Adaptive Rate Mechanism for WLAN IEEE 802.11 Based on BPA-Artificial Neural Network

Abstract

IEEE 802.11 WLANs provide multiple transmission rates to improve the system throughput by adapting the transmission rate to the current wireless channel conditions. The AutoRate Fallback (ARF) scheme is a simple and heuristic link adaptation approach and compliant with IEEE 802.11 standard, also most of commercial devices implement it but it’s suffer from random packet collisions especially when the number of nodes increases and consequently cause a decline of the over all throughput. In this paper we propose rate adaptation in WLAN 802.11 based in neural network. The proposed rate adaptation scheme, appropriately adjust the data transmission rate based on the estimated wireless channel condition, specifically by dynamically adjusting the system parameters that determine the transmission rates according to the contention situations including the amount of contending nodes and traffic intensity. Through extensive simulation runs by using the Qualnet simulator, we evaluate our proposed scheme to show that our scheme yields higher throughput performance than the ARF scheme.