2019 16TH WORKSHOP ON POSITIONING, NAVIGATION AND COMMUNICATION (WPNC)

Periodic Ambiguity Function as a Tool in

Designing Periodic Synchronization Preamble
Igor Dotlic and Michael McLaughlin
Decawave Ltd., Adelaide Chambers, Peter Street, Dublin 8, Ireland
Email: first_name.last_name(at)decawave.com

Abstract—The paper introduces the periodic ambiguity func- mode is being added to HRP UWB PHY with the mean PRF
tion (PAF) of the preamble code as a tool in designing periodic of scrambled timestamp sequence (STS) equal to 124.8 MHz.
synchronization preamble. It is shown that in the considered One proposal for the HPRF preamble codes was to use
low-complexity receiver architecture characteristics of the PAF,
namely, sidelobes peak level and main lobe phase, play an impor- binary m-sequences of length 127 with spreading factor of
tant role in the quality of the channel estimation. Furthermore, four [6]; this results in exactly the same mean PRF as in STS
it is shown that these PAF characteristics vary considerably with and keeping the preamble symbol length of 1017.63 ns, but
the circular shift of the preamble code and care should be taken the codes used are not perfect.
when selecting both preamble codes and corresponding circular Another proposal was to use length 31 perfect ternary
shifts. This is done on an example of the preamble design for
high pulse repetition frequency (HPRF) mode of the upcoming Ipatov codes with 25 non-zero elements also with spreading
IEEE 802.15.4z standard amendment for enhanced ultra wide- factor of four [7]; this proposal has a lower mean PRF
band (UWB) physical layers (PHYs) and associated ranging of 100.65 MHz and also a considerably lower preamble symbol
techniques. Similar analysis is carried-out on the preamble codes length of 248.4 ns.
already specified in the high-rate pulse repetition frequency
Compromised solution was to use perfect Ipatov ternary
(HRP) UWB PHY of the IEEE 802.15.4-2015 standard.
codes of length 91 with 81 non-zero elements (Ipatov-91 from
now on), again, with the spreading factor of four [8]; this
I. I NTRODUCTION results in the mean PRF of 111.09 MHz and the preamble
When designing a periodic preamble for a PHY usually a symbol length of 729.17 ns.
class of codes is preselected based on the system modulation The authors of this paper took upon the task of selecting
and, in the case of a UWB PHY, mean pulse repetition eight Ipatov-91 codes together with the corresponding circular
frequency (PRF) of the PHY. Moreover, selected code class shifts for the standard amendment. In the process it was found
usually has preferred periodic auto-correlation (PACR) prop- that the circular shift of any Ipatov-91 code affects greatly its
erties. When selecting a set of codes from a given class, they PAF characteristics; namely, peak sidelobes level and, to the
are usually selected based on minimizing peak periodic cross- lesser extent, main lobe phase.
correlation (PCCR) within the set. The energy of PCCR of As it will be shown here, PAF characteristic have a great
two codes in the class is fixed and equal to the energy of impact on the channel estimation performance of the con-
the corresponding PACR due to the fact that all the codes sidered low-complexity receiver architecture [9] especially
in the class have the same PACR. This can be easily shown when carrier frequency offset (CFO) is high. Therefore, it
by applying Parseval’s theorem to discrete Fourier transform was decided to take PAF characteristics criteria into account
(DFT) of PCCR which has the same amplitude spectrum as when choosing Ipatov-91 codes and their circular shifts for
DFT of the corresponding PACR. Circular shifts of the codes the standard.
in the chosen set is irrelevant in the above criterion and can The paper first gives a simplified model of the considered
be selected based upon some other criteria. receiver architecture and shows the relationship between PAF
A good example are perfect Ipatov ternary codes [1]– and the channel estimate in the receiver. Then, the process of
[3] found in HRP UWB PHY published in IEEE 802.15.4a choosing Ipatov-91 codes is documented that includes peak
standard amendment [4] and afterward integrated into PCCR as the primary criterion and also PAF characteristics
IEEE 802.15.4-2015 standard [5]. Being perfect the codes have criteria.
zero PACR sidelobes. Spreading factors, lengths of the codes At the end the approach is generalized to show how
and number of their non-zero elements were carefully selected PAF characteristics of the Ipatov codes currently found in
in order to match mean PRFs in preamble and data parts of IEEE 802.15.4 HRP UWB PHY vary with the circular shift.
the frame as closely as possible together with having preamble
symbol length of roughly 1 µs. For example, Ipatov codes of II. C ONSIDERED R ECEIVER A RCHITECTURE AND PAF
length 127 with 64 non-zero elements together with spreading
factor equal to four produce mean PRF of 62.89 MHz and the A. General description
preamble symbol length of 1017.63 ns [4], [5]. Simplified block diagram of the considered receiver ar-
IEEE 802.15.4z task group was formed in 2018 in order to chitecture [9] during the channel estimation phase is shown
make enhancements to IEEE 802.15.4 UWB PHYs [5]. HPRF in Fig. 1. The main characteristic of this architecture is
2019 16TH WORKSHOP ON POSITIONING, NAVIGATION AND COMMUNICATION (WPNC)

