Advancing Non-ADAS Vehicles with ML Based ADAS Modules and Technologies

Learn about an innovative road navigation system designed to bring autonomous capabilities to vehicles without Advanced Driver Assistance Systems (ADAS) in this 20-minute conference talk from DevConf.IN 2025. Presented by Aaryan Joshi, Taksh Dhabalia, and Samanyu Bhate, discover how YOLOv8-based object detection combined with CANBUS communication creates a comprehensive solution for safer transportation. Explore how the system identifies road hazards in real-time and uses STM32 microcontrollers with CAN shields to dynamically adjust vehicle behavior. The presentation covers the modular hardware design that allows retrofitting into existing vehicles with minimal modifications, sensor fusion techniques for reliable operation in various conditions, and how this cost-effective approach improves road safety and vehicle intelligence.