Backtracking

This week I need to focus on few things

  • anchoring problem simulation result
  • what I have done so far in the algorithm development, which is the basic algorithm for recharging capabilities (ideas coming from the granuloma formation of the immune systems) : on going
  • correction of paper that is due on th 22nd of January 15th January : just started , awww… submitted to Jon, waiting for feedback
  • Thesis audit that I belief more about my work rather than any submission of report, but I do prepare a report for this audit , will be beneficial for my thesis in the future and I think it works ( writing while reading …I cannot remember that much now, need to write something I guess): Submision date: 31 January 2010 –  I have scribbled some experimental setup and result but yet to be updated and refined, to to prepare on energy stuff
  • Finally …finally, I need to think on how to run the simulation in epuck..
  • Ouline the
    • 3rd IFIP Conference on Biologically-Inspired Collaborative Computing (BICC 2010) : Deadline – End of January
      • Plan: to desrcibe on the experimental design and result that I have so far , not sure whether to add biological algorithm ( I prefer not to write down the immune systems algorithm rather then planning ahead on how the algorithm should be design and why I choose granuloma solution)
    • 9th International Conference on Artificial Immune Systems – End of March
      • Plan : would like to describe on granuloma formation algorithm

Swarm Tolerance to Failure:

Let me explain about my work here. In my previous simulation, we simulate 10 linuxbot robots, that are attracted to the beacon (the black box) on the right of the arena.  The robots need to – in a sense – cooperate. One robot can’t make it to the beacon on its own, nor two, nor three or four. Five is about the fewest number that can get to the beacon. We than arranged that one, two and three of the robots fail: which make them stop moving. Their electronics and software still works, only the energy have failed. But because the swarm works cooperatively, the failed robots have the effect of anchoring the swarm and impeding its movement toward the beacon. In our previous experiment, we show that t he ‘anchoring force’ of the failed robots really have effects to the the overall swarm behaviour.

To tolerate those failure, before going to the development of granuloma formation algorithm, I added energy to the robots. In traditional way on swarm robots, the energy charger can be added to the simulation so that the robots that reduced their enery can move to the charger and recharged. Results are still pending and need to be discussed whether worth plotting the results or not, we might plot and look into the coherence of the robots and the overall performance ( need to get feedback from Jon though)

So this week:

  • to finish the aggregation algorithm with charger this week
  • to start granuloma formation algorithm next week
  • to look into signature algorithm presented by Jon Aitken to be extended in granuloma formation algorithm

Overall, another week, another work ….

Advertisements