Sensor Enhanced Localization for Internet of Things
Dr Anand Bhojan, Senior Lecturer, School of Computing
COM2 Level 4
Executive Classroom, COM2-04-02
Abstract:
Beacons, the backbone of the physical web and location-based services, are widely used to tag objects and places. The former typically broadcast a message containing a particular ID or a web URL. Smarter beacons, like the SensorTags, have a wide range of sensors that includes light, humidity, object temperature, ambient temperature, microphone, and gyroscope among others and even allow third-party sensors, like heart rate sensor, to be mounted. Unlike simple beacons, these tags also have Internet access, via WiFi or NB-IoT connectivity, to provide for a richer set of applications through Cloud service providers.
As the simplistic beacon's wireless transmission is limited to a ranging technology, the localization information is only available when the beacon is nearby. Moreover, a nearby beacon, without localization information, is not easily found as it can be in any of the close-by rooms. In this work, these limitations are overcome using a scheme that is based on expanding the beacons' capability in two ways. Firstly, the system builds upon crowdsourcing using data from smartphones, carried by mobile users, and infrastructure beacons to search for facilities within an area and navigate to the target region. Secondly, in order to enable room level localization, the beacon's sensors are leveraged to generate a signature unique to the room it belongs. Apart from asset localization, in case of health-care emergencies, our system can be leveraged to quickly identify the building and room in which the person requiring medical attention is located.
Next, sensor data, from wrist-worn tags, forwarded to Cloud services to enable the triggering of external facilities, especially related to health-care, give rise to privacy concerns. To this end, we propose a lightweight privacy-preserving trust model based on the observation that a large class of applications can be provisioned from simple threshold detection. Moreover, we also propose a trigger uniformization mechanism to increase the entropy of the proposed scheme when a Cloud service provider attempts to infer which sensor is possibly behind the cause of the emergency trigger.