AUTOMATIC PATH PLANNING FOR UNMANNED AERIAL VEHICLES(UAVs) IN DYNAMIC AND UNCERTAIN ENVIRONMENT

Abstract

As Unmanned Aerial Vehicles (UAVs) continue to play a vital role in diverse applications, the demand for robust and adaptive path planning algorithms becomes increasingly imperative. This study presents an innovative approach for automatic path planning tailored for UAVs navigating dynamic and uncertain environments. The proposed algorithm integrates a Random Tree structure with a Limited Tree Depth strategy, aiming to strike a balance between exploration and exploitation in complex scenarios. This is achieved via the development of an adaptive path planning algorithm, for addressing of dynamic obstacle avoidance, evaluation of robust handling of uncertainty, ensuring of real time decision-making, the utilization of random tree with limited tree depth, and the deployment of the developed algorithm in the simulation setting. The methodology encompasses the design, implementation, and evaluation of the path planning algorithm. Leveraging insights from the literature, the algorithm integrates mechanisms for dynamic obstacle avoidance and uncertainty handling. The limited tree depth approach optimizes the exploration-exploitation trade-off, ensuring real-time adaptability. The key mathematical relationships include steering the tree expansion, predicting dynamic obstacle movement, adapting to uncertainty through probability distributions, optimizing real-time decision-making, and dynamically adjusting the limited tree depth. The algorithm is rigorously evaluated within diverse simulation scenarios, featuring dynamic obstacles, uncertain terrains, and complex environments. Quantitative metrics such as path length and computation time, along with qualitative assessments of collision avoidance and adaptability, are employed for comprehensive analysis. The result presents the superiority of the improved random tree with limited tree over the traditional random tree in high convergence and generalization of data. The study contributes a novel path planning algorithm specifically tailored for UAVs in dynamic and uncertain environments. Results demonstrate the algorithm's efficacy in achieving adaptive trajectories while ensuring real-time responsiveness. The convergence analysis revealed that the IRRT exhibits a faster convergence rate, surpassing the RRT by 1.4 seconds. This signifies the algorithm's efficiency in achieving super-optimal obstacle avoidance missions.