A BLE Mesh Edge Framework for QoS-Aware IoT Monitoring Systems

Internet of Things (IoT) monitoring applications often require a flexible network design to continuously adjust device configurations and meet Quality of Service (QoS) requirements effectively. In this paper, we provide evidence of this concept in the context of the Bluetooth Low Energy (BLE) Mesh network, where performance heavily relies on tuning various network parameters. Our contributions in this study are threefold. First, we present a software framework designed for BLE Mesh monitoring, QoS metrics computation, and QoS-aware parameter selection. This framework operates on an edge device connected to the BLE gateway, enabling efficient data gathering and decision-making. Second, we propose a Reinforcement Learning (RL) algorithm to determine the optimal BLE Mesh parameters, allowing for the simultaneous fulfillment of multiple QoS requirements. Through the RL technique, we can dynamically adapt network configurations to varying conditions and demands. Third, we conduct thorough validation of our framework on a real-world BLE Mesh test-bed, demonstrating the effectiveness of the proposed technique in network performance enhancement. The results show the ability to maximize Packet Delivery Ratio (PDR), minimize delay, or determine a suitable trade-off between these two metrics, depending on the specific user's needs.