A device identification method from BLE advertising packets with randomized MAC addresses based on regression of received signal strength

In this letter, we propose a device identification method from observed Bluetooth Low Energy (BLE) advertising packets for tracking BLE devices even if their MAC addresses are changed periodically and randomly. In our proposed method, the combination of MAC addresses is formulated as a linear assignment problem. In addition, in a cost function of linear assignment, we combine two types of cost: time-based cost and received signal strength-based cost, which is calculated based on regression of received signal strength. Through experimental evaluations, we confirmed that the accuracy of our proposed method is the highest compared to traditional methods.