anandnitt / road-paradox

Smart traffic control system

Machine Vision Assisted Elimination of Traffic using Intelligent Control

The traffic density on roads has been on an exponential rise since the advent of automobile revolution. Since then there has been some approaches for traffic management such as traffic light control based on load, traffic density calculation based on IR etc. An efficient algorithm using image processing techniques and intelligent control is proposed to completely obliterate traffic. The prototype has been developed and successfully tested using a miniature model of a rover and a wall mounted camera. In real time environments, this setup is analogous to cars with GPS and satellite cameras giving live traffic data on roads. The algorithm works in such a way that the vehicles are tuned to move at a calculated speed at which they avoid stopping at traffic signals which leads to reduction of traffic. This approach also helps us in the reduction of vehicle particulate emissions and improving the fuel efficiency, battery life in case of electric vehicles.

Image seen from camera frame without perspective transformation:

Perspective Transformation:

Experiments have been carried out and data has been logged for certain conditions. Switching time of the signal has been varied as 1,1.5,2 seconds. If the switching time of the signal is 1s, and distance of rover from signal is less than 100cm, a speed of 30 rpm would be sufficient to cross the signal. When the distance of rover from signal exceeds 250 cm, then predicted would exceed the maximum speed of vehicle, so the algorithm predicts the new speed to make the rover cross the signal at the next iteration, so the new predicted speed is brought down to 40 rpm. It is sufficient that the rover maintains a minimum speed of 30 rpm for distance less than 150cm in case of signal switching time being 1.5s and distance less than 200cm in case of signal switching time being 2s, after which the speed increases gradually with the distance.

A simulation was tested and validated successfully using Anylogic software. The stop lanes are analogous to the signals in the below photo.

Simulation Output:

Instantaneous Output seen in Terminal:

Mentions :

A research paper has been accepted relating to this work in a IEEE conference.

V Jeya Maria Jose, M Anand, Vibashan V S, DC Vivek; "Machine Vision Assisted Elimination of Traffic using Intelligent Control" ; IEEE International Conference on Inventive Systems and Control 2018 ( ICISC 2018 )

Simulation Video Link :

https://drive.google.com/file/d/1YbCorRTut3Z8BakVc2NcKeiG2ujL_xa1/view?usp=sharing