Skyfield.
Autonomous air traffic control. AI-native. Mission-tested. Deployed in the airspaces the current system cannot serve.
Built for autonomous operation.
Skyfield's core capabilities are the foundation for autonomous airspace management at production scale.
Autonomous separation
Skyfield maintains aircraft separation across managed airspace using deep reinforcement learning architectures. Conflict detection accuracy exceeds 99.97% in simulation across complex multi-aircraft scenarios — performance benchmarked above human controller capability at equivalent workload.
Voice and language
Skyfield understands standard ICAO English pilot-controller communications at production word error rates. Readback verification is automated. Flight strip generation is automated. Natural language coordination with human pilots operates without controller intervention.
Mission flexibility
Skyfield is configured per deployment context. UAM corridor management, regional airspace, BVLOS unmanned operations — the underlying architecture is consistent. The deployment configuration is mission-specific.
Safety-case-first
Every Skyfield deployment is validated against millions of simulated scenarios before live operation. Safety documentation is engineered to exceed current regulatory standards rather than meet them. Skyfield is built to certify, not just to demonstrate.
Built on open infrastructure. Engineered to certify.
Skyfield is engineered against the same open data and simulation infrastructure that defines modern ATC research. OpenSky Network for ADS-B. BlueSky for simulation. ATCO2 for voice models. The technical foundation is open. The deployment capability is what Aethon builds.
Figure — Open data and infrastructure stack underlying Skyfield's R&D pipeline.
From integration to operation.
A four-phase deployment sequence calibrated to operational confidence at each step.
Integration
Skyfield is configured for the target airspace. Local ADS-B feeds, communications channels, and operational procedures are integrated.
Simulation
Skyfield runs against millions of simulated scenarios specific to the deployment airspace. Safety case is built before any live operation.
Supervised
Skyfield operates in the target airspace under human supervisor oversight. Performance is benchmarked against operational standards. Confidence is built operationally.
Operational
Skyfield operates autonomously within the deployment scope. Continuous monitoring and safety case maintenance are persistent.
Aethon Skyfield is in active development.
Skyfield is engineered for first operational deployment with international civil aviation partners. Initial pilot deployments target regulatory sandboxes and emerging-market ANSPs actively pursuing autonomous capability. Production timeline aligns with EASA Level 3 autonomous AI deployment authorization in the early 2030s and parallel international frameworks.
Operational markets →