2024-07-02 Autonomy Director Sync
- Daniel Puratich
Intro
Past Sync2024-06-06 Autonomy Director Sync
what progress has been made in the last month?
airside lots of progress
geolocation testing in progress, seems to be working, FT soon
cluster estimation making progress
decision module, unassigned
what do you and your team want to achieve in the next month?
geolocation testing and integration completed
cluster estimation ground test
decision module ready (not integrated yet)
what roadblocks do you need help clearing to make this happen?
no big roadblocks
need to mount CV camera to Houston
RPi 5 mounted last time
Testing auto-landing by detecting a pad and landing on it with the model
This is still target for end of term
This requires cluster estimation, geolocation, & decision module
Autonomy socials, wait till fall term
Sitedness
Tong & Aaron not onsite next term
Maxwell will be onsite
MEF
items are in, need project descriptions
try to bring other members, there seems to be interest in this
Geolocation
Geolocation is used for landing
The pad landing is relative to the home position
Position of pad relative to the drone
Might need to change architecture coordinate system wise?
We want to be able to land on a pad end goal
we have an MVP script that does this
Cluster Estimation
geolocaiton gives points in a scatterplot
cluster estimation determines which points are a single pad
integration is required
not required to test geolocation
ground test end of month
Decision Module
logic to make the autoland task happen
IMACS
in middle of big refactor
works with missionplanner
trying to get working with serial
showcase-able thing for end of term
Pathing
pathing code uses dronekit right now, we can’t have two instances at once connecting
auto-landing and pathing can share dronekit instance
lots of refactoring to make this happen
trying to implement 2023 pathing requirement with the current repo
all pathing repo does it send locations and land command
why not combine auto-landing and pathing into a single python app?
that’s the goal of this refactor
RPi failsafe mechanism
software mechanism in addition to the hardware?
set variable in pixhawk could be implemented?