Collision-Based Up-Link OFDMA Random Access Mechanism for Wi-Fi 6

The IEEE 802.11ax standard defines High-Efficiency WLAN (HEW), which is known as 6th Generation Wi-Fi (Wi-Fi 6) network. HEW uses an up-link OFDMA-based Random Access (UORA) mechanism for channel access in dense multi-user transmissions. Similar to legacy Wi-Fi networks, UORA recommends adopting a random OFDMA Back-Off (OBO) procedure within the OFDMA contention window (OCW) range to access the available channel resources randomly. IEEE 802.11ax follows a centralized channel resource allocation mechanism, in which an access point (AP) is responsible for calculating and configuring the OCW range and notifying the respective Random Access stations (RA-STAs). However, IEEE 802.11ax does not specify a method to determine reasonable OCW ranges based on the number of RA-STAs for uplink transmissions. In a centralized UORA approach, one of the major challenges for an AP is to accurately and quickly evaluate the exact number of RA-STAs to minimize collision probabilities without the aid of a specific signaling mechanism. The centralized approach in Wi-Fi 6 can significantly degrade the performance of the UORA owing to an inappropriate range of OCW, particularly when the number of random access stations changes dynamically. In this study, a Collision-based Distributed OBO control (CODOBO_CTRL) scheme is proposed where each RA-STA independently determines its OBO counter, instead of an AP-guided centralized mechanism. The motivation for CODOBO_CTRL is to improve the performance of UORA in highly dynamic networks using primitive network parameters. Considering the success or failure of recent UORA transmissions, in the CODOBO_CTRL scheme, each random access station actively adjusts the OBO counter using the relationship between realistic network parameters: channel bandwidth and a collision factor. The results of a simulation study validate that the achieved throughput of the proposed scheme is almost equal to that of the OPT_UORA scheme in a static network, and 11% higher specifically in highly dynamic scenarios. Furthermore, compared to the standard UORA and OBO_CTRL schemes, the CODOBO_CTRL scheme demonstrates improvements of up to 207% and 12%, respectively, in terms of throughput performance.