Synopsis 2017 01 29 - Gujarat Technological University Rajendra Parmar.pdf · Synopsis Research...
Transcript of Synopsis 2017 01 29 - Gujarat Technological University Rajendra Parmar.pdf · Synopsis Research...
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Synopsis
Research Scholar - Introduction
Name Rajendra S Parmar; M Tech, IITB; B Tech – IIT BHU (Varanasi)
Title A model for predicting & detecting vehicular congestion to achieve uniform
vehicular density
Branch Computer Science Enrollment 12 99 90 93 10 03
Hand Phone 98210 72997
Supervisor Prof. Bhushan H Trivedi, PhD
Director, GLS Institute of Computer Technology,
GLS Campus, Opposite Law Garden, Ellis Bridge,
Ahmedabad 380 006, Gujarat, India
Hand Phone 88667 43251
DPC Member Prof. Apurva Desai, PhD
Prof. & Head, Department of computer Science,
Veer Narmad South Gujarat University,
University Campus, Udhna Magdalla Road,
Surat 395 007
Hand Phone 98241 94314
DPC Member Dr. Dr. R Nandakumar, PhD
Head, Ground Software Quality Assurance Division,
Reliability and Quality Assurance Group,
Systems Reliability Area, SAC, ISRO,
SAC, ISRO, Ahmedabad 380 015
Hand Phone 94283 55379
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Content
Synopsis 1
Research Scholar - Introduction 1
Content 2
Title of the thesis and abstract 4
Title 4
Abstract 4
Preface 5
Brief description on the state of the art of the research 5
Understanding congestion 5
Sensing Technologies 6
Signaling Systems 6
Simulation Tools 7
Contemporary Systems 8
Shortest Route 9
Approaches to the Solutions 9
Definition of the Problem 10
Objective and Scope of work 11
Objective 11
Scope of work 11
Original contribution by the thesis 12
Architecture 12
Information Flow Diagram 14
Infrastructure Database – Big Data 15
Vehicle Identification 15
Congestion Index - Definition 15
Computing Signal phases 16
Vehicle Density & Cumulative Waiting Time (CWT) 16
Look Ahead Buffer 16
Dynamically adjusting Signal Simulator 16
Case 1 – Equal signal phases; Optimum Out flow to Inflow ratio 17
Case 2 - Equal Signal Phases, Cyclic assignment of signal phases 17
Case 3 - Dynamic signal phases; Asymmetric Traffic; Signal Phases Proportional to τ 18
Case 4 - Dynamic signal phases; High Asymmetric traffic; Signal Phase proportional to τ 18
Alternate Route - Shortest Path Evaluation 19
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Moving from start node Q to end node I 21
Moving from start node I to end node Q 22
Knowledge System 23
Results & Comparisons 23
Congestion 23
Shortest Route 23
Feature Comparison 23
Conclusion 24
Future Work 24
Copies of papers published, Conference attended, Patent 25
Papers 25
Conference 26
Patents 27
References 27
Other References 28
Suggested Reading 29
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Title of the thesis and abstract
Title
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Abstract
On road Vehicular Traffic Congestion has detrimental effect on three lifelines Economy,
Productivity and Pollution (EPP). With ever increasing population of vehicles, traffic congestion
has become a major challenge to EPP and humanity. The research develops accurate and precise
model in real time to detect congestion and compute dynamic signal phases to evenly distribute
vehicle density. Congestion severity is expressed as congestion index which is a ratio of road
occupancy and road capacity. Road Capacity is area of road segment expressed as number of
vehicles it can accommodate, however quality and achievable average speed on the road segment
too contribute to road capacity computation. Road capacity is pre-computed and logged in the
database. Road occupancy is evaluated by monitoring number of vehicles on road and total area
occupied by these vehicles on the road segment.
Vehicles are installed with GPS device tagged with vehicle ID to provide information on vehicle
make, model and fuel type. Vehicles transmit GPS ID and location details periodically to traffic
server which computes road occupancy and congestion index for every road segment.
Signaling algorithm assigns green phase proportional to vehicle density and serves sequentially
to all the directions in decreasing order of their vehicle densities for every signaling cycle.
Heuristics are defined to prevent a direction with higher vehicle density hogging the signal phase
for unreasonably longer period. The algorithm is devised to distribute vehicle density evenly,
fairly, ensuring avoidance of starvation and deadlock situation.
A Dynamically adjusting Signal Simulator is developed to study traffic congestion as well as
validate the algorithm. The simulator proves that the proposed algorithm eases congestion by
more than 50%, better than any of the contemporary approaches offering 15% improvement.
In case of extremely high congestion index, alternate routes are suggested based on evaluation of
cost of travel from amongst: 1. Static distance graph; 2. Dynamic graph which changes as per the
congestion index and 3. Travel history or knowledge database. The research proposes shortest
route algorithm employing dynamic node reduction schema. The maximum computation cost of
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
proposed shortest route algorithm is 7*��� vis-à-vis N2 of by classical algorithms, where N is
the number of nodes.
Lastly the proposal brings down the present implementation cost of USD 40,000 to USD 2,000
which can pave the way for larger number of installations.
Preface
The synopsis is organized in ten sections. Section I details literature survey highlighting
understanding congestion, sensing technologies, signaling systems, simulation tools,
contemporary systems, shortest route and approaches to the solutions. Section II is problem
definition, describing the problem the research is addressing. Section III describes scope of work.
Section IV enumerates original contributions defining state-of-the-art architecture, infrastructure
database (a case of big-data) pre-computing and logging road capacity, vehicle identification,
captures vehicle coordinates, explains definition of congestion index, a deciding factor for
computing signal phases and alternate routes, dynamically adjusting signal simulator
development to study different traffic conditions and validate the hypothesis, algorithm for
alternate route and lastly a knowledge system or history database. Section V is summary of
results & comparison. Section VI highlights conclusion, section VII lists future work, section
VIII details paper published, conferences attended and details of patent filed. lastly section X
enlists references.
Brief description on the state of the art of the research
The research work commenced with appreciating definition of congestion, technologies used to
detect and count vehicles, developments in signaling technology, studying various simulation
tools, evaluate contemporary solutions to find research gap, discovery of shortest routes in case
of extreme congestion and approaches considered to address the problems are studied.
Understanding congestion
Nicholson Appreciating mobility parameters namely flow q (vehicles per hour), concentration k
(vehicles per mile), number of vehicles passing a point over a time period, road capacity, wave
velocity, headway buildup and shock wave velocity laid the foundation of the research work.
