Frequency-domain oversampling based blind channel estimation for MIMO-OFDM

By using frequency-domain oversampling at the receiver, we propose a novel blind channel estimation scheme for multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. Specifically, a subspace blind channel estimation approach is developed, based only on second-order statistics of the received signal. The estimated MIMO channel is nicely determined by N_t minimum eigenvectors of a Hermitian matrix, where N_t is the number of transmit antennas. Numerical results show that the proposed scheme performs significantly better than an existing linear precoding based alternative when signal-to-noise ratio is high and/or when the length of observation window (used to obtain noise subspace) is relatively large. In addition, the proposed scheme does not impose stringent requirements on the number of transmit antennas and, therefore, is applicable to general MIMO scenarios including MIMO, multiple-input single-output (MISO), single-input multiple-output (SIMO), and single-input single-output (SISO).