BLIND BEAMFORMER FOR CONSTANT MODULUS SIGNALS BASED ON RELEVANCE VECTOR MACHINE

The blind beamforming method for constant modulus (CM) signals based on relevance vector machine (RVM) is proposed. The proposed beamforming method is obtained by incorporating the constant modulus algorithm (CMA)-like error function into the conventional RVM framework. The RVM framework formulates the parameters of beamfomer by exploiting a probabilistic Bayesian learning procedure with assumption of Gaussian prior for parameters. The simulation results show that the proposed blind beamforming method can restore the desired signals with crowded interference signals.