基于聚类与动态规划的无人机协同态势感知研究

UAV collaborative situational awareness based on clustering and dynamic programming

当前,新一代人工智能技术、无人技术、自主技术广泛运用于国防、安全、民生、经济等多维领域。基于无人机协同态势感知任务规划问题显得越发重要。文中重点研究在特定背景下,多无人机对目标搜索的最优概率。首先利用Floyd算法解决两点之间的最短距离,对不同目标运动规律的热力图数据进行聚类分析;其次利用熵权法对聚类区域进行评价排序,根据无人机模型以及雷达检测出目标的概率模型分别对F1、F2两型无人机进行模型建立;再利用贪心算法依次访问概率和最高的聚类区域,依次对F1、F2两型无人机在规定时间内访问聚类区域的概率进行求和,求得一个相对最优解;最后利用模拟退火算法依次检验每一个可行解,并与最优解进行比较,如果优于最优解,则进行替换,否则继续检验下一个可行解,直到覆盖所有可行解路径,得到全局最优解。通过实验仿真验证了结果的有效性。

At present,new⁃generation artificial intelligence technologies,unmanned technologies and autonomous technologies are widely used in multi⁃dimensional fields,such as national defense,security,people′s livelihood and economy.Mission planning based on UAV collaborative situational awareness becomes more and more important.This paper focuses on the optimal probability of multi⁃UAV searching for targets in a specific background.The Floyd algorithm is used to get the shortest distance between two points,and clustering analysis of thermal map data with motion laws of different targets is carried out.The entropy weight method is used to evaluate and sort the cluster area.F1 UAV and F2 UAV models are established respectively according to the UAV model and the probability model of radar target detection.The greedy algorithm is used to visit the probability and the highest clustering region successively,and the probabilities of F1 UAV and F2 UAV visiting the clustering region within the specified time is summed successively to obtain a relative optimal solution.The simulated annealing algorithm is used to verify each feasible solution successively and then each feasible solution is compared with the optimal solution.If it is better than the optimal solution,it will be used to replace the relative optimal solution.Otherwise,the next feasible solution is checked until all feasible solution paths are covered and the global optimal solution is obtained.The validity of the results is verified by experimental simulation.