Content uploaded by Elizabeth Eso
Author content
All content in this area was uploaded by Elizabeth Eso on Apr 20, 2022
Content may be subject to copyright.
1
Abstract—This paper investigates a vehicle-to-vehicle relay-
assisted visible light communications link with amplify-and-
forward (AF) and decode-and-forward (DF) relaying schemes and
analyze the effect of the relay vehicle’s orientation with respect to
the source vehicle. Results show up to 150 % increase in the link
span under emulated sunlight for the DF relay link compared with
AF at the end of the second hop using the same system parameters
at the forward error correction bit error rate limit of 3.8×10-3. A
method to mitigate the effect of system performance degradation
due to misalignment between communicating vehicles is also
proposed.
Index Terms—vehicle-to-vehicle, relay-assisted, visible light
communications, relaying schemes, vehicle’s orientation
I. INTRODUCTION
Visible light communications (VLC) uses luminance as a
medium for data transmission using light emitting diodes
(LEDs) and has attracted much attention for data
communications and localization [1]. Currently, the radio
frequency wireless technology is the established
communication technology for intelligent transport system
(ITS), known as the dedicated short-range communications
(DSRC), which supports several applications such as
emergency braking warning, and intersection collision warnings
[2]. However, there are some issues regarding DSRC in meeting
the low latency and high-reliability requirements in ITS
considering both network outages and security issues [3]. Most
recently, VLC has been proposed in ITS for the exchange of
safety-based information between two or more vehicles and the
roadside infrastructures i.e., traffic and streetlights.
To extend the communications range in wireless systems, the
use of relay nodes (RNs) has been proposed in [4, 5]. In [4],
relay-based systems were shown to be viable in mitigating
channel fading over short transmission ranges, with an
enhanced power margin by up to 18.5 dB using a single RN at
a target outage probability of 10-6. In [5], the viability of VLC
using multiple taillights (TLs) transmitters (Txs) and
photodiodes (PDs)-based receivers (Rxs) with orthogonal
frequency division multiplexing for direct transmission (DT)
and multi-hop transmission (MHT) was reported. It was shown
The work is supported by the EU’s Horizon 2020 research and innovation
programme under the Marie Sklodowska-Curie grant no.764461 (VISION), EU
COST Action on Newfocus (CA19111) and project SGS20/166/OHK3/3T/13.
Elizabeth Eso, Zabih Ghassemlooy and Juna Sathian are with the Optical
Communication Research Group, Faculty of Engineering and Environment,
Northumbria University, UK.
Stanislav Zvanovec and Petr Pesek are with the Department of
Electromagnetic Field, Faculty of Electrical Engineering, Czech Technical
that, with spatial multiplexing MHT offered reduced average
signal power of -59 dBm for a single hop (for link span of 16
m) compared with -27 dBm for DT (for of 12 m). Multi-user
VLC serving as a RN for other users was introduced in [6]. A
similar method with vehicles acting as RNs was proposed in [7]
for inter-vehicle communication systems.
Recently, an experimental investigation of a relay-assisted
vehicular VLC (VVLC) link was reported in [8] showing that,
the decode and forward (DF) relay scheme is the preferred
option. Consequently, in this paper, we extend the work in [8]
by further investigating relay-assisted VVLC for a range of
transmit power Pt and longer using a Monte Carlo system-
level-based simulation model, which allows considering
different communication geometries and parameters. Also, we
investigate the orientation of the RN in terms of the azimuth
(horizontal)/elevation (vertical) angles of the source vehicle.
Note, in VVLC using a combination of an imaging lens and PDs
to increase the received optical power density, the link will
experience beam spot offset (BSO) (i.e., an offset in the
projected beam spot from the centre of the focal plane away
from the PD) when there is a misalignment between the Tx and
Rx. This is due to non-stationary vehicles positions and varying
positions of TLs (Txs) and Rxs depending on the vehicle
models. The contributions of the work carried out are (i)
investigation and providing insights on the effect of
misalignment between vehicle’s Txs and Rxs in real road
conditions on the bit error rate (BER) performance of the system
resulting from the BSO on the focal plane, which is carried out
for the first time to the authors best knowledge; (ii) investigating
relay-assisted VVLC links with amplify and forward (AF) and
DF relay schemes; (iii) mitigation of the misalignment issue
considering typical vehicular geometry parameters; and (iv)
providing insights to the impact of various parameters such as
PD size, incidence angle and , on the performance of VVLC.
II. SYSTEM MODEL
A. The VLC Channel and Noise Parameters
For the line of sight (LOS), the channel DC gain and
received signal [9] can be expressed, respectively as:
University, Prague, Czech Republic. (Corresponding author: Elizabeth Eso,
email: elizabeth.eso@northumbria.ac.uk)
Copyright (c) 2019 IEEE. Personal use of this material is permitted.
