Toward AIML Enabled WiFi Beamforming CSI Feedback Compression: An Overview of IEEE 802.11 Standardization (2503.00412v1)

Abstract: Transmit beamforming is one of the key techniques used in the existing IEEE 802.11 WiFi standards and future generations such as 11be and 11bn, a.k.a., ultra high reliability (UHR). The paper gives an overview of the current standardization activities regarding the artificial intelligence and machine learning (AIML) enabled beamforming channel state information (CSI) feedback compression technique, defined by the 802.11 AIML topic interest group (TIG). Two key challenges the AIML TIG is going to tackle in the future beamforming standards and four defined key performance indicators (KPIs) for the AIML enabled schemes are discussed in the paper. The two challenges are the CSI feedback overhead and the compression complexity, and the four KPIs are feedback overhead, AIML model sharing overhead, packet error rate and complexity. Moreover, the paper presents a couple of AIML enabled compression schemes accepted by the TIG, such as the K-means and autoencoder based schemes, and uses simulated and analyzed data to explain how these schemes are designed according to the KPIs. Finally, future research directions are indicated for encouraging more researchers and engineers to contribute to this technique and the standardization of the next generation WiFi beamforming.