Real-Time Cooperative Adaptive Robust Relay Beamforming Based on Kalman Filtering Channel Estimation

In this paper, an adaptive channel estimation algorithm is proposed for the multi-user robust relay beamforming problem. We propose a norm-bounded channel uncertainty model for all of the channels. We employ the Extended Kalman Filter (EKF) and the Unscented Kalman Filter (UKF) for joint estimation of channel coefficients and beamforming weights, and propose a Markov model for source-relay and relay-destination channels as well as the beamforming weights in the relays. The channel coefficients and bemforming weights are shown to be well-estimated in order to minimize the total relays power transmission subject to worst-case signal to interference and noise ratio (SINR) criterion at each receiver. As the main contribution of this paper, we propose an adaptive method for simultaneous estimation of the beamforming weights and channel states information, and solving the associated optimization problem by estimation tools. Furthermore, we show that our algorithm outperforms the interior point based methods for non-linear optimization. In comparison to our recent work, a sub-optimal solution to the non-convex robust relay beamforming problem was provided, the proposed method has superior performance and lower complexity.