However, permission to use this material for any other purposes must be
obtained from IEEE by sending a request to pubs-permissions@ieee.org.
Vehicle-to-Vehicle Relay-Assisted VLC with
Misalignment Induced Azimuth or Elevation Offset
Angles
Elizabeth Eso, Member, IEEE, Zabih Ghassemlooy, Senior Member, IEEE, Stanislav Zvanovec,
Senior Member, IEEE, Petr Pesek, and Juna Sathian
2
(1)
(2)
where is PD’s responsivity, h(t) is the channel impulse
response, and n(t) is the additive white Gaussian noise including
the sunlight current induced shot noise source, thermal noise,
signal dependent shot noise and the PD’s dark current noise with
noise variances of
,
,
, and
, respectively.
, , and are the incidence angle, Rx’s
angular field of view (AFOV) semi-angle, transmittance of the
optical concentrator (OC) and optical filter (OF), respectively.
is the area of the projected focused image of the light
source by the imaging lens, which depends on the actual area of
the light source , the focal length of the lens f, and the ,
which can be expressed as
. Note that, the
dominant noise source in VVLC is
[10]. For
other noise sources, we have
, and
[10]. The total variance:
,
(3)
where , , , , and are the electron charge,
system bandwidth, Boltzmann’s constant, absolute temperature
in Kelvin, load resistance, dark current and the photocurrent
generated due to the ambient sunlight, respectively and
, where is the data rate.
III. SYSTEM SETUP
The schematic system block diagrams of the proposed
amplify-and-forward (AF) and DF relay-assisted VVLC
systems are shown in Figs. 1(a) and (b), respectively. It is
simulated using Monte Carlo model and considers different
communication geometries and system parameters for
evaluating the link performance in terms of the BER. At the Tx,
a 3×105 random bit stream in the non-return-to-zero on-off
keying (NRZ-OOK) data format is generated, which is
upsampled (number of samples/bit = 10) prior to intensity
modulation of the Tx. The modulated signal is transmitted over
the 1st free space channel to the RN. Noise i.e., emulated
sunlight induced noise obtained from empirical measurements,
which is the dominant source in VVLC systems during daytime,
and all other noise sources outlined in Section II are added to
the received signal. At the RNs and Rxs, OFs (a wavelength
band of 405-690 nm) and OCs with the f of 15 mm
(transmittance of 0.92 for both OF and OC) are used to improve
the signal to noise ratio. The has RNs at 25 and 50 m,
has RNs at 25 and 65 m. For the AF relay link, at RNs the
regenerated electrical signal is amplified to its original power
level prior to transmission to the next RN and finally to the end
user (Fig. 1(a)). For the DF relay link, the regenerated electrical
signal is applied to the matched filter (MF) and decoder prior to
packet generation. Note, the detection process is the same at the
DF RN and end user, where MF is composed of integrator,
sampler (with a sampling interval ts= , and slicer (the
threshold level set to the mean value of sampled filtered signal).
The decoder sets the signals from the output of the slicer that
are below the threshold level to 0 and else to 1. The system
parameters are = 298 K, = 13.3 mA, B=5 MHz, =
5 nA,= 50 and =10 Mb/s. is calculated from the
solar irradiance measurement given in [11]. Other specific
system parameters used are stated in Section IV.
IV. RESULTS
Here, we present results for the proposed VVLC with RNs
and with
. The BER as a function of for a range of is
presented in Fig. 2, where Txs and Rxs are in alignment and
have diagonal widths and of 5 cm and 1 mm,
respectively (same PD type is used in all Rxs). As shown in Fig.
2, BER improves with a 0.25 W increase in (i.e., 0.75 to 1 W
and 1 to 1.25 W) by a drop of up to 68 and 89 %, for the AF and
DF links, respectively at the end of the 2nd channel. Note, for
error free transmission, the BER floor level is set to 10-6 (in Figs.
