ISAR motion compensation based on matching pursuit with Chebyshev polynomials under low SNR

Targets of inverse synthetic aperture radar (ISAR) are non-cooperate and the motion compensation should be performed first. This paper proposes a method for ISAR motion compensation. In range-frequency domain, the motion of imaging target can be modeled as Chebyshev-polynomial phase function that is an orthogonal polynomial and has more approximation accuracy than Taylor polynomial. With pulse compression, the data are transformed into range frequency domain. The parameters in polynomial are estimated by matching pursuit (MP). Different from general treatments, simulated annealing algorithm is introduced into MP. And parameters are estimated in continuous field. The algorithm estimates the motion parameters before range compression and avoids the loss of pulse compression. Simulations testify that the proposed method work better under low SNR than methods based on entropy minimization or contrast maximization. Stepped-frequency waveforms with phase coding in subpulses are used in our experimental system. Wide-band signals are synthesized by narrow-band subpulses, which could obtain high-resolution range resolution. To acquire the similar cross-range resolution, more coherent integration interval is needed. Imaging experiments that testify our method also perform well in this case.