hi etienne, kelsey,
nice work and the surface buoy looks really cool!
We have done a number of test remote piloting BR2 over 4G a couple of years ago and it worked decently. No surprise, but the most important was to have a decent and stable 4G connections up and down on the ROV side (say >3mbps on internet speed test with 1mbps video bitrate). it was nice to have a lightweight setup on the ROV side with the ROV directly plugged into a modem.
One issue was that we didn t manage to have the controls both locally and remotely in parallel if i remember correctly (would make contingency take over from a local surface computer easier).
One other thing was that QGC was setup to render the images with 0 buffer and no way to tune that (at the time at least), so that often turned quite bad (green images and tearing), but VLC with minimal buffer worked well. I see that you’re using your own control software so you’ve probably sorted this out already!
We also achieved slightly less than 1s latency with direct rtp streaming. so very doable to fly remotely provided you could keep it slow and steady. but if the ROV was getting snagged, feeling surface waves or anything generating fast motion, it would quickly become uncontrollable…
Best of luck with your next tests!