure for
the determination of these functions is suitable, in which available information about
tyre physics is included.

A process adapted to the various properties of passenger car and truck tyres in terms
of quality and effort is the use of regionally defined, interlinked splines to
approximate tyre characteristics. Their curvature behaviour corresponds to that of
a bending beam mounted without moments at the ends. This ensures curvature-free
inlet and outlet sections and - in contrast to the representation by a single high-
order polynomial - minimum curvature energy of the approximation curve.

The latter can easily be controlled by the number of functions to be applied so that
finding an optimum for the respective trace is not a problem. In contrast to e.g.
parabolic approaches, the character of the approximation curve is not
predetermined. The only specification made by selecting splines as approximation
curves is the necessary replication of original curves. However, this specification is
not a limitation if the underlying physics of tyre behaviour are known.

A predefined number of coefficients must therefore be determined which determine

the influence of the individual polynomials on the analytical curve. In the case of
approximation, for each curve

𝑦𝑖 = 𝑓(𝑥𝑖 ), 𝑖 = 1 … 𝑚

the solution of a system of equations of the form:

(𝐴𝑇 𝐴)𝑐 = 𝐴𝑇 𝑦
Is to be determined, for each curve, the solution of a system of equations of the
form is to be determined, where A is a band matrix (Dim n,m with n < m) with the
method and the distribution of the data points in x-dependent content. If n=m, there
is interpolation, i.e. the analytical representation contains all known y(x). If there
is an equidistant distribution for x, A is only dependent on the method, thus
predefined, and an inversion for the purpose of solution according to c can be
analytically predefined.

In the general case, however, a basic grid

𝑥𝑖,𝑗 = 1 … 𝑛 < 𝑚

Is predefined, which encloses the measurements. In IPG-TIRE, the distribution of the

grid is automatically adapted to the gradient of the curve to be simulated.
Therefore, no user intervention is required to determine the coefficients.

The described operation is performed successively first for x = slip angle, then for x
= wheel load, and finally, as desired, for x=z (see Approximation by fit).

APPROXIMATION IN THE SLIDING AREA

Due to the difficult measurement of tyre characteristics in the sliding area,
measurements are often interrupted after the maximum. The source data sets then
lack data triples, so that an approximation of the characteristic curves in the
complete possible operating range cannot be achieved without further precautions.
However, since it is basically desirable to run through simulations in the sliding area
as well, the known range is supplemented by a characteristic curve section with an
assumed curve.

A plausible addition is possible, since the general course of tyre characteristics in

the entire range from 0 < s < 1 and 0° < α < 90° is known: After the maximum, the
curve drops off, runs through a turning point and turns into a parallel to the s or α
axis.
The position of the parallels is defined in IPG-Tire by:
𝜇𝑠𝑙𝑖𝑑𝑒
𝑅𝑠𝑙𝑖𝑑𝑒 = 𝑅
𝜇𝑠𝑡𝑎𝑡𝑖𝑐 𝑚𝑎𝑥

µslide … Sliding friction coefficient

µstatic … Static friction coefficient

Rslide … Reaction force or reaction moment for large slip values or slip angles

Rmax … Maximum value of the reaction force or torque

The adaptation of the spline mapping the characteristic curve to the principal curve
is achieved by adding support values.

EQUIVALENCE APPROACH FOR SLIP AND SLIP ANGLES

The tyre model is based on the central hypothesis that longitudinal slip s and slip
angle α lead to a total slip load s*, which can be represented by the geometric
dependency shown.
The slip s* representing the total load of the tyre is therefore the result of the vector
operation:

𝑠 ∗ = √((sin(𝛼)2 + 𝑠 2 (cos(𝛼)2 )

The corresponding traction angle Φ is determined by:

sin(𝛼)
𝜙 = arcsin ( )
𝑠∗
A transformation into a polar coordinate system is performed to reconstruct the real
operating range that lies between the axes to which α = 0 or s = 0 applies. The
renunciation of experimentally determined tyre data in this area makes the model
economical and avoids the overdetermination of model parameters (the axis s = 0 is
determined by pure load maps and by traction ellipses).

The new polar coordinate system is spanned by -s and sin α in the base plane and the
reaction variable R as vertical axis. In the two boundary surfaces (x = 0 and s = 0,
respectively) the circumferential force and the lateral force, respectively, are
mapped as they were originally known by measurement, now already by
approximation. These two characteristic curves are connected by an analytical
approach by means of a three-dimensionally curved surface. This follows certain
physical laws such as continuity, consistency and symmetry.

Friction cake: resulting reaction force R (α, s) for medium wheel load:
Determination of the reaction process for 0 < Φ < π/2:

The real operating point, which is marked by s ≠ 0 and α ≠ 0, now lies in the form of
a reaction force that can be divided into U and S on the intersection curve of an
upright cylinder (radius s*) with this three-dimensional surface.

This allows the reaction force R to be determined at operating point B and broken
down into its components U and S via Φ.

The basically same procedure is used to calculate the self-aligning torques under
mixed load. It is assumed that no self-aligning torques occur with pure slip loading
(α = 0).

The model is still incomplete for circumferential forces, since drive and brake forces
have to be included in the calculation in various ways. A second quadrant is missing,
which enables a variation of s in the range of - 1... + 1 and from Φ in the range 0 ...
π. Any measurements not available for this purpose are automatically determined in
the program by a similarity transformation from the known curves.

WHEEL LOAD DEPENDENCE

With the development of the equivalence approach of slip angle and slip, it is
possible to analytically represent the dependence of the circumferential and lateral
forces or the self-aligning torque on slip and slip angles occurring simultaneously.
The corresponding net surface of the friction cake illustrates this.

However, the dependencies are only known for the wheel loads for which
characteristic curves have been measured and approximated. For the universal
applicability of the tyre model, however, it is necessary to switch from network
planes with constant wheel load to a continuous representation for any wheel loads.
For this purpose, the one-dimensional spline approach used for the representation
of the S(α)-, U(s)- and M(α) characteristic curves (with approximately constant wheel
load in each case) must be extended to a two-dimensional approach for the
representation of the S(α,P)-, U(s,P)- and M(α,P) characteristic maps.
Until now, the coefficient vector c for a certain wheel load P was determined for the
splines of each characteristic curve. The originally measured dependency can also
be represented in the form of its spline coefficients over the slip respectively the
slip angle. The transition to a two-dimensional representation is made by plotting
the coefficients of all splines over the third axis of the coordinate system (wheel
load P). In the direction of P, these coefficients can be interpolated by splines, which
represent the dependence of the coefficients of the one-dimensional splines and thus
of the original curves of any wheel loads.

A section perpendicular to the S(α) plane, which is the lateral boundary of the
friction cake for s=0, illustrates the appearance of individual S(P) curves. The lateral
force values for different wheel loads are read from the diagram at a constant slip
angle and transferred to a lateral force (wheel load) diagram. The desired constant
dependence of the lateral force on the wheel load at constant α is obtained by spline
interpolation between the points.

