Robo Rescue Simulation

  I was recently invited to the German University in Cairo for (among other things) judging a masters thesis by Amr Labib Hussein that dealt with a quite successful approach to the robo rescue simulation. Together with his student team he participated as team RMAS_ArtSapience and reached the final round with a solid third place. I haven't heard of this simulation before, but found it very interesting and thought I'd share a little bit about it here. The robo resuce simulation is about finding an optimal strategy for real-world disaster scenarios. Tasks include rescuing buried civilians, carrying them to refuges, clear road blockades, and extinguish fires. The maps on which these scenarios occur are relatively realistic parts of large cities and it certainly is an involved task. Judging from the thesis I reviewed and the overall results there still seems to be a lot of improvement potential (of course not because the existing teams were bad, but because the problem is so rewardingly complex). To give you an idea of what's involved, the approach taken by RMAS_ArtSapience involved fuzzy clustering and fine-tuned route searching algorithms combined into communicating agents. Lots of fun stuff ! Let's hope that the results of future competitions will eventually be incorporated into actual disaster strategies.