wireless sensor-based traffic light control

Wireless Sensor-Based Traffic Light Control

Malik Tubaishat, Qi Qi, Yi Shang, Hongchi Shi

Problem Statement

• Comparison of performance of wireless sensor based networks(Single sensor and dual sensor) with fixed traffic cycle TLC.

• Non-Occupancy Detection(NOD) and Occupancy Detection(OD) based comparision

• Effect of sensor distance on Average waiting time.

Results

• Used GLD(Green Light District) for their simulation

• Implemented Fixed Light controller, Single Sensor based and Fixed Sensor based

• Some values:– Random Spawn Frequency : 0.1-0.35– Cycle length 5000 averaged over 5 cycles.

(a) urban-like map and (b) downtown-like map

One Sensor- Urban Network

One Sensor- Downtown Network

Key take-aways

• Two Sensor Model outperforms one Sensor Model

• Two sensor model performance is close to best first which knows the exact number of waiting vehicles.

• OD model, the closer the sensor to thejunction the better the performance is.

Sensor Placement in Real-Time Sensor Based Traffic Light Control

• Infrastructure Based Controllers• Performance based Measurements• Control Decision through Sensor data

Outline

• Agent based Controllers– Learn the expected waiting times of vehicles and

set the traffic lights accordingly

• Infrastructure Based Controllers– Change the traffic flow based on the current flow

of Traffic– They basically know the queue length and act on

this information

List of Controllers:• Random: Switched Traffic at Random• Most Cars: Move lanes regardless of the queue

length• Best First: Give priority to lanes with maximum

length• Longest Queue: Selects only the lane with

maximum length• Relative Longest Queue: No. off road users is

divided on the basis length of the lane to get relative rewards for traffic light settings

Control Decision Through Sensor Data

One Sensor Based Controller:

• NOD-Non-Occupancy Based Detection• OD-Occupancy Based Detection

• Two Sensor-based Controller

• The Gain Value is the queue length for infrastructure-based TLC.

• In every cycle the TLC adds the gain value for the possible sign configurations and selects the sign configuration with the maximum gain value

Future Work

• Focus work more towards increasing the life time of sensors

• Study performance in realtime with smaller intervals.

Reference• Wireless Sensor-Based Traffic Light Control, Malik Tubaishat, Qi Qi, Yi

Shang, Hongchi Shi, Consumer Communications and Networking Conference, 2008. CCNC 2008. 5th IEEE

Thank You

-Questions...