Researchers at the University of Trento and the University of Cassino in Italy present an approach for detecting passive targets capable of detecting multiple complex target properties, such as position and posture, using wireless networks. Wireless detection is particularly important and useful in application fields like security and healthcare when the targets do not participate in the localisation process. In contexts where such localisation is needed wireless networks are nearly ubiquitous as electromagnetic fields propagate through objects like walls and furniture. Many existing detection technologies based on WiFi networks require wearable devices to operate. This requirement limits the adoption of wireless detection to cooperative users who are aware of being localised. It is not always possible or desirable to require patients in healthcare contexts to play an active part in localisation and of course in security applications intruders cannot be expected to be cooperate. Passive target perturbation of electromagnetic fields generated by existing wireless devices (below) compared to a standard background electromagnetic fields (above). Estimation of detailed passive target location-based features using volume of interest partitioning to improve the estimation of details using a step-by-step approach. One of the main difficulties associated with wireless detection of passive targets is the high complexity of the interactions between EM fields generated by wireless networks and the targets. The current limitation is poor detection reliability, which needs to be improved in order to make the detection truly usable in real-world scenarios. High variability of EM propagation in indoor scenarios, as well as the very high number of degrees of freedom of EM fields makes wireless detection very challenging. In this issue of Electronics Letters, the Cassino team reduce the complexity of wireless detection through the iterative estimation of multi-resolution target features. Their reported approach exploits the information contained in wireless signals by focusing step-by-step on different target behaviour details with an increasing resolution level. The estimation of very detailed target features, such as posture, is feasible with this approach. The main challenges the team had to address in this work included separation of wireless signal components related to the impact of the targets from those due to the environment; feature extraction from the wireless signals showing the complex relationship between electromagnetic fields and the targets; and system reliability with high generalisation capabilities for the proper estimation of target properties with reduced parameter calibration. The proposed approach in this Letter allows the exploitation of existing WiFi networks using information that is already available in standard WiFi devices. In the short term, such an approach can be considered a new functionality of WiFi networks. In the longer-term, developments reported in this Letter will exploit future WiFi standards trends, which will enable an increasing accuracy for understanding more complex target features, including vital signs like breathing rate and heartbeat, using commercial WiFi access points. “The next-generation development of the proposed passive-based approach will enable applications, for example in the field of healthcare, that are currently implemented using complex and expensive wearable devices”, states Cassino team member Federico Viani. The reported work falls within the research topic of wireless localisation of passive targets that has been developed by the ELEDIA Research Center for more than a decade. The research group of Viani and colleagues will continue to have a focus on wireless target detection methods for real world scenarios using existing infrastructures. The main limitations that still need to be addressed are related to long-term reliability. Over the last ten years, the high interest in IoT-based systems has stimulated the application of wearable technologies, thus enabling the development of new location-based services. Heterogeneous smart sensors have been designed to monitor the health and activity of users. Over the next ten years Viani expects that “the functionalities of wearable sensors and gadgets will be substituted by the development of smart environments that passively sense the occupants and understand their features without the need for wireless devices to be worn. Security-related application fields as well as healthcare and interactive gaming will benefit from novel passive detection approaches. For example, think about your WiFi access point that is able to detect sleep apnea or cardiac arrhythmia in real-time without using invasive sensors or expensive medical devices”.
Read full abstract
May 29, 2024
Jianke Li + 5
Open Access
