Addis Ababa University

Addis Ababa institute of Technology

School of Electrical and Computer Engineering

A Thesis proposal for partial fulfillment of Masters of Science

in Control Engineering
By

Begziew getnet
ID: GSE/8108/11
Overview of swarming algorithm of UAV: applied for inspection and
environmental monitoring
Advisor: Lebsework Negash (Ph.D.)

Signature -------------

01/28/2022
 Abstract
Unmanned Aerial Vehicles (UAV) can be used for different purposes because of their
significantly reduce the risk of human life while accomplishing important hazardous missions.
This thesis presents swarm behavior of unmanned aerial vehicles because of its key feature of
decentralized processing amongst its member. First, the thesis will review the basics of drone
swarm intelligent technology and then looking in deep to investigate the swarming behaviors,
intelligence and coordination for unmanned aerial vehicle quadcopters and simulate different
swarm patterns.
Swarming behaviors of multiple vehicles which have a variable of interest by obtaining the
mathematical model, it is possible to achieve a certain working mechanism and rules for
cooperative behavior. When the value of variable interest for this multiple vehicles agree on a
specific value it is called consensus algorithm.
This paper proposes distributed consensus of multi agent UAV control by combining a particle
swarm optimization algorithm and an average consensus algorithm.

Key words: Swarm drones, Distributed consensus algorithm, unmanned aerial vehicle
 Table Contents
1. Introduction.........................................................................................................................................1
1.1 Background..................................................................................................................................1
1.2 Problem statement.......................................................................................................................2
1.3 Objective......................................................................................................................................3
1.3.1 General objective.................................................................................................................3
1.3.2 Specific objective.................................................................................................................3
2. Methodology.......................................................................................................................................3
2.1 Related work................................................................................................................................3
2.2 Scope...........................................................................................................................................5
2.3 Research Methodology................................................................................................................5
3 Schedule..............................................................................................................................................6
4 Budget.................................................................................................................................................6
5 References...........................................................................................................................................7
 1. Introduction
1.1 Background
An unmanned aerial vehicle (UAV) can be also called a drone is a type of aircraft without a
human pilot on board. UAVs are type of an unmanned aircraft system which include vehicle, a
ground-based controller, and a system of communication link between the two [7].

Drones can be classified on a different basis. Based on the type of aerial platform used, they can
be classified into four major groups.
1. Multi Rotor Drones
2. Fixed Wing Drones
3. Single Rotor Helicopter
4. Fixed Wing Hybrid VTOL

In nature different animals and insects form different coordinated shapes of clusters, such as
school of fishes, pack of wolves, flocks of birds, colonies of bees, packs of wolf etc. Members of
the group can communicate each other through the exchange of internal information in the
system, to achieve the working mechanism of external rules and orderly cooperative behavior
[4]. Large number of drones flying in coordinated process to perform coordinated tasks is known
as drone swarm. In other words, swarming mean a set of individual agents possessing
independent individual dynamics but exhibiting intimately coupled behaviors and collectively
performing a specific task. There are many engineering swarm Examples including autonomous
ground, air, underwater or surface vehicles, satellites or deep space vehicles performing a
cooperative task. We can develop potential individual behaviors; the working mechanism of the
biological colony system can be obtained through the method of mathematical modeling.
A common application of drones is for military one, but their civilian application is increasing
attention in the recent time. Multivehicle systems has a greater potential application compared to
autonomous vehicles that perform solo missions, which has greater efficiency [8].
Existing technologies of unmanned aerial vehicle swarming systems have many application as
mentioned below [7].
 Stage entertainment
 search and rescue
 Providing wifi coverage
 National security
 Space exploration
In all above applications, the ability of a networked systems in a decentralized fashion, compute
common estimates of unknown variables, and agree on a common view of the system
parameters. A consensus algorithm is a type of algorithmic rule that specifies the information
exchange between a bunch of agents and every one of its neighbors on specific network. It is
designed to achieve reliability in a network involving multiple unreliable nodes. Solving that

1
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
issue called the consensus problem is important in distributed computing and multi-agent
systems. Generally the basic idea of a consensus algorithm is to impose similar dynamics on the
information states of each vehicle, which is the interesting problem researched in this thesis [12].
In addition the communication network of a multi-agent cooperative swarming system can be
modeled by a directed graph abbreviated as digraph.

