Purpose - To describe the architecture of iPOS (short for iPAQ positioning system), a novel fault-tolerant and adaptive self-positioning system with quality-of-service (QoS) guarantees for resource-limited mobile devices. Design/methodology/approach - The iPOS architecture is based on a novel sensor modelling technique in combination with a probabilistic data-fusion engine, which is capable of efficiently combining the location information obtained from an arbitrary number of heterogeneous location sensors. As a proof of concept, the paper present a prototypical implementation for handheld devices, which was evaluated by means of practical experiments. Findings - A major advantage of the iPOS positioning system is its extensibility and flexibility; which is achieved by means of an open plugin architecture and the support of global positioning coordinates according to the WGS-84 standard. The iPOS system scales very well with respect to the number of sensor plugins that can be operated in parallel. The main limiting factor for the number of supported active plugins is the amount of available system resources on the MoD. With regard to recognition, the experimental results indicate a good accuracy of the fusion-based positioning system in comparison to the accuracy of the individual sensing technologies. Thanks to the explicit modelling of reliable sensor events, the iPOS system is capable of providing QoS guarantees to applications with regard to the achieved positioning accuracy. Research limitations/implications - During the experiments, the author recognized time synchronisation as an important challenge that should be addressed as part of future work. Practical implications - The system enables resource-restricted mobile devices and computerised objects to exploit computing resources found in their immediate physical vicinity (locality). Originality/value - The paper presents a novel lightweight sensor-fusion architecture for fault-tolerant and adaptive self-positioning that performs well on resource-limited mobile devices. A special feature of the developed data-fusion architecture is the application of a novel event modelling technique that enables the positioning system to give QoS guarantees under certain conditions.
