摘要
无人机巡逻相对于人工巡逻具有视野广、速度快、受地形影响小及无生理局限等优点,因此无人机在武警巡逻勤务中得到了越来越广泛应用。但由于目前警用无人机一般为近程无人机,扫描半径小、续航时间短,航速受限,当巡逻区域较大,巡逻任务较重时,往往涉及多基地多无人机协同巡逻问题,巡逻任务分配复杂,有必要采取科学方法手段从全局角度对各无人机巡逻路线进行合理规划以提高巡逻效率。为此,文中以离散目标巡逻覆盖问题为例,首先基于无人机扫描半径将目标区域网格化,将目标覆盖问题转化为无人机对目标网格中心访问的多车场车辆路径问题,然后以总巡逻时长最短,巡逻无人机数最少,巡逻任务分配均匀度罚金和巡逻超时罚金最省为目标建立了多目标规划模型,最后通过染色体的编码设计,适应度函数构造,变异及交叉概率的自适应设置及遗传算法和模拟退火算法结合,提出了一种混合遗传算法(HYBRID-GA),仿真结果验证了HYBRID-GA对基本遗传算法(GA)的明显优越性。
Compared with manual patrol,UAV(Unmanned Aerial Vehicle)patrol has many merits,such as wider vision,and faster speed,and it is not affected by human physiological limitations.So,UAV patrol patrols are currently being used more and more widely in PAP patrol patrols.But,because police UAV is often short-range UAV,it has some demerits,such as short scan radius,short duration and slow speed.When the patrol area is large and the task is heavy,the problem becomes a complex target coverage problem of multi-base UAV cooperative patrol.It is necessary to employ scientific methods and means to plan the UAVs'patrol routes from a global perspective.Therefore,the discrete Target patrol Coverage Problem is taken as an example in this paper.First,the target area is gridded based on the UAV's scan radius,then the target coverage problem is transformed into the multi-depots vehicle path problem,and the target coverage problem is transformed into the visit of the target grid center.A multi-objective programming model is established with the shortest total patrol time,the least patrol UAVs,and the least penalty for patrol task distribution uniformity and patrol overtime.By the design of chromosome code and fitness function,adaptive setting of mutation and crossover probability,and combination of genetic algorithm and simulated annealing algorithm,a hybrid genetic algorithm(HYBRID-GA)is proposed.Simulation results show that the HYBRID-GA is superior to the basic algorithm(GA).
作者
王书勤
黄茜
WANG Shu-qin;HUANG Qian(Military Command Department,Officers College of PAP,Chengdou Sichuan 610213,China;Basic Courses Department,Officers College of PAP,Chengdou Sichuan 610213,China)
出处
《计算机仿真》
2025年第3期55-60,共6页
Computer Simulation
基金
军内十四五规划项目(145BZB180001000X,WJ2022B 010100)。
作者简介
王书勤(1976-),男(汉族),湖南祁东人,硕士研究生,教授,研究方向为路径优化和人工智能算法;黄茜(1983-),女(汉族),河南正阳人,硕士研究生,副教授,研究方向为路径优化和人工智能算法。
