基于改进分散搜索算法的无人机路径规划

A Revised Scatter Search Algorithm for Path Planning of Multiple UAVs

针对在敌情信息不明环境中无人机侦查路径规划问题,建立了车辆路由问题模型(VRP),提出了基于分散搜索的改进混合搜索算法.基于Bayes方法计算出点到点之间的威胁概率,并生成了一个赋权图,将无人机路径规划问题转化为车辆路由寻优模型.采用混合路径规划算法求解.该算法将模拟退火嵌入到分散搜索算法的框架中,充分利用了分散搜索的全局搜索能力与模拟退火的局部搜索能力来优化无人机的侦查路径,混合算法在保证时效性的同时提升了求解的质量.仿真结果验证了算法的有效性.

The unmanned aerial vehicle（UAV） path planning problem in uncertain and adversarial environment is modeled as a vehicle routing problem（VRP）. After that a revised hybrid algorithm based on scatter search optimization was proposed.First,with the prior surveillance and experiential evaluation,Bayes rule is used to compute the probability of threats of flight across each pair of neighboring points,and then a weighted graph can be generated based on the threat probability map in the given planning area.The original problem is transformed to be a VRP.Then,a hybrid routing algorithm is adopted to solve the VRP problem.The proposed algorithm incorporates simulated annealing（SA） method into scatter search（SS）,such that it can take advantages of both the global search ability of SS and the local optimization capability of SA in order to get good paths.The proposed method can improve the quality of solutions while not（incurring） additional time.Finally,computational experiments were conducted to verify the method.