一种基于深度强化学习的测向基线布阵技术  

A Direction Finding Baseline Array Generation Technology Based on Deep Reinforcement Learning

作　　者：胡超 朱子翰 HU Chao;ZHU Zihan(Southwest China Institute of Electronic Technology,Chengdu 610036,China)

机构地区：[1]中国西南电子技术研究所,成都610036

出　　处：《电讯技术》2022年第10期1491-1497,共7页Telecommunication Engineering

摘　　要：针对现在工程项目中测向阵列基线的布阵方法存在耗时耗力、严重依赖于技术人员测向经验且复杂测向阵列难以设计等问题,提出了一种利用深度强化学习方法实现测向阵列基线自动生成技术。基于相关干涉仪测向机制,采用深度强化学习方法构建测向布阵智能体,重点突破多场景多实体仿真建模、布阵智能体构建、测向效能评估等关键技术;利用强化学习反复试错机理,迭代优化得到符合指标的最优测向阵列,大大提高布阵效率和测向质量,并通过实验证明了该方法的有效性。采用该技术设计的阵列基线已在实际项目中进行测向试验验证,各项指标均满足实际工程应用要求。In the current design methods of direction finding(DF)array baselines in engineering projects,there exist problems such as time-consuming and labor-intensive,heavy reliance on the DF experience of technicians,and difficulty in designing complex DF arrays.For these problems,this paper proposes a technology for automatic generation of DF array baselines using deep reinforcement learning(DRL)method.Based on the related interferometer DF mechanism,the DRL method is adopted to design the DF array agent,focusing on breakthroughs in key technologies such as multi-scenario and multi-entity simulation modeling,array agent construction,and DF effectiveness evaluation.By using reinforcement learning repeated trial and error mechanism,iterative optimization is performed to obtain the optimal DF array that meets the index,thus greatly improving the array efficiency and DF quality,and the effectiveness of the proposed method is proved through experiments.The array baseline designed by this technology has been verified by DF test in projects,and its specifications meet the requirements of practical engineering applications.

关 键 词：无线电测向 测向基线布阵 相关干涉仪 深度强化学习 

分 类 号：TN971[电子电信—信号与信息处理]