ECE Department Seminar
Efficient Sparse Signal Processing for Channel Sensing in
Millimeter-wave Massive MIMO
Department of Electrical and Computer Engineering
George Mason University
Thursday, March 18, 2021
10:30 am – 11:30 am
Zoom Meeting Link:
Abstract: Millimeter-wave (mmWave) massive MIMO has been acknowledged as a promising technique for 5G and future wireless systems. To realize the full promise of mmWave massive MIMO,
transceivers need to acquire accurate channel knowledge. However, the tremendous increase of antennas raises big challenges to traditional channel estimation techniques, causing large training overhead and long sensing time. Meanwhile, mmWave channels usually
experience limited scattering propagation, where the received energy arrives only in a small number of directions. This directional mmWave channel characteristic in angular domain suggests an array processing framework, but has to cope with low sample efficiency
in the presence of large antenna arrays of massive MIMO. Advances in virtual channel modeling and sparse channel sensing based on compressed sensing (CS) reap high sample efficiency, but suffer from limited angular resolution and degraded accuracy due to the
on-grid assumption of standard CS.
Aiming at accurate and efficient sensing techniques for large-size, multi-dimensional, and directional channels of mmWave massive MIMO, this talk will present recent progress on super-resolution channel parameter sensing based
on gridless CS techniques via atomic norm minimization (ANM), which utilizes not only channel sparsity but also desired structural information in array manifolds and channel statistics. This talk will review our recent research contributions in developing
low-complexity, super-resolution channel sensing techniques for various practical settings, extending the benefits of ANM from the 1D SIMO case to the 2D MIMO case, from idealized uniform antenna arrays to imperfect array geometry and hybrid hardware architecture,
and from semi-definite programming to fast first-order algorithms for low-complexity implementation.
Bio: Yue Wang is currently a Postdoctoral Researcher with the Department of Electrical and Computer Engineering at the George Mason University. Prior to that, he was a Senior Engineer
with Huawei Technologies Co Ltd. He received his Ph.D. degree in Electrical Engineering, from the Beijing University of Posts and Telecommunications, China. His general interests lie in the areas of signal processing, wireless communications, artificial intelligence,
and their applications in cyber physical systems. His current research focuses on compressed sensing, massive MIMO, mmWave communications, cognitive radios, DOA estimation, high-dimensional data analysis, and distributed optimization and learning.