Recursive Power Allocation with Interference Minimization for Time Varying Massive MIMO

Among the emerging breakthroughs for future wireless communications, massive multiple-input multiple-output (MIMO) has attracted a lot of interest. To harvest its potential, however, accurate channel state information (CSI) is required at the base station. Unfortunately, the accurate CSI is hard to be obtained when Doppler spread occurs. The mismatch between the precoder and the actual channel causes the signal to be phase-shifted, and the interuser interference arises. To overcome this problem, we propose a power allocation technique based on the predicted interference level of all users. First, we extract the Doppler shift information from the pilots sent by the mobile station (MS) to create predicted channel matrix. Based on this matrix, we conduct a recursive power allocation to minimize the interuser interference at the MSs. Computer simulation shows that our proposed technique enables the time varying massive MIMO transmission even when the precoder matrix is not frequently updated.