2024-06-01 EFS Director Sync
Introduction
EFS leads and directors need to align on direction for the EFS team after 2024 comp debrief.
From @Hardy Yu they have tasks for onsiters it’s just the offsiters don’t have a lot.
From @Derek Tang “do we have any efs projects aside from can and zp? people want to join, but i don’t want to onboard them if there’s nothing for them to do”
Table of Contents
Attendance
@Daniel Puratich
@Derek Tang
@Hardy Yu
@Nathan Green
@Yuchen Lin
Content
Intro to https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2556100616
Current Projects
What Projects is the team working on right now?
CAN Stuff
https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555904119
https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2523856907
(@Camron Sabahi-Pourkashani leading EFS
@Jerry Tian leading EE
https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2496233926
@Michael Botros leading EE
EFS unasdigned
@Stanley Tang leading EFS
@Daniel Puratich @Meghan Dang leading EE
https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2540830770
@Parker Lawrence-Valeriani EE
EFS unasdigned
Each CAN project is a small team of two or three people
@Stanley Tang is looking into this
Don’t know what we don’t know here so more fun stuff to come
continue as is
Tracking Antenna Improvements
EFS support on https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555707425
@Roni Kant working on M2 https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2541453313
Roni wants more low level, driver dev, free rtos, zephyr too
Arduino doesn't provide low level
Make pitch tracking working
Adding in wireless communication to groundside
Nathan’s board vs STM board vs something else?
Is ardupilot support something we wanna do?
Ardu has code for this that works, but it doesn't have a bunch of extra advanced features we could add
We should just use it
Can use minipix for this?
Go with minipix currently then EE could do customer later on once we figure out exactly what fancy stuff we want.
Why do this?
The ardupilot stuff is already working
ardupilot stuff might work better when we have multiple towers and need to collaborate together in one ecosystem
A lot of tracking antenna stuff we can do we can do with arduino?
theoretically ardunio should work, but it doesnt
we could continue with it
Is this a worthwhile project?
SUAS is open air, no obstacles
Keep Roni on arduino thing
baseline works for ardu
will be easier to implement new stuff on arduino than on ardupilot
switching over to ardu may be faster short term, but not rlly in long run.
Bring Roni to flight test so he can see the challenges we’re facing with the system?
not necessary as leads are designing the architecture
no big rush to get tracking antenna working
tracking antenna base doesnt have anything to do with actual data
mavproxy over wifi
at end of M2 tracking antenna will have wifi support
wifi coprocessor
everything will be seperate for now
custom board later
right now not using any custom boards
We walked through the goals in https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2541453313 and agreed on soln
We scoped out https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2540830786 for EE
ZeroPilot M2/M3
EFS support on https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555838558
we’re havin free rtos problem with system manager (weird thread behavior)
Want to try to get M2 code completed before midterms, but unlikely to happen
https://uwarg-docs.atlassian.net/wiki/spaces/ZP/pages/2251620376
M3 is waitting on IMU but is more or less
wanna test in pieces M2 then M3
we updated https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555969565 with our objectives
ZeroPilot finishing at M2/M3 then being sunsetted
Plane for ZP M2
Getting mechanical involved in a design is optimal and preferred, we just need to write requirements and tell them what hardware we wanna use.
EFS members should write requirements and draw system diagrams
EFS support on https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555969565
After midterms S24 target M2 flight
Mid F24 go for M3 flight is the best timeline guess
so we’ll do this a month or two after CONOPS for AEAC comes out essentially
New Projects Previously Discussed
In Scope Projects we kind of already discussed all of these above
Tracking Antenna - https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2541453313
choosing a software arch and pushing it
need to give mechanical new requirements from lessons learned
already discussed above
Groundside Wifi System
needs architecture
kind of a subset of tracking antenna
supporting multiple devices with mavproxy?
Move away from personal hotspot?
lte modem instead would be more cost
why? no justification, moving on
CAN Servo - https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2523856907
CAN ESC - https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2524119043
code should be ready to go as soon as we have boards …
CAN Sensor Cluster - https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2496233926
get members the NDA
get @Hardy Yu
pixart device is used in the hereflow so its probably good
confirm what sensor models and protocols we will use
cc @Michael Botros to help lock this in early
common CAN circuit
New Project Ideas
New Project Ideas - focus on lessons from last comp cycle and in the debrief in choosing these ideas.
RTK GPS
preliminary testing currently
getting mixed results in terms of rtk fixes
need to figure out how good it is
developing comp system justified?
If it works do we want it? Do we want centimeter level precision
Is this feature neccessary?
we don’t have ardupilot rtk support?
We haven’t had much success with the current ones, better ones are ardupilot supported above 200 dollar range? There’s a reason why the one we have is 50 bucks
we tried it in an open field and got poor results
task was finding an RTK GPS for cheap that worked then integrate it into our system (ardu and tracking antenna).
unless there are any other ideas or evidence that it functions we should suneset this.
Gemini 433M/2.4G
plans and next steps for Gemini/ELRS system?
