Wireless Vehicular Networks in Emergencies: A Single Frequency Network Approach

Wireless Vehicular Networks in
Emergencies: 
A Single Frequency Network Approach
Andrea Tassi - a.tassi@bristol.ac.uk
Malcolm Egan - Université Blaise Pascal, Clermont-Ferrand, FR 
Robert J. Piechocki and Andrew Nix - University of Bristol, Bristol, UK
Da Nang, Vietnam, 9th
January 2017
University of Bristol 
Communication Systems and Network Group
SigTelCom 2017

I n d e x
1. LTE-A and the Single Frequency Network Infrastructure
2. The eMBMS Framework for Vehicular Emergencies
3. Performance Modeling and Design Optimisation
4. Numerical Results
5. Conclusions

1. LTE-A and the Single Frequency Network
Infrastructure

S t a n d a rd LT E - A S F N I n f r a s t r u c t u re
BS
BS
BS
BS
M1/M2
(MCE / MBMS-GW)
SFN
4
1
2
3
UE3
UEUUE2
UE1
UE4
LTE-A Core Network
• Multiple neighboring BSs (forming the SFN) transmit the same Point-to-
Multipoint (PtM) data streams in a synchronous fashion.
• This transmission mode has become increasingly common in 4G systems,
where it is also known as the SFN-eMBMS.
• SFNs have already proved effective in vehicular communication systems.
Multicell
Coordination
Entity (MCE)

2. The eMBMS Framework for Vehicular
Emergencies

P ro b l e m M o t i v a t i o n
• The IEEE 802.11p/DSRC can achieve at most ~27 Mbps, in practice it is
hard to observe that.
• However, DSRC standards are suitable for low-rate data services (for e.g.,
positioning beacon, emergency stop messages, etc.).
• On the other hand, future CAVs will require solutions ensuring megabit-
per-second communication links to achieve proper ‘look-ahed’ services
(involving cameras, LIDARS, etc.), etc.
• The LTE-A infrastructure is already deployed in our cities.

O u r P ro p o s a l
• Municipality-owned SFN that provides emergency coverage to a small area
of a city.
• The SFN serves a target cluster of vehicles to ensure that each vehicle can
reliably receive information to support improve road safety.
• Each base station (possibly battery-powered) in the SFN is equipped with
an antenna array with a highly directional beam. We assume that the
beamwidth of the main lobe is only sufficient to cover the target cluster.
• The SFN operates on the same frequency of an operator-owned network.

O u r P ro p o s a l
Major safety hazard
Center of the
target cluster
Interfering
Base Station
SFN Base Station
• Vehicles and interfering base stations are equipped with isotropic antennas.
• Each vehicle provides its location to the SFN controller via the nearest base
station. The SFN controller can estimate the center of the target cluster.

3. Performance Modeling and Design
Optimisation

B S D i s t r i b u t i o n
• Positions of interfering BS positions
follow a 2D PPP
• Interfering BSs can be in LOS (with
prob. pL) or NLOS (with prob. pN)
with the center of the cluster.
• SFN base stations assumed in LOS
with the center of the cluster
x-coordinate ·10−3
y-coordinate·10−3
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2
−2
−1.5
−1
−0.5
0
0.5
1
1.5
2
Interfering BS
SFN BS
Center of the
Target Cluster

T h e P ro b a b i l i t y F r a m e w o r k
• We define the SINR at the center of the cluster as
hj ~ EXP(1)
PL,
thermal noise
power
• We characterize the SINR outage as follows
SINRO =
GS,TX GRX
MX
i=1
Pi hi `(S)
(dS,i)
W + GI,TX GRX PI
X
j2
hj `(I)
(dI,i)
d ↵
PT(✓) = P [SINRO < ✓] = P
" MX
i=1
Pi hi d ↵L
S,i > ✓
W + I
GS,TX GRX
#
Inst. TX pow.

S I N R O u t a g e P ro b a b i l i t y
• are indep. exponentially distributed RV with mean
PT(✓) = P [SINRO < ✓] = P
" MX
i=1
Pi hi d ↵L
S,i > ✓
W + I
GS,TX GRX
#
{Pi hi d ↵S
S,i }M
i=1
µi = Pi d ↵S
S,i
• The cumulative distribution function of a sum of exponentially distributed
random variables is [*]:
F(z) =
aY
j=1
µ ok
j
aX
k=1
okX
`=1
µok `
k
k,` µ 1
k zok `
e z/µk
(ok `)!(` 1)!
.
[*] S. Amari and R. Misra, “Closed-Form Expressions for Distribution of Sum of Exponential
Random Variables,” IEEE Trans. Rel., vol. 46, no. 4, pp. 519–522, Dec. 1997.