1.2 Problem statement

A swarm of Unmanned Aerial Vehicles during a swarm is propelled by a number of rotors. Each
drones can vertically hover, take-off, and land (VTOL). The movement of drones is controlled
by a remote controller and autonomously by using processors deployed on the drones [9].

The applications of drones have increased in the last few decades and the availability of cheap
hardwares has increased the interest for drone swarms this aerial drones can collaborate to
achieve different collective task.
The drones can divide the workload among the group so that is more capable to perform a task
by a group of drones instead of a single. Each member can be equipped for different functions in
the mission like surveillance and reconnaissance, strike, or battle damage assessment.
Surveillance missions are completed quickly by covering wider search area when the group is
spread out. They can also use for space-based interferometers; combat, surveillance, and
reconnaissance systems; hazardous material handling; and distributed reconfigurable sensor
networks. To use these applications effectively, various cooperative control capabilities need to
be developed, including formation control, rendezvous, attitude alignment, flocking, foraging,
task and role assignment, air traffic control, and cooperative search. When a single drone is out
of mission by different case: If there is defect, when it is destroyed by the enemy or when it
drops out because of mechanical failure the rest of the group will fill in and carry out the
mission. So having a network of drones the probability of success increases [4].

Cooperative control of swarm autonomous vehicles poses significant theoretical and practical
challenges. The communication bandwidth and connectivity of each member is limited, and the
information exchange among vehicles may be unreliable. It is also difficult to decide the
members, what to communicate and when and with whom the communication takes place.
Another critical problem for cooperative control is determining algorithms so that a group of
vehicles can agree on the instantiations of the coordination variable [9].

One difficulty of average consensus algorithm is information distributed for each agents across a
network that makes the mathematical computation hard to perform and agents hard to agree on a
specific combination of single quantities. There is a delay in a calculated average when the size
of a network is increased.
When more than one drone is used which are controlled individually it will have some
drawbacks. First Controlling drones individually needs more number of controlling stations,
mission planning softwares and infrastructure. More infrastructure means more money. Second

2
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
for each drone it needs a ground pilot this in turn will cost extra money and unwisely use of labor
force.
Generally drone swarming technology can be a means to improve society way of living. It can
also analyzed and implemented by most sophisticated and modern algorism which will drive the
innovation that can be used to solve our most difficult problems and open up markets that will
elevate the drone industry [5].

1.3 Objective
1.3.1 General objective
The main objective of this research paper is to identify and design different swarming patterns of
small scale unmanned aerial vehicles. Studying and presenting key results on theory and
applications of distributed consensus algorithm problems in networked systems.
1.3.2 Specific objective
The specific objectives of this thesis work will be:
 Try to understand the modelling of quadcopters.
 Representing network of agents by using a directed graph.
 Show the swarming mechanism
 Mainly focus on aggregation of quadcopters (converge to a
singular point)
 Identify best control algorithm to simulate the swarming process.
 Applying those algorithms on the derived mathematical model in
computer simulations.
 Identify and simulate different swarming patterns.

The thesis will be concluded with a comparison between the developed control algorithms in
terms of its dynamic performance and its ability to stabilize the system under the effect of
additional disturbances.

2. Methodology
2.1 Modeling Approach
In this chapter, Newton-Euler modelling approaches are used to study the behavior of the
quadcopters. Quadcopters are under-actuated systems in which four independent inputs are used
to control motion in six degrees of freedom, three translational and three rotational. Because of
aero dynamic effects, the resulting dynamics are highly nonlinear. Quadcopters are either lifted
up ward, moves along x-axis, along y-axis, in the negative direction propelled using four
vertically oriented propellers. Quadrotors use four identical rotors arranged at the edge of a cross
body, two diagonal rotors rotate a clockwise rotation while the other two diagonal rotors rotate
anti-clockwise direction. To control and move the quadcopter, independent variation in the speed
of rotor can be changed accordingly. Change in speed of each rotor affects the total thrust, which

