Overcoming Channel Aging in Massive MIMO Basestations With Open RAN Fronthaul

Massive MIMO deployed with large channel bandwidths in mid-band spectrum is fundamental to high throughput and near ubiquitous coverage required by users in Fifth generation cellular networks. Network operators are also looking past existing monolithic radio access network architectures and looking towards functionally split architectures that allow for flexible, scalable, and cost-effective network deployments. Functional split of the baseband in massive MIMO deployments requires new and innovative approaches to distribute signal processing algorithms, and to optimize the information exchange across the fronthaul while maintaining acceptable levels of network performance. We consider a baseband split standardized by the Open RAN Foundation and investigate the issue of channel information being present at one side of the split, yet needed at the other side in order to perform uplink beamforming. We provide an analysis of air-interface performance degradation caused by the delay between determining beamforming weights from uplink sounding reference signals and applying these weights to the physical uplink shared channel. We also propose a novel beamforming weight design algorithm that can provide good tradeoff between air-interface performance and fronthaul throughput overhead. Our analysis and results are supported by both simulations and a massive MIMO prototyping testbed that emulates an outdoor environment with a high-speed user.