Different mobility systems were studied and compared their strengths, weaknesses and
efficiencies to scope desired characteristics of vehicle transport system Raj. The dynamics of
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
congestion is best appreciated by definition of continuum model of fluid dynamics which defines
characteristics of a relationship between vehicle flow and concentration. However, it has
limitations during two scenarios: 1. Overtaking 2. Vehicles in front determine wave speed in real
time system as against rear waves driving the wave speed. The higher models, inspite of their
complexities cannot rid these limitations hence real time solution is sought
Sensing Technologies
Researchers proposed vehicle detection & congestion detection through cooperative V-2-V
communication Ramon, Sandor, VANET (Vehicular Ad-Hoc Networks) Fransisco, MANET
(Mobile Ad Hoc Networks), image tracking ShungTseng, RFID and infrared technologies
Koushik. Direct measurement of vehicles through TrafficHandbook in-road or on-road sensing
technologies FHWA have exorbitant cost comprising of equipment, installation and training.
Benny Hardjono RajGupta Jubair Proposed sensors at every entry and exit of a junction to
monitor cars present. Steven Proposed probe vehicle to ascertain traffic condition however,
lower penetration hampers accuracy and precision. These approaches have disadvantages in
terms of time to install, installation cost and disruptive maintenance. Besides, the sensors are also
susceptible to weather conditions.
GPS addresses these limitations GPS, detects vehicles and provides its locations. The research
proposes vehicles be equipped with GPS enabled device and relate vehicle ID to GPS ID, thus
providing precise vehicle count and road occupancy in terms of number of vehicles as well as
occupied area by the vehicles on the given road segment. The proposition provides unparallel
accuracy, reduces cost of deployment & maintenance and is oblivious of weather condition.
More importantly, the approach is flexible and scalable.
Signaling Systems
Prashant Proposed suggested speeds based on predicated speeds of vehicles, though it lacks in
accuracy and precision. Mariagrazia, Li Proposed synchronizing signals, however achieving this
in case of adjoining road segments with different lengths, is a great challenge. Jerry Suggested
Ant Colony Optimization (ACO) to deploy Distributed Intelligent Traffic System (DITS) then
again, there is vast difference between the requirements of vehicular traffic and mobility of ant
colony. Yit Approaches based on AI, fuzzy logic, genetic algorithm and database from
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
continuous learning are proposed, but these are not real time implementable. Qiwu Deployed
camera to employ closed loop system, but the vision technology is disadvantages during adverse
weather and occlusion. Mladenovic Defined guidelines for Anthropocentric design for self-
driving vehicles. Shilpa Presented intelligent Traffic Light Controllers (ITLC) based on
microcontroller and microprocessor. It has communication interface through which cycle times
can be changed dynamically.
Parameter comparison empathetically suggesting Real Time Control is the only solution
Concept Advantages Challenges Solution
Synchronized
signals
Reduced
congestion
Disciplined movement
Varying speeds
Varying road segment
Require real time road segment
data
Vehicular density
Ant colony
optimization
Swarm
intelligence
Ants don’t have alternate route
Thrive for shortest route
Speed Prediction
Speed displayed
with traffic lights
Speeds depend on
infrastructure
Present possible speed due to
congestion must be considered
With real time data about vehicle
speeds and congestion, this may be
possible
Camera based
system
Proven reduction
in waiting time
Weather conditions
Computation intensive
Accurate, real time data can
provide precise results
Probe Vehicles Accurate data Penetration of probe vehicles
Relies on historical data
Cost effective, accurate and real
time data will increase the
penetration
Fuzzy logic based
on look up data
Real time conditions may not
match history data
Real time, accurate, precise data
with quick response time is desired
Statistical approach Gives average values
Not for real time applications
Real time, accurate, precise data
with quick response time is desired
The proposed research addresses the challenges encountered in above approaches. The model
accurately detects vehicle density, congestion severity and computes signal phases based on
cumulative waiting time considering available road capacity ahead ensuring against starvation
and deadlock situation.
Simulation Tools
Study of simulation software is undertaken to appreciate parameters studied to simulate traffic
environment and intelligence acquired from the simulators. The simulation tools studied and
evaluated are AIMSUN (Advanced Interactive Microscopic Simulator for Urban and Non-urban
Networks), NEMIS, SPEACS, DRACULA Dynamic Route Assignment Combining User
Learning and Micro-simulation and SITRA B+. These tools simulate traffic conditions and
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
provide information on traffic light strategies, route guidance, messaging systems, examining
strategies of fuel efficiency, exhaust emission.
The inputs to simulators are predefined whereas real time inputs are dynamic and stochastic in
nature. Providing dynamic input to simulators analogous to real time situation is a difficult
proposition. During the research work, a dynamically adjusting signal simulator was developed
to validate the proposed hypothesis of vehicle evacuation. The research computes congestion
index and assigns signal phases incorporating congestion in the road segment ahead, ensure
against dead lock and starvation situation and communicates with signaling system.
Contemporary Systems
SCOOT - Split Cycle Offset Optimization Technique SCOOT, SCATS -Sydney Coordinated
Adaptive Traffic System SCAT, UTOPIA - Urban Traffic Optimization by Integrated
Automation UTOPIA, ACSLite and RHODES Real-time Hierarchical Optimizing Distributed
Effective System RHODES deploy in-road or on-road sensing technologies . The average cost is
more than USD 30,000 per traffic junction (ACSLite is cheaper) and maximum congestion
improvement is 15%.
Parameters SCOOT SCAT Rhodes
Developer IMTECH Traffic & Infra UK Ltd
Siemens Traffic Controls TRL
New South Wales
Government, Australia
Research team at
University of Arizona
Type On Line On Line Off Line
Software SCOOT Kernel, UTC software
Microsoft Windows
Client Server Technology
Windows OS
Installation is Platform
independent
Performance Reduces delay by 20% Reduces delay by 5% to 45%
Detectors
Technologies
On-Street Detectors; Inductive
loops
On-Street Detectors; Inductive
loops
Inductive detectors,
Video, SONAR, RADAR
Controllers SIEMENS SCAT compatible traffic
controller
Cost High; Oder USD 30,000 High; Order USD 30,000 High; Order USD 30,000
Shortest Path No No Yes
Traffic
Information
ASTRID (Automatic SCOOT
Traffic Information Database)
Historical Data No
Contemporary systems employ fixed sensors, do not have precise road occupancy data which
introduces errors, are exorbitantly costly and reduces congestion by 15% only. The proposed
system for a city with 100 traffic junction will cost less than USD 2,000 per traffic junction. Also,
the model improves congestion by more than 50%.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Shortest Route
Biswas Dong Zarembo Dorothea Martin Studies of classical algorithms for shortest path is
undertaken for A*, BFS, Dijkstra's algorithm, HPA* and LPA*. The cost of computation
effectively is square of number of nodes.