3
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
can be further controlled to create desired total torque to navigate the quadcopter in all three
directions. By giving the desired x-value and controlling the quadcopter to move on the given
trajectory, pitch angle can be generated and controlled. Similarly, by controlling the movement
along y direction of the craft, roll angle is achieved. To increase the pitch the thrust on motor 2
and motor 3 is decreased and for the diagonally placed motor 1 and motor 4, thrust is increased.
This results in a moment, which tilts the quadcopter by creating a component of thrust in the x
direction. On the same way, the thrust on motor 1 and motor 2 is decreased and for the
diagonally placed motor 3 and motor 4, thrust is increased in a similar manner to maintain the
position in the y direction.
Quadcopter Dynamics are derived and a mathematical formulation for the Quadcopter model is
discussed in detail. Maintaining the altitude of the quadcopter, i.e .its position in the z-axis is
achieved by apply equal amount of trust along each motor. Decreasing the thrust on one pair of
motors arranged in one arm and increasing the thrust by an equal amount on the other two
motors arranged on the other arm creates a moment ab out z-axis of the quadcopter triggering it
to yaw. To determine all the equation governing the motion of the quadcopter we need to analyse
it in the inertial reference frame and the fied body frame. The position and velo city of the
quadcopter is measured from inertial frame, whereas orientation and angular speeds are
determined by body-fied frame.

4
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
2.2 Swarming approach

2.3 Simulation approach

2.4 Related work

The idea of swarming intelligence and particle path tracking properties of drones has already
attracted significant research attentions for long. There are handful researches that attempted to
achieve these swarming behaviors.

For a group of agents where each agent cooperates to track the average of locally available time-
varying reference signals, in which each agent is only capable of local computations and
communicating with local neighbors [11].

Design and performance analysis of distributed consensus algorithms can be analyzed by graph
theory [10]. Introduction to Graph theory by Robin J. Wilson they provide a brief overview of
the relevant graph theoretic concepts, definition and notations.

The fusion of swarm behavior in multi robotic system specially for quadrotors was presented by
A. Bandala, P. Dadios, P. Vicerra, and Laurence A. Gan Lim at De La Salle University, Manila
in 2014. Their study directed on using robot swarms because of its main feature of decentralized
processing amongst each member. The compatibility of applying swarm algorithm on UAV
quadrotors for aerial surveillance, search and reconnaissance operations through flight
formations and reconfiguration by abiding swarming patterns and behavior [1]. They also show
some behaviors of swarming like Aggregation, Social Foraging, Flight Formation and Swarm
Tracking.

Decentralized cooperation of autonomous robots will be self-aggregated one place is a very

frequent behavioral pattern in the natural world [3, 6]. Aggregation can be studied in to two
types, static and dynamic aggregation. Static aggregation determines the formation of compact

5
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
aggregates to form multiple disjoint clusters as a large number of robots in one region, however,
cannot be scaled. But in the case of dynamic aggregation has a greater scalability, formed units
are less compact, but they continue to move through the area, and when there is a large number
of robots in the area, different units tend to merge and form a single unit [3, 6].

The reduction in size of quadcopters is helpful for agility and the ability to operate in tight
formations was argued by Alex Kushleyev, Daniel Mellinger, Vijay Kumar at the University of
Pennsylvania. They provide experimental arguments in support of this claim [2].

Two autonomous control methods of unmanned aerial vehicle: The Particle Swarm Optimization
algorithm (PSO) and a linear control method are analyzed and discussed by Frantz and Natalie. It
focused mainly on minimizing of errors between the particles and the target [4]. When the
direction of the particle is changed to head toward the target, the algorithm accelerates the
particles. They tried to compare PSO and linear algorithm.

Now let us see the swarming control of unmanned aerial vehicles published in some papers. The
first paper to see is swarm control in unmanned aerial vehicle by Henry Hexmoor and Brian
Mclaughlan in 2005. They have implemented a simulation of UAV swarms in Java [1].
Swarms of Unmanned Aerial Vehicles — A Survey by Anam Tahir, Jari Böling, Mohammad-
Hashem Haghbayan, Hannu T. Toivonen, Juha Plosila [Journal of Industrial Information
Integration, 2019]. In this paper different application areas of swarm drones are presented
Security, Survey, Monitoring, and Surveillance, Leisure Pursuit, Disaster Management, Search
and Rescue (S&R) [2].
The positive and negative sides of drone swarming are presented in “The Upside and Downside
of Swarming Drones” by Irving Lachow [2017]. In this paper it stated that Swarming drones can
be used in three ways by military forces: to attack, defend, and provide support functions such as
intelligence, surveillance, and reconnaissance [8].

2.5Scope
The research paper mainly focusses on the mechanism of distributed consensus algorithm and
designing of drone swarming for small scale unmanned aerial vehicles and analyzing control
algorithms for simulation.

2.6 Research Methodology

The idea of this thesis is to design a simple swarming pattern for drone swarms, with a special
focus on quadcopters. First nature and behavior of quadcopters is studied with identifying the
variables.
Generally, the thesis will have the following steps.
 Analyzing all the model Variables and necessary general information about swarming
will be collected.

