Range Gap between no-CSI and CSI Mode Transmission
Many-antenna MU-MIMO faces a critical, previously unaddressed challenge: it lacks a practical control channel. At the heart of this challenge is that the potential range of MU-MIMO beamforming systems scales with up to the square of the number of base-station antennas once they have channel state information (CSI), whereas the range of traditional control channel operations remains constant since they take place before or during CSI acquisition. This range gap between no-CSI and CSI modes presents a critical challenge to the efficiency and feasibility of many-antenna base stations, as their operational range is limited to the no-CSI mode.
Faros: Control Channel Design for Many-Antenna MU-MIMO
Faros leverages two key insights. (i) First, as much of the control channel as possible should be sent over the CSI mode. We find that the only control channel operations that must use the no-CSI mode are time-frequency synchronization, association, CSI collection, paging, and random access, which are required to establish the CSI mode. By implementing the remaining control channel operations over the CSI mode, we substantially increase their efficiency, as well as avoid the aforementioned gain gap. (ii) Our second key insight is that synchronization and association are not time-critical. That is, synchronization is valid for 100s of ms and association only happens once; thus by reducing the frequency of synchronization Faros is able to substantially reduce the channel overhead of these operations in the no-CSI mode, at the cost of slightly increased association latency at the cell edges.
Guided by these insights, Faros leverages open-loop beamforming and coding gains to ensure that many-antenna base stations can achieve their full potential range. Through open-loop beamforming, Faros is able to use the full diversity, power, and beamforming gains from all of the antennas on the base station, which enables it to scale with M, the number of base-station antennas. Because open-loop beamforming is never as performant as its MU-MIMO counterpart, closed-loop beamforming, Faros employs coding gains to further increase the range and to ensure that synchronization and paging are reliable even at the cell edges. To be as efficient as possible, Faros only performs these essential tasks and communication outside of the CSI mode, which offers much higher spectral capacity. Specifically, Faros uses open-loop beamforming to sweep extra-long synchronization sequences across the coverage area. This synchronization sequence not only enables users to establish time-frequency synchronization with the base station, but also encodes the base-station ID, and optionally user IDs for paging. Faros can dynamically configure important parameters, such as the beam patterns, sweep rate, and sequence length, to match the required gain for full coverage of the desired area. Furthermore, by increasing open-loop beamforming and coding gains in no-CSI mode while reducing the modulation rate or number of users served in CSI mode, Faros can be used to extend the range of the base station in remote areas.
Implementation of Faros
We implement Faros on Argos, a many-antenna MU-MIMO base station over a 2.4 GHz channel, with 108 antennas and evaluate the real-world performance and overhead of the implementation. Measurements show that our implementation provides over a 40 dB gain compared to traditional control channel operations. Anecdotally, this enables us to provide reliable synchronization to mobile users at over 250 meters with less than 100 uW of power per basestation antenna, or 10 mW of total power, using only standard low-gain 3 dBi omnidirectional antennas.
C. Shepard, A. Javed, and L. Zhong, “Control channel design for many-antenna base stations,” in Proceedings of ACM MobiCom 2015, Paris, France, September 2015.