Space-Time Channel Shortening Based Spatial Multiplexing Techniques using Uplink-Downlink Duality

We consider the problem of performing spatial multiplexing in a multiuser system with multipath channels. The proposed solution is based on space-time channel shortening, where we extend the uplink-downlink duality known for flat fading channels. We establish that the uplink-downlink duality result for flat fading channels holds for frequency selective channels i.e. both uplink and downlink share the same SINR and normalized MSE region under a sum-power constraint. We then use this result to propose spatial multiplexing schemes based on equalization and channel shortening at the transmitter under a max-min fairness framework i.e. all users achieve the same SINR/MSE targets. The advantage of channel shortening based spatial multiplexing as compared to a complete equalization at the transmitter is demonstrated using simulation results.