Training-based Adaptive Transmit-Receive Beamforming for Random Phase Radar Signals

The advent of increasingly sophisticated control over the transmitted signal has enabled the consideration of multiple input, multiple output (MIMO) radar systems wherein each transmitter transmits a different waveform. Exploiting this capability, MIMO radars can improve target detection by jointly designing the transmit signal and receive filter so as to optimize the resulting signal-to-interference-plus-noise ratio (SINR). However, the SINR depends on the clutter covariance matrix which, in turn, is a function of the transmitted signal. This paper considers the joint design of adaptive transmit and receive weights to maximize the SINR of a target at a chosen look angle-Doppler point. This is akin to extending receive-only space-time adaptive processing (STAP) to include transmit adaptivity. Previous work in joint design assumed that the required second-order statistics are known a priori. In this paper we develop a method to estimate the required statistics through a number of training sequences. The estimation is based on received data only, and does not assume any specific structure for, or a-priori knowledge of, the clutter covariance matrix. We do assume that the clutter statistics do not change during the training and detection intervals. Simulation results show that, as in receive- only STAP, the proposed method does not suffer from a large SINR loss with respect to the known-covariance case.