Mixing matrix estimation of MIMO radar based on adaptive hierarchical clustering algorithm for underdetermined blind source separation

In order to solve the problem of mixing matrix estimation in the blind source separation of discrete frequency coding MIMO radar signals, through exploiting the linear clustering characteristic of the observed signal in time-frequency domain, a method of using time-frequency independent complex argument points detection and adaptive hierarchical clustering algorithm estimation is proposed. First, we exploit the sparsity of the MIMO radar signals in time-frequency domain. Therefore, the time-frequency points which have linear clustering characteristic can be extracted from the observed signals by the time-frequency independent complex argument points detection algorithm. Second, the adaptive hierarchical clustering algorithm is applied to the precise estimation of the mixing matrix. Compared with the traditional hierarchical clustering algorithm, the proposed method improves the filtering effect, and simultaneously removes the points which time-frequency independent complex argument points detection cannot detect due to noise and filter threshold. The algorithm can effectively solve the mixing matrix estimation problem of MIMO radar signals for underdetermined blind source separation. The simulation results show the feasibility and effectiveness of the proposed method.