Quantized compressive channel estimation for OFDM systems

Compressive sensing (CS) techniques have been shown to enhance the performance of sparse channel estimation. This paper addresses CS-based sparse channel estimation in practical systems where the measurements are quantized to a finite number of bits. In these systems, quantization introduces two kinds of errors into the CS framework: quantization errors and saturation errors. To combat the effect of quantization error on estimating the sparse channel, a new complex-valued cost function is formulated based on the quantization model. To minimize this cost function, a multi-bit iterative hard thresholding algorithm is proposed. The channel estimate performance is evaluated by modeling the quantization error as white Gaussian noise. Our simulation results quantify the number of bits needed for our proposed quantized CS-based sparse channel estimate to achieve satisfactory performance.