Actively loitering issue with rover 3.2

I think a big problem we also get on rover code is the speed estimation that isn’t very reliable at low speed

Then that might be one more reason to avoid using speed corrections but only position/heading related ones :slight_smile:

Dear rmackay9
may we ask you where we can find in the rover code the place where !near_wp and !_reached_destination are determined?
We would like to play around with something like hold position near a waypoint without actually trying to reach it (as you mentioned) and without periodic change between “driving towards destination” and “turning off the motors (mode_hold)”.
Thanks for any help:)

About sensing current-, wind-, and wave-forces some thoughts:
I did some reading about dynamic positioning for large vehicles. To have very accurate position hold we will need indeed some sort of 3 dimensional thrusters (surge (forward/astern), sway (starboard/port), yaw (rotation about heave axis)). But maybe it could be tolerable to have some kind of slowly forward-backward movement (like car-parking) to correct for sway-forces without the need of an extra thruster. Wave-forces can be very powerful and can come in from different directions and therefore we could foresee an extra EKF parameter were the Fourier coefficients representing the waves would be estimated out of the IMU data to be able to counter the periodic wave-forces (we can estimate the wave and wind behavior on our platform from the IMU data).