I have some questions about RC / Mavlink failovers.
My experiments show that if I'm using radio communications, Mavlink can over-ride those commands.
For example: If I ARM the craft using radio, I can DISARM the craft using Mavlink and vice-versa.
If I set a MODE via radio, I can change the MODE via Mavlink. If (for example) the MODE is set to AUTO in Mavlink, and the radio is set to AUTO, I can switch out of AUTO using the radio if I first switch to any other mode except AUTO and then switch to the mode (on the radio) that I want.
My goal is to accomplish object avoidance in AUTO mode, something that I don't think APM code can do.
My idea is that I would fly using Mavlink (only), but when an object was detected, I would feed a PPM signal into the controller to change the MODE to STABILIZE, then I would give commands via PPM to do a hard reverse followed by a loiter, followed by a sideways movement, then followed by a slow forward .... and in the end, switch back to AUTO mode and continue on.
My reasoning is that by sending commands via PPM, I can avoid the 1 second or so delay inherent in Mavlink.
I would generate the PPM signals and interpret the object sensors using a TEENSY 3.2 board.
Does anyone think this method will work?
Replies