Channel prediction using compressed sensing in multi-user MIMO systems

In downlink multi-user multiple-input multiple-output (MIMO) systems, a base station needs downlink channel state information (CSI) for each user to eliminate inter-user interference and inter-stream interference. In wireless communication, however, signal propagation environments change over time, and CSI obtained at the base station is different from the channel at the actual transmission time because we have delay. This deteriorates communication quality, and the effect of outdated CSI is a critical issue. To overcome this problem, some channel prediction schemes have been developed. Among them, a sum-of-sinusoids (SOS) method can predict time-varying channels over a long range. The SOS method, however, needs to resolve an incident signal into individual multipath components. In this paper, we propose a compressed sensing technique for the resolution, and formulate the channel prediction scheme for multi-user MIMO systems. Also, we evaluate the performance of the proposed scheme using computer simulations.