Heuristic Motion Planning for Autonomous Outdoor Vehicle: A Lane Detection and Tracking Approach

The autonomous vehicle is a driverless vehicle that can travel from a starting point to a predetermined destination by detecting and responding to its local surroundings using various vehicle technologies and sensors. Planner modules provide collision-free waypoints on the route to the destination which serve as the foundation for the decision made by autonomous vehicles to avoid both static and dynamic obstacles. The aim of this project is to develop motion planning architecture for outdoor vehicles that can manoeuvre in indicated routes using visual based lane detection. An autonomous navigation system is developed based on image-based lane detection and heuristic algorithm to determine a suitable field of view for the vehicle to safely manoeuvre. The proposed algorithm was developed using ROS (Robot Operating System), with OpenCV tools employed for image processing. The TurtleBot 3 Waffle Pi model was utilized to evaluate the algorithm. The algorithm proposed it tested in 3 cases which are absent of obstacle, present of static, and present of dynamic objects. It was found that the results show that the vehicle can move autonomously with the successful rate of more than 95% in every case tested.