Horizon CÆLIS provides a fully visual navigation system for drones that operates without any GNSS signals, using offline maps stored directly on the device.
Funding
Funding not disclosed
Founders
Product
Problem
Most commercial drones rely on GNSS (GPS) signals for navigation and positioning. In conflict zones, urban canyons, or other GPS‑denied environments, satellite signals are jammed, spoofed, or simply unavailable, rendering conventional drones unable to operate safely or autonomously.
Solution
Horizon CÆLIS offers a fully visual navigation system that enables drones to localize and fly without any GNSS input. The platform stores high‑resolution aerial and satellite maps directly on the drone and uses on‑board AI models to recognize terrain features in real time. By matching observed visual cues to the offline maps, the system computes a position estimate with accuracy comparable to a standard GPS receiver. The visual‑inertial pipeline runs at a frequency sufficient for stable flight control, allowing autonomous missions in GPS‑denied or contested areas. All processing is performed on the device, eliminating the need for external connectivity.
Target Audience
Primary customers are defense and security agencies, as well as commercial operators that require reliable drone navigation in GPS‑denied or contested environments.
Features
- On‑board AI‑driven object and feature recognition that aligns live camera feed with pre‑loaded orthophoto maps
- Offline map storage on the drone, removing dependence on real‑time satellite links
- Visual‑inertial odometry loop delivering high‑rate pose updates for precise flight control
- GNSS‑independent positioning accuracy comparable to traditional GPS receivers
- Compatibility with existing drone autopilot hardware via a plug‑and‑play module
- Designed for operation in GPS‑jamming, spoofing, or urban canyon scenarios common in defense and security missions