Blind OFDM Channel Estimation Using Receiver Diversity

In this paper we exploit receiver diversity for blind channel estimation for OFDM systems with cyclic prefix. The developed channel estimation algorithm has the following advantages compared with existing blind methods: a) It is data efficient — the proposed algorithm can be implemented using a single OFDM block; b) It is computationally efficient — only a single singular value decomposition is required for channel estimation; c) It is symbol independent — unlike many existing algorithms, there is no restriction on the input symbol constellation. Cramer-Rao lower bound for the channel estimation is derived to evaluate the performance. We show through numerical examples that the proposed algorithm compare favorably with an existing subspace algorithm.