2024-08-30 Houston Atrium Cleanup
Introduction
Follow up to 2024-08-28 Houston Atrium Cleaning
Identical safety plan
Same flight time, 9pm to 10pm, but on 2024-08-30
Attendees
@Hardy Yu
@Daniel Puratich
Online Support throughout the day
@Nathan Green
@Smile Khatri
@Yuchen Lin
Improvements/Changes
Better Pixhawk Mount
Pixhawk is now solidly bolted to the frame on a 3d printed adapter instead of sticky taped
Optical Flow Sensor Mounted
Hoping this would improve GPS denied position hold stability
Notch Filter, Quicktune, & Autotune completed
Hoping this would improve stability
Attempted to mount Rangefinder
We did not succeed in this.
The rangefinder we have seems to be RS-485 only and doesn’t support CAN
We have a new one coming intended for another aircraft we may be able to use in the future once it arrives.
The rangefinder was not mounted for the indoor flights, we just played with in the bay for the past two days.
Debrief
Media
Videos and Photos: 2024-08-30 Houston Field & E7
Pixhawk Ardupilot Logs
Tuning Part of Flight Test: 2024-08-30 Tuning
E7 Atrium Cleaning: 2024-08-30 E7 Atrium
Some smaller media and conversations were posted in discord thread here.
We took some more photos of our outdoor test on the nice camera which we didn’t upload yet. All indoor media was captured on cell phones.
Flights
Tuning Outdoor Flight
takeoff flight test with flying in stabilize, loiter mode
Proceed optical flow sensor calibration
Proceed quick tune, and succeed quickly
Did autotune, each axis at a time. Each axis cost us one battery
Autotune aggressiveness is set to 0.1 (strong)
The batteries we have right now are pretty old, new ones are coming in
Tried flying in FlowHold flight mode without GPS
Indoor Flight 1
notch filer set to in-flight FFT
all indoor flight test is done in flowHold
Indoor Flight 2
notch filer set to in-flight FFT
Indoor Flight 3
notch filter set to static
Indoor Flight 4
notch filter set to throttle
Wins
Started up top and was able to make some tangible progress at cleaning.
Got more pilot practice.
Did not crash.
Issues
Oscillation Issue
Drone oscillation was visually worse than the previous indoor test
See media videos as evidence
Possible Causes
Autotune aggressiveness being set too high resulted in a poor tune. We really just want a very stable tune (lots of phase margin in all axes) so low aggressiveness may be best.
IMU Drift
Same issue as last time
See videos to show how offset it was when pilot landed or let go of the sticks
Unsure if this is solvable without scanning Lidar
Flowhold flight mode seemingly did not help
Pilot experience is helping a bunch here
Cornered & Large Paper Airplanes
We were not able to blow the larger airplanes
Videos show this but airplanes trapped in the corners were tough to get
A piece of debris has a heavy weight with a long streamer
there is a large concern of getting it in the propeller and we should continue to be careful
Possible solutions
Increase the weight of aircraft
increases the downforce the drone produces
would require retuning (though we probably need to do this anyways)
there are many other more complicated ideas floating around