This technology is a semi-blind algorithm for detecting signals without channel-state information in multi-user hybrid massive multiple-input multiple-output (MIMO) systems that can be used for 5G broadband networks.
Massive MIMO systems vastly improve throughput and spectrum efficiency of signals and are crucial in the development of future broadband networks such as 5G. However, typical massive MIMO systems require large numbers of high-resolution analog-to-digital converters (ADCs), which can be prohibitive in terms of cost and energy consumption. This constraint could be overcome by adopting a hybrid architecture or using low-resolution ADCs, but both methods require large training overhead in their signal channel estimation algorithms.
This technology proposes a semi-blind detection method to detect signals without channel-state information via a low-rank matrix completion formulation in multi-user hybrid massive MIMO systems. The low-rank property is obtained from the assumption that the number of users is lower than the number of antennas and length of coherence time in massive MIMO. The matrix completion problem is solved using two iterative algorithms: regularized alternating least squares and bilinear generalized approximate message passing. This detection method can be extended to work for hybrid massive MIMO systems that use low-resolution ADCs for further energy savings, thereby achieving significant performance gains with reduced training overhead.
This technology has been validated with simulation results that showed performance improvement over pilot-only methods.
IR CU20107
Licensing Contact: Greg Maskel