A Comparative Study of Path Planning Algorithms for Autonomous UAV Navigation in Dynamic Environments

The Unmanned Aerial Vehicles (UAVs) that are autonomous need reliable path-planning algorithms to be operated safely such environment of obstacles and uncertainty. This is more imperative when dealing with those cases where there is a need to make decisions quickly and adaptability to changes in the environment. This paper involves a comparative analysis with the aim of analyzing the performance of some of the most popular graph based and sampling based path-planning algorithms in three-dimensional UAV navigation. The analysis is conducted on A*, RRT*, RRT-Connect, and DG-RRT, which are all considered in the same set of circumstances involving static and dynamic obstacles in a single Python-based simulation framework. Performance is assessed by means of metrics like computation time, computational cost, and path length. The findings have shown that, A* has predictable and effective performance in real-time navigation as its computational overhead is low, sampling-based planners are likely to generate higher quality paths at the cost of high computation. The results of this paper provide useful advice in the choice of appropriate path-planning algorithms during real-time autonomous UAVs.