Gyro vs Dead Reckoning

[arhajder]

Which is better The Gyro sensor or Dead Reckoning?

Dead Reckoning is when you program the autonomous through given intervals that are calculated by the programmer.

[psy_wombats]

Almost certainly the gyroscope. I'd try to avoid dead reckoning as much as possible, just because the counters are quite 'messy,' as far as flexibility is concerned. I can see it set up to work just as well using an interrupt-driven timer, but I just like the gyro because you can use PID control to get more refined values. (And just tell the robot to turn until it's at an angle. Much easier!)

[lukevanoort]

Sensor feedback in autonomous is almost always better than dead-reckoning; in fact, the only situation I can think of where dead-reckoning would be preferred is if your sensor was really buggy and gave false reading often (in which case you probably should be using a different sensor). The question is like asking "Is it better to drive by looking at a map the whole time or by remembering where you want to go and watching the road instead?"

I have written both sensor-driven and dead-reckoning autonomous modes, but the dead-reckoning ones were always rush jobs that weren't expected to do much (like just deploy our arm). We use a hardware timer and interrupts for timing a dead-reckoning autonomous modes. If you do use dead-reckoning (many autonomous modes will have at least some time-dependent stuff, so the advice will probably apply with gyros too), I would definitely suggest using timers. Loop-counting seems like a bit of an unreliable and undesirable solution to me.

[arhajder]

unfortunately for me my team got none of the sensors and FIRST has yet to send them in to us.

[Vikesrock]

We didn't either! According to Update #5 FIRST has begun shipping replacement parts out via Fedex 2-day and will be doing so through the beginning of next week, so you should be getting them soon

[arhajder]

Originally Posted by Vikesrock

We didn't either! According to Update #5 FIRST has begun shipping replacement parts out via Fedex 2-day and will be doing so through the beginning of next week, so you should be getting them soon

Finally! We are almost done as it is!

[jacobhurwitz]

Our team purchased an ADXRS150 from Spark Fun electronics for two reasons:
(1) It is 180 degrees/sec instead of the 80 degrees/sec one that came in the kit.
(2) It works with Kevin Watson's gyro code and with WPILib, but the one in the KoP works with neither.

However, we've been having some problems with the accuracy of the sensor. Nevertheless, using a gyro is definitely better than dead reckoning. If you use timers or something, a slight change in battery voltage may affect your times. If you use loop count, adding more code will increase the time per loop.