认知雷达网联合传感器选择和功率分配  

Joint Sensor Selection and Power Allocation in Cognitive Radar Systems

作　　者：俞晓波 左燕[1] 郭宝峰[1] 谷雨[1] YU Xiao-bo;ZUO Yan;GUO Bao-feng;GU Yu(Institute of Information and Control,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]杭州电子科技大学信息与控制研究所,杭州310018

出　　处：《火力与指挥控制》2018年第10期57-62,67,共7页Fire Control & Command Control

基　　金：国家自然科学基金(61673146);浙江省自然科学基金资助项目(LY16F030009)

摘　　要：合理分配雷达有限的传感器资源,充分发挥传感器的认知能力可提高跟踪性能。提出了一种基于后验克-拉美罗下界(PCRLB)的多雷达系统联合传感器选择和功率分配方法。推导多雷达系统单目标跟踪下的PCRLB,该指标不依赖于具体的滤波算法,且多目标跟踪下的PCRLB的计算相互独立。以PCRLB的迹为代价函数,构建基于PCRLB的联合传感器选择和功率分配模型。对于构建的非凸优化问题,设计了基于凸松弛的循环最小化算法,优化求解最佳的传感器组合方式和功率分配方案。最后,仿真结果显示,所提的算法和策略具有较好的跟踪性能。A reasonable resource allocation strategy can make full use of limited capacity of radar resources in the multi-radar system to improve target tracking performance.In this paper,a Posterior Cramer Rao Lower Bound(PCRLB)based joint sensor selection and power allocation method in a multiple radar system is proposed.Firstly,the PCRLB of global state posterior estimation is derived in the multi-radar target tracking problem.The PCRLB is independent of the filtering algorithm employed and the PCRLB for each target could be calculated independently of that of other targets.Then,a PCRLB-based joint sensor selection and power allocation model is given.In this model,the cost function is the trace of PCRLB of multi-targets.A cycle minimization algorithm based on convex relaxation is then developed to solve the non-convex optimization problem to obtain optimal sensor combination and power allocation scheme.Finally,the simulation results show that the proposed algorithm can obtain good track performance.

关 键 词：多雷达系统 后验克拉美罗下界 传感器选择 功率分配 

分 类 号：TN953[电子电信—信号与信息处理]