Channel estimation versus equalization in an OFDM WLAN system

We study an OFDM system, following the IEEE 802.11a standard, with QPSK modulation, allowing a 10 Mbps data rate when using convolutional coding. We compare one-tap equalization and channel estimation structures in the frequency domain by theoretical analysis and floating point simulations. We use the minimum mean square error (MMSE) and the zero-forcing (ZF) criteria. In the adaptive case, we use the RLS algorithm. We compare data aided (using pilot tones), and decision directed modes. The analysis is made for AWGN, one path Rayleigh fading, one path Ricean fading and multipath Ricean fading channels. The ZF criterion performs better than the MMSE criterion. The channel estimator is simpler and performs better than the equalizer in Ricean fading channels. This paper proposes new insights on the ZF and MMSE criteria, equalization and channel estimation.