Determination of tyre characteristics under any wheel loads:

For wheel loads approaching zero, the curve follows a course that changes to the
(S=0, -P) axis at P=0. This form is forced by the automatic insertion of an artificial,
horizontal (P=O,S=0) characteristic curve into the S(α) diagram.

In the same way, the dependencies circumferential force - slip - wheel load and self-
aligning torque - slip angle - wheel load are also determined. Using the equivalence
approach, the friction cake for any wheel load can then be determined. Thus, the
completely constant mapping of the tyre behavior is realized in all possible real
operating points.
HORIZONTAL DYNAMIC MODULE
The behaviour of a mechanical system can in very few cases be described in good
approximation by algebraic equations. In most cases, differential equations are to
be applied here; the observable states which the system passes through within a
period of time must therefore not be regarded as isolated operating points. They
are dependent on the prehistory and cannot be approximated by stationary measured
maps.

In the case of tyres, this is mainly due to transport processes; malfunctions or

suggestions have to pass through the "Latsch" in order to have a full impact. The
measurement of such phenomena is extremely time-consuming - even the
comprehensive stationary measurement of the tyre provides a large number of
independently reproducible maps. The additional variations of velocities of change
in the measurement parameters and their respective combinations not only provoce
problems due to inadequate equipment of the measuring instruments, but would also
place a disproportionate burden on a tyre model in terms of data management.
Therefore, one always follows the path of a more or less approximated, effective
analytical description of the physical behavior.

The relevant mode of action of IPG-Tire (horizontal behaviour as a result of time-

varying input parameters) will be demonstrated in the following.

Both the analytical derivation of complicated models and the selective observation
and extrapolation of the results lead here to a relatively simple differential equation,
which already achieves a considerably improved simulation compared to the
stationary calculation.

The influence of horizontal dynamics is explained using the example of delayed

lateral force build-up as a response to a change in the slip angle. Assuming low
deflections from the rest position, i.e. relatively low loading of the tyre in the
investigated direction transversely to the main plane, one arrives at the following
correlation

𝑆 + 𝑇𝑆̇ = 𝑆𝑠𝑡𝑎𝑡 (𝛼, 𝑃, 𝑠, … )

between the lateral forces S and the slip angle α.

This is a differential equation of the first order, which contains as free parameter
the unknown lateral guiding force S and its first temporal derivative S. A system
following such an equation is characterized by the term low-pass. The system
behaviour of a low pass corresponds to the frequency response shown below.

Low-pass behaviour of a tyre:

If there is a constant excitation - this is the right side of the equation - i.e. the slip
angle is constant, the system behaviour can easily be discussed. In this simplest case,
the stationary state, there is no time-dependent change in the system and the above
correlation degenerates to the stationary model equation:

𝑆 = 𝑆𝑠𝑡𝑎𝑡

the right side of which can be identified from stationary measurements.

The response of such a system to arbitrary stimuli can be represented analytically.

In particular, the response to the slip angle jump can be determined by the
exponential approach:

𝑆(𝑡) = 𝑆𝑠𝑡𝑎𝑡 (1 − 𝑒 −𝑡𝑇 )

The lateral force approaches the stationary state asymptotically.

Lateral force build-up in response to a slip angle step:

The high quality of the replica of the transient structure shows the response to
harmonic wheel load excitation. A harmonic fluctuation of the amplitude of the
static base load is superimposed on a constant wheel load, so that the vertical load
of the tyre varies between 0 and twice the static load. At a constant slip angle, the
lateral force also develops periodically, but strongly distorted. In particular, the
mean value of the lateral guiding force is strongly dependent on the wavelength L of
the wheel load excitation.

The calculation of this effect by IPG-Tire corresponds not only qualitatively but also
quantitatively with corresponding measurements. Such a realistic mapping of the
transient tyre behaviour is achieved by considering different dependencies between
the relaxation length and the load of the tyre. If the wheel load is disappearing, for
example, it must also go towards zero, since otherwise the low-pass filtering would
still deliver lateral forces even without traction, depending on the preceding load.
Lateral force reaction to harmonious wheel load excitation:

This short discussion should essentially convey the following insight: static and
dynamic behaviour can - at least in approximation - be observed and replicated
separately. In the above simple model only the time constant T has to be determined
to simulate the time behaviour, while the stationary behaviour has to be verified by
the usual measurements according to the right side of the equations.

Due to this consideration, the horizontal dynamics were considered in IPG-Tire in

this modular way. After the stationary approximation following the measurements,
the tyre forces S and U are manipulated according to the above differential equation.
Only the time constant T has to be known. A series of measurements as well as the
results of work on complex physical models suggest that these can be determined
from the wheel diameter and the instantaneous wheel rotational speed. It has been
shown that with the step response the stationary state is always reached after a
constant fraction of a wheel revolution. The corresponding wheel circumference
portion is called the relaxation length. This fraction must be specified in the tyre
data set. A standard value is about 0.5 for lateral forces and 0.05 for circumferential
forces.

IPG-Tire later automatically takes into account that this fraction is load-dependent,
which means, for example, that with increasing skew angle and the associated
sliding, the Latsch becomes smaller. If the constants mentioned are set to 0, the
model calculates purely statically.

The additional horizontal dynamic approach has resulted in a model whose stationary
behaviour can be optimised and whose time behaviour in the lower load range is
already well approaching the real conditions.

Due to the reduced order of the differential equation, it must be accepted that belt
vibrations cannot be considered - the figure is therefore additionally limited to the
frequency range below 40Hz. On the other hand, the advantageous possibility of
solving the tyre differential equations separately from the differential equation core
of the rest of the vehicle system was created by precisely this reduction, which
benefits the flexible use of the model in the most diverse environments.

„IPG-TIRE“ TYRE APPROXIMATION

Measured and calculated S(α,P) characteristic curves:

Measured and calculated Ub(s,P) characteristic curves:

Ua(s,P) characteristics determined by means of similarity transformation:

Measured and calculated M (α,P) characteristic curves:
Traction curves S(U), P=4kN:

Traction curves S(U), α=const.:

Traction curves S(U), s=const.:

