Optimal mismatched filter bank design for MIMO radar via convex optimization

In this paper, a mismatched filter bank is designed for suppressing the autocorrelation peak sidelobe level (PSL) and the peak cross-correlation level (PCCL) of an orthogonal polyphase sequence set applied in a multiple-input multiple-output (MIMO) radar system. The mismatched filter bank is obtained by minimizing a weighted maximum of the PSL and PCCL on the basis of the convex optimization. Compared with the iteratively reweighted least squares (IRLS) method, the proposed convex method can get the optimal mismatched filter bank with the minimum PSL and PCCL, and can also control the system signal-to-noise ratio loss (SNRL). Numerical examples show that the optimal mismatched filter bank at the cost of a slight SNRL can achieve a good improvement of the PSL and a moderate improvement of the PCCL, if the filter length P and the weighting factor w between the PSL and PCCL are appropriately chosen.