Time-Variant Broadband mmWave Channel Estimation Based on Compressed Sensing

Millimeter-wave communication is a key technology to provide multi-Gbps connectivity for next generation of wireless network. Millimeter-wave leverages the degree of freedom provided by Massive MIMO systems to achieve 100x to 1000x faster data communication. Hybrid digital-analog pre-coder/combiner architecture has been recommended to overcome propagation and hardware challenges, which causes trade-offs between power dissipation and performance. As a result, mm Wave channels exhibit sparse structure with a low degree of freedom. Compressed sensing solutions have been recommended for accurate mmWave channel estimation with respect to sparse nature of mmWave system. Current works in mmWave channel estimation mainly focus on a frequency-selective broadband systems with a fixed AoD/AoA and no Doppler effect consideration. In this paper, we are considering a mmWave channel that spreads in the both delay and angular domains and has a Doppler shift in the Doppler domain. Considering the mobility of the mobile station and angular spread of AoD/AoA, we recommend a time-variant broadband compressed sensing channel estimation solution for a mmWave communication. We are using sparse common support for multicarrier systems like OFDM. We propose Enhanced Orthogonal Pursuit Matching - Partial Sparse Common Support (EOMP-PSCS) to enhance the concept of sparse common support further to support cluster birth/death in slow varying channels. Numerical results show that the suggested method outperforms existing solutions both in the accuracy of channel estimation and spectral efficiency in time-variant, low SNR scenarios while maintaining low complexity for sparse recovery.