Real-time dynamic path planning for UAVs using the improved RRT algorithm

This study presents the Improved Rapidly-Exploring Random Tree (IRRT) algorithm for path planning for UAVs in critical applications such as search and rescue, disaster management, and logistics. Although the Rapidly-Exploring Random Tree (RRT) algorithm is a popular method used in complex search and path planning, it has some limitations, such as inefficient exploration, real-time adaptability, and fixed step size. To address the deficiencies of RRT, novel variations, including RRT-Star, RRT-Connect, Bi-Directional APF-RRT*, and APF-RRT, have been introduced in the literature. It is stated that improvements in path length often increase computational cost, while reducing it may lead to longer paths. In this study, the limitations of classical RRT and its variants have been surpassed, and innovations such as dynamic step size, path smoothing strategy, and optimal objective function have been introduced. The study enhanced the RRT algorithm’s fundamental structure. The proposed algorithm aims to increase the convergence speed of the RRT algorithm, dynamically adjust the step size, and create more efficient paths with the path smoothing strategy. The proposed algorithm is tested in static and dynamic environment simulations. In the experimental results, the performance of the IRRT algorithm has been tested in both static and dynamic environments and has shown superior results compared to the existing RRT, RRT*, RRT-Connect, APF-RRT, and Bi-Directional APF-RRT* algorithms. The tests indicated that the IRRT and RRT-Connect algorithms offer times of less than 1 s when calculating a new node. Additionally, in path planning, the IRRT algorithm offers a route that is 20% shorter than the RRT-Connect algorithm in terms of route length. These results indicate that the IRRT algorithm offers a potential solution for real-world UAV applications.