-
Information Dissemination Dynamics through
Vehicle-to-Vehicle Communication over
Transportation Networks.
Alobeidyeen, Ala, Ph.D. student, University of Florida
Du, Lili Associate Professor, University of Florida
IEEE ITSC, 10/30/2019
Auckland, New Zealand
-
Background and Motivation
Vehicle to Vehicle (V2V)
DSRC protocol; Hundreds of meters (
-
Research Objectives
Develop a discrete mathematical simulation model to track
the
information spreading dynamics over a transportation network
(road
segments, intersections).
Capture the information coverage dynamics over the
transportation
network.
Explore the correlation between information spreading
dynamics
and traffic congestion.
3Ala and Du @ University of Florida
-
Preliminaries
Information Front and Wave on Network
Information-spreading front over network
Intersection:
Front disseminates to different directions
On road segments:
Follow intermittent transmission pattern
Form the information wave with two
boundaries (fronts/tails).
tail π
πππππ‘ π
π πππ‘π‘πππ
π€ππππππ π πππππ¦
Network
Fronts and tails forms the spreading
boundary of the wave.
4Ala and Du @ University of Florida
-
Methodology: IT-CTM for Road Segment
Information-Traffic Coupled Cell Transmission Model (IT-CTM)
Ξπ₯
β¦ {ππ π‘ , ππ π‘ , ππ π‘ }πΆπππ j
Each cell keeps three pieces of information including: passing
front, tail
and traffic volume: {ππ π‘ , ππ π‘ , ππ π‘ }
πΆπππ jπΆπππ j-1 πΆπππ j+1
Inner-cell front/tail propagation and delay: forward, backward
or stay in a cell
(depends on intermittent transmission state).
IT-CTM cell with Ξπ₯, covering
several intermittent patterns.
Inter-cell front/tail propagation: forward, backward or stay
(depends on the
states of two adjacent cells).
5Ala and Du @ University of FloridaLili Du and Hoang Dao (2014).
Information Dissemination Delay in Vehicle-to-Vehicle Communication
Networks in a Traffic Stream. IEEE Transactions on Intelligent
Transportation Systems 16(1), 66-80.
Lili Du, Siyuan Gong, Lu Wang, and Xiang-Yang Li (2016).
Information-Traffic Cell Transmission Model for Information
Coverage Dynamics over V2V Communication
Network on Road Segments. Transportation Research Part C:
Emerging Technologies, 73,Pages 30-48.
-
Methodology: Information Flow Network Model for
arterial intersection (IFNM-a)
The existing flow at each arm,
Various vehicle mobility,
Combinations of wireless and ferry
transmissions.
front
Hard to
enumerate
Y:Wireless transmission and delay
X: Ferry transmission and delay
IFNM-a for
Arterial
Intersection
Network components
Nodes
Links (0-1 state with delay)
Many potential
routes
Depends on
IT-CTM cells adjacent to intersection
Information front will always spread by taking the fastest/
shortest path.
6Ala and Du @ University of Florida
-
Methodology: Information Flow Network Model at
Highway-Ramp Intersection (IFNM-r)
The Front may take many routes
Mixed transmissions:
IFNM-r for a Highway-Ramp Intersection
Scope of the network
neither IT-CTM nor
IFNM-a works
Y: wireless transmission.
X: ferry transmission.
Z: intra-cell transmission between IT-CTM
cells at the connection of highway and ramp.
ππ
ππ
ππ
front from
mainline
upstream
front from
mainline
downstreamfront from
ramp
π½
π1 =ππ
π ππ π, π2 =
ππ
π ππ π, π3 = ππ
ππππ ππ
7Ala and Du @ University of Florida
-
Information flow network model Coupled with IT-CTM
(IFNM-CTM)
IFNM-a tracks the information spreading dynamics at arterials
intersection
(un/signalized)
IFNM-r tracks the information spreading dynamics at various
intersection (highway-
ramp)
IFNM-CTM integrate the IFNM-a, IFNM-r and the IT-CTM; tracks the
information
spreading dynamics over a network
A small Network example
Highway-ramp Arterial intersectionIFNM-aIFNM-r
αnodes: IT β CTM cells
LInk: {Y, X}αnodes: IT β CTM cells
LInk: {Y, X, Z}
Methodology: IFNM-CTM for Network Spreading
8Ala and Du @ University of Florida
-
Methodology: Coverage dynamics (Machine learning)
Information coverage Matrix
Front coverage matrix
Present road network by IT-CTM cells.
