Show HN: Open-Source AI Racing Harness

ddanAtElodin about 15 hours ago 4 commentsRead Article on elodin.systems

RU version is available. Content is displayed in original English for accuracy.

Hi I'm Dan from Elodin, making an open source real-time capable flight software simulation.

For AI Grand Prix contestants, the wait for the Round 1 virtual qualifier simulation has been grueling.

If you’re competing, check out our simulation harness to tide you over, built to match the published competition constraints and message format. It runs against real Betaflight, which we learned requires at least 1000 sensor samples per second to run real-time correctly.

The competition warranted introducing a new feature to generate the camera sensor directly in the simulation loop. Typically people connect to Unreal or similar game engine to create a camera sensor, which works well but is very heavy. For the simple needs of this challenge, creating sample directly in the loop is very handy and easy to use. Happy to hear your feedback on this! While it's not fancy looking currently, it uses the Rust Bevy game engine, which should allow us to improve the visual fidelity quickly.

We all should easily be able to shift our implementation to the published competition sim once it lands. Hope you enjoy and good luck!

⚡ Community Insights

Discussion Sentiment

100% Positive

Analyzed from 131 words in the discussion.

Discussion (4 Comments)Read Original on HackerNews

xspad3s•about 13 hours ago

Does the starter kit include a simple baseline agent (like a basic line-follower) that I can run immediately to see how the simulator works?

danAtElodin•about 13 hours ago

the included baseline just demonstrates that the Betaflight controller will indeed move to the hard-coded center positions of each gate. So the basic first solver task would be to replace that with a CV or ML inference based approach. We plan to add another demo solver into the sample set that shows this next, currently in progress.

coin_artist•about 13 hours ago

Looks cool! But why not just use Gazebo? Or Issac?

danAtElodin•about 13 hours ago

Both great tools, optimized for slightly different needs. Gazebo is great for simulation and testing of higher level control, or for simulation that doesn't require 1000hz sensor simulation rate. Nvidia Isaac focuses on ML model training workflows, not ideal for interacting directly with your flight software for flight testing your code before you fly. This aims to be software CI/CD for your drone builds.