Waveform Design Based on Modified Genetic Algorithm for MIMO Radar Imaging

MIMO radar imaging which integrates wideband signal processing with MIMO radar system is a potential technology for remote sensing. Waveform design is a principal important issue before radar imaging successfully in MIMO. There are two main requirements for the waveform used in MIMO radar imaging: the signals should be orthogonal and its auto-correlation function should have low side lobe noise and narrow main lobe. This paper analyses statistical characteristics of random stepped frequency (RSF) signal´s auto-correlation deeply and proves that large stepped frequency number is suitable for MIMO imaging radar. Then a method for designing random stepped frequency-coding waveforms by modified genetic algorithm is presented here. A random matrix is used for initializing the genetic algorithm, and annealing algorithm is applied here for avoiding GA´s disadvantages such as prematurity. In the end we compare the waveforms designed by GA and MGA and apply the waveforms in MIMO radar imaging simulation, the result proves that it is suitable for overcoming interfere among MIMO radar channels and imaging target.