Analog I/Q CFO removal

Correlator Accumulator
RF and BB sampling (loop)

Fig. 1. Simplified block diagram of the considered receiver architecture during periodic preamble channel estimation.

that CFO is removed after the correlation with the preamble Typical Ipatov-91 Ambiguity Function
symbol. For this reason, received preamble is deformed by
CFO in the correlation. Compared to the classic receiver
architecture where CFO is removed before the correlation, the 1.0

| (n, )| norm. level

lower complexity is achieved by: 0.8
• This architecture has considerably lower bit-width of the
received signal in the correlator which is a very large 0.6
digital block. 0.4
• RF phase locked loop in the receiver is locked onto local
oscillator and it does not depend on the received signal 0.2
which lowers its complexity.
0.0
Another advantage of this architecture is that the receiver
3540
has a precise digital estimate of the CFO which is rather useful
30
in systematic localization error compensation [10]. 25 pm)
40 20 (p
20 15
0 10 9 CFO
B. Mathematical model n 20 5 CH
40 0
Let us consider an infinite periodic preamble modulated by
the CFO; in the baseband discrete-time transmitter equivalent Fig. 2. Slices of a typical Ipatov-91 PAF normalized amplitude.
it will have a form:
s(n) = a (n mod M ) exp(2jπνn), (1) Here, χ(n, ν) represents PAF of a(n) with a single integration
period [11], [12]:
for −∞ < n < +∞. Here, a(n) is the preamble code of
length M and ν is a discrete-time frequency of CFO. s(n) goes χ(n, ν) = [a (n mod M ) exp(2jπνn)] ⊗ a∗ (−n), (5)
through the channel with the complex discrete-time baseband-
equivalent impulse response hc (n) which includes both trans- for −∞ < n < +∞. An important property of χ(n, ν) that
mitter and receiver impulse responses. At the receiver it will is easily derived from (5) is its pseudo-periodicity in time:
produce baseband-equivalent complex signal r(n): χ(n + kM, ν) = exp(jkϕ)χ(n, ν), (6)
r(n) = s(n) ⊗ hc (n), (2)
for integer k and ϕ = 2πνM .
where ‘⊗’ denotes aperiodic convolution. This property is exploited in the rotation and accumulation;
In a real receiver the sampling rate is higher than one sample the receiver rotates and accumulates K preamble symbols at
per preamble chip, hence, a realistic model is up-sampled the output of the correlator in order to get the scaled channel
with regards to the model introduced here. Furthermore, a estimate (ĥc (n)):
CFO also produces the effect of dilatation / expansion of the K−1
baseband signal in time, since the same reference clock in the
X
ĥc (n) = c(n + M k) exp(−jkϕ),
transmitter is used both for carrier and baseband signal gen- k=0
(7)
eration. Therefore, baseband receiver processing also requires = Kχ(n, ν) ⊗ hc (n),
re-sampling [9]. However, these are omitted here for the sake
of simplicity and without loss of generality. for n ∈ {0, 1, . . . , M − 1}.
The received signal goes through the correlator with the
impulse response a∗ (−n) to produce the signal c(n): C. PAF characteristics
∗
c(n) = r(n) ⊗ a (−n). (3) Slices of a typical Ipatov-91 PAF amplitude are shown in
Fig. 2. Relative CFO () is represented with the respect to
(3) can be rearranged as
HRP UWB PHY [5] channel 9 (CH9) carrier frequency (fc )
c(n) = χ(n, ν) ⊗ hc (n). (4) in parts-per-million (ppm).
2019 16TH WORKSHOP ON POSITIONING, NAVIGATION AND COMMUNICATION (WPNC)

