Comparison of WiFi interference mitigation strategies in DSME networks: Leveraging reinforcement learning with expected SARSA
3rd IEEE International Mediterranean Conference on Communications and Networking (MeditCom 2023)
Contribution to Conference
Institute of Electrical and Electronics Engineers Inc.
IEEE 802.15.4 Deterministic and Synchronous Multichannel Extension (DSME) networks have demonstrated their robustness in industrial environments, particularly in data collection scenarios. However, their performance in coexistence with other wireless technologies, such as WiFi, remains largely unexplored. In this work, we perform a simulation analysis using the OpenDSME framework to evaluate the effect of WiFi interference on a DSME network for data collection, considering different channel diversity mechanisms. The proposed strategies include an overprovisioning scheme and the adoption of the recently proposed virtual sink strategy to countermeasure the inherent funnel effect.Our findings indicate that, in general, channel adaptation outperforms channel hopping, except in scenarios with high transmission rates and limited resources, where channel hopping is more effective. When comparing the proposed strategies, the frequency selection algorithm based on reinforcement learning using Expected State-Action-Reward-State-Action (SARSA) demonstrates the most favorable overall performance in the presence of WiFi interference.
IEEE 802.15.4 DSME
Virtual Sink Strategy
004: Computer Sciences