Traffic-MAN
Smart traffic Light Everyone must have stopped a lot at traffic lights for such unnecessary reasons and wasted a lot of time. This project aims to make a system to automate the process and sort the traffic out efficiently, cutting short travel times. Also ambulances, fire brigades etc are stopped at traffic lights due to the traffic. To make them pass, the whole flow of traffic is affected which sometimes results in a traffic jam. This system tries to work around this problem and regulate the flow of traffic. Here’s a short video giving us a brief about the project The aim of this project is to: The project was divided into four parts according to the roles each part does: Image processing program detects the vehicles in images taken at the intersection. After processing them with YoloV3, the number of vehicles along with their types are stored in a database. The database is used by the machine learning subsystem to learn and predict red times for the traffic light. Machine learning program processes the given database to allot best clearing time for vehicles at the intersection. It has a threshold algorithm to train the model while the program is working. Linear regression is used to train light times. Both Image Processing and Machine Learning subsystem act together to create a Traffic Light. These traffic lights are what constitute the final program. The traffic program is the main part of the whole project. It handles all the values provided by the Traffic Lights present at the intersection to manage the traffic. It gives a green light to the intersection which is the most crowded. In order to ensure all traffic lights are covered equally, the program will make sure that all other lights are made green before repeating it again. It also handles the emergency vehicles. After receiving a emergency call at one of the traffic lights, the program will make that way green and others red until the emergency signal is turned off. In order to simulate a real traffic light, a simple UI using Django was made. One can cycle images to change the number of cars present at the intersection. Buttons were added to indicate emergency situations at any traffic light. Some of the improvements in the project could be: This project was made by Shubham Ojha (me), Mayur Ware, Akash Fokane and Tanmay Hiremath as part of the Institute Technical Summer Program 2020 conducted by IIT Bombay.Aim
Implementation
Image Processing
Machine Learning
Traffic Program
User Interface

Future Plans
Acknowledgements