The relative CFO that a standard-compliant device can 50

encounter is || ≤ 40 ppm since the prescribed frequency
tolerance is ±20 ppm [5].
40
The carrier frequency for CH9 is fc = 16fb ≈ 8 GHz
with fb = 499.2 MHz being peak chip frequency. With

Number of Code Sets

the spreading factor equal to four, the peak preamble chip 30
frequency is fp = fb /4, hence, ν =  fc /fp = 64.
We will consider PAF characteristics like peak sidelobe level
and main lobe phase at relative CH9 CFO of  = 40 ppm and 20
thus ν = 64 = 0.00256.
In the radar literature [11], [12] both periodic and aperiodic 10
ambiguity functions are traditionally depicted up to very large
values of ν; this will not be done here as such values are not
encountered in a standard-compliant device. 0
From Fig. 2 it can be seen that the main lobe level decreases 22 22.5 23 23.5
with . On the other hand, sidelobes level increases relatively
Mean Peak PCCR
uniformly with . Thus, perfect periodic autocorrelation prop-
erty, depicted for  = 0, is progressively lost as  increases. Fig. 3. Mean peak PCCR distribution between 576 code sets.
From (7) it is clear that the first path sensitivity of the
considered receiver will be directly affected by the peak
sidelobe level of the PAF defined as B. Peak PCCR criterion
max |χ(n, ν)| Before choosing a circular shift, a set of eight codes needs
n∈{1,2,...,M −1}
p(ν) = . (8) to be chosen, as the IEEE 802.15.4z standardization process
|χ(0, ν)|
required this number of codes for the standard draft [8].
A second PAF property which is important for the con- Calculated PCCR matrix of all 24 Ipatov-91 codes shows that
sidered receiver architecture is its main lobe phase ∠χ(0, ν), the minimum PCCR is 15 and maximum PCCR is 48.
since from (7) it is clear that it determines the rotation of the The criterion used to select the set of eight codes is
channel estimate for a given ν. This knowledge is especially minimizing the maximum peak PCCR within the set while
important for coherent receivers when using ĥc (n) in data having at least two codes in the set that have the minimum
demodulation. possible mutual peak PCCR equal to 15.
From (5) it follows that the PAF main lobe is given by The minimization process found 576 code sets with the
M −1
X minimum attainable maximum peak PCCR within the set equal
χ(0, ν) = |a(n)|2 exp(2jπνn). (9) to 27 while satisfying the above condition. Distribution of the
n=0
mean peak PCCR among these code sets is shown in Fig. 3;
Hence, for a ternary code PAF main lobe phase depends on it clearly varies little and there are 12 code sets with the
the position of zeros within the code and length of the code. minimum mean peak PCCR of 21.89.
For the considered maximum CFO of  = 40 ppm and
the given code length of 91 the main lobe phase always
changes approximately linearly with ν, i.e., ∠χ(0, ν) ≈ κν. C. Applying PAF criteria
The receiver complexity can be decreased if κ varies as little Since mean peak PCCR varies little within 576 code sets
as possible within the preamble code set and, thus, may be analyzed above we will consider all these sets when optimizing
regarded as a constant independent on the preamble code PAF characteristics.
index. As shown in Fig. 4, PAF peak sidelobes level varies with the
Additional criteria that might be considered are PAF RMS circular shift rather significantly up to roughly 5.5 dB, whereas
sidelobes level and PAF main lobe level loss with CFO: RMS sidelobes level varies to a much lesser extent. Codes 10
|χ(0, ν)/χ(0, 0)|. However, these variables do not change and 16, which are dual, attain the minimum peak sidelobes
considerably with the code index and the circular shift. level of -21.1 dB for the optimal circular shifts in this sense.
Codes 7 and 22, which are also dual, are worst in this sense,
III. I PATOV-91 C ODE S ELECTION
as they attain the minimum peak sidelobes level of -19.75 dB.
A. Generated Ipatov-91 codes On the other hand, PAF main lobe phase varies up to few
Ipatov-91 codes obtained by the method described in [1], degrees, as shown in Fig. 5.
[3] are given in Table I; there are total of 24 unique codes The first criterion to chose one of the 576 code sets is for
if the circular shift and the binary multiplication constant are the set to include code 16, i.e., to have one of the codes with
arbitrary. However, each code in Table I has its dual in the the minimum attainable PAF peak sidelobe level. The second
table that can be obtained by reversing the code and doing criterion is to minimize the maximum PAF peak sidelobe
a circular shift. Dual codes are displayed together due to the level within the set when the circular shift for each code is
fact that all 24 codes are equal in the process of selection that optimal in this sense. The code set chosen with these criteria
will be described below. is {1, 9, 14, 15, 16, 17, 19, 23}.
2019 16TH WORKSHOP ON POSITIONING, NAVIGATION AND COMMUNICATION (WPNC)

