A long-running personal project that combines custom hardware with a large amount of software I am writing myself. It is early, nothing has been proven on real hardware yet, and I am not writing it up in detail until it is.
PROJECT AT A GLANCE
The stage matters more than the specifics right now. Most of the work so far is software written ahead of the hardware it will eventually run on.
01 / Context
Argus is the project I work on when nothing else is due. It is personal rather than commercial, and it exists because I wanted one build that forced me to hold electronics, embedded software, and perception work in my head at the same time instead of treating them as separate hobbies.
02 / Approach
I am building the software first and buying hardware second. That order is deliberate: it keeps the expensive commitments late and lets the architecture settle while it is still cheap to change.
How that architecture actually works is the part I am keeping to myself until it runs.
03 / Result
A substantial and actively growing codebase, a set of design documents, and an honest status table that currently says almost nothing has been confirmed on physical hardware. Most of the parts have not been purchased.
There is no public repository and no public site, and I am not claiming results I have not measured.
04 / Reflection
The main lesson so far is about honesty in my own notes. Keeping a status table that plainly separates "written" from "verified on hardware" has stopped me several times from believing a subsystem was finished because the code compiled.
Next step is the boring one: buy the first real parts and find out how much of this survives.