Extension of Linear and Nonlinear Transmit Filters for Decentralized Receivers

Abstract:
For the downlink of a multi-user system with decentralized receivers, we extend both linear and nonlinear transmit filters. As a special case of the joint optimization of transmitter and receiver, transmit processing emerges from restricting the receivers to be scalar. Such a setup then corresponds to the downlink of a multi-user system with decentralized receivers that cannot cooperate and cannot combine their received signals in order to obtain the symbol estimates of the data streams associated to them. Instead of designing the transmitter based on the common assumption that all receivers use the same scalar weight, i.e. the factor of the automatic gain control (AGC), we allow different weights and end up with modified transmitter structures. Thus, performance improvements can be obtained especially for scenarios, where the receivers have different average channel powers due to path loss or shadowing. Consequently, smaller bit error rates can be allocated to users with strong channels compared to the conventional case, whereas users with weak channel powers show almost no deterioration.