A family of autonomous field platforms spanning ground-level MOOP detection, lunar-analog research, and full-size land stewardship.
The Autonomous Field Rover (AFR) is designed to operate reliably in dust, wind, temperature swings, and remote locations—conditions typical of the San Luis Valley and similar high-desert environments.
GPS waypoint following, SLAM-based mapping, and obstacle avoidance across unstructured terrain.
Temperature, humidity, barometric pressure, VOC, and soil moisture data collection.
Stereo vision and LiDAR-based 3D terrain reconstruction for geological survey.
LiFePO4 batteries with optional solar charging for 4–10 hour autonomous missions.
The Castalia rover program uses a shared autonomy and data model across three physically distinct platforms. Each vehicle is sized for a different layer of field work rather than treated as a replacement for the others.
A compact iPhone/ARKit rover developed first to scan desert ground for Matter Out Of Place.
The current AFR is a compact research vehicle for harsh-terrain navigation and lunar-analog field science.
A larger platform converted from an RC kids’ ride-on toy for useful work on gardens, land and burn sites.
The GalaxyRVR Mars Rover Kit serves as a prototyping reference—a rocker-bogie suspension system, ESP32 camera, solar charging, and Arduino-based control in a compact, terrain-ready package.
Multi-band RTK-capable receiver delivering ~2 cm positional accuracy for precise waypoint navigation and geo-referenced data collection.
Bosch BNO085 or ICM-20948 for orientation and dead reckoning between GPS fixes.
OAK-D or ZED Mini for real-time depth mapping and obstacle detection.
Sony IMX sensor with M12 lens for terrain classification and navigation AI.
Temperature, humidity, barometric pressure, and VOC gas sensing in a single module.
Wi-Fi (100–300 m primary), LoRa or LTE for long-range, optional Starlink base station. Telemetry includes GPS, battery, speed, sensor data, and system health.
Direct teleoperation via Wi-Fi. Operator drives the rover remotely with live FPV camera feed and full sensor telemetry overlay.
Human-directed with autonomous obstacle avoidance. The rover overrides commands that would cause collisions while maintaining operator intent.
Mission-based navigation: waypoint following, autonomous mapping, and survey task execution with full SLAM and path planning.
ROS2 Robotics framework Nav2 Navigation stack RTAB-Map Visual SLAM ORB-SLAM Feature-based SLAM Ubuntu Operating system ESP-IDF MCU firmware Geo-tagged stereo and monocular photography for terrain classification and visual survey.
3D point clouds and elevation models from stereo reconstruction.
Time-series measurements of temperature, humidity, pressure, VOC, and soil moisture.
RTK-precision trajectory logging for mission replay and coverage analysis.
| Component | Estimated Cost |
|---|---|
| Chassis | $300 |
| Motors | $300 |
| Wheels | $150 |
| Jetson / SBC | $300 |
| Sensors | $400 |
| Battery | $250 |
| Misc Electronics | $200 |
| Estimated Total | $1,800 – $2,500 |
Manipulator for sample collection and instrument placement.
Automated core sampling for geological and biological analysis.
Aerial survey companion with autonomous launch and recovery.
Deployable weather station for long-term monitoring.
Multi-rover coordination for large-area survey missions.