Evaluation of a class of NLFM radar signals

Abstract

AbstractSignal design is an important component for good performance of radar systems. Here, the problem of determining a good radar signal with the objective of minimizing autocorrelation sidelobes is addressed, and the first comprehensive comparison of a range of signals proposed in the literature is conducted. The search is restricted to a set of nonlinear, frequency-modulated signals whose frequency function is monotonically nondecreasing and antisymmetric about the temporal midpoint. This set includes many signals designed for smaller sidelobes including our proposed odd polynomial frequency signal (OPFS) model and antisymmetric time exponentiated frequency modulated (ATEFM) signal model. The signal design is optimized based on autocorrelation sidelobe levels with constraints on the autocorrelation mainlobe width and leakage of energy outside the allowed bandwidth, and we compare our optimized design with the best signal found from parameterized signal model classes in the literature. The quality of the overall best such signal is assessed through comparison to performance of a large number of randomly generated signals from within the search space. From this analysis, it is found that the OPFS model proposed in this paper outperforms all other contenders for most combinations of the objective functions and is expected to be better than nearly all signals within the entire search set.