Idle Detection Based Optimal Throughput Rate Adaptation in Multi-Rate WLANs

Multiple transmission rates are supported in the current 802.11 WLANs, which allows stations to exploit different rates in an adaptive manner to cope with the variability of wireless channels and achieve the higher system throughput. Many efforts have been made at rate adaptation in the literature. The design of rate adaptation algorithms needs to take four basic issues into account. In this paper, by properly addressing the four issues, we present a novel rate adaptation algorithm, idle detection based optimal throughput rate adaptation (ITRA). We apply a comprehensive approach to designing the rate adaptation mechanism and attempt to maximize the overall system throughput. ITRA is furnished with two instructive features. First, ITRA modifies the conventional exponential backoff rule to a constant one. With the help of constant contention window (CW), ITRA can discriminate collisions from channel errors, and does not need RTS/CTS handshakes which are adopted by many existing works. We derive an explicit relation between the throughput and the CW size, and choose the CW size to optimize the throughput as well as improve the fairness. Secondly, ITRA directly estimates the current network throughput and changes transmission rate according to the estimation. Since the fundamental purpose of rate adaptation is to improve the network throughput, using the throughput as the metric is more straightforward than applying indirect criteria like frame transmission statistics or PHY metrics. We evaluate our proposed scheme by simulation, and the result shows that ITRA achieves satisfactory performance.