Indoor environments---characterised by obstacles such as walls, floors, ceilings, and furniture---provide a number of challenges for location sensing. In particular, these environments prohibit the use of common outdoor systems such as GPS. Many inexpensive indoor positioning systems use the properties of ultrasound to provide spatial information to their algorithms. Traditionally, this information is in the form of range measurements, obtained using a combination of ultrasonic signals and electromagnetic signals. The main disadvantage to this dual-medium approach is that it requires more hardware than an approach that only uses one medium.
In this thesis, we demonstrate that it is possible to position mobile computers in an indoor environment using only narrowband ultrasound. We describe the design, implementation and evaluation of two novel systems: the Synchronous Buzz and the Asynchronous Buzz positioning systems. In both systems, the only form of measurement is from narrowband ultrasound signals originating from beacons in the environment. Positioning is made possible through the use of transmission patterns, which communicate timing information from the infrastructure to receiving wearable devices. Compared with traditional systems, our approach saves on component costs and power consumption, while improving on form factor.
The evaluation of the systems is done through the use of a custom built simulator and two real-world experiments. The simulator uses a constrained random-walk to generate realistic paths of a wearable user. Sensor noise, reflections, occlusions, as well as beacon locations can all be controlled within the simulator's environment. The proposed real-world experiments compare the paths generated by the positioning algorithms to two different ground-truths: a known static path followed by the wearable and one captured by a high-speed camera. Based on the results of the evaluations, we provide recommendations for the uses of our systems.