73 lines
3.0 KiB
Markdown
73 lines
3.0 KiB
Markdown
---
|
|
aliases:
|
|
Tags:
|
|
- daily
|
|
day_grade:
|
|
Dehnen:
|
|
Sport:
|
|
Ernährung:
|
|
---
|
|
# 2025-07-18
|
|
|
|
[[2025-07-17]] <--> [[2025-07-19]]
|
|
|
|
|
|
Error generating daily quote
|
|
|
|
---
|
|
## Planning
|
|
|
|
|
|
___
|
|
## Reflection
|
|
Ugurs feedback
|
|
- Not happy with software stack
|
|
- port configuration
|
|
- performance counter
|
|
- logging rates and logging system setup with multi-sink logging
|
|
- parameter system
|
|
- proper config file: everything in one place, don't pass config into the factory
|
|
- not good to do everything in the main loop --> RTOS
|
|
- he doesn't like the teensy platform --> RTOS
|
|
- unhappy with state machine: problems with kill switch, etc
|
|
- unhappy with control sequence: cannot separately enable/disable modules
|
|
- would like a common module interface to handle starting, stopping, etc
|
|
- no loose coupling --> not include `#include <arduino>`everywhere
|
|
- he would like to run entire stack on computer for testing and fast iteration
|
|
- unhappy with how rarely floris syncs in and then focuses on minute details (variable naming, etc) instead of the actual functionality
|
|
- Office is a mess
|
|
- better system
|
|
- cleaning
|
|
- he want's a baseline to build upon
|
|
- thinks we want to do everything ourselves instead of outsourcing (e.g. windtunnel, prop testing, manufacturing)
|
|
- hesitation about building a flight controller which can have permission by goverments to actually fly --> follow standards, etc?
|
|
- Doesn't think that arduino and teensy is the good platform for this--> hobbyist framework
|
|
- thinks the last year was to understand the control algorithm and learnings, but not for actually building important things
|
|
|
|
|
|
Omar
|
|
- has doubts about manufacturing: assembly and manufacturing of propulsion system is way too complicated --> outsourcing?
|
|
- Doesn't know what to prioritize: is not very good at having agency --> dannick needs to lead more.
|
|
|
|
Both are waiting for instructions too much. I think that is also because the feedback model that e.g. Floris uses punishes agency if it wen't slightly into a different direction that Floris intended --> they learn to just follow instructions and thus don't be self sufficient. This in combination with having no leader in Istanbul is a bad combo.
|
|
|
|
|
|
### Next steps
|
|
- [ ] fix ESC issues
|
|
- [ ] reconfigure
|
|
- [ ] test full throttle thrust (at least 2x the maximum takeoff weight, rather 3x)
|
|
- [ ] then redo propulsion characterisation and flap characterization for physics model
|
|
- [ ] if escs are bad --> find new ones.
|
|
|
|
___
|
|
## Notes
|
|
- Smarter [everyday](https://youtu.be/cRaKMTK7ea0?si=lNWswp3NHyqOeMJe)
|
|
#business-idea : a sensor for many things that extends our senses (example: radiation Digital Alarming Dosimeter (DAD) or Dosimeter of Legal Records (DLR), fungal sporren, etc). AR glasses, notification system
|
|
|
|
- Veritaseum: Math video: https://youtu.be/x32Zq-XvID4?si=xQmNjVa4lybMgK13: do what you love, bot what you think is best
|
|
-
|
|
|
|
- asimov AI: bet that understanding existing codebases is more important than generating new code —> for the simulator project? Understand the existing codebase to help build a better one?
|
|
- Firefly AI
|
|
|
|
- how does an RTOS really work? |