Wireless sensing devices: from research to real applications in logistics and healthcare

Wireless Sensor Networks have the potential to enable design of innovative applications and reinvent existing ones. To this effect, the scientific community has made many contributions. However many researchers relied on strong assumptions leading to existing gaps between theory and practice. Many of the proposed solutions can work only in a laboratory setting. This thesis provides two examples on how to bridge this gap by developing two distinct real world applications in the fields of logistics and healthcare; the first application overcomes the limits of rangefree localization techniques, whereas the second application provides a solution for efficiently monitoring human movements and includes a usability study. In logistics, the problemof localization of shipping containers is phased starting from the description of currently adopted systems along with their limits. Sensing nodes are placed on containers to detect the signal strength and thus the presence, of other nodes placed on other containers. The system exploits geometrical constraints and an integer linear programming solution to localize even in presence of real faulty nodes. The sensor network’s functional and their non-functional issues, such as energy consumption, scalability and fault tolerance are studied. In the healthcare application, sensors are used in a monitoring system to detect falls among elderly. The approach starts from a survey of fall detection techniques to the design of our fall-detection algorithm which reduces false positives. A conceptual, minimally invasive monitoring sensor platform and reusable architecture is designed for the deployment and testing of the algorithm. The usability and acceptability study of this application in our test sites revealed some interesting insights about human aspects and time of adaptation to the technology.