6
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
  Different swarming algorithms and patterns will be analyzed.
 Mathematical equation of the system is formulated.
 Decide the best possible swarming pattern and its algorithm.
 Simulate the result on Matlab and discussion.

 Measuring the performance of the model and analyze the output.

 Finalize
The detail of each of the above steps are decided though out the thesis time.

3 Schedule
Task 2021 2022

Jun July Aug Sep Oct Nov Dec Jan Feb Mar Apr May
literature
Review
preparing
basic
documents
Proposal
writing
Proposal
defense
Formulating
mathematical
equations
Developing
model and
simulate
Analyze
decide the best
possible
options
Conclusion

Thesis paper
writing

7
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
 Final defense

*My admission date was July, 01, 2021

4 Budget
Computer simulation use more computing power so we need high spec computers.

No Item Price (in Birr)

1 Registration cost 18,240.00

2 Laptop 27,000.00
3 Internet cost 4,576.00
4 Transport cost 5,000.00
5 Inflation cost 1,000.00
total 55,816.00

5 References
[1]. A. Bandala, E.Dadios, P.Vicerra, A. Gan Lim “Swarm algorithm for unmanned aerial vehicle (UAV)
quadcopters: swarm behavior for aggregation, foraging, formation and tracking”, Journal, De La
Salle University, may 2014.
[2]. A. Kushleyev, D. Mellinger, V. Kumar “Towards a swarm of agile micro quadcopters”, University
of Pennsylvania.
[3]. Veysel Gazi, k. passin “Swarm stability and optimization”, Marmara University, January 2011.
[4]. Frantz, Natalie, “Swarm intelligence for autonomous UAV control”, Naval postgraduate school,
California, June 2005.
[5]. Y. Zhou, Bin Rao, Wei Wang, “UAV Swarm Intelligence: Resent Advances and Future Trends”, Sun
Yat-sen University, Guangzhou, China, October 2020.
[6]. A. Alfeo1, Mario. Cimino, Nicoletta De Francesco1, Massimiliano Lega, Gigliola Vaglini. “Design
and simulation of the emergent behavior of small drones swarming for distributed target
localization.”, University of Naples, Italy,
[7]. Electronicsforu
URL (https://www.electronicsforu.com/technology-trends/drone-swarm-kingmaker-technology)
[8]. Wei Ren, Randal W. Beard, “Distributed Consensus in Multi-Vehicle Cooperative Control Theory
and Applications”, Utah State University and Brigham Young University, USA, 2008.
[9]. Anam Tahir, Jari Böling, Mohammad-Hashem Haghbayan, Hannu T. Toivonen, Juha Plosila
“Swarms of Unmanned Aerial Vehicles — A Survey”, Journal of Industrial Information Integration,
University of Turku, Finland, 2019.
[10]. ROBIN J.WILSON “introduction to Graph theory”, 4rth edition, Edinburgh Gate, Harlow
England, 1996.

8
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
[11]. S. S. Kia B. Van Scoy J. Cortes R. A. Freeman K. M. Lynch S. Martinez “Tutorial on Dynamic
Average Consensus: The problem, its applications, and the algorithms”, March 2018
[12]. What ls
URL(https://whatis.techtarget.com/definition/consensus-algorithm)

6 Chapter one
1. Introduction
Included in this chapter
[Definition of Distributed consensus algorithm, application of multi vehicle autonomous
vehicles, single integrator and double integrator dynamics]
The dynamic average consensus problem is for a group of agents to cooperate to track the
average of locally available time-varying reference signals, where each agent is capable only of
local computations and communicating with local neighbors [11].
[Definition of cooperative control]
When multiple vehicles agree on the value of a variable of interest, they are said to have reached
consensus.
Consensus algorithms are designed to be distributed, assuming only neighbor-to neighbor
interaction between vehicles. Vehicles update the value of their information states based on the
information states of their neighbors. The goal is to design an update law so that the information
states of all of the vehicles in the network converge to a common value [8, 6].
The most common continuous-time consensus algorithm is given by

(1.1)

a ij ( t ) is the (i, j) entry of adjacency matrix

x i is the information state of the ith vehicle.

