AI Autonomous Vehicle Simulator
Real-time optimal path planning for an autonomous vehicle.
This project simulates autonomous vehicle navigation using a ROS2 publisher-subscriber architecture built on the DDS protocol. The system generates a dynamic cost-map from LiDAR and odometry data in real time, enabling the vehicle to detect and reason about obstacles as the environment changes.
Navigation combines the Pure Pursuit algorithm for smooth trajectory following with A* path planning for optimal route generation, allowing the vehicle to adapt its path on the fly when obstacles are encountered. The cost-map is continuously updated from sensor input, ensuring the planner always works from a current representation of the environment.
The project was built as part of the WATonomous Autonomous Systems Division, focusing on scalable, modular ROS2 node design for perception, planning, and control pipelines.