Congestion/Fatigue after some time

[simonschmeisser]

In a current project we observe an incremental slowdown of the driver. While similar trajectories take around 1s with a freshly restarted controller, they will need up to 4s after prolonged use. This seems to be mostly due to an increasing delay before starting the execution, once the robots start the speed is as commanded.

I checked if there are offsets between the start of the trajectory and the current position but it is around 0.00015 typically, so the is_near should skip the first step

Any Ideas how to debug this further?

[simonschmeisser]

"prolonged" means after around 180 trajectories, each having around 10-14 sparse points

[gavanderhoorn]

I'm afraid I would not know.

At its core, the RAPID scripts basically manage a static array/buffer of points, which gets overwritten for every new trajectory.

I would not immediately know where such increasing delays / slowdowns could occur.

Have you had any luck debugging this in the meantime?

[simonschmeisser]

No, unfortunately not. It was just a demo for a customer but if we win the actual project I certainly will have to debug it.