PID robot heading--comments?

[nathanww]

Okay, this is an idea that I've been thinking of over the past few days. Basically the problem is that obviously steering is going to be quite hard this year, specificlly in regard to the fact that to steer the robot effectivley one will likely have to make a lot of course corrections due to oversteering and the effect of the trailer behind the robot

My idea:Implement the steering control as a PID loop, where the ideal robot heading was determined by a joystick, and the actual heading was determined by a gyro mounted on the robot. Then use this to determine the offsets sent to the motors in order to turn the robot. Each time the PID calculation was run, the gyro would be zeroed.

The thing is, I'm not sure if this would actually hold water. Has anyone used a similar system, or does anyone have any advice about it?