Near-optimal pilot allocation in sparse channel estimation for massive MIMO OFDM systems

Inspired by the success in sparse signal recovery, compressive sensing has already been applied for the pilot-based channel estimation in massive multiple input multiple output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. However, little attention has been paid to the pilot design in the massive MIMO system. To obtain the near-optimal pilot placement, two efficient schemes based on the block coherence (BC) of the measurement matrix are introduced. The first scheme searches the pilot pattern with the minimum BC value through the simultaneous perturbation stochastic approximation (SPSA) method. The second scheme combines the BC with probability model and then utilizes the cross-entropy optimization (CEO) method to solve the pilot allocation problem. Simulation results show that both of the methods outperform the equispaced search method, exhausted search method and random search method in terms of mean square error (MSE) of the channel estimate. Moreover, it is demonstrated that SPSA converges much faster than the other methods thus are more efficient, while CEO could provide more accurate channel estimation performance.