Back to browse
GitHub Repository

An open framework to simulate and deploy perception-based PX4/ArduPilot drone swarms with ROS2, YOLO, LiDAR, NVIDIA Jetson

464 starsC++

Aerial-autonomy-stack–Simulate and Deploy Perception-based Drones

by SufficientFix42·Feb 19, 2026·1 point·0 comments

AI Analysis

●●●BangerWizardrySolve My ProblemBold Bet

Sim-to-real drone stack with PX4/ArduPilot integration, real working proof.

Strengths
  • Faster-than-real-time Gymnasium simulation with YOLO + LiDAR closes research-to-deployment gap
  • Dual-subnet architecture (sensors vs comms) is clean design, not obvious
  • Actually deployed on real hardware (JetPack Orin); backed by arxiv paper with rigorous eval
Weaknesses
  • Requires NVIDIA GPU and Docker; high barrier to entry limits audience
  • Niche to research robotics; no commercial drone integration yet
Target Audience

Roboticists, drone researchers, multi-agent systems developers

Similar To

AirSim · Gazebo · Isaac Sim

Post Description

Sim2real is a recurrent challenge in robotics: despite the wide array of open-source solutions, end-to-end deployments often require painstakingly long integration of many different tools and frameworks.

This repo aims to provide a "batteries included" solution for flying multi-robot systems (in simulation and real-life).

Aerial autonomy stack (AAS) is an all-in-one software stack to:

- Develop multi-drone autonomy—with ROS2, PX4, and ArduPilot

- Simulate faster-than-real-time perception and control—with YOLO and 3D LiDAR

- Deploy in real drones—with JetPack and NVIDIA Orin

Runs on Ubuntu22/24, WSL. Requires Docker and an NVIDIA GPU.

Rationale and design choices: https://arxiv.org/abs/2602.07264

Happy to hear feedback/answer questions/provide clarifications.

Similar Projects

Developer Tools●●Solid

ARB – A Differentiable Rigid Body Simulation Library Using C++23

C++ rigid body simulation with end-to-end automatic differentiation for ML.

WizardryNiche Gem
wbyates777
214d ago
Developer Tools●●Solid

Vibe Deploy... Deploy full-stack apps to your own servers via AI

RunOS turns the AI from 'write the code' into 'ship the app' by running an MCP server that lets Claude Code request and provision PostgreSQL, Redis-compatible Valkey, MinIO and more, then build and deploy — all without Docker, Git, or local runtimes. It's a neat reduction of plumbing friction, but handing the deployment dialog to a model raises sensible questions around reproducibility, secrets, and audit trails.

WizardrySolve My Problem
didierbreedt
144mo ago
AI/ML●●Solid

Docker AI Stack, self-hosted LLM/STT/TTS/MCP in one compose file

One-command local AI stack bundling Ollama, Whisper, and MCP Gateway.

Ship ItCozy
hwdsl2
101mo ago