Blockchain-Based Resilient Pairing and Bonding of BLE Devices Using Deep Reinforcement Learning

In the day-to-day deployment of the Internet of Things (IoT), IPv6-over-Bluetooth Low Energy (BLE) devices became significant enablers due to their low power, low range, and effortless connectivity. To overcome the latency during BLE device migration, we have proposed a Blockchain-based resilient pairing and bonding of BLE devices using a lightweight authenticated encryption scheme. The bonding information is stored on the local ledger of the intermediate IPv6 over Low power Wireless Personal Area Networks (6LoWPAN) for BLE (6LoWPAN for BLE or 6LoWBLE) over gateway. Whenever the BLE device migrates from one gateway to another, the bonding information is transferred to the local ledger of the current gateway through the global ledger with the Proof-of-Voting consensus algorithm. The resilience of the proposed authentication scheme is analyzed using BAN logic. In addition, we proposed a Man-in-the-Middle (MITM) detection framework using Deep Reinforcement Learning (DRL) due to the escalation of MITM attacks on BLE devices. The DRL-based model is implemented in Python, and its performance is evaluated using the Kitsune Network Attack Dataset. The performance of the overall framework is tested on the Hyperledger Fabric, which results in low latency and a low average pairing time in comparison with existing frameworks.