Robust Adaptive Beamforming Based on Fuzzy Cerebellar Model Articulation Controller Neural Network

摘要

To solve the problem of degraded adaptive beamforming performance of smart antenna caused by array steering vector mismatch and array manifold errors, a robust beamforming algorithm based on Fuzzy Cerebellar Model Articulation Controller (FCMAC) neural network is proposed. The proposed algorithm is based on explicit modeling of uncertainties in the desired signal array response and a FCMAC neural network. The calculation of the optimal weight vector is viewed as a mapping problem, which can be solved using FCMAC neural network trained with input/output pairs. Our proposed approach provides excellent robustness against some types of mismatches and keeps the mean output array SINR consistently close to the optimal value. Moreover, the FCMAC neural network avoids complex matrix inversion operations and offers fast convergence rate. Simulation results show that the proposed algorithm can significantly enhance the robustness of the beamformer in the presence of array steering vector mismatch and array manifold errors, and the output performance is superior to the current methods.