A Centralized MIMO Radar Signals Under-determined Separation Algorithm

The source signal separation problem of the centralized multi-input multi-output (MIMO) radar is studied in this paper and an under-determined blind source separation algorithm with more source signals than mixed signals is proposed. Firstly, this paper introduces the reconnaissance model of centralized MIMO radar. Secondly, the linear clustering variables are constructed by using the short-time Fourier transform results of different mixed signals. Then, the number of source signals and clustering variables are obtained by the clustering algorithm. Finally, all the source signals are separated by constructing an appropriate model and using the complex independent component analysis algorithm. Both formula derivation and simulation analysis prove that the proposed algorithm can estimate source signals of the centralized MIMO radar well under the condition of under-determination.