Robust channel estimator for MIMO-OFDM systems with FPGA implementation

Multiple-input multiple- output (MIMO) and orthogonal frequency division multiplexing (OFDM) are two key techniques for broadband wireless mobile communications. In order to provide accurate channel state information (CSI) to support MIMO-OFDM transmission in doubly-selective fading channels, we propose the robust channel estimator based on two-dimensional enhanced discrete-time Fourier transform interpolation (2D-EDFTI), which needs no channel statistics and can improve the channel estimation accuracy significantly. In order to validate its performance in practice, we implement it into our real-time FPGA testbed of 4-by-4 MIMO-OFDM. Also, channel estimator based on two-dimensional linear interpolation (2D-LI) is implemented for comparison. Experiments demonstrate that our proposed 2D-EDFTI channel estimator is robust to doubly-selective fading channels and its performance is always superior to 2D-LI channel estimator. Therefore, MIMO-OFDM transmission by spatial multiplexing with different kinds of MIMO detector, i.e., zero-forcing (ZF) and dynamic layer-ordering m-path detection (DOM), can be well supported.