AgiPIX: Bridging Simulation and Reality in Indoor Aerial Inspection
Open-source drone platform achieves zero-shot sim-to-real transfer for indoor industrial inspection — no calibration tax.
The Thesis
AgiPIX is a fully open hardware-software drone stack that transfers autonomy policies from simulation to real flight without retraining, targeting industrial indoor inspection (think power plants, refineries, data centers). The killer feature is hardware-synchronized active sensing plus onboard GPU compute in a compact form factor — the combination that has historically forced teams to choose between agility and perception quality. If zero-shot sim-to-real actually holds at scale, this compresses the deployment timeline for inspection drones from months to days.
Catalyst
Photorealistic digital twins and containerized ROS 2 stacks have matured enough in 2024-2025 to make simulation environments faithful proxies for real sensor noise and lighting — the gap that historically killed sim-to-real transfer for vision-based indoor flight. Onboard GPU modules (Jetson-class) are now small and power-efficient enough to fit on sub-500g platforms.
What's New
Prior indoor inspection drone research (e.g., MIT Agilicious, ETH Zurich's Flightmare) either lacked open hardware designs or required significant manual tuning to bridge sim-to-real. AgiPIX ships a photorealistic digital twin of the exact physical platform, closing the fidelity gap that forced prior systems to retrain or hand-tune controllers for real deployment.
The Counter
Indoor inspection drones have been 'almost there' for a decade. Flyability already sells hardened, cage-protected drones to industrial clients, and their closed stack is a feature — certified, liability-managed, and supported. An open-source academic platform, no matter how elegant, faces an enormous gap between 'works in a lab demo' and 'approved for use inside an operating nuclear facility or oil refinery.' The zero-shot sim-to-real claim is the most aggressive assertion in the paper, and it is demonstrated in environments the authors designed — cherry-picking favorable conditions is endemic to this subfield. The real test is a third party dropping this platform into an unfamiliar site and achieving comparable results, which has not happened. Finally, the competitive moat for hardware platforms is brutally thin: any well-funded drone company can replicate a hardware design in six months; open-sourcing it helps academia more than it helps commercialization.
Longs
- NVDA
- FLIR
- HON
- BA
Shorts
- Flyability (Elios series — proprietary closed stack for indoor inspection faces open-source commoditization pressure)
- Traditional NDT inspection contractors who rely on human-piloted systems
Enablers (Picks & Shovels)
- NVIDIA Jetson (onboard GPU compute)
- ROS 2 ecosystem maintainers (Open Robotics)
- Matterport / NVIDIA Omniverse (photorealistic sim environments)
- FLIR (thermal active sensing)
Private Watchlist
- Skydio
- Exodigo
- Percepto
- Firestorm Labs
The Paper
Autonomous indoor flight for critical asset inspection presents fundamental challenges in perception, planning, control, and learning. Despite rapid progress, there is still a lack of a compact, active-sensing, open-source platform that is reproducible across simulation and real-world operation. To address this gap, we present Agipix, a co-designed open hardware and software platform for indoor aerial autonomy and critical asset inspection. Agipix features a compact, hardware-synchronized active-sensing platform with onboard GPU-accelerated compute that is capable of agile flight; a containerized ROS~2-based modular autonomy stack; and a photorealistic digital twin of the hardware platform together with a reliable UI. These elements enable rapid iteration via zero-shot transfer of containerized autonomy components between simulation and real flights. We demonstrate trajectory tracking and exploration performance using onboard sensing in industrial indoor environments. All hardware designs, simulation assets, and containerized software are released openly together with documentation.