Intrusion Detection System for Bluetooth Mesh Networks using Machine Learning

Bluetooth Low Energy (BLE) mesh network is gaining popularity in the IoT sector due to its simple network configuration, network extension and security features. Device security is a concern for IoT systems and is crucial in detecting device attacks. This paper proposes a machine learning based intrusion detection system (IDS) targeted for the BLE mesh networks. The setup consists of nine ESP32 BLE mesh devices and one raspberry pi. The IDS detects known threats, such as black hole attacks and grey hole attacks, in the BLE mesh network using a single watchdog node. The watchdog setup continuously sends network layer data to the Message Queuing Telemetry Transport (MQTT) broker. And the database stores the BLE mesh packets with the timestamp. The target attack node uses a push button to switch between attacks for evaluating IDS. The system uses machine learning algorithms to detect network attacks. The project analyses the performance of attack detection when the location of the watchdog is changed. The machine learning model classifies legitimate, grey hole attack and black hole attack. Finally, a web application shows the BLE mesh network detection status.