Interference Rejection for Parametric Channel Estimation in Reuse-1 Cellular OFDM Systems

This paper proposes a joint channel-estimation method for the desired and interfering channels for cell edge users in reuse-1 orthogonal frequency-division multiplexing (OFDM) cellular systems. The following assumptions are made in proposing the algorithm: 1) the desired and interferer's channel multipath delays do not overlap; 2) the same pilot sequence is sent from the desired and interfering base stations (BSs); and 3) reuse-3 preamble symbols, as in the preamble structure in IEEE 802.16d/e (Wimax) systems, are used to obtain initial channel estimates without interference. If we make these assumptions, then it is possible to estimate the desired and interfering channels, even with reuse-1 pilots. The proposed pilot-based channel estimation technique exploits the delay subspace structure to reduce the impact of cochannel interference (CCI) on channel estimation. Delay subspace refers to the set of basis vectors spanning the frequency response of the desired and interfering multipath channels. This enables us to jointly estimate and track the desired and interfering channels with a lower mean-squared error (MSE), when compared with the conventional modified least-squares (mLS) technique, which ignores the structure of interference. The significance of accurate channel estimates in symbol detection schemes is demonstrated for systems employing two or more receiver antennas. The proposed channel estimator significantly improves the performance of symbol-detection schemes based on either interference nulling combiner (INC) or minimum MSE diversity combiner (MMSE-DC), when compared with detection schemes using mLS-based channel estimates. Analytical expressions are derived for the MSE of the estimated multipath delays. Simulation results are also provided to show the improved performance for the proposed channel-estimation method compared with the mLS-based channel estimation method, when used in conjunction with INC and MMSE-DC detectors.