search

CUEDOS / helix_framework

Cascade Drone Swarm Physical Demonstration Project
3 stars 1 forks source link

Figure out why the drone moved so slowly in offboard mode during flight test #27

Closed vcrann closed 2 years ago

vcrann commented 2 years ago

Drone didnt appear to move when simple flocking was activated however the simulated drones moved as expected. Have checked the flight logs and it appears that offboard mode failed to activate.

vcrann commented 2 years ago

Further investigation has shown a sleep statement between starting offboard and the first offboard command may be the issue. Also possiby a firmware issue as the simulted drones are on PX4 1.12 and the real pixhawk was on 1.11 so have updated firmware. Need to test at field

vcrann commented 2 years ago

Have lowered position and velocity update rates to 10Hz from 50Hz, current theory is that too many messages were being sent at once through MQTT. With one SITL drone this appears to generate the correct velocity commands, will test at field tomorrow.

vcrann commented 2 years ago

fix explored in previous comment did not work with accelleration limit at 1m/s^2 however when increasing the accelleration limit to 5m/s^2 simple flocking worked as expected!!!! Current theory is that having such a small accelleration limit produced initial velocity commands in the around or under 0.1m/s range which is probably in the same order of magnitude as the variability in the velocity and position data from px4. This explains why it worked in SITL where there is no variability in sensor readings.