In reality, vehicular traffic network exhibits duality of network: 1. Static network is a network of
fixed distances; 2. Dynamic network is a network where edge weights change continuously
depending on congestion levels. Static network is evaluated once for shortest distances between
nodes whereas dynamic network is evaluated every time the edge weight or traffic conditions
changes, the periodicity may be as high as every few minutes. This necessitates high optimization
of the graph to reduce computation cost. The algorithm is based on dynamic node reduction
schema which expedites shortest route discovery. The cost of computation of proposed algorithm
is 7*��� as against N2 in the classical algorithms.
Approaches to the Solutions
Mingqui Proposed using acceleration data from undedicated mobile phones which also reduces
power consumption of GPS based technologies. Andreas Differentiated idle and active phones
by extracting CDRs (Call Data Record) obtained from telecom service providers. WeiHun The
paper discovers traffic bottle necks in spatiotemporal coordinates (Spatial – Location; Temporal
– Time). The sensing is done through location based services. Fransisco The LocHNESs
(Localized Handling Network Event Systems) platform takes inputs from fixed sensors as well
as GPS enabled devices. Afshin Employed a simulator based approach. The prediction errors are
2% to 12% for a 5 to 30 minutes window respectively. José The paper employs IEEE 802.11
network beacon frames sent periodically. To detect the vehicle and its location, a smart phone
with IEEE 802.11 is used which detects the frame even at low power. It deploys road side
sensors which pose space, power and cost related challenges.
Summary of Literature Review
Principle Advantages Limitations
Undedicated
Mobile Phones
Relies on acceleration; Energy Efficient
Measures speed, congestion
Accelerometer measurements prone to inaccuracies
Dependency on multi-cellular service providers
Anonymized
Signaling
Active and Idle handsets; Energy Efficient
Large sample size
Measures average speeds
Vehicle passing cell boundaries with a traffic
junction, encountering green or red phase, which
introduces error
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Dependency on multi-cellular service providers
Traffic flow
prediction
Simulator on history data
Real time data submitted to get prediction
Not in real time
Simulation
COTraMS Employs IEEE networks
Energy Efficient
RSU installed on road side
Requires IEEE 802.11 infrastructure
SBTM Identifies traffic bottlenecks
Location based services
Accuracy of Prediction > (Congestion
converge OR Congestion drop)
Dependency on multi-cellular service providers
Not in real time
LocHNESs Accepts fixed as well as GPS enabled
sensor
Large datasets; Computation intensive
Dependency on multi-cellular service providers
AFRC Network of Controllers communicate
Compute congestion levels
Disseminates data to traffic junction ahead
Physical installation of controller
Data acquisition from roadside sensors
Proposition GPS Sensors; Real time system Accurate, Economical, No deployment time
The proposed research detects and measures vehicles on road, area occupied by vehicles on
road, computes signal phases incorporating spare road capacity ahead of the vehicles to check if
the road ahead can accommodate additional vehicles, advises on alternate shortest route in case
of congestion and provides fair opportunity to every direction avoiding starvation and dead lock
situation.
Definition of the Problem
The problems to address traffic congestion are:
1. Vehicle detection sensor technology, installation and training is prohibitively costly,
disrupts traffic during installation or maintenance.
2. Vehicle detection technology does not detect vehicle foot print and vehicle location
3. Present systems don’t compute road capacity, road occupancy and available road capacity
ahead. This limits the accuracy in computing signal phases which restricts achievable
decongestion.
4. For dynamically changing traffic situation, dynamically changing signal phases are
required instead of static or pre-programmed signal phases. Static signal phases are major
reason for dead lock situation and signal starvation to a direction, resulting in long hours
of traffic congestion.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
5. Computation cost to discover shortest alternate route is N2, which is very high, and needs
to be reduced. N is number of nodes.
6. Achieved decongestion of 15% by present systems is much lower than current severity,
resulting in worsening congestion with time.
7. Present installation cost is prohibitively high, USD 30,000 per traffic junction, limiting
the number of installations.
The problem is to develop a model which is affordable, efficient and accurate with ability to
predict, detect traffic congestion and distribute vehicular densities evenly to ease out congestion.
Objective and Scope of work
Objective
The objective is to:
1. To evolve economical, accurate, efficient, flexible, scalable and user friendly system to
detect & compute vehicle, vehicle location, vehicle density, vehicle velocity, road
capacity, road occupancy, congestion index and recommend alternate routes
2. To define methodology to measure road capacity, monitor road occupancy, compute
congestion index, determine available road occupancy ahead and evaluate signal phases
3. To build vehicle detection mechanism which is cost effective, easy to disseminate, has
near nil installation cost and is highly user friendly. Congestion index to be computed
from road capacity and road occupancy
4. Define algorithm for vehicle evacuation at a traffic junction incorporating fair strategy for
waiting vhicles at traffic junction ensuring against dead lock and starvation
5. To define algorithm to compute shortest alternate route with cost of computation far
lower than N2
6. Finally realize a system which is highly cost effective compared to the present cost of
USD 30,000 per traffic junction
Scope of work
1. To create infrastructure database to compute road capacity. The database parameters for
road segment are its length (geometric and actual), breadth, latitude, longitude, quality,
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
road segment name (demographic), traffic signals at the two ends and merging road
segments.
2. To compute road capacity, parameters required are road length, breadth, latitude,
longitude, traffic signals and road quality. This is logged in the infrastructure database.
3. To consider directed graph evaluation for right of way (one way road segments). The
system is ideally suited for metros and class A cities.
4. To create flexible and open platform of data acquisition of vehicle location such that it
can ingest data directly from vehicles, telecom service providers, social networks Chen
Lefei or installed native applications.
5. In case of direct data acquisition from vehicles, a database of vehicle ID with GPS ID is
created to provide information about vehicle manufacturer & model, foot print area, fuel
type and location in terms of latitude & longitude. This data is used to compute road
occupancy, ARCA (Available Road Capacity Ahead) and congestion index
6. To define an algorithm for vehicle evacuation at traffic junction, ensuring avoidance of
starvation & deadlock situation and evenly distribute reduced vehicular densities.
