Abstract: Aiming at the problems of complex channel coupling, large channel scale, large pilot cost and difficult estimation of multi-hop cascade channel estimation in the millimeter wave communication system assisted by multiple intelligent reflecting surfaces(IRS), an efficient cascade channel estimation method is proposed based on compressed sensing theory. Firstly, based on the inherent sparsity of millimeter wave channel, the sparse representation of a large-scale multi hop cascade channel is found by using the properties of Katri-Rao and Kronecker product. Then, the large-scale cascade channel estimation problem is transformed into the recovery problem of multiple sparse signals. Finally, an improved OMP algorithm is used to recover the large-scale cascade channel. Simulation results show that the proposed method can greatly reduce the pilot cost of multi hop cascade channel estimation, and can accurately estimate the state information of multi hop cascade channel.

Key words: millimeter wave communication, intelligent reflecting surface, compressed sensing, sparse reconstruction, channel estimation