Autonomous Rules

[petec]

Discussion about alterations to rules and arenas for autonomous systems.

[BeligerAnt]

To save reinventing the wheel, we should at least consider existing autonomous robots. The ones closest to what we are all interested in are mini sumo and micro sumo, although neither quite fit with what we have at the moment.

Mini sumo: must fit in a 100mm x 100mm footprint, no height restriction. 500g maximum weight. Fights on a 770mm (30") circular arena, 25mm high, black with a 25mm white border and a 500mm "outer area" which is not white and is free of any objects.

Micro sumo: must fit in a 50mm x 50mm footprint, no height restriction. 100g maximum weight. Fights on a 385mm (15") arena with a 12mm border.

For the really adventurous, there is nano sumo - robots must fit in a 25mm cube and weigh less than 25g!

From some web research, it seems that mini sumo is quite well-supported in the US and Canada. I also know of a few mini sumo robots in the UK. There are a few mini sumo kits available both in the UK and the US.

I think micro and nano are less well supported and I do not know of any in the UK at the moment.

There are several sets of rules out there, it seems that every robotics society in the US has its own slightly different version. Some allow remote-controlled robots, but most make the competition autonomous-only. All the rules are based on the original Japanese (robotic) sumo rules for a 3kg robot that fits in a 200mm square footprint.

A fairly coherent set of rules can be found at:
http://www.sorobotics.org/RoboMaxx/sumo-rules.html

A useful description of the rules with some enlightening comments and some nice illustrations can be found at:
http://www.robotroom.com/SumoRules.html

It appears that mini sumo is a more genteel sport than the destructive combat we are used to, but maybe if it ain't broke don't fix it, and we should just adopt the mini sumo rules to start with. We could certainly attract the existing mini sumo builders and maybe even convert them to "the way of the antweight" in time

I have a paper design for a mini sumo robot and some of the parts. If it ever gets any further than the drawing board the PCB design and software will be published on my website to help others. Unfortunately I do not have the time to make or assemble boards for other people.

A couple of kits/parts suppliers I found:
http://www.totalrobots.com/beam.htm#minisumo (complete kit for ?56!)
http://www.robotmaker.co.uk/index.htm

Would this be a good starting point? Is it the direction we want to go in?
I think sumo is difficult enough for us all to start with, we can always progress to weapons when we collectively have some more experience.

Discuss...

[Remote-Controlled Dave]

So, are we talking autonomous v. autonomous only? We talking about another separate comptetition as opposed to pitting them in with existing ants?

[BeligerAnt]

We could do either, or both. It's all part of the discussion at the moment.
However, I think there are some practical reasons why we should consider separate competitions for now.

1. The current developments in autonomous fighting robots are in sumo. AFAIK there are no autonomous combat robots (ie with active weapons) in any weight class.

2. Including autonomous robots in the current rules would require changes to the arena (autonomous robots have problems with corners). Also, I suspect the battlebox side-screens would play havoc with infra-red "vision" systems, due to reflections. A battlebox big enough to overcome these problems would be very big and expensive.

3. I do not think autonomous robots (at the hobbyist level) are anywhere near sophisticated enough to take on a human-controlled ROV (remote operated vehicle).

What do people want? What are people hoping/planning to build? Does anyone have experience of building/programming autonomous robots? Where are we starting from? Where do we want to get to?

Enough questions...

[Adam Hargreaves]

well, I was hoping to chuck my ATR in against a few ants as an exhibition fight at Dave's AWS (need the photographs for DT coursework). The theory behind it is sound, but it's gonna be over ant weight limits :S

[Remote-Controlled Dave]

I was gonna ask about considering a bigger weight limit for auto's.

[BeligerAnt]

Why should an autonomous ant be allowed a weight advantage? It would certainly help with a pusher, but is it necessary? The electronics needn't weigh very much, even the sensors are quite small and light.

Weight is always a big advantage in ants as traction is very dependent on weight, at at this level every gram counts! This is why walking spinners have proved so popular and (sometimes) effective. The extra weight doesn't just allow more mass in the disc, it also means the walker is less likely to fly out of the arena when the disc hits its target.

It's unfortunate that the existing autonomous sumo classes don't align with ants. A 100g micro-sumo bot would struggle to compete against 150g ants (due to the weight disadvantage), whereas as 500g mini-sumo bot is really quite a different beast to a 150g ant, although their size limits are similar. Adding 350g of lead to an autonomous ant would not make an effective mini-sumo bot!

Is anyone else actually interested in autonomous bots, or am I just idly chatting to Dave?

[Simon Windisch]

I can't see myself building an autonomous robot anytime soon. I'd probably build a featherweight or a super, or possibly a fleaweight first. It sounds too hard.

Simon

[EpicentrE]

In my opinion, the driving and tactics part of a fight is as much a deciding factor on the outcome as the robots themsevles. Pitting autonomous robots against human-driven ones would take a degree out of the fight that I, personally, think is very important. Then again, given that there is no autonomous ants atm, and will likely never be more than a handful of them, I don't supose I have anything to worry about

P.S. Simon, build a Flea, there good fun

[josh]

im lookin at buildin an autonomous 1 prob not battle worthy but should be ok if u really wanted a battling autonomous ant