2024-06-06 Autonomy Director Sync

  • Attendance

    • @Daniel Puratich

    • @Nathan Green

    • @R D

    • @Tong Zhang

  • Small fixed wing no autonomy stuff

    • Pixhawk and MP

    • trying to avoid using RFDs and LTE if we can get it to work

    • 5.8GHz VTX

  • Pegasus Path Forward

    • locking in UART with RPi 5, minimal comms to ground

    • Pegasus sync next week for this

  • Lots of new members

    • slow getting started

    • lots of new members for summer

    • have plenty of tasks available

  • Testing auto-landing by detecting a pad and landing on it with the model

    • model hasn't been trained yet

      • we have data and pipeline, train on the desktop

      • before midterms have a model ready

    • goal is test end of term

    • lots of integration

    • modules need to be integrated to main system

      • Geolocation

        • biggest and most complicated

        • 2 months integration effort so far

        • just a few people on it, tough to have people work together

        • discuss in subteam meeting

        • needs ground testing to prove then flight test

        • 3 weeks from now completion?

        • has never worked before

        • would be nice to be able to run this on the ground to hold Houston up to things

      • cluster estimation

        • can’t integrate it until geolocation is tested

        • 1 weeks to complete after geolocation is done

        • there is integration tests but they’re simulated

      • decision

        • module is roughly 4 weeks out from completion (this is unblocked right now)

        • 2 weeks

    • everything is getting tested on Pi with “200$ CV camera”

      • we collected data with this so need to use it for next competition cycle

      • we only have one (potentially another is around)

        • Can we buy a few of these and cables?

          • Nathan identified the cable here

          • We have one Pi 5 and a few Pi 4, SD Cards

          • MEF for these

    • plan is to use Houston for this testing

  • Op Flow Project

    • give to EFS if we did it bc devving on STM32

    • would not be a

  • Logging

    • especially if we’re running airside we want logs

    • LTE to stream logs to help with debugging

    • could SSH into Pi?

      • wifi access point on ground could connect to the Pi’s wifi chip to SSH

      • SSH over ELRS airport

    • right now we have logging to the Pi itself, if we SSH in we can read these

      • exists but not implemented for all modules

  • Socials?

    • not a lot of autonomy people in the bay

      • maybe wait to fall?

    • introductory explanation for architecture and project for new members?

      • updated slide deck in onedrive

      • update onboarding doc?

      • workshop code review?

  • Other Projects

    • IMACS

      • people are getting used to the code base

      • Recreating MP features: map, upload commands,

      • log processing?

    • Pathing

      • Aaron is on this

      • approach right now is not live,

      • working on qr functionality

        • it can be scanned, but it can’t come up with the path yet

      • right now Pathing is intended to run groundside only

      • pathing runs groundside

        • can update waypoints slowly so groundside is ok

    • LTE

      • ongoing

      • streaming images

      • streaming logs is important

      • airside system doesn’t know to work well with ground station, we can fix this

      • guided mode ← might be used already

        • there’s a lot of diff modes

    • Obstacle Avoidance

      • isolated project

      • co-op working on it

  • Architecture