WiMax Capacity Enhancements introducing MIMO 2×2 Diversity and Spatial Multiplexing

Abstract

The IEEE802.16d/e standardization initiative for wireless metropolitan area networking (MAN) aims at portable and mobile broadband wireless access (BWA) in cellular deployments offering high data rate and high capacity scenarios for both mobile and fixed wireless access. The OFDM based WiMax air interface covers a state-of-the-art feature package including e.g subchanneling, turbo coding as well as multiple antenna systems (MIMO) promising huge capacity gain. This paper focuses on the mobile IEEE802.16e performance presenting capacity and spectrum efficiency figures for both the basic SISO and advanced 2x2 MIMO scenarios. Detailed link level simulation results have been presented for SISO, MIMO diversity and MIMO spatial multiplexing assuming a pedestrian Ped B 3 channel model characterizing slow moving subscriber scenarios. The evaluation of the channel capacity and spectrum efficiency is based on a semi-analytical approach for rapid performance estimation in cellular packet data networks. In this paper two generic traffic models are used to evaluate the spectrum efficiency for the upcoming new broadband OFDM IEEE 802.16 WiMax system. The first traffic model assumes an equal mean packet data call duration (EMPC-D) whereas the second one assumes an equal mean packet call data volume (EMPC-V) per subscriber. The study of mobile WiMax performance in 1x3 frequency reuse shows a considerable difference of roughly 40% in terms of spectrum efficiency between both traffic models. The comparison of mobile WiMax with SISO and MIMO reveals a considerable performance advantage for advanced MIMO antenna systems in the order of 120% for EMPC-D and up to 270% for EMPC-V. It will be shown that MIMO requires a dynamic switching between the diversity and spatial multiplexing modes to realize the full capacity gain. Finally the achievable benefits of spatial multiplexing massively depend on the applied traffic model and are rather negligible for EMPC-V but really significant for the EMPC-D traffic model.