I am currently working on an underwater ROV for the MateROV competition, using a 6-thruster vectored frame (4 horizontal at 45 degrees, 2 vertical) running ArduSub on a Navigator Flight Controller, interfaced through BlueOS and Cockpit.
When testing with an Xbox One S Wireless Controller, the ROV arms and disarms correctly, all button functions work as expected, and the axis inputs are being read properly in both Cockpit and QGroundControl. MAVLink MANUAL_CONTROL x/y/z/r values change correctly when the sticks are moved, confirming the controller input is reaching the flight controller.
Individual motor tests in BlueOS also work perfectly. All 6 motors respond correctly in their expected physical positions, with one vertical motor reversed as needed. FRAME_CONFIG is set to 0 in ArduSub parameters.
However, when actually piloting the ROV with the controller sticks, the movement has no discernible pattern relative to stick input. The problem persists across both Cockpit and QGroundControl, ruling out a software-specific issue.
Any guidance on what could cause correct motor test behavior and correct MAVLink input readings, but completely wrong and unpredictable thruster response during actual piloting would be greatly appreciated. We are on a tight timeline with regionals in 3 weeks, so any help is very much appreciated.
Hi @kingcoolance -
It sounds like you have everything setup correctly. Does the vehicle “go crazy” in stabilize or depth-hold modes? If so, the compass may not be calibrated correctly, or the AHRS_ORIENTATION has not been set to match how the autopilot is mounted.
Can you give more detail on what you mean by:
In air, if you are in manual mode and push fwd on the left stick (or however you have it mapped) do all 4 horizontal thrusters push air to towards the back of the vehicle? Do you have an equal number of CW and CCW props intalled?
When I am piloting with the xbox controller, in air or in water, there is no pattern that I can see to which movement of the joystick moves which. Like, if I push fwd, one of the fwd two turn on fully, the other one stays stagnant, and the back ones go slower than the fwd ones. We do have an equal number CW and CCW props installed, just like the tutorial on the Bluerobotics website
Hi @kingcoolance -
We’re still missing some critical troubleshooting information to assist you. What flight mode is the vehicle in when you test? Are the props all possible to turn easily by hand?
What is your AHRS_Orientation set to vs. how is the autopilot mounted in the vehicle?
The more context you can share, including pictures and your autopilot parameters file, the better chance we have to assist you!
The behavior you describe is normal for the 6 thruster standard BROV2 frame - the front propellers work harder than the front to minimize the pitching affect when starting from a stop, I believe… however if all 4 horizontal thrusters don’t respond, something is likely wrong with the one that isn’t!
1 Like