基于MOEA/D-ARMS的无人机在线航迹规划

无人机(unmanned aerial vehicle, UAV)在线航迹规划是UAV协同控制关键技术之一, 在线航迹规划问题本质上是一种动态多目标优化问题。为了求解该问题, 提出了一种基于自适应应答机制选择的动态多目标进化算法(multi-objective evolutionary algorithon based on decomposition-adaptive reaction mechanism selection, MOEA/D-ARMS)。多种应答机制构成应答机制池, 以应答机制最近一次的整体表现赋予应答机制一定的奖励, 并采用基于概率的方法从应答机制池中选择应答机制。MOEA/D-ARMS分别在静态环境情况、突发威胁情况、突变威胁情况和偏好改变情况下进行仿真实验。仿真结果表明, MOEA/D-ARMS可有效求解UAV在线航迹规划问题。

UAV online trajectory planning based on MOEA/D-ARMS

Unmanned aerial vehicle (UAV) online trajectory planning is one of the key technologies of UAV collaborative control. The problem of online trajectory planning is essentially a dynamic multi-objective optimization problem. To solve this problem, a dynamic multi-objective evolutionry algorithm based on decomposition-adaptive reaction mechanism selection (MOEA/D-ARMS) is proposed. A variety of response mechanisms constitute a response mechanism pool. The response mechanism is given a certain reward based on the most recent overall performance of the response mechanism, and the response mechanism is selected from the response mechanism pool using a probability-based method. MOEA/D-ARMS is applied to four simulation instances, such as static environment instance, sudden threat instance, mutation threat instance and preference change instance. The simulation results demonstrate MOEA/D-ARMS is effective in solving the UAV online trajectory planning problem.