[FTC]: Field Control Issues / Analysis

[billw]

At the VA tournament this past weekend, we experienced field control issues four separate times. Each time, it appeared that the robot would just keep running any motors that were active at the time control dropped out.

In three of these instances, the field management system reported that it was still in communication with the robot (I am unsure about the fourth). We experienced this behavior twice at the same station (controller #1, far left blue field). I also know that this controller was pronounced "dead" in between these occurrences and was swapped out for another. BTW, we use only a single joystick controller.

In one of these instances, our partner also lost control later in the match (controllers#3/4, same field).

We ran with absolutely no issues for more than 60 minutes on the practice field. (For those that were not there, this field was in the same venue as the tournament fields and would have been subject to roughly the same amount of Bluetooth traffic.)

Our code is written in RobotC using the current versions of firmware, etc.

I know that other teams experienced problems also, and I have heard several suggestions that Bluetooth was at fault. It did appear to give the organizers a difficult time when setting up some connections, but it has occurred to me that the problem could be in the link between the joysticks and the field management system. This would explain RF link still being reported as operational and the fact that we could run without problems in the same RF environment when disconnected from the FMS.

I was hoping that we could document these problems and perhaps use our collective experience to help identify the underlying issue. This is in no way meant to turn into a negative commentary, so please refrain from posting anything of that nature.

--Bill Wiley
Coach, ACME Robotics
FTC #12, 14, VRC #12a-e

[jbbjjbt]

Bill,
Our two teams never had any issues with the field management system or bluetooth that only affected only our robots. There was at least one match where all four robots on the field were disabled in teleop.

We use LabView, so I can't comment on RobotC. But we've had many issues when practicing and also on the field at the tournament where the hitech motor controllers go bizerk (sp?). I don't know if it is a hardware or a software issue. Our 177 bot stopped during autonomous during its second match and never recovered. The NXT brain was still running and communicating, it's just the motor and servo controllers that were messed up. The team came up with a fix or we wouldn't have had a very successful day. Our 3165 bot decided on its own to do a little spin dance at the end of autonomous. It sure wasn't in the programming. Fortunately it always recovered when teleop started.

Again, I don't know about robotC, but in Labview, the original FTC Move Motors commands had issues and would stay sometimes stay running for 2.5 seconds. Also if an encoder didn't reset, all sorts of weird results can happen. Most of these issues were corrected with updates.

Many teams had not updated their software and were having issues that had been taken care of earlier in the season.

There are no RF links that I am aware of. The Logitech controllers are hooked directly to the computer running the Field Management System using USB cables.

[PhilBot]

The "running on" problem is classic indicator of software that has not been updated. I sure hope that's what it was.

However, I've recently noticed a disturbing occurance on my NXT's that have the latest firmware installed. (1.21)

While running a program, the NXT will just lock up and stop responding to all it's buttons. There is no way to exit the program, or even turn off the brick.
The only way to reset the NXT is to either pull the battery or press the secret reset button.

It usually hapens after the NXT has been running a program for several minutes.

I first noticed this while I was running a plain vanilla (FLL) NXT-G program on a brick that had updated firmware. It happened three times in one evening, on two different NXT's. I went and re-installed the older 1.05 rev and it has not happened since.

I just had it happen to me on a third NXT while I was running some Labview code for FTC.

This is a SERIOUS problem if it really is hapenning.

Now that there have been several events using this firmware, as anyone else noticed a simmilar problem?

Phil.

[billw]

Jon,

I was not suggesting that there was an RF issue with the Joysticks--only that a problem in that part of the chain would still allow the Bluetooth connections to report that they were operational.

Phil,

Our NXTs run the firmware supplied with the 1.46 release of RobotC. If there is a later version, I am unaware of it. We have encountered a couple of anomalies with the encoders, but this may be a "learning" issue on our part. That said, we have never encountered a problem like the ones we experienced while on the field.

There is also (what I would consider to be) a weirdness in the RobotC development environment that requires you to keep the joystick debug window open; otherwise no joystick commands reach the robot--even though the program is running. Next time we are in the lab, I think we will try closing the window while a motor is running; this might duplicate the behavior seen on the field.

Thanks for your thoughts.

--Bill