Multidimensional approximation: lateral force as a function of slip angle and height

h of a water film on the road surface at constant wheel load:
FURTHER LITERATURE
 Schieschke, R. - RALPHS – ein effizientes Rechenmodell zur Ermittlung von
Reifenkräften auf physikaler Basis - Automobil-Industrie 4/86
 Schieschke, R.; Gnadler, R. - Modellbildung und Simulation von
Reifeneigenschaften - VDI-Berichte 650, VDI-Verlag Düsseldorf 1987
 Schieschke, R. Wurster, U. - IPG-Tire – ein umfassendes, effizientes
Reifenmodell zum Einsatz in Simulationsumgebungen Automobil-Industrie
5/88
 Schieschke, R. - Zur Relevanz der Reifendynamik in der Fahrzeugsimulation -
VDI-Berichte 778, VDI-Verlag Düsseldorf 1989
 Schieschke, R. - The Importance of Tire Dynamics in Vehicle Simulations - 9th
Annual Meeting and Conference on Tire Science and Technology -March 20-21,
1990 in Akron, Ohio
 Schieschke, R. - The Relevance of Tire Dynamics in Vehicle Simulations - XXIII
FISITA Congress - May 7-11, 1990 in Torino, Italy
 Schieschke, R. - Reifendynamik, Fahrzeugstabilität und Allradlenkung – eine
Unterschulung mit IPG-TIRE - Automobil-Industrie 3/91
 Schieschke, R.; Hiemenz, R. - The Decisive Role the Quality of Tyre
Approximation Plays in Vehicle Dynamics Simulations Proceeding 1st
International Colloquium on Tyre Models for Vehicle Dynamics Analysis held in
Delft, the Netherlands, October 21-22, 1991 Supplement to Vehicle System
Dynamics, Volume 21
 Schieschke, R. - Zusammenwirken von Reifendynamik und Allradlenkung zur
Verbesserung der Fahrzeugstabiltität - Tagung: Allradlenksysteme bei
Personalwagen - Haus der Technik, Essen, 3-4. Dezember 1991
 Wang, Y,Q.; Gnadler, R.; Schieschke, R. - Two-Dimensional Contact Area of a
Pneumatic Tire Subjected to a Lateral Force Vehicle System Dynamics to be
published
DIMENSION DATA

1 Tyre width
4 Height-width ratio
5 Tyre design
6 Rim diameter
7 Load Index (LI)
8 Speed symbol (GSY, also called "Speedindex")

For example: 195/65 R 15 91H

195 Tyre width 195 mm

65 Tyre flank 127 mm
15 Rim diameter 381 mm
2 * 127 + 381 Tyre diameters 635 mm
91 Load capacity index (LI) 618 kg
H Speed index 210 km/h

Speedindex [km/h]
 M 130
N 140
P 150
Q 160
R 170
S 180
T 190
U 200
H 210
V 240
W 270
Y 300
ZR > 240

LOAD-INDEX
The load index indicates how much load the
individual tyre can carry. The axle load of
the front and rear axles is noted in the
vehicle documents. The tyre should have a
higher load capacity than the axle load/2
(since two wheels are fitted per axle).

Load N Load N Load N Load N Load N

Index Index Index Index Index
50 1900 60 2500 70 3350 80 4500 90 6000
51 1950 61 2570 71 3450 81 4620 91 6150
52 2000 62 2650 72 3550 82 4750 92 6300
53 2060 63 2720 73 3650 83 4870 93 6500
54 2120 64 2800 74 3750 84 5000 94 6700
55 2180 65 2900 75 3870 85 5150 95 6900
56 2240 66 3000 76 4000 86 5300 96 7100
57 2300 67 3070 77 4120 87 5450 97 7300
58 2360 68 3150 78 4250 88 5600 98 7500
59 2430 69 3250 79 4370 89 5800 99 7750
100 8000
Calculation of the load capacity from the load index:
80
𝑇 [𝑘𝑔] = 45 ∙ √10𝐿𝐼
Influence of speed on load capacity:

The load capacity of V, W and Y tyres depends on the maximum vehicle speed, V
tyres up to 210 km/h, W tyres up to 240 km/h and Y tyres up to 270 km/h have the
maximum load capacity assigned to the tyre. At speeds above 210 km/h or 240 km/h
/ 270 km/h, the load capacity of the tyre decreases continuously, which is why
sufficiently dimensioned tyres should always be used.

SERIAL COLLISION (MODULE)

CHANGE OF VEHICLE VELOCITY

In rear-end collisions or serial collisions, there is usually a collision with a large
overlap. In addition, only collision speeds in the range of small speed changes (less
than 20 km/h) are of interest. In these accidents it is sufficient to start from the
equations for the straight central impact.

Especially with these small velocity changes, which are related to the threshold for
cervical spine injuries, it is particularly important to consider the frictional force of
the wheels during the rush hour. Even rough estimates of the forces exerted by
impact and tyre force show that neglect would lead to errors. The equations
therefore used for the principle of linear momentum are as follows:

Momentum before impact - momentum after impact = momentum change due to

frictional force at the wheels (force impact FI)

This force is effective during the compression time tKo and the restitution time tRe.

𝑚1 𝑣1 + 𝑚2 𝑣2 − (𝑚1 𝑣1′ + 𝑚2 𝑣2′ )

= (𝑚1 𝑎1 𝑡𝐾𝑜1 + 𝑚2 𝑎2 𝑡𝐾𝑜2 ) + (𝑚1 𝑎1 𝑡𝑅𝑒1 + 𝑚2 𝑎2 𝑡𝑅𝑒2 ) = 𝐹𝐼

In general, the times tKo + tRe can be assumed to be the same for both vehicles. Only
if a vehicle comes to a standstill or changes direction before the end of the collision
phase are case distinctions necessary.
Force - distance - deformation characteristics and idealisation assumed here:

If a linearization is carried out, there is a correlation between force and deformation

or restitution path analogous to an elastic spring. However, the structural stiffness
of the phases differs. Thus the following applies in general:

𝐹 = 𝑓 𝑠 = −𝑚 𝑎 = − 𝑚 𝑠̈

𝑓
=> 𝑠̈ + 𝑠=0
𝑚
𝑠̈ … 2. derivation of the distance after time

The solution of the differential equation is as follows:

𝜋
𝑠(𝑡) = 𝑠𝑚𝑎𝑥 sin (2 𝑡)
𝑇

𝜋 𝑓
2 =𝜔= √
𝑇 𝑚

In this model 2 springs encounter each other in an idealized way. The springs
themselves can easily be regarded as massless. They can therefore be regarded as a
single spring. The structural stiffness of this resulting spring is calculated as follows:

𝑓1 𝑓2
𝑓𝑟𝑒𝑠 =
𝑓1 + 𝑓2
This results in the following for ω:

2𝜋 𝑓𝑟𝑒𝑠
𝜔= = √
𝑇 𝑚𝑟𝑒𝑠

Whereby the reduced mass mres is defined by:

𝑚1 𝑚2
𝑚𝑟𝑒𝑠 =
𝑚1 + 𝑚2

T is the period of a complete vibration, which of course does not take place here.
The duration of the compression phase is ¼ of T. The duration of the restitution
phase can be calculated analogously.

In the impulse equation the four velocities are contained as unknown quantities.
With the law of conservation of energy and the definition of the impact factor two
further equations are available. A velocity must therefore always be known.

The law of conservation of energy for the problem at hand is as follows:

1
(𝑚1 𝑣12 + 𝑚2 𝑣22 − (𝑚1 𝑣12̇ + 𝑚2 𝑣22̇ )) = 𝐸𝐷 + 𝐸𝑅𝑒𝑖𝑏
2
The impact factor k is defined as follows:

Δ𝑣̇
𝑘=
Δ𝑣
Since Δ𝑣̇ is constant over a wide range and averages about 5 km/h (spread ± 3 km/h),
this value can be reckoned with well.

