Before coming to University, I spent time working in a number of industries. Most recently I worked in the roof maintenance industry as a technical consultant. While doing this work I realised that there should be technologies that could greatly simplify some tasks. In particular, how do we determine what maintenance should be done, given the condition of a roof? This is challenging for sales staff without a technical background, and especially challenging to new people.
During third year study, the problem was addressed, and the underlying engine for such an expert system was developed and tested. Later, a prototype system was created using this engine, implemented on a hand-held device.
Getting results from Artificial Neural Networks, in real-time, on a hand-held device? Can't be done. Can it?
Another major issue is the social problem; Roofing staff admit to being technophobic. How can we design an interface on a small device that is so natural that they don't even think about it? Be amazed at the simplicity (and speed) of the solution created in response to this complex problem.