Rookie team here. Have not competed yet, but have been watching matches. We are surprised by how much trouble people are having during sandstorm.
It is making me wonder if the latency we experience with our camera when practicing is comparable to the latency experienced on a real field. If we set up our radio correctly with the configuration utility, can we expect similar results at competition with FMS?
We installed and configured our radio after bag day (before that we were tethered). We have been practicing with a practice chassis and a manipulator we removed as our withholding allowance. We did select bandwidth limiting. We understand that driving with a camera is not easy, but with a little practice our driver seems decent (at least better than most of what we’ve seen from other teams on twitch). It made us wonder if there was something we were missing.
If your radio is really configured with the correct bandwidth (and optionally, firewall), then you can reasonably hope for similar conditions at events, but be prepared to simply be unlucky with the FMS. (Nobody likes it when the FMS breaks, but it happens.)
All the events so far have been weeks one and two. Weeks one and two are always underwhelming. Just how it is.
Looking at your team’s TBA page, you have until week four to answer these kinds of questions:
Are you using lots of bandwidth on your stream? If so, can you lower the framerate/resolution (or use something beyond CameraServer, shameless plug, to cut bandwidth?)
Can your drivers reliably cycle with only a camera on a practice field with everything in correct positions?
Can you just do it autonomously and bypass potential issues with cameras altogether?
We competed week 2 and there were many teams (us included) having camera issues. Cameras would show up during systems tests in the pits just like they did on wifi while practicing, but not show up when connected to FMS for a variety of reasons (take advantage of practice matches, if for no other reason than to verify everything is working when connected to FMS the same as when you were practicing).
So you are seeing teams drive without cameras (or not moving because they can’t see) as well as camera latency and field of view (if your cameras are not positioned in an optimal spot(s), the driver may need to turn to try to get their bearings – particularly since now there are multiple robots as opposed to during practice you are often by your self, so you don’t have to worry about driving into someone). I know we’ve relocated cameras to try to improve the driver’s view of the robot/field.
Also, we’ve seen through the years, the first time a driver is out in front of a crowd they tend to not drive as well as after they’ve done it a few times. One year we had a new driver forget what was the front/back of the robot and drive straight into the alliance wall at full speed when they intended to drive away from the alliance wall.
We competed week 2 with our dashboard and camera well below the bandwidth limit and found that the FMS often induced almost 1s of lag time in the camera feed. It was very difficult for our drivers to try and be effective during the sandstorm period since the programmers relied solely on the camera feed instead of adding some autonomous assistance.
Other teams may observe this same camera lag or they may not. It depends on your camera setup, the field, and the arena you’re in (WiFi interference at your event can affect your bandwidth). In any case, I recommend planning for some field latency by adding some autonomous assistance or driver training if you want to be more effective during the Sandstorm period.