The friction work on the wheels depends on whether and to what extent the driver
actuates the brake pedal during the collision. The following applies for the
calculation:

𝑊𝐹𝑟𝑖𝑐𝑡 = 𝑚1 𝑎1 𝑠𝐾1 + 𝑚2 𝑎2 𝑠𝐾2 = 𝐸𝐹𝑟𝑖𝑐𝑡

The distances during the collision (sK) generally follow from:

1 2
1 2
𝑠𝐾 = 𝑣 𝑡𝐾𝑜 + (𝑎𝑚𝑎𝑥 − 2 𝑎) 𝑡𝐾𝑜 + 𝑣̇ 𝑡𝑅𝑒 + (2 𝑎̇ − 𝑎𝑚𝑎𝑥 ) 𝑡𝑅𝑒
6 6
Acceleration-time diagram:
a

amax

a'

Ko Re
Pre Crash Crash Post Crash t

A determination equation für 𝑣1̇ can now be derived:

−𝐾𝑆 + 𝑚2 Δ𝑣̇ + (𝑚1 + 𝑚2 )𝑣1 + 𝑚2 Δ𝑣

𝑣1̇ =
𝑚1 + 𝑚2

The equations are now inserted into the equation of the law of conservation of
energy, resulting in a determination equation for Δv:

1 𝐶32 − 𝐶42
Δ𝑣 = 𝐶1 + √𝐶12 + Δ𝑣 2̇ + 2 Δ𝑣̇ (𝑎1 𝑡𝑅𝑒1 − 𝑎2 𝑡𝑅𝑒2 ) + 𝐶 +
𝑚𝑟𝑒𝑠 2 𝑚1 𝑚2

With:

𝐶1 = 𝑎2 𝑡𝐾𝑜2 − 𝑎1 𝑡𝐾𝑜1

1 2 2
𝐶2 = (𝑚1 𝑎1 ((𝑎𝑚𝑎𝑥1 − 2 𝑎1 )𝑡𝐾𝑜1 + (2 𝑎1̇ − 𝑎𝑚𝑎𝑥1 )𝑡𝑅𝑒2 )) + 2 𝐸𝐷
3
𝐶3 = 𝑚1 𝑎1 𝑡𝐾𝑜1 + 𝑚2 𝑎2 𝑡𝐾𝑜2

𝐶4 = 𝑚1 𝑎1 𝑡𝑅𝑒1 + 𝑚2 𝑎2 𝑡𝑅𝑒2

The change in velocity of the two vehicles during the collision can be calculated by
inserting:

𝑚1 𝐶3 + 𝐶4
Δ𝑣1 = (Δ𝑣 + Δ𝑣̇ − )
𝑚1 + 𝑚2 𝑚2

Δ𝑣2 = Δ𝑣1 − Δ𝑣̇ + Δ𝑣

With the described systems of equations the velocity changes caused by collisions
can be calculated. The following must be known:
 Iputs: Available from:

One of the four velocities before and For example, a velocity can be used
after the collision from information from participants. A
velocity can also be determined, e.g.
from skid marks after the collision.
Δ𝑣̇ From mean value and/or wingspan
5 ± 3 km/h
Brake decelerations during and after Based on empirical values or
collision information from the parties involved
Remaining deformations and structural Measurement of the permanent
stiffness. Alternatively, EES values can deformation in the direction of the
be specified. For both vehicles at least shock drive
1 size must be specified. Structural stiffness from literature.
Evaluation of the damage patterns.
The following approximate relationship applies between Δ𝑣, 𝑘 𝑎𝑛𝑑 Δ𝑣̇ :

Δ𝑣 (km/h) 𝑘 Δ𝑣̇ (km/h)

0 1 0
10 0,65 6,5
20 0,32 6,4
30 0,22 6,6
40 0,16 6,4
50 0,11 5,5
60 0,09 5,4
70 0,08 5,6
80 0,07 5,6
90 0,06 5,4
In the range between 2 and 20 km/h:

k v'(km/h)
v (km/h) min medium max min medium max
2 0,99 1 1 2,0 2,0 2,0
4 0,70 0,97 1 2,8 3,9 4,0
6 0,55 0,90 0,98 3,3 5,4 5,9
8 0,43 0,66 0,95 3,4 5,3 7,6
10 0,35 0,54 0,85 3,5 5,4 8,5
12 0,28 0,45 0,72 3,4 5,4 8,6
14 0,23 0,39 0,63 3,2 5,5 8,8
16 0,20 0,33 0,54 3,2 5,3 8,6
18 0,17 0,29 0,48 3,1 5,2 8,6
20 0,14 0,26 0,45 2,8 5,2 9,0
CALCULATION OF PASSENGER ACCELERATION DUE TO COLLISIONS
The collision-related force acting on a person sitting in a vehicle is primarily
dependent on the acceleration of the vehicle, and in addition to the posture, the
seat construction also has a significant influence.

In principle, the system vehicle - seat - passenger can be idealized as a system

consisting of 3 springs (spring model):

f1 Spring in the rear (impact area) of the vehicle.

f2 Spring in the backrest.
f3 Spring in the upholstery or initial structural rigidity.
fres Currently effective structural stiffness in the backrest area -
upholstery.
m2 Mass of the passenger's upper body, i.e. the relevant proportion of
the mass of the person pressed against the seat.
m1 Total vehicle mass less m2.

The deformation zone in the rear of the vehicle is simulated by spring 1 and in the
backrest of the vehicle seat by springs 2 and 3. Spring 2 simulates the seat
construction and spring 3 takes into account the influence of the upholstery. Springs
2 and 3 are arranged in series.

The mass m2 is only connected to the vehicle via spring 2 and spring 3, f2 and f3 must
be adapted to the body height.

The deformation of the backrest (bending backwards) depends on the effective

torque, i.e. on the product force multiplied by the distance of the force from the
pivot point (force arm). The smaller the body height, the smaller this distance is and
thus the harder the backrest appears. This is taken into account by replacing the
structural stiffness of the backrest by:

𝐿
𝑓𝑟𝑒𝑠 => 𝑓𝑟𝑒𝑠
𝐹𝑆 (𝑆𝐻 + 𝑆𝐹)

SH Seat - Shoulder height

SF Distance seat surface - pivot point of the backrest (= 0 - 5 cm)
FS The SH in which the force center of gravity is assumed (75% of SH)
L In the backrest characteristic curve used, the force - deformation
relationship is measured at a distance of 49 cm from the pivot point.
This makes it possible to use the model of a linear spring instead of
a torsion and the torsion module, which is difficult to estimate. The
 value 49 cm results from the characteristic curve from which the
values of the presetting were taken.

It generally follows from the seat construction that the upholstery is deformed first
and only slightly the backrest. In this phase, the currently effective structural
stiffness fres has the value f3. This is actually the result of two springs connected one
after the other.

