I think I might have been a bit too harsh on M366 in my post a couple of weeks ago; I’ve been working through TMA02 this week, and I particularly enjoyed writing about subsumption architecture in the question about McSCOR, the “M366 course Subsumption COntrolled Robot”. I’ve become quite interested in the vertical decomposition way of doing AI systems, but having to write about the drawbacks of the SENSE-ACT cycle in Question 4 certainly stopped me from getting overly smitten with the approach.

I found the presentation called “Behaviour-Based Control in Mobile Robotics” by Paul Fitzpatrick from MIT’s Computer Science and Artifical Intelligence Laboratory very useful in getting my head around the characteristics of horizontal and vertical decomposition. I’m interested in the “lateral” architecture referred to in that presentation, in which the priority of the various layers/modules seems to be dynamic rather than fixed, but I don’t think this is covered in M366, so I might have to do a bit of extra-curricular reading about that one!

I’m pretty much done with M366 TMA02, except for one question that has been driving me mad – the question asks for a 400-word description of the emergent behaviour of McSCOR, but the most I could manage was about 80 words, and to be honest that was fairly waffley! So, I’ve emailed my tutor to ask for a nudge in the right direction, and in the mean time I’ve been getting back into MS221 B3.

It took me a while to get into the swing of things with MS221 again after taking a week or so away from it, but I’ve been loving all the material about eigenlines and eigenvalues, particularly the section about the long-term behaviour of sequences of points generated by iterating linear transformations.

The Mona Lisa before and after a y-shear, with an eigenvector in red. They did a similar transformation in the episode of Algebra Workout that accompanies Chapter B3, but instead of the Mona Lisa they used a photo of the presenter, which was quite amusing in a puerile way!

The Mona Lisa before and after a y-shear, with an eigenvector in red. They did a similar transformation in the episode of Algebra Workout that accompanies Chapter B3, but instead of the Mona Lisa they used a photo of the presenter, which was quite amusing in a puerile way!

I was very intrigued to read the margin note in B3 Section 2 about complex eigenvalues, which aren’t covered in MS221; I’m looking forward to the chapter in Block D about complex numbers very, very much, so I’m quite curious about complex eigenvalues too. As far as I can tell from the contents page of the M208 unit about eigenvectors (which is here), they’re not covered at all in M208, but it looks like they are covered in Unit 7 of MST209: Mathematical methods and modelling, which is available on OpenLearn here. I’m not planning on actually taking MST209, but I might dip into some of the units you can access on OpenLearn during the slow period between MS221/M366 and M208 this winter…

Advertisements