With high spectral efficiency, large scale antenna array, also referred as massive multiple input multiple output (MIMO), has been adopted as a key enabling technology for 5G-and-beyond wireless communications to meet the strict requirements on high transmission rate, massive connectivity and ultra-reliability in various mobile Internet and Internet of Things (IoT) applications. However, the inherent high-dimensional signal processing and complex channel structure in massive MIMO brought great challenges in performance analysis and transceiver design. This project solved the crucial and challenging problems in practical massive MIMO systems. It developed innovative approaches for capacity and reliability analyses, and quantitatively revealed the impacts of MIMO channel spatial correlations and various system parameters on communication performance. The analytical results have provided effective guidance for practical design. By exploiting the spatial correlation in the channels, sparse correlation representation of the high-dimensional channels was developed, and novel channel estimation and reconstruction methods were proposed with high accuracy but low overhead and complexity. Last but not least, the project further discovered the intrinsic commonality of various design objectives and a unique matrix-monotonic property in matrix-variable design problems, and developed a general optimization framework for the MIMO transceiver design. The findings and results have laid a solid foundation and provided effective solutions for 5G-and-beyond wireless communications.

Fig 1 Impact of spatial correlation on massive MIMO communications, and sparse correlation representation and channel reconstruction. (a) transmission capacity versus spatial correlation; (b) outage probability versus spatial correlation; (c) joint probability density function of angle of arrival (AoA) and angle of departure (AoD); (d) comparison of the massive MIMO channel reconstruction methods

Fig 2 Massive MIMO communication testbed