Device Mobility Detection Based on Optical Flow and Multi-Receiver Consensus

The signal strength of wireless transmissions from WiFi-enabled devices, such as smartphones, is related to their distance from the receiver. While this relation is environment-dependent and generally noisy, fluctuations of the signal strength can be used as an indicator of whether a device is moving relative to a static receiver. People tracking in video sequences is also the subject of many Computer Vision (CV) studies. In our work, we combine the inferred motion from these two distinct sensing modalities to identify the time intervals during which people are moving in an indoor environment and to relate to them the WiFi-enabled devices they carry. Deciding whether a device is moving or not is based on the Received Signal Strength Indicator (RSSI) from single and multiple receivers scenarios. We carry out several experiments to corroborate the efficacy of the proposed techniques (both CV and RSSI-based) in a real-world setting.