Impact evaluation and detection of malicious spoofing attacks on BLE based occupancy detection systems

Occupancy detection is beneficial for applications such as emergency management and building energy management, as it provides information on the location of occupants. Internet of Things (IoT) devices such as Bluetooth Low Energy (BLE) beacons installed in a building can benefit the performance of occupancy detection systems, by providing information on an occupant's location. However, BLE beacons operate by broadcasting advertisement messages, and this renders them vulnerable to network attacks. Here, we evaluate the effect of two types of malicious spoofing attacks on a BLE based occupancy detection system, and propose an attack detection method. The building blocks of the system include BLE beacons installed inside the building, a mobile application installed on occupants' phones, and a remote control server where we perform occupancy detection using machine learning. Our real-world experimental results indicate that the attacks can significantly affect the system's performance. Also, our proposed detection method is able to accurately detect an attack by an adversary with physical access, with accuracy ranging from 84% to 91%