Parrot Minidrone - stable flight

I am using the yaw pitch roll configuration and I want to achieve stable flight. So I want to use the optical flow values in the x and y direction to achieve this as my drone keeps drifting. I think it works when i put feedback loops in for the roll direction. But when i use pitch, the drone in the simulation just flips out? What am i doing wrong or how else can keep my drone aligned to the track or the arena around it?

Answer 1

[Pranav Murali]

Hi Keith,

It’s a nice idea to use the optical flow values for a good, controlled flight. Here’s a couple of things you can try:

1. Tuning the PID controller of the pitch with different coefficients.
2. Process the values x and y of optical flow and provide feedback to the roll axis alone during position control. The pitch can be controlled during orientation control.

For a better understanding of the position and orientation control refer this:

https://www.mathworks.com/matlabcentral/answers/493866-position-or-orientation-control-parrot-mambo?s_tid=answers_rc1-2_p2_MLT

[Débora Oliveira]

That's a great idea Pranav.

I would recommend updating the optical flow computation. The last version I used was the 20a, but as for the release page, no consideration was done to the model in the newest version.

The equations represented in the block diagram is a partial approximation of the Lucas-Kanade method. I plotted the output signals and saw an oscillation on the velocity estimation value at axis Y. There may be a signal mistype at the term of angular axis X velocity.

I replaced with the optical flow equations of the Horn-Schunck method and got a much more stable drone in the simulation.