关键词: 无人机轨迹规划, A^*算法, 高斯伪谱法, 多目标动态均衡, 障碍物运动模式

Abstract:

To address the challenges of high dynamics, strong real-time requirements, and multiple constraints in unmanned aerial vehicle trajectory planning within complex battlefield environments, and to overcome the limitations of traditional methods, such as their sensitivity to initial solutions and poor numerical stability, an integrated trajectory planning framework combined the A* algorithm with the Gaussian pseudospectral method （GPM） is proposed. Firstly, a multi-objective weighted joint optimization function is designed, with balanced optimization achieved via interior point optimizer. Then, the A^* algorithm is used to rapidly generate an optimal initial path, which is smoothed by interpolation to serve as a high-quality initial solution for GPM. Combined with Legendre-Gauss node and constraint coupling strategy, the Runge phenomenon is avoided. Finally, for the first time, the influence of three typical dynamic obstacle motion patterns—random walk, directed interception, and periodic patrol—on planning performance is systematically analyzed, filling the research gap where existing hybrid algorithms ignore dynamic environmental characteristics. Simulations show that compared with traditional GPM, the proposed algorithm shortens the path by 2.69%, reduces optimization iterations and function evaluations by 49% and 49.1%, respectively, and exhibits superior economy, real-time responsiveness, and obstacle avoidance success rate in dynamic scenarios, providing an efficient solution for resource-constrained battlefields.

Key words: unmanned aerial vehicle （UAV） trajectory planning, A^* algorithm, Gaussian pseudospectral method （GPM）, multi-objective dynamic equilibrium, obstacle motion pattern