7. To build a dynamically adjusting signal simulator for traffic monitoring and control, with
a view to study impact of different models of traffic evacuation and validate the
hypothesis by measuring extent of decongestion achieved by the algorithm
8. To define an algorithm to discover shortest route from amongst static distance graph,
dynamically changing congestion graph and knowledge database created from travel
history
9. To construct a state-of-the-art architecture, encompassing all elements of the solutions.
The architecture must be cost effect, offer geographic redundancies and have high
reliability.
Original contribution by the thesis
Architecture
RajBhushanICECCS Architecture comprises of computing engine, which houses database for
infrastructure, vehicle data, travel history to compute road occupancy, ARCA, congestion index,
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
signal phases and shortest routes. The results are communicated to vehicles and signaling
systems.
A m
odel
for
pre
dic
ting &
det
ecti
ng v
ehic
ula
r co
nges
tion t
o a
chie
ve
unif
orm
veh
icula
r den
sity
Info
rma
tio
n F
low
Dia
gra
m
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Infrastructure Database – Big Data
RajBhushanAleks A test vehicle equipped with application software traverses through the length
and breadth of the city to capture infrastructure parameters. Latitude, Longitudes, average
achievable speed on road segment are captured automatically by GPS whereas demographic
names, number of lanes, presence of traffic signals, converging road segments are input to the
application. This information provides road capacity, computed as:
RC = L * N (1)
L = Length of road segment; N = Number of lanes
For a city with 5,00,000 vehicles, having a road presence of 4 hours a day, and a data frame of
100 bytes transmitted every minute generates 12GB of data per day, not accounting other
infrastructure data, qualifying the application for a big data approach, which brings in
computational efficiencies.
Vehicle Identification
A GPS enabled device with application software is attached with each vehicle, which transmits
its location. The vehicle ID and GPS device are paired to create a unique identification providing
manufacturer, model, fuel type and economy, area, length and breadth. The location coordinate
data is time stamped. Consecutive time stamped location coordinates provides speed and mean
velocity. Data acquired from different vehicles are mapped to road segments with vehicle area to
provide RO (road occupancy).
RO = � ������ (2)
Congestion Index - Definition
Congestion Index defines Degree of Congestion, is a ratio of Road Occupancy to Road Capacity.
Congestion index is a parameter of significance to decide when to compute alternate routes.
Road occupancy is computed from vehicle detection whereas road capacity is extracted from the
infrastructure database. The data is available both units, number of vehicles as well as area
occupied by the vehicles.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Computing Signal phases
Vehicle Density & Cumulative Waiting Time (CWT)
The signal phases are assigned based on CWT τ of vehicles in a direction RajBhushanAleks. A
direction with maximum CWT gets priority. Maximum value of green phase is limited by upper
bound Km to avoid starvation to other directions. Also, fair distribution is ensured by assigning
green phase to every direction in a signal phasing cycle.
� �� � �� ����
��� (3)
T1 = Entry time; T2 =Time at computation; N = Number of vehicles; K is modulating factor
In order to incorporate area A of waiting vehicle, the above equation is modified as follows:
� �� � ��� ����
��� �; (4)
If τ > Km, then τ = Km
Look Ahead Buffer
Green phase is assigned based on as well as ARCA to accommodate oncoming vehicles. If a
green phase is issued without considering ARCA, it leads to deadlock situation. ARCA is
computed from RC and RO. RC is known from the infrastructure database and RO is derived
from the vehicles on the road segment, hence ARCA = RC – RO
ARCA=L*N – � ���� (5)
A direction is assigned green phase by optimizing the following two situations:
1. Potential of evacuating maximum number of vehicles, as this ensures even density
distribution
2. Maximum CWT
The parameters to compute are maxima of cumulative time and maxima of ARCA. It describes
the three road conditions:
i. RO < ARCA → No congestion
ii. RO = ARCA → Normal traffic
iii. RO > ARCA → Congestion
Dynamically adjusting Signal Simulator
A dynamically adjusting signal simulator is developed in windows using C++ and MS SQL
database is designed to:
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
1. Understand the present traffic congestion
2. Study the impact of different outflow to inflow ratio and evolve an optimum ratio
3. Validate the signaling algorithm based on different traffic situation and appreciate
extended green phase cycles and vehicle evacuation
Case 1 – Equal signal phases; Optimum Out flow to Inflow ratio
A single junction is considered with 4 directions, with fair opportunity of signal phases. The
variables set are inflow and outflow rate. Congestion is monitored with different outflow rates. It
is observed that congestion reduces with higher ratio of outflow to inflow. To increase the
outflow, green phase or road capacity has to be increased. Logical solution is to increase green
phase.
Graph shows vehicle accumulated at a traffic junction for different outflow to inflow ratios
From the graph it is evident that optimum outflow to inflow ratio is 5:1 and increasing this ratio
does not reduce congestion proportionately.
Case 2 - Equal Signal Phases, Cyclic assignment of signal phases
The simulator was then deployed to evaluate situation for four directions. The results of the
vehicles in queue are plotted in the figure below. After a few initial signal phases cycles,
congestion reduces and the waiting vehicles vary between 0 to a maximum of 60 waiting
vehicles.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Above graph confirms the results of case 1 albeit it takes few cycles to arrive at equilibrium
It demonstrates, even distribution of vehicles and emphasizes 5:1 is optimal ratio. All the
directions are cyclically assigned green phases and waiting vehicles are between 0 & 60.
Case 3 - Dynamic signal phases; Asymmetric Traffic; Signal Phases Proportional to τ
During initial condition the waiting vehicles are 50, 40, 25 and 20 in north, east, west and south
directions respectively. The graphs plotted are vehicles in queue and evacuating vehicles. The
software assigns priority to a direction with maximum CWT.
Evacuating Vehicles
North direction is assigned green phase 6 times, followed by east 5, west 4.5 and south 4 times
and waiting vehicles reduce between 12 & 16 vehicles.
Case 4 - Dynamic signal phases; High Asymmetric traffic; Signal Phase proportional to τ
Now to increase congestion, initial waiting vehicles are set to 100 for north direction keeping the
waiting vehicles for other directions same. It is observed that the green phase is granted for
extended duration to evacuate initial congestion confirming to the upper bound of green phase.
Also it can be seen that a total of five green phases are granted with the first green phase being
extended.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Graph with 100 waiting vehicles at initial condition
It demonstrates extended green phase in terms of duration as well as number of cycles assigned
proportional to the CWT
Employing static signaling phases the waiting vehicles vary between 0 & 60, average value of
waiting vehicles being 30.
While employing dynamic signal phases the waiting vehicles vary between 5 & 20, average value
of waiting vehicles being 12.5.
