autonomous rules

[coldabert]

I was just wondering if there are any limitations on dynamic autonomous programming. I have started to design a program that acts in real-time and changes according to its enviroment. Is there any rule against not having a set program running?

[Cory]

Nope, what you sound like youre doing is what Wildstang did last year where the robot knew where it was on the field, even if it got pushed off course, and could find it's way back to it's objective. It is perfectly legal.'

Cory

[shyra1353]

I never knew about that being possible .. how exactly do you plan to do it ?? and how did Wildstang do it last year ?? And did it work ??

[EvanG]

It's perfectly legal. To see what wildstang did, I suggest you consult their web-site and view their flash video. http://www2.wildstang.org/2003/video/StangPS/StangPSOverview.html

Essentially, it requires reading in variables from various sensors and then reacting to them autonomously. The difficulty level of doing this depends on what you decide to do. Some teams have spent thousands of man hours on this, tried, and failed, some teams just read wheel counts so they know their speed. What you do depends on what you can handle. Its not a small task to make something completely responsive, but various levels leading up to completely responsive can be done.

A simple idea of how to react to it autonomously, take a look at this "basicified" code.

Wheelcount = sensorreadin
If wheelcount > 50 then slowdown else speedup
end if

slowdown:

speedup:

[coldabert]

Thanks for the reply all. I have done alot of research in artificial intelligence. I am going to try a method that I have used in some video games where the robot will create its own methods of navigating using neural networks. I have used this before in a lisp machine I wrote for an rpg, so I think I can easily implement it in our robot.

[Dave Scheck]

Thanks for answering their questions Evan.

coldabert, based on your original question and last respone, it sounds like you have some interesting ideas. I think we'd all be interested to see how it works out. In the meantime, if you have any questions about our system last year, we'd be more than happy to answer them.

[coldabert]

I have looked at wildstang's ideas, and I think they are pretty good. However, random events such as electromagnetic interference and collisions will really obscure his calculations. The tactometer approach doesn't really work well for traveling over long distances. I think I'm going to use a collection of sensors and algorithms to calculate position and formulate a response.W hen I have a near finished idea about my approach with a prototype, I will post it here and to my website.

[Air Soarer]

How did StangPS get the angle of the robot relative to the field?

[Pat Fairbank]

How did StangPS get the angle of the robot relative to the field?
Correct me if I'm wrong, but I think they used the kit-provided yaw rate sensor (it's not in this year's kit). If you're using tank drive, you can also calculate this angle without the gyro, and with just the distance each side has traveled.

[coldabert]

I'm sorry to announce that due to mechanical failures, I wasnt able to try my autonomous ideas on our robot. I did have a simulator to test them on, and I think they might have worked. Hopefully next year and over the summer the mechanics will get their stuff together and we can try it out in a competition setting. I think the way I designed the code generator, the bot will have enough intelligence to complete goals given to it by the human elements. If anyone wants the specifics of my method, email me at fredderfchicken@hotmail.com