DOA Estimation in MIMO Radar With Non-Perfectly Orthogonal Waveforms

In direction of arrival (DOA) estimation of multiple input multiple output (MIMO) radar, many existing publications deal with perfectly orthogonal waveform signals (POWS). But in practice, it is a real challenge to generate POWS. In this letter, we study DOA estimation in MIMO radar with the case of non-perfectly orthogonal waveform signals (Non-POWS). A two-step algorithm is proposed. The first step is to blend the covariance matrix of Non-POWS into the steering matrix to set up a new signal model similar to the POWS model. The second step is to perform pre-whitening processing for the colored noise produced by Non-POWS utilizing known transmitted signal covariance matrix. Simulation results demonstrate that: 1) the Non-POWS degrades the DOA estimation performance and 2) our method can solve this problem well.