Road Pulse

Abstract

According to the traffic conditions in Pakistan, traffic violations and anomalies occur daily. Rush drivers do not respect traffic rules and regulations which is a major concern. According to a survey each family in Pakistan owns a vehicle whether it’s a car or bike. This results in issues like traffic congestion as the number of vehicles are increasing. Transportation demand exceeds the road capacity. So, we provided an intelligent system to improve traffic management and control Road Pulse. Building new roads, hiring new police officers is not a good option. Fast reaction and prevention of traffic violations play a key role to ensure the safety of vehicles and citizens. For this purpose, in Pakistan surveillance cameras are installed everywhere to prevent traffic violations but they are unable to detect events like a Lane violation, vehicle crash events etc. due to the faulty architecture. In Pakistan, traffic issues are rising day by day which are causing many causalities like accidents, loss of innocent lives, and compromise in the safety of citizens. Our project, Road Pulse, is an application based upon Web as well as Android modules that detects the traffic violations traffic anomalies that are happening daily in Pakistan. Violations and anomalies like lane violations, vehicle crash, etc. which results in many tragic events. The app uses a pre-trained model that’s embedded within the Web app. The pre-trained model YOLO v3 is a Fully Convolutional Neural Network that uses machine learning techniques to help us in predicting the violations accurately. The model is trained on a relevant dataset to achieve accuracy. Efforts directed at enforcing traffic laws lack due to no improper surveillance. On the other hand, surveillance systems are limited in duration and intensity, thereby producing weak halo effects on drivers' behaviour. Also, the public attitudes towards compliance are largely conditioned by the visibility of Automated surveillance systems. The effects of the development on driver’s behaviour are reviewed and proposals for the improvement of a modified traffic policing structure are urged out. By the presentation of such a framework many innocent lives can be saved and citizens safety can also be restored. The burden of policing departments can also be reduced by introducing such an intelligent and automated surveillance system.