Analysis of Effects of System Mismatches on the Performance of Adaptive Generalized Sidelobe Canceller and Compensation Methods

The effect of system mismatches on an adaptive linear constrained generalized sidelobe canceller （LC-GSC） is discussed in this paper. Based on the array gain index, two classic system mismatches, the direction of arrival （DOA） mismatch and the mismatches arising from array disturbance, are studied, respectively. To obtain the effective methods for compensating for the system mismatches, we analyze the performance of the improved LC-GSC with the diagonal loading and additional constraints （such as the directional constraints and derivative constraints）. The computer simulations show that the techniques of diagonal loading and additional constraints can effectively compensate for the system mismatches. The loss of array gains can be controlled within 3 dB in the presence of 20% of array disturbances or DOA mismatch when the signal-to-noise ratio is less than 10 dB. The analysis illustrates that the proposed compensation methods are valid and feasible.

Robust generalized sidelobe canceller based on eigenanalysis and a MaxSINR beamformer

A robust generalized sidelobe canceller is proposed to combat direction of arrival(DOA)mismatches.To estimate the interference-plus-noise(IPN)statistics characteristics,conventional signal of interest(SOI)extraction methods usually collect a large number of segments where only the IPN signal is active.To avoid that collection procedure,we redesign the blocking matrix structure using an eigenanalysis method to reconstruct the IPN covariance matrix from the samples.Additionally,a modified eigenanalysis reconstruction method based on the rank-one matrix assumption is proposed to achieve a higher reconstruction accuracy.The blocking matrix is obtained by incorporating the effective reconstruction into the maximum signal-to-interferenceplus-noise ratio(MaxSINR)beamformer.It can minimize the influence of signal leakage and maximize the IPN power for further noise and interference suppression.Numerical results show that the two proposed methods achieve considerable improvements in terms of the output waveform SINR and correlation coefficients with the desired signal in the presence of a DOA mismatch and a limited number of snapshots.Compared to the first proposed method,the modified one can reduce the signal distortion even further.