Embedded Broadband Ultrasonic Sensing for Robust and Scalable Positioning

Important applications spanning the fields of ubiquitous computing and sensor networking require the spontaneous provision of reliable, accurate, and up-to-date location information in unprepared environments. Examples include simultaneous localisation and mapping of new environments using resource-constrained devices; impromptu location-aware collaborative work and data sharing between proximate users; and emergency search and rescue and live coordination of personnel in disaster sites. Embedded, ad hoc positioning systems are a crucial ingredient for such applications. In such systems, wireless and battery-powered sensing devices can be deployed in a matter of minutes to collaboratively measure and estimate their locations, and supply the required information.While solutions for minimal deployment and ad hoc positioning have been put forth for ubiquitous computing and sensor networks, none of the systems to date have been able to provide the high-fidelity information the above applications require. Due to their interactive and real-time nature, they tend to require location readings which are robust (consistently delivered accuracy of a few tens of centimetres or better) and up-to-date for all participating nodes (several location readings per second for each locatable device). The aim of this project is to develop broadband ultrasonic signalling and processing methods for ad hoc, embedded positioning systems. This will allow them to (1) produce robust estimates (through noise-resistant coding and measurement of multiple physical quantities such as range, bearing, and velocity) even when only a small number of devices are present; and (2) maintain high location update rates (through multiple access signalling) when a large number of devices are present. Our approach is to couple broadband ultrasonics (piezopolymer transducers) with real-time signal processing implemented using reconfigurable fabrics (field programmable gate arrays).