Version 1

The overview of the proposed contraband wireless device detection system consists of hardware and software components. The hardware component is a receiver with multiple antennas, each of which receives the wireless signal at different locations. The received wireless signals carry unique characteristics of the device, known as device fingerprints and unique information (e.g., arrival strength and phase angle) of their propagation paths, known as location fingerprints. The software component takes the received multi-channel signals as input and applies RF signal analysis approaches including mathematical models and machine learning algorithms to differentiate devices and locations. To achieve the project goals and objectives, the study team developed and implemented the following three types of methods: device fingerprinting, location fingerprinting, and content summarization. For further technical details for each method please refer to the related publications.