9
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
The above consensus algorithm can be written in a matrix form as follows

Ẋ ( t )=−Ln ( t ) x (t )

Where x is the information state L is Laplacian matrix

| ||
1 −1 0 xi
ẋ= 0 1.5 −1.5 x 2
−2 0 2 x3

| |
−x 1 + x 2
ẋ= −1.5 x 2 +1.5 x3
2 x 1−2 x2

A GUIDE TO HOW DRONES ARE USED FOR INSPECTIONS

These days, drone inspections are being performed in almost every industry that requires visual
inspections as part of its maintenance procedures. By using a drone to collect visual data on the
condition of an asset, drone inspections help inspectors avoid having to place themselves in
dangerous situations. 
Inspection drones are transforming manual inspection processes, allowing inspectors to speed up
their collection of inspection data while removing slower manual steps that place them in danger.
For example, inspecting a cell tower manually may require climbing several stories in the air on
a tower to take a close look at a guy-wire. And inspecting an industrial boiler manually may
require climbing fifty feet or more into the air on scaffolding.
But inspection drones are changing all of this.
Now, a drone inspection service provider can fly an inspection drone up the cell tower, or
throughout the boiler, capturing all the visual data the inspector needs to complete their
inspection.
Types of inspection

10
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
  ROOF INSPECTIONS
 CELL TOWER INSPECTIONS
 Industrial inspection

Let me take 3 vehicles strongly connected.

1
1 2

2 1.5

3
The Laplacian matrix of the above topology is

| |
1 −1 0
L= 0 1.5 −1.5
−2 0 2

For the agents to achieve convergence and consensus the simple eigenvalue must be zero.

Substituting the above Laplacian matrix to equation

Chapter
Graph theory and representation
Introduction
graph theory is the study of graphs, which are mathematical structures used to model pairwise
relations between objects. A graph in this context is made up of vertices (also called nodes or points)
which are connected by edges (also called links or lines). A distinction is made between undirected

11
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
graphs, where edges link two vertices symmetrically, and directed graphs, where edges link two
vertices asymmetrically. [wikipidea]

Definition

It is natural to model information exchange among vehicles by directed or undirected graphs.

The points are called vertices, the lines are called edges, and the whole diagram is called a graph.
The degree of a vertex is the number of edges with that vertex as an end-point. The study of
directed graphs (or digraphs, as we abbreviate them) arises from making the roads into one-way
streets.
Graphs can be divided into two directed and undirected
1. Undirected Graph

An undirected graph is a graph in which the vertices are connected by undirected edges.
An undirected arc is an edge that has no arrow. Both ends of an undirected arc are
equivalent--there is no head or tail.

2. Directed graph

Directed graph has arrows……

A simple graph is a finite undirected graph without loops and multiple edges [10].

Relationship between Matrices and Graphs

Graph theory is closely related to Matrix theory.
A matrix A is positive definite (semi definite) if and only if all eigenvalues of span style="font-family:
Times New Roman;font-size:12pt;color:rgb(0,0,0);font-style:normal;font-variant:normal;">span
style="font-family: Cambria Math;font-size:8pt;color:rgb(0,0,0);font-style:normal;font-
variant:normal;"> are positive (non-negative) [periapoilan].

Kronecker matrix
In mathematics, the Kronecker product, sometimes denoted by ⊗, is an operation on two matrices of
arbitrary size resulting in a block matrix. It is a generalization of the outer product (which is denoted by
the same symbol) from vectors to matrices, and gives the matrix of the tensor product linear map with
respect to a standard choice of basis. The Kronecker product is to be distinguished from the usual matrix
multiplication, which is an entirely different operation. The Kronecker product is also sometimes
called matrix direct product [wikipidea].

12
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
Adjacency Matrix

Let G = (V, E) be a graph and assume that V = {v1,…,v2}. The adjacency matrix of
G is an n × n matrix A defined as:

Let us consider the following two graphs, the first one is strongly connected weighted graph and the
second one is connected unit edge graph.

13
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
The adjacency matrix of a simple graph is symmetric with diagonal entries as 0.

Laplacian matrix L = LG is the n × n matrix given by

Where di is the degree of the ith vertex

Example

14
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
Swarm agent model can be divided in to two.
1. Single integrator model: It is the simplest mathematical representation of unmanned vehicle
agents. In this model the motion of a vehicle ⅈ , ⅈ = 1, 2, .... N is given by
ẋ i=ui
Where x i is state of the agent and ui is control input.
2. Double integrator model: Double integrated swarm agents are connected with N agents, each
agent has mass dynamics of
ẋ i=vi
1
vi = ui
Mi

Where x iis position vi is velocity M i is mass ui is the control input of agent i

When we rearrange the above equation ui=M i ẍ i when we integrate acceleration twice we will have
position of individual agents that is why we said it “double integrator model”. Each individual agent has
its own dynamics.