The upper limit is reduced by 66% where as the average value is 58%
Alternate Route - Shortest Path Evaluation
Vehicular traffic network exhibits duality with static distance graph and dynamic travel time
graph. For static graph, shortest distances are pre-computed and made available as a look up
table. The travel time graph weightages keep on changing, and are continuously computed
dynamically.
The graph is reduced from a 64 node to a 36 node graph.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Network with coordinate & edge lengths
A bounding box is created between the start and end node to optimize the network to reduce
computing cost. The nodes outside the bounding box are ignored. From the start node, the next
preferred node amongst available nodes is selected based on:
Min (Edge weight to next node + Cartesian distance of next node to the destination) (6)
This is evaluated for all possible next nodes, amongst them the one with minimum value is
selected and other are stored for evaluation later. The next node now becomes the new start node.
A new bounding box is defied between start node and end node. This helps further reduce the
network size. The algorithm computes all options and arrives at shortest distance.
One start to end node is computed, the nodes are swapped to compute from end node to start
node.
In case the shortest distance is not found, the bounding box is increased considering Cartesian
coordinates and the network is re-evaluated.
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Moving from start node Q to end node I
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Moving from Q to I
1. Start node Q
2. Bounding box between Q to I is drawn
3. Q is connected to R only
4. Bounding box between R to I
5. R connects with M
a. (Distance from R to M + Distance from M to I) < (Distance from R to S + Distance from S to I)
b. S is kept OPEN to be dealt later
6. Bounding box between M to I is drawn
7. M connects to G as N is out of bounding box
8. G connects to H
9. H connects to I
10. Route QRMGHI 16.32
11. OPEN S
12. Route so far QRS
13. S connects to N
14. N connects to H
15. H connects to I
16. Route QRSNHI 17.44
Moving from start node I to end node Q
The shortest route exploration is done starting from both the ends Q as well as I.
Starting from Q the two shortest routes are QRMGHI is 16.32 units and QRSNHI is 17.44 units.
Starting from I the two shortest routes IHNMRQ is 17.48 units and IHGMRQ is 16.32 units. It
demonstrates that from both the ends, we have consistent results. All the three results are feasible
and consistent. The deviation from the shortest route offers alternate shortest routes.
Further reduction of nodes is possible by
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
The maximum computing cost of the proposed solution in a square matrix is computed as:
Distance of diagonal node is ���, each node has 7 directions to evaluate, giving us maximum
computing cost of 7*��� which is less than N2.
Knowledge System
Travel path of every vehicle is logged and stored in the data base with time stamp. This creates
knowledge database to refer to for history of preferred routes.
Results & Comparisons
Congestion
The maximum congestion is reduced by 66% where as the average congestion is reduced by 58%
which is much higher than contemporary solutions offering 15% congestion reduction. The
proposed solution brings down the waiting vehicles between 5 and 20 making average number of
waiting vehicles as 12.5 whereas in case of static signaling the average waiting vehicles is 30.
These results are obtained with an outflow to inflow ration of 5:1
Shortest Route
The maximum computing cost of the proposed solution in a square matrix is 7*��� which is
less than N2 achieved by classical algorithms.
Feature Comparison
Parameters ATCS Google Google Proposal
Sensors Embedded in road GPS, Crowd sourcing GPS
Infrastructure data No No Yes
Vehicle Size No No Actual, Incorporates vehicle size
Signaling Interface Yes No Yes
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Cost Per Traffic
Junction
More than USD 25,000 No signal interface
Not available
Less than USD 2,000
Congestion resolution 15% Less than 15% More than 50%
Alternate Route
Computing cost
Cost proportional to N2 Cost proportional to N
2 Network duality, Accurate,
Lower cost 7*���
Scalability
Traffic offenses No No Yes
Toll collection No No Yes
Post accident analyses No No Yes
Emission
Measurement
Yes; Low accuracy No Yes – Best
Traffic Signal
Avoidance
No No Yes
Conclusion
1. The signaling system achieves more than 50% decongestion, which is better than 15%
claimed by contemporary system.
2. The cost of shortest route algorithm is less than 7*��� which is far better than N2, the
computation cost of classical algorithms.
3. Proposed GPS assisted approach is best suited for the application which not only detect
detects vehicles, but also provides foot print area information to accurately compute
congestion index.
4. The proposed system is flexible, scalable and costs USD 2,000 per traffic junction as
compared to USD 30,000 for other systems.
5. The simulator results confirms to our algorithm which is based on road capacity, road
occupancy and ARCA
Future Work
1. Global Warming
Vehicle travel is precisely monitored with knowledge of fuel type and travel time which
enables accurate computation of CO2 emission responsible for global warming.
2. Traffic Signal Avoidance (TSA)
This paper proposes to duplicate the information of signal phases in the vehicle display
thereby rendering physical traffic junctions redundant. Computing systems transmit
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
signal phase information to vehicles, thus avoiding the necessity of dedicated traffic
signals, thereby eliminating cost of equipments, power, and costly installation and
maintenance efforts.
Abdus Proposed VICS (Vehicle-Intersection Coordination Scheme) where it is proposed
to do away with traffic signaling systems. The system coordinates through vehicular
communication network to navigate traffic.
3. Post Accident Analyses
A frame work can be developed to provide parameters (speed, directions of vehicles)
prior to the accidents.
4. Road Quality
The load on road in terms of number of vehicles plied vis-à-vis reducing average speed is
a good indication for deteriorating road. The minimum weight of the vehicle is had from
vehicle registration data.
5. Tax and Toll Collections
Zones based toll collection and Pay-Per-Use Road Tax can be realized automatically
6. Traffic Offenses
Guovu A vehicle jumping traffic lights, travelling in the wrong direction (one way) or
prohibited direction, vehicles parked in no-parking zone or extended parking hours can
be easily detected.
Copies of papers published, Conference attended, Patent
Papers
[1] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, C. Aleksandar Stevanovic, PhD,
“A Model with Traffic Routers, Dynamically Managing Signal Phases to Address Traffic
Congestion in Real Time”, Submitted
[2] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, “Modulating Traffic Signal Phases
to Realize Real-Time Traffic Control System”, Journal of Transportation Technologies,
vol 7, no. 1, pp 26-35, Jan. 2017
[3] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, “Shortest Route – Domain
Dependent, Vectored Approach to Create Highly Optimized Network for Road Traffic”,
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
IJTTE International Journal of Traffic & Transportation Engineering – IJTTE, vol. 5, no.
1; pp 1 – 9, Jan. 2016.
