Channel Estimation in mmWave System with One-bit Compressed Sensing

Abstract

Considering the cost and large power consumption of high-resolution analog-to-digital converters in mmWave massive Multiple Input Multiple Output (MIMO) system, we develop a channel estimation algorithm with one-bit quantizer. Given the quantized received signal, the conditional expectation is first calculated to estimate the unquantized received signal. Based on the sparsity of mmWave channel in the angular domain, the column most correlative to the current residual is selected for least square estimation in each iteration. We also propose a solution to select multiple atoms in each iteration for the efficiency improvement. The simulation results show that the average normalized mean square error of the proposed algorithm and the running time of each iteration are 1/10 and 1/20 of the Expectation Maximization (EM) algorithm. It can be seen that the algorithm has improved performance and higher reliability, with certain practical value in the channel estimation application of mm-Wave large-scale MIMO systems.

Keywords

• mmwave
• one-bit quantizer
• channel estimation
• orthogonal matching pursuit