Cognitive Phase-Only Sequence Design With Desired Correlation and Stopband Properties

This paper considers the phase-only sequence design problem for a cognitive radar in order to achieve both the desired autocorrelation and the stopband properties, which is useful to obtain good range compression as well as the spectral compatibility. We develop a new objective function accounting for the weighted integrated sidelobe level and the spectral power within specific stopbands. An iteration algorithm based on pattern search (PS) is proposed to minimize the nonconvex multidimensional objective function with the continuous phase case, which is split into multiple one-dimensional problems that are simplified and solved with closed-form solutions. In particular, the PS algorithm can be applied to the discrete phase case. Finally, we evaluate the effectiveness of the PS algorithm compared with the weighted-stopband cyclic algorithm new algorithm via numerical simulations in terms of the autocorrelation function, the spectral power function, and the convergence speed.