Published Paper Details:

Globally, more than 135 million people die annually as a result of road traffic accidents. Of those fatalities, India experiences more than 155,000 each year.These staggering statistics indicate that the current strategies in place to increase road safety are less in preventing road collisions. The system has been created with one purpose in mind -- to prevent road collisions/accidents before they happen rather than after they occur. This project has brought together experts in two different areas of science: computer science (specifically through machine learning and still-image preprocessing) and electronics & communications engineering (specifically through sensor design, wireless communication, and signal processing). We are engaging in a collaborative effort among these two groups of experts within an entirely new paradigm called transdisciplinary research. The proposed system consists of five layers in its design. At the bottom layer are three sensors, an ultrasonic proximity detection array and a LiDAR, which facilitate non-contact detection of a vehicle's approximate location. The next layer above the sensors is an AI-powered collision prediction engine that can predict possible collisions with vehicles or objects up to several seconds before they would otherwise occur. A vehicle-to-vehicle (V2V) communications module allows for real-time sharing of hazard information between two vehicles; and an automated emergency response interface completes the loop by acting on the information provided by the V2V communications system without waiting for human input. To substantiate the design with empirical evidence, we conducted four case studies of deployed smart safety systems: the Euro NCAP AEB program in Europe, the ITS Connect V2V system in Japan, the IHMCL smart highway pilot in India, and Tesla Autopilot's collision avoidance functions. The collective lessons learned from these four studies were used to inform the technical architecture and policy recommendations of the proposed system. The key outputs of this project include a conceptual system prototype, a multi-dimensional feasibility analysis, and a specific set of recommendations regarding India's road infrastructure and automotive laws/regulations necessary for the successful implementation of smart vehicle safety technologies.