Design of Sequences with Low Autocorrelation Sidelobes using Genetic Algorithms

Unimodular constant modulus sequences with desirable autocorrelation function properties are widely used in radar and communication systems. One of the performance metrics that determine the goodness of a designed sequence is the integrated sidelobe level (ISL) of the autocorrelation function of the sequence. Studies in the literature have focused on minimizing the metric of ISL. In this study, we utilize genetic algorithm (GA) to minimize ISL via the frequency domain representation of the designed sequence. The proposed algorithm is initialized by either a random sequence or the Golomb sequence whose autocorrelation is known to have good properties. By this way, a radar transmit signal with minimum ISL was designed using GA. Finally, the performance results of the GA are compared against the already existing cyclic-new (CAN) algorithm.