15
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
This paper only focus on single integrator agent model that means for simplicity I assume individual
agents have no mass or considered them as a point.

Single integrator agent model

ẋ i=ui [3.1]

For the above agent dynamics equation (3.1) I will develop control algorithm to obtain aggregation.

I assume all individual swarm agents move simultaneously and know the exact position of other
individuals.

The control input ui of individual agent i have the form [3].

ui=−∇ x J (x )i
[3.2]

Where J is a potential function which represents the repulsion and attraction between individual agents
and needs to be chosen by the designer.

!!!What is gradient? What is the advantage of gradient?

For aggregation the potential function J can be selected as J ( x )

N −1 N
J aggregation ( x )= ∑ ∑ [ j a (‖ x i−x j ‖¿ ¿)− j r (‖ x i−x j ‖)]¿ ¿ [3.3]
i=1 j=i+1

Where j a attraction component j r repulsion component

Substitute [3.3] into [3.2]

The control input of individual agents can be calculated as

N
ui=− ∑ ¿¿¿ [3.4]
j=1 , j ≠i

Depend on [3.1] the motion of individual agent is along the negative gradient and leads to a decent
motion towards a minimum of a potential function J .

J a (‖ y‖) and J r (‖ y‖) create a potential field of attraction and repulsion around each individual the
above property restricts the motion of the individual toward each other along the gradient of these
potentials.

Because of the chain rule

∇ y J a (‖ y‖)= y g a (‖ y‖)

16
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
 ∇ y J r (‖ y‖)= y g r (‖ y‖)

y g a (‖ y‖) = Actual attraction

y g r (‖ y‖) = Actual repulsion

Let me define the function g(.) as

We assume the above function satisfies the following three assumptions.

1. On large distance attraction dominates.

2. On short distance repulsion dominates.
3. There is a unique distance where attraction ad repulsion balances.

In other words there exist a unique distance δ at which we have

1. ga ( δ )=gr ( δ ) balanced
2. ga (‖ y‖) > gr (‖ y‖) for ‖ y‖>δ when attraction is greater than repulsion
3. gr (‖ y‖) > g a (‖ y‖) for ‖ y‖<δ when repulsion is greater than attraction.

Let me choose a potential function which satisfies the above assumptions [3].
2
N −1 N
−‖x −x ‖
J ( x )= ∑ ∑ [ a2 (‖x i−x j‖ ¿¿)+ bc2 exp( ic j )]¿ ¿
2
(3.6)
i=1 j=i+1

Its attraction and repulsion function can be calculated by the above equation

∇ y J a (‖ y‖)= y g a (‖ y‖)

∇ y J r (‖ y‖)= y g r (‖ y‖)

After a long derivation we will reach as

g ( y )=− y ¿ (3.7)

The equilibrium distance δ can be calculated by making the above equation to zero.

√ b
δ = cln ( )
a
(3.8)

Question!! What is the value of a, b and c? Refer [92]

“You can refer and take a sample potential function from other papers and verify it by your own.”
bilognal

17
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
 A veh=diag
(| 0 1
1
0
1 ………… n
a21 a22 | 1
n
a21 a 22 | |)
Bveh =I n ⊗ ( 0 )
1 ||
3√3 2
Area= d
2
2
Area=r

h=h p ⊗ 1
0 ()
A=I N ⊗ A veh

B=I N ⊗ B veh

( x p ) ( t ) −( h p ) ( t )=q
The vehicles converge to formation h if there exist Rn-valued functions q(·)
ẋ i= A x i

ẋ=( I i ⊗ A ) x

N xN
A ⊗ In

1
y i=( x i−hi )− ∑ ( x j−h j )
ji

ⅈ=1 … . n
y i=L(x−h)
L=LG ⊗ I n

LG

18
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
Structure of the paper
Chapter 1 introduction (14 pages)
application area, cooperative control, advantage, disadvantage, challenges, problem statement,
objective

Chapter 2 graph theory (10 pages)

Chapter 3 consensus algorithm (6 pages)

Chapter 4 formation control (6 pages)

Chapter 5 conclusion and future work (4 pages)

19
AAU, AAiT, School of Electrical and Computer Engineering, November 2021
 20
AAU, AAiT, School of Electrical and Computer Engineering, November 2021