Intelligence Beyond Conventional Flight.
Developing the intelligence layer for autonomous flight. Silvereye is advancing drone navigation, environmental perception and mission intelligence to enable more capable and adaptable autonomous systems.
- What we are
- A software company. We do not build drones.
- Initial focus
- GPS-resilient positioning and navigation
- Built on
- PX4, ROS 2, visual SLAM, visual-inertial odometry
- Stage
- Pre-commercial. Seeking development partners.
Autonomous flight needs more than autopilot.
Most drone operations rely on satellite positioning, and in open sky it works well. Many of the places where drones are most useful are not open sky.
Close to bridges, towers and buildings, satellite signals can be blocked or reflected. Indoors and under cover they may not be available at all. Interference can degrade them without warning. When position quality drops, a conventional autopilot has fewer good options, and many operations fall back to manual flying or are not attempted.
More capable autonomy needs three things: a position estimate that does not depend on a single source, awareness of the surrounding environment, and navigation that adapts when conditions change. It also needs to know how confident it is, and to tell the operator.
- Obstructed
Signals blocked
Under bridge decks, inside structures and in dense urban areas.
- Degraded
Signals unreliable
Reflections near large surfaces, and interference that reduces accuracy.
- Unavailable
No signal at all
Indoor and enclosed industrial spaces where GPS was never an option.
An intelligence layer for autonomous systems.
Our intended architecture is a modular AI Pilot that runs on compatible drone platforms alongside their existing flight controllers. It is organised into four capability areas. These describe our development direction, not products available today.
Navigation Intelligence
Estimating where the aircraft is and how it is moving, including when GPS is unavailable or unreliable.
- Position estimation when GPS is unavailable or unreliable
- Visual-inertial navigation
- Environmental mapping and spatial awareness
- Position tracking and drift management
Environmental Perception
Turning camera data into an understanding of the surroundings the aircraft can act on.
- Computer vision
- Environmental understanding
- Obstacle awareness
- Recognition of relevant environmental features
Mission Intelligence
Interpreting a mission and choosing how to fly it as conditions change.
- Mission interpretation
- Route planning
- Navigation decisions
- Adaptation to changing environmental conditions
Safety and Reliability
Knowing how much to trust the system at any moment, and keeping a human in charge.
- Confidence monitoring
- Fault detection
- System health monitoring
- Human supervision and intervention
- Safe operating boundaries
Building towards GPS-resilient autonomy.
Our first engineering work is on positioning and navigation using visual and inertial sensor information. The aim is to help compatible drones keep a reliable estimate of their movement and position when GPS is unreliable.
A camera sees how the surroundings move between frames. An inertial measurement unit senses acceleration and rotation many times a second. Combined, they can track how the aircraft is moving without any outside signal. This is visual-inertial odometry.
The estimate still drifts slowly, so it has to be corrected and its confidence reported. When GPS is good, it can be one of those corrections. When it is not, the aircraft should keep a usable estimate and the operator should know how far to trust it.
The first milestone is to investigate and demonstrate this in simulation, using existing open-source robotics infrastructure such as PX4 and ROS 2.
- First milestoneSimulationGPS-resilient positioning demonstrated in simulated environments.
- ThenRecorded dataEvaluation against real sensor recordings.
- ThenHardware in the loopRunning on representative on-board compute.
- ThenPartner flight trialsControlled testing with partner platforms.
- 01 · Sense
Sensors on the aircraft
Camera, inertial measurement unit (IMU) and barometer. GNSS is used when it is available and passes checks.
- 02 · Perceive
Environmental perception
Stable visual features are found in each camera frame and tracked as the aircraft moves.
- 03 · Estimate
Position estimation
Tracked features and inertial data are fused into an estimate of position and velocity, together with a confidence value.
- 04 · Navigate
Navigation
The estimate is passed to the flight controller, which keeps flying the aircraft as it does today.
Feedback Confidence monitoring decides how much to trust each source, flags drift, and tells the operator when the estimate is weakening.
Silvereye software. Unshaded stages are existing hardware and flight-control systems.
Where resilient navigation could matter.
These are potential applications we are exploring with industry, not existing deployments.
Infrastructure inspection
Bridges, viaducts, tunnels and dense urban structures.
Energy and utilities
Transmission lines, substations, wind and solar assets.
Industrial operations
Plants, storage facilities, warehouses and semi-enclosed sites.
Commercial drone systems
Manufacturers and integrators building the next platforms.
Help shape the next generation of autonomous flight.
We want to understand where positioning and navigation limit real operations, find the technical gaps, and build development partnerships around them. We would like to hear from:
Start a Conversation- Drone manufacturers
- Commercial operators
- Robotics engineers
- Technology partners
- Organisations investigating GPS-resilient operations