• The cumulative distribution function of a sum of exponentially distributed
random variables is:
F(z) =
aY
j=1
µ ok
j
aX
k=1
okX
`=1
µok `
k
k,` µ 1
k zok `
e z/µk
(ok `)!(` 1)!
.
k,` (t) =
@` 1
@t` 1
8
<
:
1
t
aY
j=1,j6=k
✓
1
µj
+ t
◆ oj
9
=
;
⌦k,` (t) = ( 1)ok ` @ok `
@xok `
(
e
µ
1
k
✓W
GS,TX GRX
x
LI
✓
µ 1
k ✓I
GS,TX GRX
x
◆ )
x=1
.
often these
terms refers to  
“0-derivatives”

PT(✓) = P [SINRO < ✓] = P
" MX
i=1
Pi hi d ↵L
S,i > ✓
W + I
GS,TX GRX
#
PT(✓) = EI

F
✓
✓
W + I
GS,TX GRX
◆
(a)
=
aY
j=1
µ ok
j
aX
k=1
okX
`=1
µok `
k
k,` µ 1
k
(ok `)!(` 1)!
·
µok `
k EI0
⇥
Uok `
e U
⇤
U µ 1
k ✓ W+I
GS,TX GRX

S I N R O u t a g e a n d R a t e C o v e r a g e P ro b .
• … after some manipulations we optain:
PT(✓) =
aY
j=1
µ ok
j
aX
k=1
okX
`=1
µok `
k
k,` µ 1
k ⌦k,` µ 1
k
(ok `)!(` 1)!
,
• From above we define the rate coverage probability as
RC() = P[ log2(1 + ·SINRO) > ] = P
⇥
SINRO > 2

1
⇤
system BW

P o w e r A l l o c a t i o n M o d e l
(PA) min
P1,...,PM
MX
i=1
Pi (1)
subject to PT(ˆ✓)  ˆT (2)
0  Pi  ˆP i = 1, . . . , M. (3)
• The sum of the TX pow. is minimised.
• Eq. (2) provides a QoS constraint, while Eq. (3) is a design constraint
• Easily solvable by means of a water-filling strategy. For further details please
refer to A. Tassi, I. Chatzigeorgiou and D. Vukobratović, "Resource-
Allocation Frameworks for Network-Coded Layered Multimedia
Multicast Services," in IEEE Journal on Selected Areas in
Communications, vol. 33, no. 2, pp. 141-155, Feb. 2015. Andrea Tassi - a.tassi@bristol.ac.uk

4. Numerical Results

C o n s i d e re d S c e n a r i o
• Simulated scenario with radius
equal to 1E3 m
• System BW equal to 50 MHz
• TX pow. of the interfering BSs equal
to 10 W
• Max. TX pow. of an SFN BS set
equal to 30 W
x-coordinate ·10−3
y-coordinate·10−3
−2 −1.5 −1 −0.5 0 0.5 1 1.5 2
−2
−1.5
−1
−0.5
0
0.5
1
1.5
2
Interfering BS
SFN BS
Center of the
Target Cluster

θ (dB)
PT(θ)
6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5 22.5 24.5 26.5 28.5 30.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
λBS = 0.1 · 10−5
λBS = 0.2 · 10−5
λBS = 0.3 · 10−5
Simulation
Theory

λBS · 105
PT
0.1 0.2 0.3 0.4 0.5 0.6
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
θ = 10.3 dB
θ = 15.3 dB
θ = 20.3 dB
θ = 25.3 dB
Simulation
Theory

PA M o d e l
ˆθ (dB)
M
i=1P∗
i(W)
6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5
0
10
20
30
40
50
60
λBS = 0.1 · 10−5
λBS = 0.2 · 10−5
λBS = 0.3 · 10−5
λBS = 0.4 · 10−5
λBS = 0.5 · 10−5
λBS = 0.6 · 10−5

PA M o d e l
λBS · 105
M
i=1P∗
i(W)
0.1 0.2 0.3 0.4 0.5 0.6
0
10
20
30
40
50
60
ˆθ = 6.3 dB
ˆθ = 8.3 dB
ˆθ = 10.3 dB
ˆθ = 12.3 dB
ˆθ = 14.3 dB

5. Conclusions

F i n a l R e m a r k s
• We characterised the performance of a SFN suitable for vehicular
emergencies
• We obtained performance guarantees of SFNs in terms of bounds on
outage probabilities using techniques from stochastic geometry.
• These bounds form a basis for optimizing the power allocation of each
base station in the SFN, which is important when these base stations rely
on off-grid power sources.
• In the considered scenarios, we have shown that the proposed PA model
can ensure and overall transmission power footprint that: (i) can be up to
20 times smaller than a static PA solution, and (ii) meets target SINR outage
constraints. Andrea Tassi - a.tassi@bristol.ac.uk

Wireless Vehicular Networks in
Emergencies: 
A Single Frequency Network Approach
Malcolm Egan - Université Blaise Pascal, Clermont-Ferrand, FR 
Robert J. Piechocki and Andrew Nix - University of Bristol, Bristol, UK
Da Nang, Vietnam, 9th
January 2017
University of Bristol 
Communication Systems and Network Group
SigTelCom 2017
Thanks for your attention!

Wireless Vehicular Networks in Emergencies: A Single Frequency Network Approach

More Related Content

What's hot (17)

Similar to Wireless Vehicular Networks in Emergencies: A Single Frequency Network Approach (20)

More from Communication Systems & Networks (19)

Recently uploaded (20)

Wireless Vehicular Networks in Emergencies: A Single Frequency Network Approach