🤖 AI Summary
This work addresses the transmit power allocation problem in IEEE 802.11bn multi-AP coordinated spatial reuse (Co-SR) by proposing a per-TXOP power optimization framework based on proportional fairness (PF). Through theoretical analysis, the original two-dimensional optimization is reduced to two one-dimensional line searches, and closed-form solutions are derived under discrete modulation and coding schemes (MCS). The study proves that in any Pareto-optimal solution, at least one AP operates at maximum transmit power, substantially reducing computational complexity. It further reveals a critical distinction between continuous and discrete rate models in shaping fairness-oriented strategies. Komondor-based simulations demonstrate that joint evaluation of dual-AP links is essential to unlock Co-SR gains, yet current 802.11bn signaling supports only selfish strategies, necessitating enhanced per-TXOP channel feedback mechanisms to achieve fair and optimal performance.
📝 Abstract
IEEE 802.11bn (11bn) introduces Coordinated Spatial Reuse (Co-SR), a Multi-AP Coordination (MAPC) scheme in which two Access Points (APs) coordinate to control the transmit power for a simultaneous transmission. This letter introduces a Proportional Fairness (PF)-driven framework for allocating Co-SR transmit power on a per-Transmission Opportunity (TXOP) basis. We prove that any Pareto-optimal power pair keeps at least one AP at its maximum power, reducing the joint two-dimensional search to two cheap one-dimensional line searches that fit comfortably within TXOP timing, and show that this collapses to an exact closed-form expression in a real, discrete-rate system. We validate the resulting policies, together with a low-complexity selfish baseline, against a brute-force oracle and through packet-level simulations in Kom8ndor, an 11bn network simulator. Results show that jointly evaluating both APs' links is key to unlocking Co-SR's spatial reuse gains, that the coordinated AP's fairness-optimal power depends critically on whether the rate is modeled continuously or through real, discrete Modulation and Coding Scheme (MCS) steps, and that current 11bn signaling supports the selfish policy but not the fairness-optimal ones, which would need additional per-TXOP channel reporting.