Abstract:
Road accidents continue to be a major problem across the globe with overspeeding and driver fatigue being some of the major reasons for fatal accidents. Vehicle safety systems at present are such that each system focuses on one risk factor like speed control or driver state monitoring and does not handle multiple risk factors together using one single system. There is a need for a new system which can handle multiple risk factors together and act in a real time mode without any intervention from the driver. The Auto Sentinel aims to bridge this gap by integrating a driver awareness system with a brake reinforcement system. It comprises three key components, namely, a GPS module to constantly update the vehicle’s current speed, a vision detection system which is camera based and lights and sounds to provide warnings and alerts. The prototype developed on a Raspberry Pi platform has a camera using real-time facial behavior detection using YOLOv8, a NEO-6M GPS component that measures the vehicle’s speed, and the brakes is a servo hub-based actuator. The system runs continuously taking in the speed as well as driver state information in to make onthe-fly decisions without human intervention. Cost effective, modular design is suitable to be validated in the laboratory using low speed studies. Planned additional features involve auto, real-time connection to the vehicle’s ECUs for direct application of the brakes and throttle, cloud upload and access for logging events and accidents from anywhere, and experimentation with more sophisticated AI models for more detailed classification of driver behavior. This Auto Sentinel prototype lays the groundwork for intelligent, accident-resilient transportation
