FPCSR: A Fixed-Point-Based Resource Allocation Scheme for Multi-AP Coordinated Spatial Reuse in IEEE 802.11bn Networks

The upcoming IEEE 802.11bn (Wi-Fi 8) standard emphasizes multi-access point (AP) coordination to enhance spectral efficiency and network performance in dense environments, where coordinated spatial reuse (CSR) stands out for its practical and deployable. However, existing CSR methods are limited to intra-extended service set (ESS) scenarios, overlooking practical multi-ESSs environments where severe cross-ESSs interference poses significant challenges. Although deep learning-based methods exhibit strong adaptability, their limited interpretability hinders practical deployment. To this end, this paper proposes a fixed-point-based resource allocation scheme for multi-AP coordinated spatial reuse in Wi-Fi downlink (FPCSR) to improve throughput by optimizing power and the user association. FPCSR mitigates cross-ESSs interference through alternating iterative optimization. Specifically, we propose a knowledge-driven fixed-point network to optimize power control, where the weighted minimum mean square error-inspired optimization steps are embedded into the fixed-point network to enhance the interpretability. Moreover, We further adopt the genetic algorithm to select user association. Simulation results demonstrate that FPCSR achieves satisfactory performance compared to other methods. Code will be released at https://github.com/lily-qiuzz/FPCSR.