【目的】多固定翼无人机(UnmannedAerialVehicle,UAV)的任务分配和路径规划问题中,将固定翼UAV简单地视为质点,未考虑自身航向角、初始和终止速度向量等问题。针对多固定翼UAV任务规划问题,将改进的蚁群算法(AntColonyOptimization,ACO)和自组织映射(Self-OrganizingMap,SOM)算法相结合,提出ACO-SOM算法应用于多无人机系统。【方法】通过SOM进行任务分配,引入蚁群状态转移概率方法,调整SOM获胜神经元的权值,使得固定翼UAV按照最优Dubins路线依次到达各个目标点。根据巡航能力和运动学约束合理分配任务,构造多目标优化函数,实现多固定翼UAV的协同执行任务能力。【结果】分别在障碍物、不同起落点场景下对ACO-SOM算法进行仿真验证,实验结果表明,该方法能使多固定翼UAV合理任务分配并规划出最优路径,具有较高的避障和规划能力。

【Purposes】The task allocation and path planning of multi-fixed-wing Unmanned  Aerial Vehicle (UAV) simply regard it as a particle, thus the problems of ignoring its own heading  angle, initial and terminal velocity vectors exist. Aiming at the problem of multi-fixed-wing UAV  mission planning, improved Ant Colony Optimization (ACO) algorithm and Self-Organizing Map  (SOM) algorithm are combined to propose ACO-SOM algorithm for multi-UAV system. 【Methods】 The SOM was used for task allocation, and the ant colony state transition probability method was in⁃ troduced to adjust the weights of SOM winning neurons, so that the fixed-wing UAV reached each  task point in turn according to the optimal dubins route. According to the cruise ability and kinematic  constraints, the tasks were reasonably allocated, and the multi-objective optimization function was  constructed to realize the cooperative task execution ability of multiple fixed-wing UAVs. The ACO- SOM algorithm was simulated and verified in the scenarios of obstacles and different landing points.  【Findings】Experimental results show that, compared with SOM, the proposed method can make the  multi-fixed-wing UAV reasonably assign tasks and plan the optimal path, and has higher obstacle  avoidance and planning abilities.