Coverage Matrices: Front (F) tail (Ξ€) 1: front/tail passed cell
j
0: Otherwise
Searching connected clusters
Track cell clusters by depth first
search (DFS).
Determine the numbers of clusters.
Size of each cluster (how many cells
inside each cluster).
9Ala and Du @ University of Florida
-
Results: Spreading Dynamics
Information coverage dynamics
The avg MAE for the locations of
15 fronts < 6% for 132 runs.
10Ala and Du @ University of Florida
2
3 4 5
8 7
12
14
11 10
15 19
20212413
1
23 22
6
9
17 18
β¦
β¦
β¦
β¦
β¦
β¦β¦
β¦
β¦β¦
π0 @ π‘=1
π2 @ π‘=3
π1 @ π‘=2
π6 @ π‘=7
π3 @ π‘=4π4 @ π‘=5π5 @ π‘=6
π7 @ π‘=8π8 @ π‘=9π9 @ π‘=10π10@ π‘=11π11@ π‘=12π12@ π‘=13π13@
π‘=14π14@ π‘=15π15@ π‘=16
πΆπππ‘π π‘1 = 0.092
πΆπππ‘π π‘2 = 0.132
πΆπππ‘π π‘3 = 0.194
πΆπππ‘π π‘4 = 0.289
πΆπππ‘π π‘5 = 0.322
πΆπππ‘π π‘7 = 0.451
πΆπππ‘π π‘8 = 0.531
πΆπππ‘π π‘9 = 0.620
πΆπππ‘π π‘6 = 0.378
πΆπππ‘π π‘10 = 0.683
πΆπππ‘π π‘11 = 0.715
πΆπππ‘π π‘12 = 0.731
πΆπππ‘π π‘13 = 0.787
πΆπππ‘π π‘14 = 0.827
πΆπππ‘π π‘15 = 0.91
πΆπππ‘π π‘16 = 1
Testbed: Sioux Falls city network
24 intersections; 38 link
8, 10, 11, 16, 14, 15, 23, 22
18, 7, 20, 13, 2, 3, 12
Signalized arterial
Metering ramp
Highway Urban roads
VISSIM simulation for traffic data 45 mins; 13 mins warm up
-
Information coverage obtained from the simulation and the
mathematical
model of IFNM-CTM are very close at each individual time
stamp
The average MAE is 4.72% with a very small standard deviation
equal to
0.01152
11
Results: Coverage Dynamics
Ala and Du @ University of Florida
-
12
Results: Congestion and Information
Coverage Correlation
The number (a) and size (b) of clusters under congestion
CG(t) and passed by information front CF(t)
The similarity/consistence of the
cells in informed clusters CF(t) and the congested cells CG(t)
is 80% on the average.
The number and size of the
informed clusters CF(t) and the congested cells CG(t) show high
correlation
IFNM-CTM approach to capture
the Congestion and Information
Spreading Correlation over time in
a roadway network
Ala and Du @ University of Florida
-
Conclusions & Summary
Understanding information spreading dynamics in
transportation
network is essential to sustain the mobility application of
V2V
technology.
Our approach IFNM-CTM (IFNM-a, IFNM-r, IT-CTM) can capture
information spreading over transportation network.
The accuracy of the proposed approach is satisfied by < 6%
MAEand < 5% MAE for tracking network information front location
andcoverage dynamics.
The strong correlation between traffic congestion and
information
spreading coverage shows a promising index to identify
congestion over a network.
13Ala and Du @ University of Florida
-
THANK YOU!
Acknowledgement: This research is partially supported by the
National Science Foundation awards CMMI-1436786 and
CMMI-1554559.
14Ala and Du @ University of Florida
[email protected] University of Florida
mailto:[email protected]
