Compute at the edge
The current hardware direction uses Jetson Orin Nano compute with ROS 2 Humble, depth perception and modular sensing. The rover connects to the platform through the robot bridge.
RANGER / THE TECHNOLOGY
Intelligence that
moves with you.
A geospatial intelligence rover, connected from hardware to interface.
Explore the systems behind sensing, remote operation and spatial understanding.
THE CONNECTING PLATFORM
One platform. A view of the field.
RANGER connects rover observations, positions and sensor records in a geospatial interface. Position and reading provenance remain separate, so users can inspect where information came from.
Select a component to explore its role.
01 / FOLLOW AN OBSERVATION
One reading.
A connected journey.
A sensor value becomes useful when its location, time and source travel with it. Select each stage to follow the information through RANGER.
→→→
01 / ACQUIRE
Start with a measurement.
A payload observes its surroundings. Edge compute and ROS 2 connect the rover’s software with its hardware. The sensor’s integration and calibration determine what its reading means.
EXAMPLE RECORDACQUIRED
- Payload
- Temperature example
- Value
- 24.6 °C
- Reading source
- Simulated
Synthetic value for this explainer; not a field measurement.
Stage 1 of 4
THE ENGINEERING STACK
Built across
hardware and software.
A spatial backbone
Django and Django REST Framework serve the platform, with Channels for real-time communication. PostgreSQL/PostGIS stores spatial records, supported by Redis.
A view into the mission
React and TypeScript power the interface, with MapLibre for maps. Dashboard, data, visualisation, satellite and mission screens form the implemented application routes.
02 / SPATIAL CONTEXT + PROVENANCE
A point on a map.
A story behind it.
Switch layers and select an observation. A location, a reading and a source each answer a different question. A live position would not automatically make every attached reading live.
SCHEMATIC / NOT TO SCALE
Rover path and observations visible.
EXAMPLE / OBSERVATION A
An available reading.
- Reading
- 24.6 °C
- Position source
- Simulated
- Reading source
- Simulated
- Freshness
- Current in example
The value is present and its source is explicit. This entire record is synthetic, including its position.
LiveDirectly acquiredReportedSource-attributedModelledDerivedSimulatedSynthetic
RANGER V3 tracks position and individual reading provenance separately. The context layer above is illustrative, not satellite imagery. Scientific satellite matchups and uncertainty analysis remain planned capabilities.
ENGINEERING IN THE FIELD
Useful information.
Visible limits.
Missing and stale data
A missing sensor value should remain missing. An old position should appear stale. The interface distinguishes unavailable information from a valid zero.
Satellite data as supporting context
Sentinel-2 and Landsat retrieval and rendering are documented in V3. Scientific matchups and uncertainty analysis are planned extensions of the wider geospatial workflow.
Modular sensing
The platform distinguishes readings by source. Sensor integration, calibration and field checks determine how a payload’s records can be interpreted.
Axon: a research direction
The Axon Intent Engine is a concept for turning a human objective into a sequence of robotic tasks. Autonomous planning and execution require further development and validation.
BITS AND BYTES FOR QUALITY LIVES.