[Radical Pi]

We had this exact same problem. The root of it is there is something with your robot that is causing WPILib errors at an extremely high rate, which slows down the code enough to cause the jags to miss their internal watchdog and freeze up (as far as I could tell). Our exact problem was our camera mount was constantly pressing the camera reset button, causing the cRIO's camera init to fail and fail again until the system just gave up.

Also, if any jags are messed up (reset IDs, no breakers, broken cable, etc) and you are sending commands to them, if enough are down then I've found the intermittent trouble and sluggish controls (another symptom btw) shows up too.

To the folks at WPI if you are reading this: Please cut down on your errors. At the very least, put a timer on them to limit the rate of errors being thrown. They seriously are a problem for anything that can cause an error on every single Teleop loop. In the time between cRIO finishing boot and camera finishing boot, we can still sometimes have this problem (NetConsole definitely shows the overflow of errors).