TABLE I
G ENERATED I PATOV-91 CODES .

Index Ternary Code

0 --+0--+0+++++--+++--0--0+-+----0-+-+-+0+--++-+-+++0-0++++--+--+-++++-+---++--++++++-++-+00+
1 0+-++-++------0+-+++-+++-+++++0+-+-+0-+-+++---+++++--+-++-++--00-++0--0++--+++-+0+0-+-+----
2 0-++++-+---+++--++00+--+-0-----++-0-+-++++-+-0++-+++-+-+++--+-++-++++-++++0---0--++-++0-0-+
3 +---+++0+-0--+00++++++++-+--+--+-+++--+-+0+-+--0++---+--+-+----0+-+++-+++-++-0+++---++0+0-+
4 -0+++-+++++--+-++---+-+---+-+0-+--0+0+-++-++0+-+--0+++++-+-++0+++--++-0++00--++--+++-+-----
5 00+-++-++++++--++---+-++++-+--+--++++0-0+++-+-++--+0+-+-+-0----+-+0--0--+++--+++++0+--0+--+
6 -+--0-+-++-+-+-0++---0+---++-0-0-+-+0+++-+++++++--+++-+--++-+--+0-++++++-----++-++00++0--++
7 +--0++-+-+0-++++++++++--+--++-++0++-+-+--+--+++---0-+++-0+00---++-+--0-++0+++--+++---+-+-0-
8 -0+0++---+++0-++-+++-+++-+0----+-+--+---++0--+-+0+-+--+++-+--+--+-++++++++00+--0-+0+++---++
9 +-0+0-+-++-++--+--+++++++++++---+-++--++00-+-+0-+-0-++--+++-+++0--0++-+-+-0---++-0----++++-
10 +-+00++-0-0++++-++-+--+-+--+++0+-----+0+-+---++0--+-+---++-+++--++++++0+++---+--0+-0+-++-++
11 +-+++----0+-+++--+++---++-++--+0-+-++-0---+0+++-+-+-++-+---0---0+-++--+++00+0++++-0+-+-++++
12 +++----0-+0+--+++-++-00-++0+0+-+-+-+++++-+++----0+--+-++0++-++++-0-++++-++---+--+--++---0+-
13 +--0++00++-++-----++++++-0+--+-++--+-+++--+++++++-+++0+-+-0-0-++---+0---++0-+-+-++-+-0--+-+
14 ++0--0++-00--++-++-+--+++++---+++-+-0+-+-+0+++++-+++-+++-+0------++-++-+0----+-+-0+0+-+++--
15 +++-0+-0-0++-++--0---0++++-++++-++-+--+++-+-+++-++0-+-++++-+-0-++-----0-+--+00++--+++---+-+
16 +++--+++-++---+-+--0++---+-+0+-----+0+++--+-+--+-++-++++0-0-++00+-+++-++-+0-+0--+---+++0+++
17 -+-+-+0-+-----+0--+++00++-+--++++++-++-+0--0+0+-++++-0+++0-+---++--+--++-+-+++++---+-+0++--
18 +++-+-+0-++++0+00+++--++-+0---0---+-++-+-+-+++0+---0-++-+-0+--++-++---+++--+++-+0----+++-++
19 +++++-+-+-+0+0++-00-++-+++--+0+-0----+++-+0---++--+--+---++-++++-0-++++-++0++-+--+0----+++-
20 +++++0--+-+0++-++-+0+0--+-0+-+---+-+---++-+--+++++-+++0------+-+++--++--00++0-++--+++0++-+-
21 ++-++++++--+-++00+++--0+-----+-0+-+-+---++0+-+---+++++-+-++--+--++---+-0+++0-++++-+0+0--0+-
22 +-+-++0++-++--+--++++++++++-0+-+-++0--+-0-+-+---+++--+++0++-0--+-++---00+0-+++-0---+++--+--
23 ++-+---+++++++++++--+--++-++-+-0+0-+-++++----0-++---0-+-+-++0--0+++-+++--++-0-+-0+-+-00++--

