Data rate-based grouping using machine learning to improve the aggregate throughput of IEEE 802.11ah multi-rate IoT networks

作者：Miriyala Mahesh, Badarla Sri Pavan, V. P. Harigovindan · 年份：2020 · DOI：10.1109/ants50601.2020.9342758 · 被引用次数：11 · 研究领域：Wireless Networks and Protocols、IoT Networks and Protocols、Bluetooth and Wireless Communication Technologies

IEEE 802.11ah, a standard for Internet of Things (IoT), uses restricted access window (RAW) mechanism to minimize the impact of collisions and to improve the aggregate utility of the network. However, in IEEE 802.11ah based multi-rate IoT networks, it is observed that throughput of higher data rate devices degrades below the level of lower data rate, due to performance anomaly. To overcome this problem, we propose data rate-based grouping using machine learning (DGML) in this paper. The proposed scheme exploits the self-organizing map neural network to classify the devices as per their data rates. Then, every group is allocated a RAW slot for channel access. Results show that the DGML scheme outperforms the default uniform grouping scheme.