Smart Traffic Light Controller with Computer Vision

Traffic light automation is no longer a foreign topic in Malaysia as the country adopt automation years ago. Other than conventional method by using timer, automation in Malaysia works by placing pressure/weight sensor underneath roads with couple of meters gap to determine numbers of queue ca...

Full description

Main Author: Effendy, Muhammad Shahrin Nidzam
Format: Final Year Project
Language: English
Institution: Universiti Teknologi Petronas
Record Id / ISBN-0: utp-utpedia.20837 /
Published: IRC 2019
Subjects:
Online Access: http://utpedia.utp.edu.my/20837/1/Muhammad_Shahrin_Nidzam_24181.pdf
http://utpedia.utp.edu.my/20837/
Tags: Add Tag
No Tags, Be the first to tag this record!
Summary: Traffic light automation is no longer a foreign topic in Malaysia as the country adopt automation years ago. Other than conventional method by using timer, automation in Malaysia works by placing pressure/weight sensor underneath roads with couple of meters gap to determine numbers of queue cars. The problem is cost installation and maintenance is high. The sensor must be installed after construction of road finish as they need to make sure tar and other chemicals run its course. The process is similar if the sensor failure occurred and need for replacement/repair. Furthermore, maintenance process creates another problem, namely, road block. The idea of computer vision in traffic light system is to determine numbers of cars and calculate interval between traffic light in one intersection. Combination of Internet of Things (IoT) and Artificial Intelligence (AI) able to reduce numbers of components to develop intelligent traffic light system. Hence, reduce cost of installation exponentially. Expected results of the system is communication between software running as intended. Dashboard is developed to monitor system operation and performance running smoothly and detect defect within overall ecosystem.