Fix our gemini board, run more testing
at flight test everything just died for no reason ….
https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2539915187 ← already planned improvements
get it working before we get craazy
Daniel’s thoughts discord chat
why do we need more power if no issues? we dont …
we should put antennas on opposite corners of drone on the EP1
https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2524086284 allows for four so we can space them out?
might be able to run multiple EP1/EP2 devices? need to look into this as a possibility
Gimbal System
base for final stage of tracking antenna
doesnt rlly provide any advantage and isn’t needed so don’t actively pursue essentially
a gimbal for a camera and a gimbal for an antenna can be essentially the same thing
pivot-able camera justified? (auto cv, manual piloting)
depends on next competition requirements
gimbal could help CV for scanning around?
depends on competition objectives
for a lot of things we don’t rlly need a moving gimbal
every comp in the past they’ve given GPS coordinates for pads so likely they do again
wait for CONOPS before pursuing this?
mech has a lot of stuff on their plate, don’t wanna burden with this as well
Houston Gimbal?
use as chase drone for recording?
we can solve this with a DJI
port to comp drone if needed?
mech loading makes this not really viable
prototype before we decide to go for it?
CAN Magnetometer - https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2551119891
there are good COTS ones
CAN LED Board - https://uwarg-docs.atlassian.net/wiki/spaces/EL/pages/2540830770
solid EFS project
low level development
designed specifically for our system
improve mav streaming (aka telem) over RC (aka elrs)
https://github.com/ExpressLRS/ExpressLRS/pull/2113
this one is higher baudrate than airport?
you can do mavlink over airport but its sketch/bad so the idea is this higher baudrate makes it usable
Why?
could use for other drones/planes
LTE hat / phone is expensive to fly on untested system
depending on the size of the drone the phone might be too big, elrs is small
could use as a backup
historically thsi kinda project hasn’t rlly stuck
whoever gets assigned to ELRS projects need to have used it so they understand the system, otherwise it will not go well
Is this a reasonable task?
Not for Roni bc leads don’t rlly have knowledhe
it is reasonable todo for us but want someone to understand system before going for it
improve ardupilot logging?
helps us just extremely indirectly
is AP_LOGGING too far out?
no end goal for this so not worth it
ROS Stuff
lots of teams were doing this at competition
what does doing this giving us?
unknown
helps for integration between subsystems of the drone
doing mavlink commands through ROS
Adding support for new type of frame to ardupilot
The weird folding wing things aren’t supported in ardupilot right now
we could add it into ardupilot and is really low level if mech commits
wait for mech to commit to what airframe they want. we should get mech to commit to VTOL type this term S24 then have EFS get started.
In the past to get the folding thing to work they added in a mixing system for the transition
Knowledge Transfers
Knowledge Issues
not a lot of senior members understand elrs and ardu so it’s hard for people to assign specific tasks
people will need a lot of mentorship to understand it
Solutions?
good documentation and workshops, get people participating in workshops
We will only be able to provide good mentorship if tasks are really simple?
we just gotta go for it? senior members can learn and go for it as long as we have a good direction
lot of senior members are busy with managing
right now ZP telem manager requires a lot of hardy and derek attention
only PM on EFS is for TM
what is role of directors?
Knowledge Transfer Topics
Flight school plan?
Getting EFS members comfortable with the whole system is important
EFS paper foam workshop?
I, @Daniel Puratich, think this should be handled somewhat by mech.
Get EFS members comfortable setting up ArduPilot from scratch
suggested by @Nathan Green
this is really useful for comp ← wanna get more people useful at competition
have people configure and fly stuff
do the second houston https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555969579 with EFS people
we need more people who know ardupilot
we can do a lot of software stuff but this is the stuff that matters
Get people comfortable reading and processing log files from flights
we should encourage EFS members to pull useful data from logs and help develop debriefs to help mech understand the flight better
we could have autonomy do scripting to proccess the logs?
automatically produce more readable forms of the data for use in reports and debriefs?
tools to automatically read through the log and find anomolies?
Logs are more important when we have massive problems
thsi is kinda fine as is
@Daniel Puratich
Workshops
Main Workshops
Create new https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2253815900with fixes described in https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555969565
Create new Houston https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2555969579
Flight School https://uwarg-docs.atlassian.net/wiki/spaces/ARCHS22/pages/2167440108 (better resource in onedrive made by Anni)
Target Audience
everyone ?!
In the future core members might not be available, need to pass on the knowledge.
Philosophy
We should have people understand the basics of the systems we are using (ardupilot, elrs, tracking ant, etc.) before people can start to improve and develop the more advanced features and functionality.
Need to pass on the knowledge for future generations of WARGians.
More flight test coordinators
@Hardy Yu to onboard two new flight test coordinators bc theyre interested in aviation and are interested in controlling planes and drones
may be brought up for next leads.
More Project Ideas
We didn’t get any new project ideas from this ….
embedded software side is one of hardest subteams to figure stuff out for here
We don’t have tasks for people who are offsite right now.
It is really difficult to come up with projects that are easy enough to be doable quickly but haven’t already been solved before.
possible solution paths
do things that are difficult but will be useful if we can pull them off
challenge is the we have low odds of success
this will always be a challenge
continue ZP milestones as something to do and recognize they will never be competition viable?
ZeroPilot has been in development for more than three years and not achieved fixed wing fly by wire A so like if we do things that are difficult the odds we pull them off are low
ZP has two problems: it’s rlly difficult and it’s already a solved problem.
Project ideas
Do FOC algorithm position control for BLDC instead of servos to actuate wings or rotors in a VTOL
essentially implement stuff like https://discord.com/channels/776618956638388305/776618957138034690/1246184979743113236
Color OSD
develop firmware that supports custom OSD for analog video
would be very difficult but doesnt exist nicely COTS
there are attempts but none of them work well
this would be cool, neccessity is kinda weak
would be nicer looking
Controller Lua Scripts
could write our own scripts to do things we want
not rlly low level