Channel estimation for MIMO-OFDM systems employing spatial multiplexing

Multiple-input multiple-output (MIMO) communication methods based on orthogonal frequency division multiplexing (OFDM) and spatial multiplexing (SM) can lead to a substantial improvement in the throughput performance of many existing wireless channels. The main strength of MIMO systems based on SM is their ability to support very high data rates. Accurate channel estimation is essential in realizing the full performance potential of MIMO-OFDM. Channel estimation also becomes a major challenge as the number of parallel wireless links increases substantially with the number of antennas at the transmitter and receiver. In this paper we develop a channel estimation algorithm well suited for MIMO-OFDM based on the same preamble and signal field structure as the IEEE 802.11a standard, the current single-input single-output wireless local area network (WLAN) standard. This constraint is motivated by the backward compatibility requirement of the upcoming IEEE 802.11n standard, the next generation WLAN standard that is currently being developed. In addition to the training symbol and the signal field symbol, we make use of the soft symbol information available in the data portion of the packet in updating the channel estimate in a sequential and decision directed fashion. It is shown that when this approach is tested against a 4/spl times/4 (4 transmit antennas and 4 receive antennas) SM configuration, the channel responses of all sixteen parallel wireless links can he estimated with sufficient accuracy, starting with only one OFDM training symbol.