Research Article
A Q-Learning Approach to Energy-Efficient Routing in BLE Mesh Network Based on Duty Cycle Scanning
@INPROCEEDINGS{10.1007/978-3-031-70507-6_27, author={Longrong Jiang and Jing Liu and Lan Wang}, title={A Q-Learning Approach to Energy-Efficient Routing in BLE Mesh Network Based on Duty Cycle Scanning}, proceedings={IoT as a Service. 9th EAI International Conference, IoTaaS 2023, Nanjing, China, October 27-29, 2023, Proceedings}, proceedings_a={IOTAAS}, year={2024}, month={10}, keywords={Bluetooth Low Energy mesh Q-learning energy consumption}, doi={10.1007/978-3-031-70507-6_27} }- Longrong Jiang
Jing Liu
Lan Wang
Year: 2024
A Q-Learning Approach to Energy-Efficient Routing in BLE Mesh Network Based on Duty Cycle Scanning
IOTAAS
Springer
DOI: 10.1007/978-3-031-70507-6_27
Abstract
The Bluetooth Special Interest Group (SIG) introduced the Bluetooth Low Energy Mesh (BLE Mesh) network specification in 2017, enabling multi-to-multi communication capability for devices operating on the Bluetooth Low Energy protocol. This specification has made BLE mesh network versatile for a range of Internet of Things (IoT) applications, particularly in building lighting and smart home systems. However, the existing BLE mesh network specification employs a managed-flood-based mechanism at the network layer for message dissemination, resulting in both message redundancy and unnecessary energy expenditure. This paper makes two innovative contributions to address these shortcomings: 1) Introduction of a broadcast routing protocol based on Q-learning algorithms. This approach enables network nodes to optimally select the next-hop relay node utilizing localized Q-value tables, thereby substantially mitigating data packet redundancy within the network. 2) Formulation of a comprehensive set of scanning-broadcasting strategies. These strategies not only ensure the reliable transmission of data packets but also facilitate a low-power standby mode for the majority of the network nodes’ operational time, thereby enhancing the overall energy efficiency of the network. Based on the results of our simulation experiments, the proposed methodology significantly enhances the longevity of nodes while concurrently minimizing message redundancy within BLE mesh network.
