Can I Make Something Clear?

[JamesBrown]

Originally Posted by pSYeNCe

I read somewhere that within 50 years, a single desktop computer will have the same computational power of a human brain. Albeit, it will not have the same intelligence, but nonetheless the power will be there. You suggest that this will never happen, at least within our lifetimes. You are wrong.

This is really not pertinent, sure with enough time and sufficient processing power a computer may be able to out perform a person, however there are serious limitations that prevent this from being possible in FIRST. We are severely limited in our development time. We have at most 3 and a half months to design, code, and test all game specific software for the robot (assuming a team is competing at Championships, and has a second robot to develop code on after ship, also assuming that their are no rules restricting software development after ship) We are limited in our choices for sensors, and computers. While you can certainly use co-processors and sensors that are more powerful than what teams use now, you are still severely restricted by the cost accounting rules.

There are research groups at top Universities all over the world that are working on robots that play games and perform tasks much simpler than playing a typical FRC game. These teams are creating some incredible robots, however none of them are working on short deadlines, with limited hardware, and with teams driven by HS programmers with limited knowledge in the fields of Perception and Planning.

I am not trying to discourage David, or any one else, I am simply speaking as some one with more experience than most of those posting (3 years of research between RPI's Center for Automation Technology and the Rensselaer Artificial Intelligence Research Lab) and saying that this is not a practical goal, this like any problem needs to be attacked progressively. We should not be jumping from 15 seconds of unreliable Autonomy in a relatively static environment to a fully autonomous system.

[Daniel_LaFleur]

Originally Posted by pSYeNCe

This is a good idea. Sometimes in web design I'll just make a frame of the page until I get needed resources. It works quite well. You could set it up to reference some not-yet-existing image, and then just fill it in when the time comes.

The only thing is identifying characteristics of the object. That's the tough part that takes time to program, and you won't know that until you know the object.

Maybe.

Creating a script file that defines the objects would allow you to create objects with varying numbers of properties. Also some objects are more easily defined (such as an initial map of the playfield) than others (placement, size, and orientation of an unknown game piece). All-in-all it would be an interesting and fun project.