2 and 7). Notably, links with DF show a more stable BER
compared with AF but starts to degrade after certain
depending on the inter hop distance. For comparison, at of
0.75 W, links are shown with and without
with only of
21 and up to 66 m achieved, respectively at the forward error
correction (FEC) BER limit of 3.8×10-3. Moreover, at the FEC
BER limit at the end of the 2nd hop and for = 1 W, is 150
% higher for DF compared with AF under emulated sunlight
noise. This is because, in an AF relay-based link the system
Fig. 1: System block diagram for the VVLC relay-assisted links with: (a) AF, and (b) DF relay schemes (just relay node (RN) shown)
Packet
generator
Tx1
Tx2
Source
Rx-signal
Tx1
Tx2
Amplifier
OF OC
AF relay node
n(t)
End user
Rx-signal MF
OF OC
Rx-signal MF
OF OC
DF relay node
n(t)
Tx2
Tx1
Packet
generator
Decoded
bits
Decoded
bits
(a)
(b)
AF relay
DF relay
n(t)
Channel Channel
3
performance is affected more by the noise accumulation over
the transmission span since the received signal at the RN is
simply amplified with no regeneration or reshaping in contrast
to the DF. Note, in VVLC systems (i) for practical
implementation compact devices such as field programmable
gate arrays can be used for encoding/decoding,
modulation/demodulation driver for the Tx, Rx, etc., at various
stages; and (ii) maintaining the exact alignment between the Tx
and the Rx is not possible with and without RNs due to vehicles
movement, which result in variations in , see Fig. 3. Hence,
links using an imaging lens at the Rx will experience BSOs in
the horizontal, vertical or both directions on the focal plane at
the PD, see Fig. 4. The horizontal incidence angle for at least a
Tx to be captured for a road lane width is given by:
Fig. 2: BER performances as a function of for a range of
,
(4)
where , , , , , and are the widths
of the Tx and Rx vehicles, distance between the centres of Txs
1 and 2 and Rxs 1 and 2, distance of Tx and Rx vehicles from
the edge of road lane and the length of the Tx, respectively. On
the focal plane at the PD and using a converging rectilinear lens,
the maximum BSOs is given as:
,
(5)
(6)
where the offset between the Tx’s and Rx’s vertical heights is
= , is the vertical incidence angle,
and are the horizontal and vertical BSOs on the focal
plane. Using (4)-(6), and = 1.7 m, and
=1.2 m, and , the maximum (i.e., the worst
case scenario for inter-vehicle communications at the opposite
edge of the road lane) for = 2.5 and 3.75 m (which are the
minimum and maximum for a lane) are 21.8 and 45.7o,
respectively and the maximum for = 0.5 m for
m is 14o. Consequently, Fig. 5(a) depicts the plots of BSOs as
Fig. 3: VVLC configuration (a) top view showing Tx-Rx horizontal incidence
angle and (b) side view showing Tx-Rx vertical incidence angle
Fig. 4: (a) Parallel, and (b) non-parallel light beams through a converging lens
a function of for a range of f, which shows that, BSOs
exponentially increase with for all values of f. E.g., for f = 35
mm, BSO is increased by more than 4 times to 12.8 mm for
increased from 5 to 20o, which is orders of magnitude larger
than the diagonal width of a high-speed PD (e.g., 0.1 mm).
Notably, AFOV of VLC depends on f of the lens and the size of
the PD as described in [11] and depicted in Fig. 5(b). Note, the
width of the projected focused Tx’s image (i.e., the beam spot
size (BSZ)) in terms of the Tx’s diagonal widths is
[12]. Thus, the size of Tx’s beam spot is a critical
factor in considering BSZ- and BSO-induced power losses at
the PD, see Fig. 6. As shown BSZ plots exponentially drop with
for all values of and increases with f, reaching below
Fig. 5: (a) BSO vs. and (b) AFOV as a function of for a range of f
2 mm at > 10 m and converging to < 1 mm for of 25 m.
Note, at shorter , BSZ is in orders of magnitudes larger than
at longer E.g., at f of 50 mm, = 15 cm, and at of 5 and
(a)
(b)
f
Converging lens
(rectilinear)
Light rays
f
Converging lens
(rectilinear)
Light rays
(a)
(b) bhor-of
bver-of
4
20 m the BSZs are 1.5 and 0.375 mm, respectively. Also, the
link performance is not degraded due to BSO provided the PD
size is greater than the BSO and BSZ to ensure capturing of the
projected light beam spot. Note, larger size PDs can be used,
since the data transmission rate is low in VVLC [12].
Furthermore, for the first time to the authors’ best knowledge,
we investigate the effect of the orientation of the relay vehicle
with respect to the source vehicle considering BSO due to the
on the BER performance at the first RN. System parameters
used are = 1.25 W, f = 15 mm and =5 cm. Note that, for
same horizontal/vertical misalignment offset distance on the
road between the Tx and Rx, will be higher at shorter . Fig.
7 shows the BER as a function of (angle of incidence with
Fig. 6: Beam spot size at the focal plane versus for a range of f and
respect to Tx1 and Rx1 (Fig. 3), which will be same for Tx2
with respect to Rx2) for a range of and . As can be
observed, BER plot shows step responses i.e., rapid degradation
due to resulting from BSO. Note, the centre of the road lane
is considered as the origin of the cartesian and spherical
coordinate systems for calculating the positions of the Tx and
Rx, and incidence angles, where values to the left and right
of the road lane are positive and negative, respectively.
