New flight mode

I need help. We have a tri-copter that transitions into forward flight, then transitions back to a tri-copter for landing. The problem arises when we want to land on a moving boat. The landing mode can’t keep up with the boat. How can I create a new mode where the motor-tilting mechanism/hover allows for more forward speed to complete the landing while allowing the left stick to control the altitude? I have experience modifying and compiling the code, and we’re using a Pixhawk cube.

Someone suggested changing Q_ANGLE_MAX but that affects the body angle. I need to stay close to flat in order to land.

That might be manual landing, unless you can do some new code with a Flow sensor or something…

EDIT, actually I’m sure Randy demonstrated Precision Loiter and IR Lock in Copter following a Rover - there might be scope there. And Precision Land might help.

I’ll ping @kd0aij and @tridge on this one.
This particular use case might still be problematic in quadplane even though it’s available for copter. There are ways to do it, but at the moment I think you’d need a little help from Lua or a companion computer.

Which mode is “landing mode”?
QLOITER supports tilt vectoring for forward flight; look at parameter Q_VFWD_GAIN

If you are manually controlling the landing, QLOITER should work well for a Convergence style tricopter.
I’ve also tried adding an RC channel to manually control forward tilt in QHOVER and QSTABILIZE, and that works well with the Convergence also. E-Flite mini-Convergence with ArduPilot

The PR for that has gone a bit stale though:
Let me know if you’d like to try it and I’ll rebase it.

We are using qstabilize for landing. Throttle controls climb/descend.

The forward tilt capability is not active in qstabilize, but is in qloiter. Check the wiki for documenation on that:

Our thought is to use another channel from the radio hooked to a knob. When the knob is turned, I would inject an offset into the motor tilt so that it would get a higher speed in the forward direction. Do you think this might work, or would the autopilot just crank it back to compensate?

I’ve been working on a PR for that. This is the latest version:

This is a flight test of an earlier version:

Testing would be appreciated

@cbrviking I just noticed your user name. Are you in Canberra?

No, Southern California.