Interference rejection in non‐Gaussian noise for satellite communications using non‐linear beamforming

Abstract In this paper, non‐linear beamforming based on the method of radial basis functions (RBF) is proposed to combine with constant modulus algorithm (CMA) for interference cancellation in non‐Gaussian noise. First, a signal space is constructed from the input array data vectors utilizing RBF. A set of exponential functions is used to transform the Euclidean norm of the input array data vectors into the signal space. Then, CMA is employed to eliminate the interference. Both Gaussian and non‐Gaussian noise are considered in the experiment. The numerical results indicate that this scheme is effective and robust in terms of interference cancellation. A deep null can be formed to the direction of interference for both Gaussian and non‐Gaussian noise. Furthermore, these results are compared with that of DBF with CMA based on FFT technique, a widely used linear beamforming technique. The comparison shows that a DBF array antenna with CMA based on the method of RBF outperforms its counterpart based on FFT technique. The most interesting result is that the proposed method has a robust interference cancellation performance in non‐Gaussian noise whereas the FFT technique fails to do so. This observation can be very useful in satellite communications because that non‐Gaussian noise such as atmospheric and man‐made noise are often encountered in satellite communications. Copyright © 2003 John Wiley & Sons, Ltd.