Subspace-Based Semi-Blind Channel Estimation in Uplink OFDMA Systems

This paper presents a new approach for subspace-based semi-blind channel estimation in the uplink channel of orthogonal frequency division multiple access (OFDMA) systems. In the proposed approach, the orthogonality between the signal subspace and the noise subspace induced by the virtual carriers (VCs) is exploited to estimate the channel status of each user. The redundancy provided by the sub-carriers of the non-active users is then adopted to enhance the channel estimation performance. It is shown that while the proposed subspace-based channel estimation scheme is feasible for both sub-band and interleaved OFDMA systems, it is particularly effective for interleaved OFDMA systems due to their improved frequency diversity characteristic. In addition, a novel scheme, designated as the virtual carrier recovery (VCR) scheme, is proposed to further enhance the performance of the channel estimation process by recovering the VCs to zeros at the receiver end. The simulation results demonstrate that the VCR scheme yields a significant improvement in the channel estimation performance; particularly for the case of partially-loaded OFDMA systems. It is also shown that the proposed VCR scheme has a comparable bit error rate (BER) performance to that of the case when the receiver has perfect channel state information (CSI).