As with all stories of inspiration, my imagination was captured as a child watching science fiction shows and films, such as Buck Rogers in the 25th Century, Silent Running and, of course, Star Wars. The one aspect which always appealed to me was the robot companion/worker, which always succeeded in sparking as much of an emotional reaction as any other character.
So the question is, why didn’t I follow up on this years ago when I had the chance?
For the most part, it would have to be confidence – or a lack of. However, all this changed when I hit the mid-life crisis and realised that if I didn’t make the effort to get involved in the field now, then it would never happen, so I took the plunge and enrolled at university.
For the final year of my BSc in Computer Science (Artificial Intelligence), I needed to consider a subject matter for the required project. Immediately, the robot companion idea came to mind, though now with a much better understanding of the implications. However, I had been told by a PhD student that I needed to have a question I wanted to answer, but the only question arising in my mind at that point was: Can I do it?
Initially, the idea was to develop a robot which would understand its location based on a range of clues, from visible to invisible landmarks, such as feature extraction to wifi access point signal strengths. However, a chance viewing of a TED presentation by Michael Dickinson inspired me to start wondering if the compactness and efficiency of an insect brain would be a suitable approach for embodied agents to localise and orient themselves in their environment.
Soon after this, a chance viewing of a documentary about honey bees and studies into the effects of Neonicotinoids on their ability to function, highlighted an uninfected bee’s localisation and orientation capabilities and inspired me to consider this important creature as the model for my project … and therein lay the question I now want to answer.
Does the brain of the Honey Bee provide an appropriate structure and function to be modelled for an embodied agent to localise and orient itself in its environment?
They did say we should be ambitious!