Once the maximum upholstery deformation has been reached, only the backrest
frame can still be deformed. From this point, the momentarily acting structural
stiffness (fres) assumes the value of f2 until the elastic limit of the backrest is reached
and fres decreases to usually about 1/3 (input in % possible) of the value of f 2.

The elastic limit is usually reached at 75 % (this value can be changed in the mask)
of the force limit. From the force limit fres assumes the value 0, i.e. the force remains
constant.

If a collision of a moving vehicle against the rear occurs, first the deformation of
spring 1 takes place and then the acceleration of the vehicle due to the acting spring
force. The deformation of spring 2 and spring 3 occurs due to the forward movement
of the seat which is moved with the passenger cell.

Because the velocity of the passenger cell is initially lower than that of the rear, the
deformation of spring 2 and spring 3 is also initially smaller and therefore the
acceleration of m2 as well. If the seat back is not extremely hard, its maximum
deformation is reached much later and thus the person in the seat also reaches the
maximum acceleration later.

The simulation of the seat with 2 springs connected in series offers the possibility of
limiting the cushion deformation or the flatter part of the structural stiffness. When
the initial structural stiffness is reached, usually when the upholstery is completely
compressed, spring 2 alone must be used for further calculations, so a much harder
spring comes into play.

Mathematically the problem can be formulated with differential equations:

𝑚1 𝑥̈ 1 = 𝑓1 𝑑𝑒𝑓1 + 𝑚2 𝑥̈ 2 + 𝑎𝑅𝑒𝑖𝑏 (𝑚1 + 𝑚2 )

𝑚2 𝑥̈ 2 = 𝑓𝑟𝑒𝑠 𝑑𝑒𝑓2 + 𝜏 𝑥̇ 2

𝑎𝐹𝑟𝑖𝑐𝑡 (𝑚1 + 𝑚2 ) Corresponds to the friction of the tyres on the road

𝑎𝐹𝑟𝑖𝑐𝑡 Brake deceleration during collision
𝑥1 Local coordinate of the centre of gravity of the vehicle
𝑥̈ 1 Acceleration of the vehicle
𝑑𝑒𝑓1 Vehicle deformation
𝑥2 Location coordinate of the centre of gravity of the fuselage
𝑥̇ 2 Velocity of the fuselage relative to the vehicle
𝑥̈ 2 Acceleration of the fuselage
 𝑑𝑒𝑓2 Back deformation
𝑓𝑟𝑒𝑠 Currently predominant spring constant of the system
Upholstery - backrest
𝜏 Damping factor

Physically, the equation corresponds to the equilibrium of forces prevailing at any

point in time :

Spring force + Friction = Inertial force

The mathematical solution to the above problem becomes simpler if, instead of a
vehicle colliding from behind, the vehicle model reverses against a fixed obstacle,
or, equivalently, if this obstacle is moved against the stationary vehicle at a constant
speed.

Only the structural stiffness of spring 1 and the impact velocity have to be adjusted
so that the same vehicle acceleration and deformation results.

For the resulting structural stiffness of the two vehicles, the following then applies:

𝑓𝑉𝑒ℎ1 𝑓𝑉𝑒ℎ2
𝑓1 =
𝑓𝑉𝑒ℎ1 + 𝑓𝑉𝑒ℎ2

This value has yet to be corrected, as only 1 vehicle is considered. One uses the
approach that the common spring is divided into two parts. The spring that is valid
for this one vehicle is an apparently shorter spring, the other part of the spring
belongs to the other vehicle. The lengths of these partial springs behave the other
way round as the masses. The division corresponds to the centre of mass. From this
follows:
𝑚2
𝑓1 => 𝑓1 (1 + )
𝑚1

For the aliquot deformation part of the common spring of both vehicles (s 1) the
following applies:
𝑚2
𝑠1 = 𝑠
𝑚1 + 𝑚2

s… Sum of the dynamic deformation of both vehicles.

Solution of the equation of vibration with friction:

𝑎𝑚
𝑥 = 𝑥0 sin(𝜔 𝑡) +
𝑡
The following applies to the velocity of an elastic spring:

𝑥̇ = 𝑥0 cos(𝜔𝑡)
x0 … Amplitude of the oscillation

The following applies at the time of maximum deformation:

𝑎𝑚
𝑠 = 𝑥0 +
𝑓

And further:

𝑚1 𝑥̈ 1 = −𝑓1 (𝑥1 − 𝑣 𝑡) + 𝑚2 𝑥̈ 2 + (𝑚1 + 𝑚2 ) 𝑎

𝑚2 𝑥̈ 2 = −𝑓2 (𝑥2 − 𝑥1 ) + 𝜏 𝑥̇ 2

𝜏 𝑥̇ 2 … Damping term of the backrest or upholstery

This system of equations does not offer a closed solution and must therefore be
numerically integrated. One possible solution is a derived Runge-Kutta method. The
function is approximated by a lineament, i.e. it is divided into a large number of
lines (intervals). The velocity in the interval i can be expressed from the previous
and the following position as follows:
𝑥𝑖+1 − 𝑥𝑖−1
2 2
𝑥̇ 𝑖 =
Δ𝑡
Or:
𝑥𝑖+1 − 𝑥𝑖−1
𝑥̇ 𝑖 =
2Δ𝑡
i + ½ or i - ½ means the middle of the following or preceding interval. The analog
equation is valid for acceleration and can be transformed into:

𝑥𝑖+1 − 2 𝑥𝑖 + 𝑥𝑖−1
𝑥̈ 𝑖 =
(Δ 𝑡)²

Or:

𝑥𝑖 − 2 𝑥𝑖−1 + 𝑥𝑖−2
𝑥̈ 𝑖−1 =
(Δ 𝑡)²

It follows:

(Δ𝑡)2 (𝑓1 𝑣 (𝑡 − Δ𝑡) − 𝑎 (𝑚1 + 𝑚2 ) + 𝑓𝑟𝑒𝑠 (𝑥1𝑖−1 − 𝑥2𝑖−1 ))

𝑥1𝑖 = 2 𝑥1𝑖−1 − 𝑥1𝑖−2 +
𝑓
𝑚1 (1 + 𝑚1 (Δ𝑡)2 )
1

The error can be kept small by selecting a correspondingly small increment t. A
step size of t = 0.0002 s should normally suffice.

It should be noted that the springs cannot be assumed to be elastic. The integration
must therefore essentially be carried out in 3 steps:
 1. First, a possible distance from the backrest must be taken into account. In
this case the integration has to be carried out by the following equation until
the distance becomes 0.
𝑚1 𝑥̈ 1 = −𝑓1 𝑥1

2. The system is then integrated. With spring 1, a small back deformation is to

be assumed, so that due to the backward movement the speed exceeds the
starting value v. Subsequently, the vehicle continues to travel with a constant
deceleration, so that mass 2 collides with the passenger cell, which is now
slowing down, under further deformation of springs 2 and 3. The integration
must continue until the vehicle deformation (including re-deformation) is
complete..

3. Since now mass 1 is no longer accelerated by spring 1, only integration

continues until mass 2 also reaches acceleration 0.

