Abstract - This paper presents the design and development of an autonomous vehicle system aimed at enhancing smart mobility and public safety. The proposed system integrates multiple technologies, including deep learning-based lane detection, LiDAR and camera sensors for perception, GPS-enabled navigation and real-time communication through a Raspberry Pi platform. To ensure robust performance, the system employs DeepLabV3 for semantic segmentation, optical flow for smooth motion tracking and a Kalman filter for prediction and stabilization of the vehicle’s path. A Proportional-Integral-Derivative (PID) controller regulates steering and speed, ensuring accurate maneuverability in various road conditions. Obstacle detection and avoidance are achieved using LiDAR, enabling the system to respond to both static and dynamic objects. A Flutter-based mobile application provides GPS tracking, real-time monitoring and manual override control, enhancing usability and safety. Experimental validation on a scaled prototype demonstrates the effectiveness of the proposed system in handling lane detection, sharp turns, obstacle avoidance and traffic sign recognition under diverse conditions. The results highlight the feasibility of deploying low-cost, embedded autonomous systems capable of addressing urban transportation challenges. This work provides a foundation for future research and large-scale implementation of intelligent transportation systems in smart city environments. Key Words: Autonomous Vehicles, Intelligent Transportation Systems, Computer Vision, Machine Learning, Sensor Fusion, Smart Mobility.
Bhat et al. (Wed,) studied this question.