A Micro-Doppler Analysis Method With Joint Parameter Estimation and Signal Separation Based on MIMO Array

In traditional human micro-Doppler signal applications, people usually move towards or away from the radar. There may be more situations in reality. This paper explores a new type of human micro-Doppler signal application and proposes a corresponding processing algorithm. In the new application, multiple input multiple output (MIMO) radar is used to replace traditional real aperture radar to reduce costs. At the same time, MIMO radar observes human motion on the side of the human body to obtain micro-Doppler curves. Compared to traditional applications, the new application can better serve security checks and human screening. Based on the wavenumber domain transformation, L-statistic, and the image contrast maximization, this paper proposes an algorithm for Joint Parameter Estimation and Signal Separation (JPESS) to analysis the micro-Doppler curves with translational phase under the side view. In this paper, the L-statistic algorithm is used to separate signals into rigid body parts and micro motion parts. By maximizing the image contrast, JPESS can utilize the separated signal of rigid body parts for optimal estimation of the speed and acceleration. After using estimated parameters to compensate for the micro-Doppler curve and searching for parameters by minimizing the total energy of the time-frequency map, the parameters of each micro motion part can be obtained sequentially. Simulations have verified that the algorithm can obtain human motion parameters and separate micro motion parts’ signals with high accuracy and strong robustness.