44
PAF Sidelobe Levels for  = 40 ppm (dB)

−16
∠χ(0, ν) for  = 40 ppm (deg)

43
−18

−20 42

−22
41
Max Peak
−24 Min Peak
Max RMS
40 Maximum attained
Min RMS
−26 Minimum attained
RMS for Min Peak
Min Peak SLs
39
0 2 4 6 8 10 12 14 16 18 20 22
0 2 4 6 8 10 12 14 16 18 20 22
Code Index
Code Index

Fig. 4. Different attained PAF sidelobe levels with varying circular shifts for
Fig. 5. Different attained PAF main lobe phases with varying circular shifts
all Ipatov-91 codes in Table I.
for all Ipatov-91 codes in Table I.

When choosing the circular shifts the simplest approach IV. PAF A NALYSIS OF THE C URRENT IEEE 802.15.4 HRP
would be to choose those that produce the minimal PAF peak UWB PHY P REAMBLE C ODES
sidelobe levels for each code within the set. However, the main Analysis of PAF characteristics with varying circular shifts,
lobe phase variation within the set can be reduced if the peak similar to the one done above for Ipatov-91 codes, has been
sidelobe level criterion is relaxed by 0.1 dB and maximum performed for the ternary Ipatov codes which are currently in
absolute PAF main lobe phase difference from the mean HRP UWB PHY of IEEE 802.15.4-2015 standard [5]. This has
is minimized. By using these criteria the phase variation is been done in order to see if these codese performs similarly
reduced from 1.59 deg. to 1.14 deg, as shown in Figs. 6 and 7. in the sense of peak sidelobe level and main lobe phase and,
The resulting code set is given in Table II; besides max hence, if the results presented above can be generalized to
phase variation of 1.14 deg, it has mean peak PCCR of some extent.
23.57, minimum peak sidelobes within the set of -21.10 dB, The standard includes eight length 31 codes (indices 1 to
corresponding with the code 16 and maximum peak sidelobes 8) with 16 non-zero elements. The preamble symbols are
of -19.97 dB, corresponding with the code 19. generated with the spreading factor of 16, hence, producing the
2019 16TH WORKSHOP ON POSITIONING, NAVIGATION AND COMMUNICATION (WPNC)

TABLE II
C HOSEN I PATOV-91 CODES WITH THE CHOSEN CIRCULAR SHIFTS .

