School of Science and Technology 科技學院
Computing Programmes 電腦學系
| Programme | Bachelor of Computing with Honours in Internet Technology |
| Supervisor | Prof. Vanessa NG |
| Areas | Intelligent Applications |
| Year of Completion | 2019 |
| Award | 16th IEEE(HK) Computational Intelligence Chapter FYP Competition First Runners-Up |
In Hong Kong, the increasing number of vehicles and population has caused high usage of roads, which has been a major concern. The aim of this project is to implement the Intelligent Traffic Light System (ITLS) with machine learning, which increases the efficiency of the road space usage.
The project has also defined a number of sub-objectives as follows:
TensorFlow
TensorFlow is a machine learning open source framework, which provides a variety of machine learning function to the system.
Pygame
Pygame is a free and open source python programming language library for making multimedia applications like games built on top of the excellent SDL library.
Record traffic condition videos
Record traffic condition videos
Cameras will be set at vantage points in bridges. The resulting footage will then be used as the input of the object detection application.
Object detection application
The object detection application tells users the object type (vehicle or person) in the road and show the number of vehicles and pedestrians at the road intersection per hour.
Intelligent Traffic Light Traffic Model
This traffic model will simulate road result which implements ITLS, and also include the realistic feature of the fixed cycle traffic light traffic model.
Simulation
The pygame simulation will simulate the traffic models visually. It will use the data of traffic models to build a simulation that visualizes the concept of the system.
Simulation result on normal traffic hours
The red part is the average pedestrian waiting time in seconds, and the blue part is the average car waiting time in seconds.
The above figures show that the average waiting time of ITLS was a lot lower than the fixed cycle traffic light system about 50% to 55%, showing that ITLS served well at normal and peak hours.
