Optimal pilot design with CSI feedback in massive MIMO systems

In this paper, we study the optimal pilot design problem and CSI feedback method in a massive MIMO system. We get the asymptotically optimal solution for pilot design based on minimizing the Mean Square Error (MSE) of channel estimation with both the Sum Power Constraint (SPC) and the Per Antenna Power constraint (PAPC) utilizing a t-norm approximation method. To further improve the transmission efficiency and reduce the computation load, we propose a CSI estimation feedback method with pilot weight to gradually discover the CSI. The CSI fed back to the BS is collected to design beamforming matric for different users aiming at improving users' achievable rate. Simulation results validate that the method can provide an acceptable achievable rate consuming much less transmission and computation resources.