Abstract: For multiple-input multiple-output over-the-horizon radar (MIMO-OTHR), the minimum variance distortionless response (MVDR) is usually used to solve the impact of the ionospheric multi-mode clutter in the received echo on the detection of the target. The output signal-to-clutter-plus-noise ratio (SCNR) of traditional MVDR algorithm is low due to the omnidirectional emission of MIMO-OTHR, the inaccurate estimation of interference-noise covariance matrix, and a few training samples within a finite coherent accumulation time. Therefore, the MVDR algorithm is improved to enhance the multi-mode clutter suppression ability of MIMO-OTHR. A transmit beamspace preprocessing is realized by a second order cone programming (SOCP) to focus the energy on the principal mode angle, which is the transmit angle of the target path reflected by the ionosphere. The echo signal received by array is spatially smoothed to suppress the expected signal of principal mode and improve the estimation accuracy of interference-noise covariance matrix. A bi-iterative MVDR beamforming algorithm based on the beamspace preprocessing is proposed to reduce the computational load and the requirement of training samples. Simulated results show that the improved MVDR algorithm can be used to effectively suppress multi-mode clutter, and obtain higher output SCNR with fewer training samples and less computational load. Key

Key words: multiple-inputmultiple-outputover-the-horizonradar, multi-modeclutter, beamspacepreprocessing, arraysmoothing, bi-iterativeminimumvariancedistortionlessresponse