An Improved DMSQPSO Algorithm For 3D UAV Path Planning
Zhenghan Zhou, Xing Yang, Ben Niu · 2023
Unmanned aerial vehicle (UAV) path planning problem requires obtaining an optimal path in complex terrains. To ensure smooth and stable paths, this study measures the stability of the horizontal direction by using the mean squared deviation of horizontal turning angles in the UAV path planning model, which is based on using height mean square deviation to measure vertical stability. Particle swarm optimization (PSO) is often used for UAV path planning due to its simplicity. However, PSO easily traps in local optima. Therefore, this study proposes dynamic multi-swarm quantum PSO (DMSQPSO) to address this problem which is based both on dynamic multi-swarm PSO (DMSPSO) and quantum PSO (QPSO). The proposed algorithm is as follows: we split the population into several sub-swarms and use the updating strategy from QPSO to improve the diversity of particle behaviors to avoid premature convergence to local optima. Additionally, we use a cubic B-spline curve to smooth the path, which makes it more suitable for real-world UAV flights. Simulation results demonstrate the superior performance of the DMSQPSO algorithm in obtaining feasible paths compared to other algorithms.