RAPID RANDOM TREE AND DUBINS PATH ALGORITHM FOR DRONE OBSTACLE AVOIDANCE

Unmanned aerial vehicles (UAVs) have been commonly used for domestic and military surveillance. Path planning algorithms are one of the many tools that can be used to control UAVs to make the flight autonomous and decrease the involvement of human pilots. The purpose of this study is to develop and implement path planning algorithms to enhance the navigation capabilities of drones in complex environments. Rapidly-exploring Random Tree (RRT) is one of the algorithms that uses sampling based methods to find possible flight paths in highly-dimension space. This research presents the simulation of the algorithm testing in various maps with randomly placed obstacles in MATLAB software to see whether the algorithm managed to manoeuvre around the obstacle successfully with minimum collisions. To further refine the generated path, Dubins path smoothing techniques are implemented in the simulation, to ensure a smoother and more feasible trajectories for drone navigation. The path generated is observed through the result shown in the simulation in both 2D view and 3D view for 3 different random cases. To summarize, the application of Dubins path smoothing significantly improves the efficiency of the path-planning algorithm and reduces the time to complete the obstacle navigation by 10-27% depending on the complexity of the map. The smoothed paths are more direct, involve fewer abrupt turns, and consistently reduce the time to reach the goal across different maps.