Emerging
May 28, 20261
50%
Elodin Open-Sources AI Racing Harness for Autonomous Drone Competition

Elodin has open-sourced a practice racing harness for Anduril's AI Grand Prix, a $500K autonomous drone-racing competition, allowing developers to begin writing autopilot code before the official simulator launches. The harness integrates physics simulation, real flight control software, and sensor simulation into an accessible, open-source tool for both competition contestants and aerospace developers.





Quick Facts
Who
Elodin
What
Open-sourced practice racing harness for autonomous drone competition
When
2026-05-28
Where
GitHub
- Open-sourced practice racing harness for autonomous drone competition
- Integrated physics simulation with Betaflight flight controller
- UDP bridge synchronizing simulation components at 1 kHz
- GPU-rendered camera and multi-rate sensor simulation
- Elodin
Elodin, a simulation software company, has open-sourced a practice racing harness for Anduril's AI Grand Prix, a $500,000 autonomous drone-racing competition. The release, published on GitHub, allows contestants and developers to begin writing autopilot code before the official Virtual Qualifier 1 simulator becomes available. The open-source tool runs on macOS and Linux, with setup requiring only standard package management (uv sync) and a straightforward Betaflight build.
The harness represents the culmination of Elodin's multi-year effort to bring professional-grade simulation tooling to aerospace development. The company, which was founded after the founder's experience working on simulation systems for The Sims 4, identified a significant gap in how aerospace teams approach flight software development. Traditionally, teams cobble together MATLAB/Simulink, Gazebo, and custom Python harnesses—a fragile process that often requires weeks of setup. Elodin developed an integrated stack built on Rust, featuring a JIT-compiled physics engine (nox) with Python bindings, a 3D editor connected to a time-series telemetry database, and a process runner that coordinates multiple simulation components.
The AI Grand Prix practice rig integrates three key components: Elodin's physics simulation, which handles 6-degree-of-freedom rigid-body dynamics, motor behavior, drag, and multi-rate sensor simulation; Betaflight's real Software-in-the-Loop (SITL) build, which handles flight control logic at its native PID rate; and an 80-line UDP bridge that synchronizes both systems at 1 kHz. The simulated environment includes a GPU-rendered forward camera and realistic sensor suites (IMU, barometer, magnetometer) matching the official competition specification. Contestants write a single Python function—the autopilot controller—that receives sensor data and outputs RC commands, with no access to GPS, depth sensors, or motor RPM data, reflecting the official simulator's constraints.
Elodin decided to release the practice rig early because the official simulator's launch date had slipped, and the team wanted to enable competitors to begin development rather than wait passively. The stack is designed to be both powerful for professional teams and accessible to newcomers, reflecting lessons learned from working with drone, satellite, and other aerospace customers. The underlying Elodin engine and editor are available under the Apache 2.0 open-source license, extending the company's commitment to making advanced simulation tooling more accessible to the aerospace engineering community.
Topics
Why This Matters
This release democratizes access to professional-grade aerospace simulation tools, enabling a broader community of developers to participate in autonomous drone competition and innovation. By open-sourcing the practice harness early—ahead of the official simulator—Elodin reduces barriers to entry, allowing contestants to begin algorithm development immediately rather than waiting passively. For the aerospace industry more broadly, it demonstrates how integrated, modern simulation stacks can replace fragile cobbled-together workflows, potentially accelerating the development timeline for flight software across research institutions and commercial teams.
Timeline & Sources
May 28, 2026
WireElodin open-sources AI Grand Prix practice harness on GitHub