Data-Driven Optimizations for Downlink MU-MIMO with Client Mobility in WLAN
Doctor of Philosophy
Multi-user MIMO (MU-MIMO) is a technique in 802.11ac and 802.11ax that improves spectral efficiency by allowing concurrent communications between one access point (AP) and multiple clients. In practice, the expected gain is not always achieved and is sometimes even negative. Using commodity 802.11ac devices, we experimentally demonstrate that the downlink MU-MIMO performance in a practical network not only depends on the client's channel but is also influenced by factors that are not captured by conventional models, such as client motion and device type. To optimize the MU-MIMO performance in 802.11ac networks the presence of client motion, we develop a data-driven algorithm that determines whether a client should operate in MU mode and the MU-MIMO group for clients in MU mode. Such an algorithm is based on a sequence of channel state information (CSI), SNR, and client device type. The algorithm can automatically adapt to the motion and characteristics of individual clients. Experimental results using implementation on a commodity 802.11ac AP show that the proposed data-driven mode and group selection algorithm can improve network throughput by up to 35% over existing algorithms based on conventional models. We also show that the proposed data-driven algorithm has limited sensitivity to environmental changes and can be deployed into new environments without retraining. While downgrading the moving clients to single-user mode is effective in 802.11ac networks with a limited number of transmit antennas, it may not achieve the optimal performance in 802.11ax networks where the potential gain from MU-MIMO is significant. Effectively enabling MU-MIMO with moving clients is essential for such networks to achieve optimal performance. In this thesis, we identify that the sounding period, number of spatial streams, and client grouping with the consideration of the client density of the network are essential in effectively enabling MU-MIMO in networks with moving clients. We propose an algorithm that determines the sounding period, number of spatial streams, and MU-MIMO group of each client. Using a commodity 802.11ax network, we experimentally demonstrate the significant impact of such factors on MU-MIMO performance. Based on experimental data, we develop an emulation model to evaluate network performance with different density and client mobility. Emulation results show that our proposed algorithm outperforms conventional dynamic sounding and mode selection schemes by over 20% in MU-MIMO networks with moving clients.