Utilizing technology to prevent road accidents: design and implementation of a sleep detection system for drivers

Abstract

India is one of the most impacted nations when it comes to road accidents. With more than 1.5 lakh annual deaths, it is a serious concern for our nation. Out of the many reasons contributing, fatigue causes sleep and loss of attention resulting in more than 10,000 accidents. This is one of the factors that can be controlled and completely avoided if the driver is timely alerted in such a state. In the insurance sector, India is one of the most prominent players, with 10% of the global market share, but insurance penetration is extremely low, especially in the rural population. With the introduction of government initiatives, private players gaining market share (already the majority market share in non-life insurance), and rising awareness among consumers, the penetration has steadily increased over the last couple of decades. The project focuses on developing a solution to detect sleep while driving and explores the use case in the motor insurance sector. The current solutions are either too complex or require high diagnosis time, limiting their utility in detecting sleep while driving. The solution proposed utilizes Machine Learning, Computer Vision and Image processing capabilities and an audiobased alert mechanism to detect sleep and alert the driver in case they tend to fall asleep by detecting a closure of an eye for longer than 4 seconds. Finally, to evaluate the practical feasibility of such a solution, a business Model has been constructed using Business Model Canvas (BMC).