[4] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, “Real Time Computation of
Optimal Signal Timing to Maximize Vehicular Throughput for a Traffic Junction”, 3rd
International Conference on Eco-friendly Computing and Communication Systems
(ICECCS 2014), NITK Surathkal, Mangalore, India, pp 194– 199, Dec. 8–21, 2014.
[Online] Available at
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=7208991&url=http%3A%2F%2Fi
eeexplore.ieee.org%2Fiel7%2F7051380%2F7208942%2F07208991.pdf%3Farnumber%
3D7208991
[5] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, “Identification of Parameters and
Sensor Technology for Vehicular Traffic - A Survey”, IJTTE International Journal of
Traffic & Transportation Engineering, vol. 3, no. 2; pp 101 – 106, Apr. 2014.
Conference
[1] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, “Modulating Traffic Signal Phases
to Realize Real-Time Traffic Control System, “3rd International Conference on Gujarat
Model of Governance: Lessons & Future Scope”, ICGS-2015, Gandhinagar, Mar 20 -21,
2015
[2] Rajendra S Parmar, M. Tech., Bhushan Trivedi, PhD, “Real Time Computation of
Optimal Signal Timing to Maximize Vehicular Throughput for a Traffic Junction, “3rd
International Conference on Eco-friendly Computing and Communication Systems
(ICECCS 2014)”, NITK Surathkal, Mangalore, India, Dec. 8–21, 2014.pp 194– 199.
[Online] Available at
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=7208991&url=http%3A%2F%2Fi
eeexplore.ieee.org%2Fiel7%2F7051380%2F7208942%2F07208991.pdf%3Farnumber%
3D7208991
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
Patents
“System for Monitoring and Decongesting Traffic Congestion with GPS Mobility Sensors”,
Application Number 2121.MUM/2015; Date of filing Jun 1, 2015; Date of publication Jun 12,
2015
References
[1] Raj S Parmar, Prof. Bhushan Trivedi, Apr 2014, Identification of Parameters and Sensor Technology for Vehicular Traffic - A Survey,
International journal of Traffic and Transport Engineering, Vol. 3. [2] T A J Nicholson, Applied Mathematics Group, Theoretical Physics Division, U.K.A.E.A., Atomic Energy Research Establishment,
Harwell, Didcot, Berkshire. “Finding the shortest route between two points in a network”, The computer journal 1966 (9) 3, pp 275 –
280 [3] Ramon Bauza, Javier Gozalvez and Joaquin Sanchez-Soriano, “Road Traffic Congestion Detection through Cooperative Vehicle-to-
Vehicle Communications,” IEEE 35th Conference on Local Computer Networks (LCN), October 2010
[4] Sandor Dornbush and Anupam Joshi, “Street-Smart Traffic: Discovering and Disseminating Automobile Congestion Using VANET’s,” IEEE 65th Vehicular Technology Conference, 2007 (VTC2007-Spring), April 2007
[5] Francisco M. Padron, “Traffic Congestion Detection using VANET,” A Thesis Submitted to the Faculty of The College of
Engineering and Computer Science in Partial Fulfillment of the Requirements for the Degree of Master of Science, April 2009 [6] Shung-Tsang Tseng and Kai-Tai Song, “Real-Time Image Tracking for Traffic Monitoring,” The IEEE 5b International Conference
on Intelligent Transportation Systems; 3 - 6 September 2002, Singapore
[7] http://en.wikipedia.org/wiki/Radio-frequency_identification, “RFID Description, Technology,” Accessed February 2014 [8] Koushik Mandal, Arindam Sen, Abhijnan Chakraborty, Siuli Roy, Suvadip Batabyal, Somprakash Bandyopadhyay, “Road Traffic
Congestion Monitoring and Measurement using Active RFID and GSM Technology,” 14th International IEEE Annual Conference on
Intelligent Transportation Systems, April, 2011 [9] http://www.kcscout.net/downloads/Announcements/CongestionReport.pdf; Chapter 1, Traffic Detector Handbook: Third Edition—
Volume I
[10] US Department of Transportation, “Federl Highway Administration. FHWA-HRT-06-108,” May 2006 [11] Benny Hardjono, Adi Nurhadiyatna, Petrus Mursanto and Wisnu Jatmiko, “Development of Traffic sensor system with Virtual
Detection Zone,” International Conference on Advanced Computer Science and Information Systems (ICACSIS), December 2012
[12] Hardjono B., Wibowo A., Rachmadi M.F., Jatmiko W., “Mobile phones as traffic sensors with map matching and privacy considerations,” International Symposium on Micro-NanoMechatronics and Human Science (MHS), November 2012
[13] Raj Gupta and Biplav Srivastava, “Sensor Subset Selection for Traffic Management,” 14th International IEEE Conference on
Intelligent Transportation Systems Washington, DC, USA. October 5-7, 2011 [14] http://en.wikipedia.org/wiki/GPS_signals, “GPS Explanation, Technology, Purpose, Accuracy,”, Accessed March 2014
[15] Jubair Mohammed Bilal, Don Jacob, 24-27 November 2007, Intelligent Traffic control system, IEEE International Conference on
Signal Processing and Communications (ICSPC 2007), Dubai, United Arab Emirates [16] Steven E. Shladover, Jing-Quan Li, October 5-7, 2011, Evaluation of Probe Vehicle Sampling Strategies for Traffic Signal Control,
2011 14th International IEEE Conference on Intelligent Transportation Systems, Washington, DC, USA.
[17] http://en.wikipedia.org/wiki/GPS_signals, “GPS Explanation, Technology, Purpose, Accuracy,”, Accessed March 2014
[18] Prashant Borkar, L.G. Malik, 2013, Speed Range Prediction for Subsequent Intersections, Proceedings of 7th International Conference
on Intelligent Systems and Control (ISCO).
[19] Mariagrazia Dotoli, Maria Pia Fanti, Carlo Meloni, March 21-23, 2004, Coordination and Real Time Optimization of Signal Timing Plans for Urban Traffic Control, Proceedings of the 2004 IEEE international Conference on Networking, Sensing 8 Control.
[20] Li Yinfei, 2009, Research on Synchronizing traffic signals for an Urban Arterial Road, Third International Symposium on Intelligent
Information Technology Application.
[21] Jerry John Kponyo, Yujun Kuang, Zejiao Li, October 2012, Real Time Status Collection and Dynamic Vehicular Traffic Control
Using Ant Colony Optimization, 2012 International Conference on Computational Problem-Solving (ICCP).
