一种基于目标声信息改进概率图的多UUV协同搜索方法  

A Cooperative Search Algorithm Based on Improved Probability Map of Target Acoustic Information for Multiple UUVs

作　　者：杨惠珍[1,2] 周卓彧 李源 YANG Huizhen;ZHOU Zhuoyu;LI Yuan(School of Marine Science and Technology,Northwestern Polytechnical University,Xi’an 710072,China;UnderwaterInformation and Control Laboratory,Xi’an 710072,China)

机构地区：[1]西北工业大学航海学院,陕西西安710072 [2]水下信息与控制全国重点实验室,陕西西安710072

出　　处：《水下无人系统学报》2023年第5期715-724,共10页Journal of Unmanned Undersea Systems

基　　金：水下信息与控制全国重点实验室基金项目资助(2021-JCJQ-LB-030-03)。

摘　　要：针对未知环境中多无人水下航行器(UUV)协同目标搜索问题,提出一种基于目标声信息改进概率图的多UUV协同搜索方法。建立了包含目标声场信息、UUV占用信息、目标存在概率及环境确定度的改进概率图,使UUV对动态搜索环境及目标信息的感知更加准确、全面;提出了一种基于学习机制和自适应参数调节机制的改进粒子群优化(PSO)算法,将基于线性种群规模减小和广泛学习机制的自适应差分进化算法的突变策略引入PSO算法,通过生成自适应调整参数的突变粒子,增加粒子多样性,在多UUV目标搜索应用中,减少了局部最优,提高搜索效率;设计开发了仿真程序,应用蒙特卡洛仿真方法验证分析了多UUV搜索效能。仿真结果表明,所提出的多UUV协同搜索方法与基于传统概率图的PSO搜索算法相比,同样条件下找到目标所花费的时间更少、找到的目标数量更多,对动态目标搜索具有较明显的优势。In view of the cooperative target search of multiple unmanned undersea vehicles(UUVs)in unknownenvironments,a cooperative search algorithm based on an improved probability map of target acoustic information for multipleUUVs was proposed.An improved probability map based on target acoustic information,UUV occupancy information,targetexistence probability,and environment certainty was established,making the UUV’s perception of dynamic searchenvironment and target information more accurate and comprehensive.In addition,an improved particle swarmoptimization(PSO)algorithm based on a learning mechanism and adaptive parameter adjustment mechanism was put forward,which introduced the mutation strategy of adaptive differential evolution algorithm based on linear population size reductionand extensive learning mechanism into the PSO algorithm.By generating mutated particles with adaptive adjustmentparameters and increasing particle diversity,local optimality was reduced,and search efficiency was improved in multi-UUVtarget search applications.The simulation program was developed,and the Monte Carlo method was employed to analyze themulti-UUV search efficiency.Simulation results show that compared with the PSO search algorithm based on traditionalprobability maps,the proposed multi-UUV collaborative search method takes less time and finds more targets under the sameconditions,so it has obvious advantages in dynamic target search.

关 键 词：多无人水下航行器 目标搜索 概率图 未知环境 

分 类 号：TJ630.1[兵器科学与技术—武器系统与运用工程] U666[交通运输工程—船舶及航道工程]