Dr James Pinchin from Horizon Digital Economy Research Institute will present:

Title: Unsupervised Location Labelling

Determining position indoors has been the subject of much research in recent years and many approaches have been proposed. Most approaches rely on prior knowledge or survey of the environment to be navigated. This requirement quickly becomes difficult to achieve when the spaces grow to cover an increasingly large area, cover multiple organisations and are subject to changes over time.

In this talk we present indoor positioning within unknown environments as an unsupervised labelling task on sequential data. A user’s location at each measurement update is assigned a label from a set of labels drawn from contextual information. In our test case the contextual information consists of the locations of tasks assigned to hospital staff. We develop a probabilistic framework relying on wireless network radio signals and the contextual information. Thus, we form an informative spatial classifier without resorting to a pre-determined maps or infrastructure knowledge and show the potential of the approach using both simulated and real data sets.

Results demonstrate the ability of the procedure to segregate structures of radio signal observations and form clustered regions in association to areas of interest to the user; thus, we show it is possible to differentiate location between closely spaced zones of variable size and shape.

Dr James Pinchin is a Research Fellow in Geospatial Science in the University of Nottingham’s Horizon Digital Economy Research Institute. He is interested in the incorporation of application and context into the design of positioning and navigation systems. James has experience of working with a wide variety of technologies for the measurement of position and attitude and has worked with platforms ranging from aircraft to doctors. He is currently working on the relationship between location and activity and the application of that knowledge to the measurement of workplace behaviour.