On Receive Filter Design for Coherent MIMO Radar with Non-Ideal Waveforms

In this paper, we consider the problem of coherent multiple-input multiple-output radar receive filter design. If the transmitted radar waveforms do not have the desired ideal auto- and cross-correlation properties, the signal at the output of a matched filter (MF) can suffer severe interference due to targets in neighboring range-Doppler bins. Existing receive filter design approaches, e.g., instrumental variable (IV) filter designs, employ so-called peak sidelobe level (PSL), integrated sidelobe level (ISL), or zero sidelobe (ZS) constraints in order to suppress potential range-Doppler interference. In contrast to receive filters that are designed by considering sidelobe suppression only, the receive filters derived in this work aim at an optimization of the tradeoff between range-Doppler interference suppression and noise amplification due to the receive filter. To this end, two design criteria are presented and closed-form expressions for the respective receive filters are derived which allow a computation at low computational cost. In addition, the performance of the proposed receive filter designs is evaluated using numerical simulations. Our simulation results demonstrate that the proposed receive filters are superior to IV filter designs in scenarios with high noise power or strongly shaped noise covariance matrices, whereas they perform similar in scenarios with low noise power.