一种基于最小化广义Rayleigh商的无源定位算法研究

 为改善只测角无源定位的性能,提出了一种基于最小化广义Rayleigh商的无源定位算法.该算法利用扰动观测矩阵和扰动观测向量的乘性结构,将约束总体最小二乘问题转化为最小化广义Rayleigh商问题,从而只需对一对矩阵束进行广义特征值分解即可求得全局最优定位解.仿真结果表明所提算法性能稳健且计算量较小,定位收敛精度逼近克拉美罗限(CRLB),远优于最小二乘(LS)算法和总体最小二乘(TLS)算法,实用性强.

Research on a Passive Localization Algorithm Based on Minimizing the Generalized Rayleigh Quotient

In order to improve the performance of bearings-only passive localization,a new passive localization algorithm based on minimizing the generalized Rayleigh quotient was proposed,which uses the multiplicative structure of perturbation part of the observation matrix and observation vector.The constrained total least squares(CTLS) problem was converted to the generalized Rayleigh quotient minimization problem,which only needs to do the generalized eigenvalue decomposition for a pair of matrixes to obtain the globally optimal localization solution.Simulation results indicate the proposed algorithm possesses robust performance and low computation load.Its localization solution can achieve the CRLB performance and is far better than those of the least squares (LS) algorithm and total least squares (TLS) algorithm,which implies good practicability.