The maximum acceleration (deceleration) occurs when the speed of the vehicle
reaches 0. If the calculation of the passenger acceleration is carried out according
to the above procedure, the spring constant of the vehicle may not simply be used
at f1, but must be corrected accordingly (slightly), since here the total mass does
not take effect immediately, but only m1 - i.e. the total mass reduced by m2 - is
used.

𝑓𝑟𝑒𝑠
𝑚1 + 𝑚
𝑓1𝑎𝑙𝑡 2
𝑓1 => 𝑓1 = 𝑎𝑚𝑎𝑥
𝑠𝑑𝑦𝑛
𝑓𝑟𝑒𝑠
The term considers, that during collision via spring 2 also part of mass m1
𝑓1𝑎𝑙𝑡

becomes effective. The diagram below shows the backrest curve used for the preset
values. This curve does not include upholstery.
RELATIONSHIP BETWEEN EES, PERMANENT DEFORMATION AND
STRUCTURAL STIFFNESS, CALCULATION OF COLLISION DURATION

STRUCTURAL STIFFNESS IN THE CONTACT ZONE

When two vehicles collide, the impact forces act on the one hand, idealized by
springs, and on the other hand inertia force and friction force of the wheels.

Idealized mathematical model of a collision:

x
x1 x2

1 2

Explanation of terms:

Springs usually have a linear behaviour (F = c • x) where F is the compression force,

c is the spring constant and x is the deformation distance. This law is also used as
an approximation for the structural stiffness of vehicles.
When determining the structural stiffness of vehicles in the contact zone, it
depends on where the contact zone lies and how large the area expansion is. The
position of the surface on the vehicle and its formation are decisive here. The
three identical surfaces on the following vehicle result in different indentations
with the same contact force, which results in different structural stiffnesses.

While a spring usually has a linear force-displacement relationship, this is usually not
the case with a vehicle structure. Depending on the depth of the indentation, there
is a different approximate slope of the curve F(s) and thus a different structural
stiffness. To a small extent, the force-distance relationship is also dependent on the
deformation velocity, i.e. on the collision velocity. From this it can be seen that
even the loading of the same vehicle structure can produce different average
structural stiffnesses depending on the penetration depth.

The mass of the vehicles affects the structural stiffness in several ways. The vehicle
manufacturers strive to achieve a very specific deceleration for the passengers in
the event of the legally prescribed wall impact with 100% overlap. This has led to
similar deformation distances in small cars as in luxury cars. The heavy luxury cars
must therefore have greater structural stiffness than the smaller, lighter vehicles.
Example of permanent deformation in mm:

From the example it follows that in the case of a wall impact with 100% overlap and
good approximation it could be said that 10 cm permanent deformation corresponds
to 10 km/h impact velocity. If such a car is deformed at the entire front by 50 cm,
then the impact velocity was about 50 km/h. Even with rust-weakened vehicles, this
correlation is largely retained.

Modern vehicles have special properties with regard to internal safety, vehicle
structure and structural rigidity. For example, miniclass vehicles are manufactured
with a particularly rigid body to prevent the interior from deforming. The braking
distance (formerly the crumple zone) for the passengers is transferred to the
interior. Large vehicles should deform more easily so that this results in more braking
distance for the smaller vehicles. So far, there are no uniform rules or procedures
for designing the structural stiffness of vehicles over their entire circumference.

CALCULATION FORMULAS
For the penetration of the vehicles it is assumed that at the point of impact at each
vehicle there is a linear relationship between force and distance. Damping is not
taken into account, which means that the collision duration tends to be too short.
After the vehicles have reached their maximum penetration, there is still a low
springback.
The collision duration corresponds approximately to twice the value of the
penetration duration or half of a complete oscillation period. The duration until the
maximum penetration is reached is calculated from a quarter of the period of the
complete oscillation. The period duration is:

𝑚
𝑇 =2𝜋√
𝑐

This results in the following for the collision duration:

𝑇 𝑚
𝑡𝑘 = 2 =𝜋√
4 𝑐

In this formula, m is the mass of a body vibrating on a spring with stiffness c. m and
c cannot be used directly. Firstly, the vehicle mass must be reduced to the impact
normal, and secondly, it must be taken into account that two vehicles are involved.
This would be comparable to two masses located at the ends of a spring.

Spring model:
The common spring stiffness results from the series connection of two springs of
different stiffness as follows. A collision situation should be considered, as shown in
the following figure:

One can see from the picture that the springs act in the direction of the contact
normal. This has several meanings:

 The vehicle mass acts completely on the spring in vehicle 1, but only partially
in vehicle 2. This means that the vehicle mass must be reduced to the contact
normal.
 The collision duration will only be calculated correctly if it is an impact
without slipping. In the case of a collision with slipping, the spring stiffness is
calculated correctly from the penetration in normal direction within the
framework of idealisation, but not the duration of the collision. The value GEV
is defined as the characteristic value for slip. If this drops below 0.85, the
collision duration is no longer calculated; it must be determined by the user
from the simulation by observing how long the vehicles remain in contact with
each other.

The spring constants (structural stiffness) c of the two vehicles must be calculated.
The permanent penetration depths, which are derived from the damage patterns of
the two vehicles, are available for this purpose. These permanent penetration depths
must be determined in the normal direction, which happens automatically in
AnalyzerPro if the vehicles have been correctly brought into collision position
according to their damage. The calculated EES values are also available. Thus the
structural stiffnesses can be calculated from these values. In a new reference
system, which is installed in the point of contact (centre of mass), i.e. between
spring 1 and spring 2, the problem can be examined in an idealized way.
Mass proportional re-deformation:

The basic idea is that the springback behaviour is largely proportional to the mass.
This would have the consequence that the contact surface in the coordinate system
moved with the center of mass remains relatively calm.

As a general rule, the spring forces in the contact area must be the same at all times
(actio = reactio). Thus, the reaction forces (inertia force and tyre force) at the other
ends of the springs must be equally big, opposite and equal to the spring forces. The
following applies to the amounts:

𝑐1 𝑠𝑑𝑦𝑛1 = 𝑐2 𝑠𝑑𝑦𝑛2 = 𝑚1 (𝑎𝑚𝑎𝑥1 + 𝑎1 ) = 𝑚2 (𝑎𝑚𝑎𝑥2 − 𝑎2 )

The inertia and friction forces act in the opposite direction in the case of the vehicle
bumping into the front vehicle (vehicle 2) and in the same direction in the case of
vehicle 1. The respective structural stiffness c = F/sdyn can be expressed by:
𝑚
𝑐= (𝑎𝑚𝑎𝑥 ± 𝑎)
𝑠𝑑𝑦𝑛

The deformation work from the permanent deformation follows from the area
enclosed by the spring characteristic curves:

FRe FD
s

1 1 1
𝐸𝐷 = 𝑐 𝑠𝑑𝑦𝑛 𝑠𝐷 = 𝑚 (𝑎𝑚𝑎𝑥 ± 𝑎)𝑠𝐷 = 𝑚 𝐸𝐸𝑆 2
2 2 2
This results in the maximum acceleration for each vehicle at the end of the
compression phase:

𝐸𝐸𝑆12
𝑎𝑚𝑎𝑥1 = − 𝑎1
𝑠𝐷1

𝐸𝐸𝑆22
𝑎𝑚𝑎𝑥2 = + 𝑎2
𝑠𝐷2

And:

𝑠𝐷1 𝑚1 𝐸𝐸𝑆12
=
𝑠𝐷2 𝑚2 𝐸𝐸𝑆22
The structural stiffness could be calculated from the above equation if the dynamic
deformation were known. The deviation between dynamic and permanent
deformation can be considerable, especially at low collision speeds, and must not be
neglected under any circumstances.

