一种基于多普勒频率的恒模信号直接定位方法

相比于传统的差分多普勒(DD)两步定位方法,以Amar和Weiss提出的基于多普勒频率的单步直接定位方法在低信噪比和小样本条件下具有更高的定位精度。在该类新型定位体制的基础上,提出了一种基于多普勒频率的恒模信号直接定位方法。首先,依据最大似然(ML)准则以及恒模信号的恒包络特征,建立相应的直接定位优化模型。然后,根据目标函数的代数特征将全部未知参量分成两组,并提出一种有效的多参量交替迭代算法,用以获得该优化问题的最优数值解。新算法包含了针对这两组未知参量的Newton型迭代公式,用以避免网格搜索,并能实现多维参数的"解耦合"估计。最后,推导出针对恒模信号的目标位置直接估计方差的克拉美罗界(CRB)。数值实验验证了新方法的优越性。

Direct localization method for constant modulus source based on Doppler frequency shifts

Compared with the conventional Differential Doppler (DD) localization method,the Direct Position Determination (DPD) method proposed by Amar and Weiss has higher position estimation accuracy under the condition of low Signal-to-Noise Ratio (SNR) and small number of samples.Based on this novel localization mechanism,a new DPD method using Doppler frequency shifts is presented for the constant modulus source.The DPD optimization model is constructed based on the Maximum Likelihood (ML) criterion as well as the constant modulus property of the source.All the unknowns are then classified into two groups according to the algebraic characteristic of the cost function,and an effective alternating iteration algorithm is presented to solve this DPD optimization problem numerically.In the proposed algorithm,two Newton-type iterative steps are devised for the two groups of unknowns,and then the grid search can be avoided and the multidimensional parameters are decoupled.The Cramér-Rao Bound (CRB) on the direct position estimation variance for constant modulus source is derived.Simulation results corroborate the good performance of the proposed method.