System-Level Performance for MMSE Multiuser Detection in Asynchronous Mixed Service UMTS Networks

The main objective of this report is to investigate linear multiuser detection (MUD) in asynchronous mixed service UMTS networks in terms of system capacity as well as power reduction at mobile stations. The service mix composes voice users and high data rate users being expected for upcoming UMTS networks. Note that system capacity and required power consumption are both strongly related to load control algorithms. Thus the impact of load control on these system performance indicators cannot be neglected and has to be taken into account for system level performance assessments. Load control decides to admit new or to drop already establish users and it is based in UMTS on noise rise measurements. In order to carry out meaningful results we apply an iterative two-stage power control multiuser detection algorithm being suitable for UMTS application. The users are detected by minimum mean square error (MMSE) block detection to take into account asynchronous received user signals. Moreover, a proper and intensively studied inter-cell-interference model is applied in order to ensure correct modelling the interaction between UMTS cells. By detailed system level simulations we evaluate the UMTS load control parameter noise rise and show system capacity gains as well as power reductions compared to simple single-user matched filtering (SUMF). It is important to note that for UMTS the signal-to-noise operating points of the MMSE detector are very low (less than 0 dB). For these UMTS constraints, MMSE multiuser detection system capacity gains are fairly modest depending on the mix of data rates. With respect to mobile station power, MMSE multiuser detection indicates fair to high improvements. For instance, for medium loads we observe a power reduction in the order of 2-3 dB leading to coverage gains in the range of 11% to 14%. The authors like to point out that within this report it is assumed that the SINR requirement is the same for all detectors applied. The impact of the interference statistics on the channel decoding performance has been completely neglected. Investigations on the interference statistic may lead to diverse SINR requirements. Moreover, the investigations are exclusively carried out for UMTS dedicated channels. Investigations on SINR requirements as well as on recent enhanced uplink channels for packet data transmission are interesting tasks for future work.