The calculation of the dynamic deformation is based on the area enclosed in the
figure above. For the derivation of the equation a coordinate system attached in the
common mass center is used. This has the advantage that the following can be said:

 The kinetic energy of the system is zero at the end of the compression phase.
 The kinetic energy at the end of the restitution phase corresponds to the area
under the spring characteristic curve of the restitution phase.
 The following equations apply, whereby the letter u is used for the velocities
in the new reference system

Velocity of the centre of mass:

𝑚1 𝑣1 + 𝑚2 𝑣2
𝑣𝑚 =
𝑚1 + 𝑚2

Centre-of-mass velocities of vehicles:

𝑢1 = 𝑣1 − 𝑣𝑚 𝑢2 = 𝑣2 − 𝑣𝑚
𝑢̇ 1 = 𝑣̇ 1 − 𝑣𝑚 𝑢̇ 2 = 𝑣̇ 2 − 𝑣𝑚
The velocity differences remain unchanged.

If the law of conservation of energy is applied in such a system, the total kinetic
energy is converted into deformation during compression and is then 0 at the end of
the compression phase. Energy then flows back into the system again during the re-
deformation.

The velocity difference of the two bodies at the end of the restitution phase
corresponds to the separation velocity of the contact points in normal direction
calculated from the collision analysis (𝑣𝐵𝑛1
′ ′
− 𝑣𝐵𝑛2 ). Thus the law of conservation of
energy for the restitution phase and the case that vehicle 1 is the pushed is:

1 1 ′ 2
𝐸1 = 𝑐1 𝑠𝑑𝑦𝑛1 (𝑠𝑑𝑦𝑛1 − 𝑠𝐷1 ) = 𝑚1 𝑣𝐵𝑛1
2 2
This results in the equation for determining the structural stiffness of both vehicles.

MASS PROPORTIONAL RE-DEFORMATION

If the vehicles behave approximately in proportion to the mass during the restitution
phase, the formula is:
 (𝑚𝑟𝑒𝑑1 + 𝑚𝑟𝑒𝑑2 )2 𝑚12 𝐸𝐸𝑆14
𝑐1 = 2 2 ′ 2
𝑠𝐷1 (𝑚𝑟𝑒𝑑1 𝑚𝑟𝑒𝑑2 Δ𝑣𝐵𝑛 + (𝑚𝑟𝑒𝑑1 + 𝑚𝑟𝑒𝑑2 )2 𝑚1 𝐸𝐸𝑆12 )

The structural stiffness of vehicle 2 can be calculated by exchanging the indices in

the equation above.

Applying the formula means that of the variables v' and the six "structural
variables", namely EES1, EES2, SD1, SD2, c1, c2, only three must be given in addition to
v', the remaining two can be calculated.

NOT MASS PROPORTIONAL RE-DEFORMATION

Linear characteristic curve:

1
𝑐1 =
𝑠
(𝑚𝑟𝑒𝑠 Δ𝑣 ′2 + 𝑚1 𝐸𝐸𝑆12 (1 + 𝑠𝐷2 )) 𝑠𝐷1
2
𝐷1 1
−𝑐
𝑚12 𝐸𝐸𝑆14 2

Applying the formula means that only two of the variables v' and the six "structural
variables" need to be given in addition to v', the remaining two can be calculated.
However, a structural stiffness must be given. It is interesting, however, that due to
the constraints that the dynamic deformation cannot be smaller than the permanent
deformation, a relatively small possible range results for the possible structural
stiffness. The smaller v' is, the smaller this range is.

Definition of a structure with nonlinear characteristic curve:

In cases where very different impact partners are present, for example in crash tests
against a rigid barrier, it will be necessary to define a non-linear characteristic curve
for a vehicle. In this case, however, it is not always possible to assume a mass-
proportional re-deformation behaviour.

In some crash tests, an acceleration (force) time curve was measured, where it is
noticeable that the duration until reaching the maximum force is only slightly longer
than the duration from the maximum to reaching the value 0. Since the extent of
the damage can be assumed to be largely plastic, the end of the compression phase
may not have been reached at the highest point. The area under the acceleration-
time curve corresponds to the velocity change. In a plastic collision, the velocity
change during the compression phase is greater than during the restitution phase.
The restitution phase therefore begins towards the end of the acceleration-time
curve, i.e. at a point where the maximum has already been exceeded. This can only
be explained by an area of the force-deformation curve with a negative slope starting
at the maximum.
While vehicle 1 behaves "normally", it is assumed that at the end of the compression
phase vehicle 2 enters an area where the structure collapses. It seems largely
unlikely that exactly the same force will occur in the collision partner. For this reason
one can assume that the force increase of this vehicle can be described by a linear
function. Also with this vehicle it can come during the deformation to a partial
collapse of the momentarily involved structure. In the course of the further
deformation, however, a fixed structure will again be achieved. The linear function
represents the average course as an approximation.

Deformation:

At the end of the compression phase of a vehicle, an area can be reached which can
only be approximated inaccurately by a linear function.

For a short phase, the force can remain approximately constant (µ 2s2) and then the
structure can become unstable. This means that the force-distance curve gets a
negative slope (decrease of the curve up to k2F). k2F is the force at the end of the
compression phase. During the section with constant force at vehicle 2, vehicle 1
remains at the highest point of the force, then vehicle 1 undergoes a re-deformation
up to k2F, while vehicle 2 continues to deform at the same time. Since this takes
place at the expense of vehicle 2, the energy released in vehicle 1 is not converted
into kinetic energy, but used for the deformation energy of vehicle 2. Therefore, the
deformation up to k2F has no effect on the kinetic energy after the collision ( velocity
difference of the contact points). Therefore, only the energy corresponding to the
area of the small right-angled triangle below k2F flows back into the system.
Force:

The figure shows the course of the force as a function of the total deformation. The
total maximum deformation is slightly smaller than the sum of the maximum
deformation of each individual vehicle.

The structural stiffness of vehicle 2 must be defined. The increase from the zero
point to the upper left corner of the deformation figure would be possible. This would
correspond to the average structural stiffness. However, this definition does not
correspond to the increase in force to the maximum of the force. The recommended
definition is therefore the increase at the front of the curve. In the formulas below,
c2 corresponds to this definition. The difference lies in the factor d2 defined below.

