An Improved LAMP Algorithm for Beam Space Channel Estimation in MIMO-OFDM mmWave Systems

This paper explores downlink communication in a multiple-input multiple-output Orthogonal Time Frequency Space (MIMO-OTFS) system. By leveraging the prior information of the beam space channel (BC), we derive a novel Laplace shrinkage function. We propose a BC estimation scheme, Laplace Distribution Enhanced LAMP (LDE-LAMP), by integrating the Learned Approximate Message Passing (LAMP) network with this new shrinkage function. Simulation results demonstrate that the proposed LDE-LAMP network algorithm achieves superior channel estimation accuracy compared to traditional LAMP networks.