Potential and pitfalls of Multi-Armed Bandits for decentralized Spatial Reuse in WLANs

Abstract

Spatial Reuse (SR) has recently gained attention to maximize the performance of IEEE 802.11 Wireless Local Area Networks (WLANs). Decentralized mechanisms are expected to be key in the development of SR solutions for next-generation WLANs, since many deployments are characterized by being uncoordinated by nature. However, the potential of decentralized mechanisms is limited by the significant lack of knowledge with respect to the overall wireless environment. To shed some light on this subject, we show the main considerations and possibilities of applying online learning to address the SR problem in uncoordinated WLANs. In particular, we provide a solution based on Multi-Armed Bandits (MABs) whereby independent WLANs dynamically adjust their frequency channel, transmit power and sensitivity threshold. To that purpose, we provide two different strategies, which refer to selfish and environment-aware learning. While the former stands for pure individual behavior, the second one considers the performance experienced by surrounding networks, thus taking into account the impact of individual actions on the environment. Through these two strategies we delve into practical issues of applying MABs in wireless networks, such as convergence guarantees or adversarial effects. Our simulation results illustrate the potential of the proposed solutions for enabling SR in future WLANs. We show that substantial improvements on network performance can be achieved regarding throughput and fairness.

Target Wake Time: Scheduled Access in IEEE 802.11ax WLANs

“…• Spatial reuse improvement: Spatial reuse is a key aspect in future high-density WLANs in order to improve the area throughput [12,13]. By reducing the uncertainty with respect to the number of stations in a given area that may transmit in a given period of time, techniques to improve the spatial reuse by adapting the transmit power, tuning the Clear Channel Assessment (CCA) threshold, and selecting the channel to operate can be further improved.…”

The TMB path loss model for 5 GHz indoor WiFi scenarios: On the empirical relationship between RSSI, MCS, and spatial streams

“…This operation was repeated 9 times, one per each possible BW-P TX combination from Table II. 11 RSSI values at the receiver laptop were used to compute the corresponding path loss (PL) according to…”

“…is proposed. In this case, W i, j value is replaced by the distance-dependent expression W (d i, j ) = W · d i, j , where W is the average number of traversed walls per meter in the N L analyzed locations [11]. Additionally, and for comparison purposes, the ITU-R indoor site-general model PL ITU (d i, j ) = 20 · log 10 ( f c ) + N · log 10 (d i, j ) + L f − 28, (7) is considered [12], where f c is the employed frequency, N is the distance power loss coefficient (in our particular case and according to the model guidelines, N = 31), and L f is the floor penetration loss factor (which was removed as experimentation was performed on a single floor).…”

Damysus: A Practical IEEE 802.11ax BSS Color Aware Rate Control Algorithm

“…where Interference is the recorded interference level based on the color mismatch. The parameter diff is confined between [1,6] to avoid high fluctuations on the rate selection. If Interference > OBSS∕PD max then diff is deducted by 1 to account for the high interference level.…”

“…In general, SR comprises those schemes that adapt carrier sensing or use a transmit power control (TPC) to eliminate the number of exposed nodes. However, tuning the Clear Channel Assessment (CCA) threshold or the transmit power level has implications on the network performance and may lead to poor performance due to the increased number of hidden or exposed nodes [6].…”