Device-free radio vision is an augmented functionality provided by radio transceivers - typically heterogeneous, densely distributed and networked - that monitor the fluctuations of the electromagnetic (EM) field across the space. These monitoring devices may be pre-existing, deployed at arbitrary (or optimized) locations for communication purposes in the area of interest, and exchange digital information by any wireless communication protocol. Radio vision systems leverage diffraction, reflection and scattering phenomena that affect the radio-frequency (RF) propagation for ubiquitous sensing. RF signals (namely channel quality information - CQI) can be either narrowband or wideband, in licensed or unlicensed frequency bands, with carrier frequencies ranging from MHz to GHz, and above. The presence, position and motion of a human body in the network area affect the EM field in a predictable way, making it possible to estimate and track its activity without the need to deploy and calibrate any additional wearable sensor (sensor-free detection), nor to ask for specific user actions (non-cooperative detection). This passive sensing approach has been experimented with in heterogeneous networks but it is also appropriate for most of the emerging low-power wireless standards, and for personal and device-to-device (D2D) communication, including WiFi, Bluetooth low energy (BLE), ZigBee and D2D enabled long term evolution (LTE Advanced).
The adoption of device-free radio vision technologies is highly attractive in the context of assisted living as a fallen person might not be able to get to activate a personal emergency response system, if not forgetting how to use it. Experimental studies have been carried out by the IEIIT researchers of the Milano site, specifically focused on the real-time processing of RF channel quality for detection of the impact shock during body fall and "long lie" detection conditions (e.g., the time the elderly remain lying on the floor after falling).
The video shown in the link below, demonstrates the use of wireless device-free localization technology to detect an operator in a shared human-robot industrial workspace. The design and the validation of a sensor fusion framework (carried out in cooperation with the ITIA researchers of the Milano site) is also shown to integrate the device-free human body sensing technology with industrial image sensor sources.
Link (youtube): https://youtu.be/dS9MgFMWfl4
Future research on radio vision systems is expected to combine the use of localized RF signal inspection with large-scale big-data analytics. Running real-time analytics from massive volumes of possibly incomplete RF data, measuring channel responses over different scales (PHY, link-layer and upper layers) will be therefore of paramount importance, while posing new scientific problems, as conventional signal processing and statistical learning tools will need to be re-designed when applied to these unprecedented high-dimensional data structures.
Focus