The problem Traffic Control Using AI solves

When it comes to traffic, from big metropolitan cities to small towns, everyone is facing the same problem.

Even in the largest cities like Mumbai, New Delhi, and Bangalore, unregulated traffic causes huge traffic jams which lead to a loss of a lot of man-hours and are responsible for extensive unnecessary usage of fuels due to the slow movement of vehicles.

We propose to solve this issue by implementing the use of AI and ML.
The traffic density will be measured by the video recognition algorithm, and depending on the outcome the traffic signal throughout the city will work in sync to provide appropriate red and green lights.

The algorithm will be specifically designed to identify and prioritize emergency vehicles so that In case an Emergency vehicle is detected, that roadway will be given a priority green light as soon as possible.

The algorithm for vehicle detection through a live feed of traffic signals video camera is written in python and uses libraries such as cv2 and NumPy. These will detect the number of vehicles in front of a particular signal and thus measure the traffic density. The data from one signal will be matched to the data from other signals present at the road intersection and AI will allot the order and timing to the traffic signals accordingly. Further, this complete data of a road intersection will be matched to the other road intersections of the city and the same process will repeat.

Since the program will use AI & ML the results generated by the algorithm regarding timings and coordination of traffic signals throughout the city will act as a dataset for the next execution of the program and subsequently its dataset will grow. All of this data will be managed and stored by Database Management System(DBMS). For this databases like MySQL or MongoDB will be put to use and linked through the python program. As the dataset grows, the efficiency and accuracy of the AI algorithm to generate a viable solution will improve.

Challenges we ran into

we were not able to get a valid prototype up and running and as expressed by the judges feel that there are some flaws in the revenue model and there will be problems in generating accurate results where the cameras and the required infrastructure is not already present.