Index Ternary Code

1 +-+----0+-++-++------0+-+++-+++-+++++0+-+-+0-+-+++---+++++--+-++-++--00-++0--0++--+++-+0+0-
9 -+++++++++++---+-++--++00-+-+0-+-0-++--+++-+++0--0++-+-+-0---++-0----++++-+-0+0-+-++-++--+-
14 -++0--0++-00--++-++-+--+++++---+++-+-0+-+-+0+++++-+++-+++-+0------++-++-+0----+-+-0+0+-+++-
15 ---0++++-++++-++-+--+++-+-+++-++0-+-++++-+-0-++-----0-+--+00++--+++---+-++++-0+-0-0++-++--0
16 +++++--+++-++---+-+--0++---+-+0+-----+0+++--+-+--+-++-++++0-0-++00+-+++-++-+0-+0--+---+++0+
17 ---+-+-+0-+-----+0--+++00++-+--++++++-++-+0--0+0+-++++-0+++0-+---++--+--++-+-+++++---+-+0++
19 ++-00-++-+++--+0+-0----+++-+0---++--+--+---++-++++-0-++++-++0++-+--+0----+++-+++++-+-+-+0+0
23 -+--++-++-+-0+0-+-++++----0-++---0-+-+-++0--0+++-+++--++-0-+-0+-+-00++--++-+---+++++++++++-

43

PAF Sidelobe Levels for  = 40 ppm (dB)

Min Peak SLs
Reduced Phase Var. −10
∠χ(0, ν) for  = 40 ppm (deg)

−12
42

−14

Max Peak
−16
41 Current Peak
Min Peak
−18 Max RMS
Min RMS
−20 Current RMS
40

0 2 4 6 8 10 12 14 16 18 20 22 1 2 3 4 5 6 7 8
Code Index Code Index

Fig. 6. PAF main lobe phases for the chosen code set and two considered Fig. 8. Different attained PAF sidelobe levels with varying circular shifts for
circular shifts. IEEE 802.15.4 HRP UWB PHY length 31 codes.
PAF Peak Sidelobe Level for  = 40 ppm (dB)

−20 Min Peak SLs

Reduced Phase Var. 65
∠χ(0, ν) for  = 40 ppm (deg)

−20.2
60
−20.4

55
−20.6

50
−20.8 Maximum attained
Current
45 Min Peak SLs
−21
Minimum attained

0 2 4 6 8 10 12 14 16 18 20 22 1 2 3 4 5 6 7 8
Code Index Code Index

Fig. 7. PAF peak sidelobe levels for the chosen code set and two considered Fig. 9. Different attained PAF main lobe phases with varying circular shifts
circular shifts. for IEEE 802.15.4 HRP UWB PHY length 31 codes.

mean PRF of 16.10 MHz and the preamble symbol duration are generated with the spreading factor of 4, hence, produc-
of 993.59 ns. PAF sidelobes levels and PAF main lobe phases ing the mean PRF of 62.89 MHz and the preamble symbol
performances for length 31 codes are shown in Figs. 8 and 9, duration of 1017.63 ns. PAF sidelobes levels and PAF main
respectively. lobe phases performances for length 127 codes are shown in
The standard also includes 16 codes of length 127 with 64 Figs. 10 and 11, respectively.
non-zero elements (indices 9 to 24). The preamble symbols From Figs. 8 and 10 it is clear that for both length 31
2019 16TH WORKSHOP ON POSITIONING, NAVIGATION AND COMMUNICATION (WPNC)

PAF Sidelobe Levels for  = 40 ppm (dB) to much greater extent compared to Ipatov-91 codes which
−14 have only 11% of zeros. The PAF main lobe phases for the
length 31 and 127 codes are generally larger compared to
−16 Ipatov-91 codes due to the preamble symbols being roughly
30% longer in both cases.
−18
V. C ONCLUSIONS
−20
Max Peak The paper introduced PAF as a tool for designing preamble
Current Peak code sets which are CFO friendly.
−22 Min Peak Probably the most important contribution is the observation
Max RMS that that the PAF peak sidelobe level varies considerably with
−24 Min RMS the circular shift for all considered ternary Ipatov codes; this
Current RMS is useful in choosing preamble codes for the considered low-
−26 complexity receiver architecture. Moreover, this knowledge
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 can also be very useful in designing transmit signals used in
Code Index other systems like coded continuous-wave radars and sonars.
Future work should examine if PAFs of other codes with
Fig. 10. Different attained PAF sidelobe levels with varying circular shifts
for IEEE 802.15.4 HRP UWB PHY length 127 codes. favorable PACR properties, like m-sequences or Zadoff-Chu
sequences, also exhibit similar variation of peak sidelobe level
with circular shift.

