基于群稀疏整体最小二乘的多节点探测阵高精度被动定位方法

High-accuracy Multi-node Passive Localization Based on Group Sparse Total Least-squares

随着人类经济社会的发展,海洋环境噪声逐年增加,再加上各国针对舰船辐射噪声防震减噪技术的研发,孤立节点水下探测阵作用距离和探测精度随之逐年下降,且其仅能获取目标方位或距离信息,无法准确获知目标位置,已经越来越难以满足实际应用。因此,基于多节点的协同探测定位技术正在成为水下被动定位的热门研究方向。如何构建适配多节点的接收信号模型、设计高精度定位算法,是多节点协同探测亟待解决的难题。针对此,本文基于目标在空间中的稀疏特性,通过将目标所处区域网格化,构建基于群稀疏的多节点目标被动定位模型,并采用群稀疏优化技术实现直接定位。为解决网格失配引起的定位精度缺失问题,本文对联合阵列流形矩阵进行一阶泰勒展开,构建离格定位模型,采用群稀疏整体最小二乘算法,实现对目标位置的高精度估计。仿真实验表明,对于长宽均为1 km的探测区域,在4组节点情况下,所提模型和算法均是正确有效的,能对多目标位置进行正确解算。大网格间距下,相较于稀疏模型,离格模型能够进一步降低10 m左右的定位误差。不同网格间距下的对比试验表明,本文所提出的离格定位模型对于网格间距有较强的鲁棒性。相较于传统两步法及经典的直接定位法,单源及多源情况下,信噪比不低于-5 dB时,本文所提出的基于群稀疏整体最小二乘的离格直接定位算法具有更高的定位精度。

With the development of human economy and society,the noise of the ocean environment increases year by year,coupled with the development of noise reduction technology in various countries,the range and detection accuracy of a single node(underwater array)decreases year by year.On the other hand,a single node can usually only obtain target direction or distance information,and cannot accurately know the target position,which has become increasingly difficult to meet the practical applications.Therefore,the cooperative detection with multiple nodes is becoming a popular research topic for underwater passive positioning,where how to build the received signal model and design the high-precision local-ization algorithm for multi-nodes is difficult to be addressed.Based on the sparseness of the target in space,this paper es-tablishes a multi-node passive positioning model based on group sparsity,and adopts group sparse optimization technology to achieve direct positioning.To solve the problem of missing localization accuracy caused by grid mismatch,this paper constructed an off-grid localization model by first-order Taylor expansion,and propose the group sparse total least-squares algorithm to achieve high-precision estimation of the target position.Simulation experiments show that the proposed model and algorithm are effective and robust,and can correctly solve the multi-target position in the case of four groups of nodes.Compared with the on-grid model,the off-grid model can further reduce the localization error by about 10 m.Compared with the traditional two-step methods and the classical direct localization methods,the proposed group sparse total-least square algorithm has higher localization accuracy when the signal-to-noise ratio is not less than-5 dB for single-source and multi-source cases.