Chris Alexander

On Engineering

Building a robot swarm for £100

19th January, 2012

First of all, it has come to my attention that a number of people read my blog. Upon reflection with a friend and colleague today this was a surprise to both of us (he was referring to surprise over his own blog having readership). He pointed out that we both treat them like private diaries, but in public, which I think is true and potentially dangerous if I ever get caught napping thinking nobody is reading.

On to the topic I was intending to talk about - swarm robotics.

Other Chris and I have been having a bit of fun with this topic - not because we are working on swarm robotics, but because it seems like 50% of the MEng finalists are doing something to do with swarm robotics.

This is a bit of a joke for us. There are a number of reasons which I shall enumerate:

1) There is not the budget to do a properly working robot swarm. The budget is £100. I have spent that on 25% of a robot (it’s all my own money - no I don’t mind this, no I don’t intend to let the Uni get their hands on my robot, yes it was far easier to spend my own money than go through the Uni procurement process; any more questions?). How you intend to build a swarm of robots with £100 and 9 months I’m not sure. You may be able to build 1, 3, 5, 10, or even 20 at a real push, but as one of my lecturers made the point today, 20 isn’t a swarm of robots. I would say 100 isn’t. It’s probably when you’re getting well into the 1000′s that it is actually a swarm. Good luck making robots for 10p each.

2) As one of my colleagues pointed out, what’s the point in doing a swarm project, if you can’t make one robot do what you want yet?

3) Today our lecturer listed around 10 acronyms for various swarm intelligence algorithm families. I could get one or two (but I am not doing swarm robotics as my project, hence why I am on this module to learn), most of the class got one tops. One guy got nearly all of them - I assume his project is swarm robotics, and I do wonder what the others have been up to if they couldn’t list those algorithm classes. Presumably not background research.