Inspiration
It seems like many hackathon participants wanted to do Theme 4: AI-Enhanced Telemetry and Sensing but balked at the lack of an "official" data set to hack on. Let's create a simulation (of high fidelity) to create those data sets!
What it does
High fidelity simulation of a 6 degrees of freedom (it moves, and it rotates) vehicle (satellite, rocket, drone) that also simulates the sensors on it.
How we built it
Built on top of a high-fidelity simulation written in C/C++ called "42" in github, and modifying it to capture the telemetry out of a simulation run.
Challenges we ran into
Long C code is hard to understand so I wrote unit tests for it.
Accomplishments that we're proud of
Getting the OpenGL graphics to work on it.
What we learned
Doing ML on sensor and telemetry data, essentially a numerical time sheet, doesn't lend itself to LLMs immediately.
What's next for Vehicle sensors and telemetry sim and ML
More understanding of the sim and explore more ML techniques we can apply onto it.
Log in or sign up for Devpost to join the conversation.