A study of performance and complexity for IEEE 802.11n MIMO-OFDM GIS solutions

This paper investigates the point-to-point PHY layer performance of two newly proposed IEEE 802.11n candidate solutions by suggesting extensions to 802.11a. The new solutions incorporate a MIMO architecture for the wireless links between the network terminals. The FEC block in the PHY layer chain is enhanced with a turbo encoder to provide interleaving gains for the larger sized PSDUs (1-4095 bytes). Due to the ambitious data rate targets identified in 802.11n, where ´on-air´ data rates of 108-320 Mbit/s are quoted, a large MIMO architecture comprising of 6 transmit and 6 receive antenna elements is adopted. ML detection using such large configurations is often prohibitively complex due to an unrealistic enumerated symbol list. In this paper we propose ZF-GIS and MMSE-GIS solutions to ease complexity at the receiver. The PER performance offered by both proposals is compared together with the added computational requirements in constructing their matrices. Link throughputs are compared to demonstrate the viability of the chosen solutions.