Improving Throughput and Efficiency for WLAN: Sounding, Grouping, Scheduling


TR Number



Journal Title

Journal ISSN

Volume Title


Virginia Tech


Wireless local area networks (WLANs) have experienced tremendous growth with the proliferation of IEEE 802.11 devices in past two decades. Wireless operators are embracing WLAN for cellular offloading in every smartphone nowadays [1]. The traffic over WLAN requires significant improvement of both WLAN throughput and efficiency.

To increase throughput, multiple-input and multiple-output (MU-MIMO) has been adopted in the new generation of WLAN, such as IEEE 802.11ac. MU-MIMO systems exploit the spatial separation of users to increase the network throughput. In an MU-MIMO system, efficient channel sounding is essential for achieving optimal throughput. We propose a dynamic sounding approach for MU-MIMO systems in WLANs. We analyse and show that the optimal sounding intervals exist for single user transmit beamforming (SU-TxBF) and MU-MIMO for given channel conditions. We design a low-complexity dynamic sounding approach that adjusts the sounding interval adaptively in real-time. Through our collected over-the-air channel measurements, we demonstrate significant throughput improvements using our proposed dynamic sounding algorithm while being compliant with IEEE 802.11ac standard.

Subsequently, we investigate the user grouping problem of downlink WLANs with MU-MIMO. Particularly, we focus on the problem of whether SU-TxBF or MU-MIMO should be utilized, and how many and which users should be in a multi-user (MU) group. We formulate this problem for maximizing the system throughput subject to the multi-user air time fairness (MU-ATF) criterion. We show that hypergraphs provide a suitable mathematical model and effective tool for finding the optimal or close to optimal solution. We show that the optimal grouping problem can be solved efficiently for the case where only SU-TxBF and 2-user MU groups are allowed in the system. For the general case, where any number of users can be assigned to groups of different sizes, we develop an efficient graph matching algorithm (GMA) based on graph theory principles with near-optimal performance. We evaluate the proposed algorithm in terms of system throughput using an 802.11ac emulator using collected channel measurements from an indoor environment and simulated channel samples for outdoor scenarios. We show that the approximate solution, GMA, achieves at least 93% of the optimal system throughput in all considered test cases.

A complementary technique for MU-MIMO is orthogonal frequency-division multiple access (OFDMA), which will be the key enabler to maximize spectrum utilization in the next generation of WLAN, IEEE 802.11ax. An unsolved problem for 802.11ax is maximizing the number of satisfied users in the OFDMA system while accommodating the different Quality of Service (QoS) levels. We evaluate the possibility of regulating QoS through prioritizing the users in OFDMA-based WLAN. We define a User Priority Scheduling (UPS) criterion that strictly guarantees service requests of the spectrum and time resources for the users with higher priorities before guaranteeing resources to those of lower priority. We develop an optimization framework to maximize the overall number of satisfied users under this strict priority constraint. A mathematical expression for user satisfaction under prioritization constraints (scheduler) is formulated first and then linearized as a mixed integer linear program that can be efficiently solved using known optimization routine. We also propose a low-complexity scheduler having comparable performance to the optimal solution in most scenarios. Simulation results show that the proposed resource allocation strategy guarantees efficient resource allocation with the user priority constraints in a dense wireless environment. In particular, we show by system simulation that the proposed low-complexity scheduler is an efficient solution in terms of (1) total throughput and network satisfaction rate (less than 10% from the upper bound), and (2) algorithm complexity (within the same magnitude order of conventional scheduling strategy.



Wireless local area networks, Graph theory, Optimization