[22] Yit Kwong Chin, Wei Yeang Kow, Wei Leong Khong, Min Keng Tan, Kenneth Tze Kin Teo, 14-16 Nov. 2012, Q-Learning Traffic Signal Optimization within Multiple Intersections Traffic Network, 2012 Sixth UKSim/AMSS European Symposium on Computer
Modeling and Simulation (EMS).
[23] Qiwu Ran and Jianguo Yang, March 23-25, 2012, A Novel Closed-Loop Feedback Traffic Signal Control Strategy at an Isolated Intersection, 2012 IEEE International Conference on Information Science and Technology Wuhan, Hubei, China.
[24] Mladenovic, M.N. ; Abbas, M. ; McPherson, T, 2014, Development of Socially Sustainable Traffic-Control Principles for Self-
Driving Vehicles: The Ethics of Anthropocentric Design, IEEE International Symposium on Ethics in Science, Technology and Engineering
[25] Shilpa S. Chavan (Walke), Dr. R. S. Deshpande, J. G. Rana, 2009, Design of intelligent traffic light controller using embedded
system, Second International Conference on Emerging Trends in Engineering and Technology, ICETET
[26] http://www.scoot-utc.com/
[27] http://www.scats.com.au/
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
[28] UTOPIA (Urban Traffic Optimization by Integrated Automation) was combined with SPOT to account to changes at the network
level (Mauro and DiTaranto 1990).
[29] RHODES - Real Time Hierarchical Optimized Distributed Effective System, http://ocw.nctu.edu.tw/course/sc011/2012-08-23-1.pdf, 2008
[30] P. Biswas, P. K. Mishra and N. C. Mahanti, “Computational Efficiency of Optimized Shortest Path Algorithm”, International Journal
of Computer Science & Applications, Vol. II, No. II, pp. 22 – 37, 2005
[31] Dong Zhang, ZuKuan Wei, Jae-Hong Kim, ShuGuang Tang, “An optimized Dijkstra algorithm for Embedded-GIS”, International
Conference on Computer Design and Applications (ICCDA), 2010, Year: 2010, Volume: 1, Pages: V1-147 - V1-150
[32] Imants Zarembo, Sergejs Kodors, “Path finding Algorithm Efficiency Analysis in 2D Grid”, Proceedings of the 9th International Scientific and Practical Conference. Volume 1I,, 2013
[33] Dorothea Wagner, Thomas Willhalm, “Geometric Speed-Up Techniques for Finding Shortest Paths in Large Sparse Graphs”,
Universit at Karlsruhe, Institut fur Logik, Komplexit at und Deduktions systeme, D-76128, Karlsruhe, Springer Berlin Heidelberg, 2003.
[34] Martin Holzer, Frank Schulz, and Thomas Willhalm, “Combining Speed-Up Techniques for Shortest-Path Computations”, Universit
^at Karlsruhe, Fakult, Postfach 6980, 76128 Karlsruhe, Germany, "Experimental and Efficient Algorithms. Springer Berlin Heidelberg, 2004. 269-28, Germany
[35] Kittelson & Associates, Inc., Texas Transportation Institute, University of Maryland, Siemens ITS, Purdue University, & the Institute
of Transportation Engineers, June 2008, Signal timing manual.
[36] Rajendra Parmar, Bhushan Trivedi, “Modulating Traffic Signal Phases to Realize Real-Time Traffic Control System”, 3rd
International Conference on Gujarat Model of Governance: Lessons & Future Scope, ICGS-2015, 20th – 21st March, 2015
[37] Rajendra Parmar, Bhushan Trivedi, “Real Time Computation of Optimal Signal Timing to Maximize Vehicular Throughput for a Traffic Junction”, in ICECCS 3rd International Conference on Eco-friendly Computing and Communication Systems , NITK
Surathkal, Mangalore, India, December 18 – 21, 2014, Available at
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=7208991&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel7%2F7051380%2F7208942%2F07208991.pdf%3Farnumber%3D7208991
[38] Rajendra S Parmar, Bhushan Trivedi, C. Aleksandar Stevanovic, “A Model with Traffic Routers, Dynamically Managing Signal
Phases to Address Traffic Congestion in Real Time”, IEEE Transcation on ITS (TBP)
[39] Mingqi Lv, Ling Chen, Xiaojie Wu, and Gencai Chen, “A Road Congestion Detection System Using Undedicated Mobile Phones,”
IEEE trans. Intelligent Transportation Systems, vol. 16, no. 6, Dec. 2015.
[40] Andreas Janecek, Danilo Valerio, Karin Anna Hummel, Fabio Ricciato, and Helmut Hlavacs, The Cellular Network as a Sensor: From Mobile Phone Data to Real-Time Road Traffic Monitoring,” IEEE trans. Intelligent Transportation Systems, vol. 16, no. 5, Dec. 2015.
[41] Wei-Hun Lee, Member, IEEE, Shian-Shyong Tseng, Member, IEEE, Jin-Lih Shieh, and Hsiao-Han Chen, “Discovering Traffic
Bottlenecks in an Urban Network by Spatiotemporal Data Mining on Location-Based Services,” IEEE trans. Intelligent Transportation Systems, vol. 12, no. 4, Dec. 2011.
[42] Francesco Calabrese, Member, IEEE, Massimo Colonna, Piero Lovisolo, Dario Parata, and Carlo Ratti, “Real-Time Urban Monitoring
Using Cell Phones: A Case Study in Rome,” IEEE transactions on intelligent transportation systems, vol. 12, no. 1, Mar. 2011.
[43] Afshin Abadi, Tooraj Rajabioun, and Petros A. Ioannou, Fellow, IEEE, “Traffic Flow Prediction for Road Transportation Networks
With Limited Traffic Data,” IEEE trans. Intelligent Transportation Systems, vol. 16, no. 2, Apr. 2015.
[44] José Geraldo Ribeiro, Jr., Student Member, IEEE, Miguel Elias Mitre Campista, Member, IEEE, and Luís Henrique M. K. Costa, Member, IEEE, “ COTraMS: A Collaborative and Opportunistic Traffic Monitoring System,” IEEE trans. Intelligent Transportation
Systems, vol. 15, no. 3, Jun. 2014.
