Even as an undergrad, I had a paid part-time job (as a lab tech) in one of Europe’s premier electrical machines and drives research groups. It was great to be immersed in that culture, so when my final year undergrad project came round, I threw myself into building a lab rig to run some of the new fangled Artificial Intelligence (this was 1996) to implement real-time control. I got 12/14 for my project write-up, so overall pretty pleased. I started my PhD that summer.
Fast forward to 2002, and by now I’ve just started as a junior lecturer after a couple of post-doc jobs, and starting to look at publishing some papers on those projects. By chance, I had a flick through my undergrad thesis, and with new eyes, realised that the project was novel and I didn’t think anyone else has done it. Luckily, no-one had published anything like it, so I got……..
- Stewart P., Stone D.A. and Fleming P.J. “Design of robust fuzzy-logic control systems by multi-objective evolutionary methods with hardware in the loop” IFAC Journal of Engineering Applications of Artificial Intelligence, Vol.70, no.3, pp.275-284, May 2004.
- Stewart, P.; Stone, D.A.; Fleming, P.J., “On-line multiobjective automatic control system generation by evolutionary algorithms,” Industrial Electronics, 2004 IEEE International Symposium on , vol.2, no., pp.1459,1464 vol. 2, 4-7 May 2004, doi: 10.1109/ISIE.2004.1572029
- Stewart P., Stone D.A. and Fleming P.J., “On-line design of robust fuzzy-logic motion control systems by multi-objective evolutionary methods”, 2004 American Control Conference
which led to….
- Gladwin D., Stewart P., Parr M. and Stewart J., “Multiobjective evolutionary-fuzzy augmented flight controller for an F16 Aircraft.” Proceedings of the Institution of Mechanical Engineers, Part G: Journal of Aerospace Engineering. Vol.224, No.3, pp 293-309, 2010. Awarded IMechE Charles Sharpe Beecher Prize for best Aerospace Paper 2010.
The big message here is that if I hadn’t become an academic, I would have lost the publication of those first three papers. I just wouldn’t have known ;-). So I begant to treat my undergrad project groups and MSc projects in the same vein as my PhD students. Concentrating on novelty, formal methods. It doesn’t always work out, but the 4th paper above came out of an MSc project. It looked great on Matt Parr’s CV when he graduated and took a job with Rolls-Royce. I think this is one of the most positive contributions we can make for our students.
Here’s the journal paper from my undergrad project. Out of interest, it was a very early example of multiple AI heuristics interacting to optimise performance in real-time with hardware in the loop. I’m just about to develop an MRI research project based on this principle. I think it looks a bit crappy now, but at the time it was great…