On the basis of these considerations, the following relationships can be established:

𝐴 = 𝑚𝑟𝑒𝑠 Δ𝑣 ′2 𝑑2 = 1 − 𝜆2 − 𝜇2
𝑏2 = 1 + 𝜇2 + 𝜆2 𝑘2 − 𝑘2 𝑔2 = 1 + 𝜇2 + 𝜆2 𝑘2
The following shall apply:

1 1
𝐸𝐷1 = 𝑐1 𝑠1 𝑙1 = 𝑚1 𝐸𝐸𝑆12
2 2
𝐸1 = 𝑐1 𝑠1 𝑙1

1 1
𝐸𝐷2 = 𝑐2 𝑠2 𝑑2 (𝑘2 𝑙2 + 𝑏2 𝑠2 ) = 𝑚2 𝐸𝐸𝑆22
2 2
𝐸2 = 𝑐2 𝑠2 𝑑2 (𝑘2 𝑙2 + 𝑏2 𝑠2 )

With the equilibrium of forces:

𝑐2 𝑑2 𝑠2 = 𝑐1 𝑠1

𝐴 = 𝑘2 𝑐1 𝑠1 (𝑠2 − 𝑙2 + 𝑘2 (𝑠1 − 𝑙1 )
It follows:

𝐸1 𝑚1 𝐸𝐸𝑆12 𝑙1
= 2 =
𝐸2 𝑚2 𝐸𝐸𝑆2 𝑘2 𝑙2 + 𝑏2 𝑠2

In addition, the following must apply:

𝑠1 > 𝑙1 𝑢𝑛𝑑 𝑠2 > 𝑙2

This results in the condition:

𝐸1 𝑙2 (𝑘2 + 𝑏2 ) 𝐸1 𝑔2 𝑙2 𝑘2 + 𝐴 𝑙1 𝑏2
≤ 𝐸2 ≤
𝑙1 𝑙1 (𝑔2 − 𝑏2 )

With the help of equations 1 to 4, an equation can be established in which s 1 and s2

are eliminated, plus one of the six variables c 1, l1, E1, c2, l2, E2. From this equation
one can be calculated from four known variables. If l1, E1, l2, E2 are given, then a
check has to be made. If E2 is set to the minimum value, there is no re-deformation
for vehicle 2. If the largest possible value is set, there is no re-deformation for
vehicle 1.

For example, E2 can be eliminated:

𝐸1 𝐸1 𝐸1
𝐴 = 𝑘2 ( − 𝑙2 + 𝑘2 ( − 𝑙1 ))
𝑙1 𝑐2 𝑙1 𝑑2 𝑐1 𝑙1

Or c2:

𝐸1 𝑑2 𝑙1 𝐸2 − 𝑘2 𝑙2 𝐸1 𝐸1
𝐴 = 𝑘2 ( − 𝑙2 + 𝑘2 ( − 𝑙1 ))
𝑙1 𝐸1 𝑏2 𝑐1 𝑙1

If this formula set is used in conjunction with the momentum method, it must be
taken into account that the momentum method supplies the total deformation
energy ED. E1 + E2 = 2 ED is therefore available as an equation. In addition to the
deformations, there must also be structural stiffness. Here, too, only a relatively
small range is possible for the structural stiffness.

Collision duration:

In the model used here, two massless springs meet. The forces acting at the ends of
the springs are the same, so the two springs can be considered as one spring with a
resulting spring stiffness.
One can see from the picture that the springs act in the direction of the contact
normal. This has several meanings: The vehicle mass acts completely on the spring
in vehicle 1, but only partially in vehicle 2. This means that the vehicle masses must
be reduced to the contact normal. The vehicle mass is replaced by its reduced mass
in each case.

Calculation of the reduced mass:

𝑖2
𝑚𝑟𝑒𝑑 = 𝑚
𝑖 2 𝑎²
m… Mass of the vehicle

a… Distance of the centre of gravity from the shock drive

𝐽
i… Radius of gyration: 𝑖 = √𝑚

J… Moment of inertia

Further considerations again are based on a coordinate system, which is located at

contact point (= same time mass center of system of both springs). The centre of
mass should be at rest during the impact and the vehicles collide with each other
with their reduced masses. The position of the centre of mass shall be as follows:
𝑚1 𝑙1 = 𝑚2 𝑙2 . If the centre of mass remains at rest, then one can think of a rigid wall
in its place, against which each vehicle hits on its own. However, the single vehicle
has a spare mass and a spare spring with greater spring stiffness. For this situation,
for example, the following applies to vehicle 1:

𝑙1 + 𝑙2 𝑚𝑟𝑒𝑑1 + 𝑚𝑟𝑒𝑑2
𝑐𝐸1 = 𝑐𝑟𝑒𝑠 = 𝑐𝑟𝑒𝑠
𝑙1 𝑚𝑟𝑒𝑑2

The equation for the other spring is obtained by swapping the indices. The collision
duration can now be calculated:
 𝑚𝑟𝑒𝑠
𝑡𝐾 = 𝜋 √
𝑐𝑟𝑒𝑠

With:

𝑚𝑟𝑒𝑑1 + 𝑚𝑟𝑒𝑑2
𝑚𝑟𝑒𝑠 =
𝑚𝑟𝑒𝑑1 𝑚𝑟𝑒𝑑2

An alternative for the calculation of the impact duration is given by the following
formula:

𝜋 𝑠𝑑𝑦𝑛1 + 𝑠𝑑𝑦𝑛2 𝑠𝑑𝑦𝑛1 − 𝑠𝐷1 + 𝑠𝑑𝑦𝑛2 − 𝑠𝐷2

𝑡𝑘 = ( + ′ ′ )
2 |𝑣𝐵1 − 𝑣𝐵2 | |𝑣𝐵𝑛1 − 𝑣𝐵𝑛2 |

vBn or v'Bn are the values of the contact point velocities in the direction of the impact
normal before or after the collision. The above formula also follows from the spring
model and also provides the same value in the case of a linear characteristic.
However, it can also be used in the case of a self-defined structure. The factor π/2
is a consequence of the linear characteristic and will probably have to be increased
in the non-linear case. If the factor π/2 is used, this means a change of the velocity
according to a cosine function: at the beginning a little less, towards the end a little
more. If the factor 2 is used, then this means a change of the speed according to a
linear function - thus an steady change.

The collision duration will only be calculated correctly if it is a collision without

slipping. For a collision with slip, the structural stiffness is calculated correctly from
the penetration in the normal direction within the framework of idealization, but
not the duration of the collision.

In AnalyzerPro, the calculation is more precise, the time is calculated separately for
each area with different behavior. For the area with a negative slope of the force-
deformation curve, an iterative calculation method must be used. No closed solution
for time can be derived from the given distance and the acceleration as a function
of the distance. If a collision duration could be calculated, the mean center of gravity
acceleration can also be calculated. The maximum acceleration of the center of
gravity can be assumed to be about twice as large.