UAVs Path Planning Using Visual-SLAM Technique Based Hybrid Particle Swarm Optimization

Abstract

Abstract Due to their excellent mobility on various robotics platforms, unmanned-aerial vehicles (UAVs) are becoming extremely popular. We trace the UAV poses while simultaneously creating an iterative and progressive map of the surrounding area using a cutting-edge VSLAM technique, termed as visual simultaneous localization and mapping. In this case, a single UAV initially creates a map of the area of interest using a monocular vision-based method. In order to determine the best pathways for several UAVs, the created map is treated as an input for the optimization method. UAVs need to execute missions effectively, and they need to access the best route quickly in a challenging environment. This necessitates solving the automatic path planning problem. In this paper, a new hybrid particle swarm optimization (HPSO) technique is suggested as a solution to this issue. The proposed algorithm enhances the optimization capability and prevents dropping into local convergence by combining the simulated annealing algorithm; each particle integrates the advantageous information of the optimization method in accordance with the dimensional learning approach, which reduces the occurrence of particles fluctuation during the transition process and improves the convergence speed. Additionally, we proposed dynamic fitness function (DFF) in order to assess the path planner’s planning approach while taking into account a variety of optimization parameters, including the calculation of flight risk, energy usage, and operation completion time. The efficiency of our proposed H-PSO-VSLAM system, as shown by the simulation results, is validated by the recommended planner’s high fitness value and safe arrival at the destination while avoiding all unanticipated dangerous events and restricted locations.