65
R EFERENCES
∠χ(0, ν) for  = 40 ppm (deg)

[1] V. P. Ipatov, “Ternary sequences with ideal autocorrelation properties,”

Radio Engineering and Electronic Physics, vol. 24, pp. 75–79, Oct 1979.
60 [2] T. Hoholdt and J. Justesen, “Ternary sequences with perfect periodic
autocorrelation (Corresp.),” IEEE Transactions on Information Theory,
vol. 29, no. 4, pp. 597–600, July 1983.
[3] V. P. Ipatov, Spread Spectrum and CDMA: Principles and Applications,
1st ed. Wiley, 2005.
55 [4] Amendment 1: Add Alternate PHYs, IEEE Amendment to IEEE Std
Maximum attained 802.15.4-2006 802.15.4, Rev. 802.15.4a-2007, Aug 2007.
Current [5] IEEE Standard for Low-Rate Wireless Networks, IEEE Standard
Min Peak SLs 802.15.4, Rev. 802.15.4-2015, April 2016.
50 [6] F. Leong, “Text proposal for HRP UWB SRDEV PPDU,”
Minimum attained IEEE 802.15 TG4z, Tech. Rep. 15-18-0211-00-004z, May 2018.
[Online]. Available: https://mentor.ieee.org/802.15/dcn/18/15-18-0211-
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 00-004z-text-proposal-for-hrp-uwb-srdev-ppdu.docx
[7] B. Verso, “Proposed enhancements to the HRP UWB PHY,”
Code Index IEEE 802.15 TG4z, Tech. Rep. 15-18-0108-04-004z, May 2018.
[Online]. Available: https://mentor.ieee.org/802.15/dcn/18/15-18-0108-
Fig. 11. Different attained PAF main lobe phases with varying circular shifts 04-004z-hrp-uwb-phy-enhancements.pptx
for IEEE 802.15.4 HRP UWB PHY length 127 codes. [8] B. Verso, F. Leong, J. Hammerschmidt, J. Niewczas, B. Ibrahim,
T. Shah, A. Marquez, T. Reisinger, and D. Knobloch, “HRP
UWB PHY enhanced mode converged consensus,” IEEE 802.15
and 127 codes PAF peak sidelobes levels vary significantly TG4z, Tech. Rep. 15-18-0590-00-004z, Nov 2018. [Online].
Available: https://mentor.ieee.org/802.15/dcn/18/15-18-0590-00-004z-
with the circular shift, whereas PAF RMS sidelobes levels hrp-uwb-phy-enhanced-mode-converged-consensus.pptx
vary much less. This is similar to what has been observed [9] C. McElroy and M. McLaughlin, “Receiver for use in an ultra-wideband
with Ipatov-91 codes. communication system,” U.S. Patent US10 177 809B1, Jan 8, 2019.
[10] I. Dotlic, A. Connell, and M. McLaughlin, “Ranging Methods Utiliz-
On the other hand, it is clear from Figs. 9 and 11 that ing Carrier Frequency Offset Estimation,” in 2018 15th Workshop on
PAF main lobe phase for both length 31 and 127 codes varies Positioning, Navigation and Communications (WPNC), Oct 2018, pp.
much more compared to Ipatov-91 codes. The reason for this 1–6.
[11] A. Freedman and N. Levanon, “Properties of the periodic ambiguity
is the fact the phase of the PAF main lobe depends on the function,” IEEE Transactions on Aerospace and Electronic Systems,
position of zeros in the ternary code (9). Both length 31 and vol. 30, no. 3, pp. 938–941, July 1994.
length 127 codes have close to 50% of zeros. Hence, circularly [12] N. Levanon, “The periodic ambiguity function — Its validity and value,”
in 2010 IEEE Radar Conference, May 2010, pp. 204–208.
shifting the position of zeros influences the main lobe phase