Moreover, the system performance is highly dependent on the
geometry of the Tx and the Rx. As can be seen, the Rx with a
larger allow higher degrees of misalignment between the
Tx and the Rx i.e., larger . Note, at certain values of the BER
tends to0.5 as the optical signal or the beam spot from the lens
is not captured by the PD (only the ambient light is). At certain
and the following capturing scenarios are possible: (i)
both Txs; (ii) a single Tx only; (iii) part of the Tx(s); and (iv) no
Txs being detected by the PD. To minimize/overcome this
effect, we propose the use of a spherical PD array Rx to provide
a range of varying angular PD orientations with respect to the
Tx with an azimuth and elevation curvature deflection angles of
91.4 o (i.e. to cover value of 45.7o to the right or left) and
28o (for coverage of value of 14o to the top or bottom)
considering a minimum of 2 m, = 0.5 m and = 3.75 m,
which is the largest width for the single road lanes. Note, with
this method, lower values of can be achieved using a spherical
PD array, as lower results in lower BSO see Fig. 5(a) and
consequently a minimum BER degradation due to BSO.
V. CONCLUSION AND FUTURE WORK
We investigated a vehicle-to-vehicle VLC link with AF and
DF relaying schemes and results showed that DF is a potential
option providing up to 150 % increment in the achievable link
distance than AF at the FEC BER limit. Furthermore, we
investigated the effect of misalignment between vehicles in
Fig. 7: BER degradation as a function of for a range of and
real road conditions on the BER performance. Results showed
a sharp BER degradation after a certain range of , but a 10
times improvement in this achievable range at certain values
(i.e., when the beam spot is incident on the PD) as the PD size
increased from 0.1 to 1 mm. Consequently, to mitigate this
effect, we proposed the use of a spherical multi-PD array Rx
with the following merits (i) a range of varying PD angular
orientations that can minimise the misalignment offset angle;
and (ii) a multi-PD array providing a larger surface area (like a
camera-based Rx with several PDs) compared to a single PD-
based Rx. The future scope is to also look at the effect of the
non-LOS signal and multi-hop links.
REFERENCES
[1] P. Sharda and M. R. Bhatnagar, "Diversity-Multiplexing Tradeoff for
Indoor Visible Light Communication," 2020 16th Int. Conf. on Wireless
and Mobile Comp., Netw. and Comms. (WiMob), Greece, 2020, pp. 1-6.
[2] S. Gao, A. Lim, and D. Bevly, "An empirical study of DSRC V2V
performance in truck platooning scenarios," Dig. Comms. and Networks,
vol. 2, pp. 233-244, 2016.
[3] A. Memedi, C. Tebruegge, J. Jahneke and F. Dressler, "Impact of Vehicle
Type and Headlight Characteristics on Vehicular VLC
Performance," 2018 IEEE VNC, Taiwan, 2018, pp. 1-8.
[4] M. Safari and M. Uysal, “Relay-assisted free-space optical
communication,” IEEE Trans. Wireless Commun., vol. 7, no. 12, pp.
5441–5449, Dec. 2008.
[5] O. Narmanlioglu, B. Turan, S. C. Ergen, M. Uysal, Cooperative MIMO-
OFDM based inter-vehicular visible light communication using brake
lights, Computer Communications, Volume 120, 2018, Pages 138-146.
[6] Z. Wu, Free space optical networking with visible light: A multi-hop
multi-access solution. Boston University, 2012.
[7] C. B. Liu, B. Sadeghi, and E. W. Knightly, “Enabling vehicular visible
light communication (V2LC) networks,” in Proc. 8th ACM int. Workshop
on Vehicular Inter-Netw. ACM, 2011, pp. 41–50.
[8] E. Eso, P. Pesek, P. Chvokja, Z. Ghassemlooy, S. Zvanovec and J. Sathian.
‘A Relay-Assisted Vehicular Visible Light Communications Network’.
IEEE Photonics Conference, 2020.
[9] J. M. Kahn and J. R. Barry, "Wireless infrared communications,
"Proceedings of the IEEE, vol. 85, pp. 265-298, 1997.
[10] R. Hui ‘Introduction to Fiber-Optic Communication’ Academic Press, pp.
125-154, 2020
[11] E. Eso, Z. Ghassemlooy, S. Zvanovec, J. Sathian and A. Gholami,
"Fundamental Analysis of Vehicular Light Communications and the
Mitigation of Sunlight Noise," in IEEE Trans. on Veh. Tech., 2021
[12] E. Eso, S. Teli, N. B. Hassan, S. Vitek, Z. Ghassemlooy and S. Zvanovec,
"400 m Rolling Shutter based Optical Camera Communications Link.
Optics Letters, 2020.