[45] Po-Ta Chen, Feng Chen ; Zhen Qian, Road Traffic Congestion Monitoring in Social Media with Hinge-Loss Markov Random Fields, IEEE International Conference on Data Mining, Pages: 80 – 89, 2014
[46] Lefei Li, Social Media in China, Intelligent Transportation System, IEEE 2013
[47] http://www.theconnectivist.com/2013/07/how-google-tracks-traffic/
[48] Md. Abdus Samad Kamal, Member, IEEE, Jun-ichi Imura, Member, IEEE, Tomohisa Hayakawa, Member, IEEE, Akira Ohata, and
Kazuyuki Aihara, “A Vehicle-Intersection Coordination Scheme for Smooth Flows of Traffic Without Using Traffic Lights,” IEEE
trans. Intelligent Transportation Systems, vol.16, no. 3, Jun. 2015.
[49] Guoyu Ou, Yang Gao, Ying Liu Real-Time Vehicular Traffic Violation Detection in Traffic Monitoring Stream, 2012,
IEEE/WIC/ACM International Conferences on Web Intelligence and Intelligent Agent Technology
Other References
1. James Hardy, Lu Liu, Department of Computing and Mathematics, University of Derby, United Kingdom; “Reducing vehicular traffic congestion using available forward road capacity detection,” 2015 IEEE International Conference on Computer and Information
Technology; Ubiquitous Computing and Communications; Dependable, Autonomic and Secure Computing; Pervasive Intelligence
and Computing, Oct.26-28, 2015, pp 2092-2097
A model for predicting & detecting vehicular congestion to achieve uniform vehicular density
2. Sigurður F. Hafstein, Roland Chrobok, Andreas Pottmeier, Joachim Wahle, and Michael Schreckenberg, “A High Resolution Cellular
Automata Traffic Simulation Model with Application in a Freeway Traffic Information System,” Computer-Aided Civil and
Infrastructure Engineering Volume 19, Issue 5, pages 338–350, September 2004.
3. VTPI - Transportation Cost and Benefit Analysis II – Congestion Costs Victoria Transport Policy Institute www.vtpi.org.
4. Juan Pan, Iulian Sandu Popa, Karine Zeitouni, and Cristian Borcea, “Proactive Vehicular Traffic Rerouting for Lower Travel Time,”
IEEE Transactions on vehicular technology, Vol. 62, No. 8, October 2013.
5. Christoph Sommer, Ozan K. Tonguzy and Falko Dressler, “Adaptive Beaconing for Delay-Sensitive and Congestion- Aware Traffic
Information Systems,” IEE Vehicular Networking Conference (VNC), December 2010.
6. http://www.epa.gov/otaq/about/faq.htm#question1 “Idling Vehicle Emissions for Passenger Cars, Light-Duty Trucks, and Heavy-Duty Trucks,” 2008, Accessed March 2014.
7. SB Dikshit, “Road statistics of India, Problems & Solutions, State Quality Monitor, UPRRDA,” July 2009.
8. David Schrank, Bill Eisele, Tim Lomax, “ TTI’s 2012 Urban Mobility Report,” Dec 2013.
9. Jin-Jia Chang Yi-Hua Li, Wanjun Liao, Ing-Chau Chang, “Intersection based routing for urban vehicular communication with traffic
light consideration”, IEEE wireless communication 2012
10. Shu-Chuan Chu, John F. Roddick, Jeng-Shyang Pan, “Ant colony system with communication strategies”, ELSEVIER - Information Sciences 167 (2004) 63–76
11. Sven Koenig a, Maxim Likhachev b, and David Furcy, “Lifelong Planning A*”, Artificial Intelligence, Vol 155, Issues 1–2, May
2004, Pages 93–146
12. Juan C. Herrera1, Alexandre M. Bayen2, Departamento de Ingeniería de Transporte y Logística, Pontificia Universidad Católica de
Chile, Chile1;, Systems Engineering, Department of Civil and Environmental Engineering, University of California, Berkeley, United
States2; Incorporation of Lagrangian measurements in freeway traffic state estimation 13. Wei Zhang, Guozhen Tan, Nan Ding, and Guangyuan Wang; Traffic Congestion Evaluation and Signal Control Optimization Based
on Wireless Sensor Networks: Model and Algorithms; School of Computer Science and Technology, Dalian University of Technology,
Dalian 116023, China, [email protected], 2012 14. http://people.umass.edu/ndh/TFT/Ch15%20High.pdf Chapter 15, High-Order Models
15. Raj S Parmar, Prof. Bhushan Trivedi, “Shortest Route – Domain Dependent, Vectored Approach to Create Highly Optimized Network
for Road Traffic,” IJTTE International Journal of Traffic & Transportation Engineering – IJTTE, vol. 5, no. 1; pp 1 – 9, Jan. 2016..
16. Andrew P. Nichols, PhD, PE, “Adaptive Traffic Signal Control”, WVDOH/MPO/FHWA Planning Conference 2012,
http://www.transportation.wv.gov/highways/programplanning/plan_conf/Documents/2012PC/Adaptive%20Signal%20Control.pdf
17. Matt Selinger, P.E., PTOE and Luke Schmidt, “Adptive Traffic Control Systems in the United States”, HDR Engineering, Inc., September 2009
18. http://www.hdrinc.com/sites/all/files/content/white-papers/white-paper-images/3729-adaptive-traffic-control-systems-in-the-united-
states.pdf 19. Aleksander Stevanovic: Advanced Transportation Concepts, LLC, Salt Lake City, Utah;, Consultant, Adaptive Traffic Control
Systems, Domestic and Foreign State of Practice, National Cooperative Highway Research, NCHRP (www.national-academies.org)
20. Steven G. Shelby, Darcy M. Bullock, Doug Gettman, Raj S. Ghaman, Ziad A. Sabra, Nils Soyke, “An Overview and Performance Evaluation of ACS Lite – A Low Cost Adaptive Signal Control System.”, Submitted to the 87th TRB Annual Meeting in Washington,
DC, January 2008.
21. http://www.westernite.org/annualmeetings/sanfran10/Papers/Session%209_Papers/ITE%20Paper_9A-Fehon.pdf
22. http://www.siemens.co.uk/traffic/en/index/productssolutionsservices/signalsandcontrollers/MOVA.htm http://www.traffic-signal-
design.com/microprocessor_optimised_vehicle_actuation_mova.htm
Suggested Reading
[1] http://www.consystec.com/fampo/web/html/opscon/rr43.htm
[2] http://www.navipedia.net/index.php/Traffic_Management
[3] http://gpsworld.com/software-gnss-receiver-an-answer-for-precise-positioning-research/
[4] http://www.tomtom.com/en_us/services/live/hd-traffic/
[5] http://www.theconnectivist.com/2013/07/how-google-tracks-traffic/
[6] http://en.wikipedia.org/wiki/Satellite_navigation