2023-03-23 IMACS Sync
Brief
This is the first IMACS sync meeting in preparing for competition in early May 2023. The goal is to synchronize EFS and Autonomy on what needs to be done with regards to the different aspects of the ground station and ensure that everyone has action items/is on the same page
Â
Attendance
Â
Major Points
Updates from ZP Sync
Not flying ZP or Ground Station at Competition
GUI progress should continue but for purposes of competition, everything is being done through mission planner
Antenna towers, enclosures for laptops, etc. are still being used at competition
No use of Jetson either
QR Code scanning and waypoint sequence generation needs to be done on laptop that runs Mission Planner, and should be loaded directly onto Mission Planner from that laptop
Â
Discussion/Next Steps
Tasks
QR Code scanning module
Waypoint generation and parsing from QR Code
Encoding into Mission Planner format
General Discussion:
Mission planner takes in a file on a disk and uses that for waypoint generation - usage of desktop for that was discussed but was deemed not required
Task 1 - Done live, on the spot
Task 2 - All 50 routes need to be optimized within one hour
Â
Action Items and Deadlines
Task 1: Ready to fly by April 1st @Aaron Wang
Includes QR Code parsing and waypoint generation on Ground Station laptop
Encoding and loading onto Mission Planner (Mission planner mission)
2 QR Codes need to be scanned:
Press start, scan first code, stop the QR Code application
Start application again for second set of QR codes, submit, and start next set of tasks
Task 2: Ready to fly by April 8th @Aaron Wang
Include QR Code and waypoint sequence generation on Ground Station laptop
Ardupilot Mission Planner website/documentation will have information on encoding into Mission Planner
Identify Mission Planner system better - ICARUS Sync @Dhruv Upadhyay :
Investigate whether there is an API/more effective mechanism to upload missions
Uploading missions when the drone is already flying
Best done on Houston
Use last year’s QR Code scanning mechanism - outputs string that needs to be parsed
Ensure text output from QR Code is able to be sent to Task 1/Task 2 parsing system
Work on parsing the outputted text into something that can be optimized for Task 1/2
Landing Pad Selection
WARG laptop not powerful enough to run inference in realtime
Use analog video signal with multiple receivers to split video feed
Pipe one signal to USB port of Jetson:
Lower frame rate
Perform Jetson inference, draw bounding boxes, and display on monitor
Additional Hardware - VRX, Jetson, Monitor, power supply for Jetson and Monitor
Power monitor with inverter, Jetson with battery:
Can also use power supply at competition flight line
Updates for Landing Pad
April 2nd Flight Test - Collect data from the video transmission that will be used to run Landing Pad inference at competition.
The data will include both blue and orange landing pads and the video will contain the grass, field, and asphalt
In the video feed it is expected to see the OSD and analog breakout. Having the Landing Pad model train on this will make it more prepared for what it can see in competition.
the video feed should contain images of the landing pad when the drone is at different elevations
Â
Â
Â