Enabling a 'high degree of autonomy' (hΓΆg grad av autonomi) for a 'safer society' (ett sΓ€krare samhΓ€lle).
This repository serves as a limited 'Source-Available' showcase intended strictly for research and education.
The Commercial Enterprise Edition of the Green Automated Platform (GAP) is a proprietary enterprise product developed by Corax CoLAB AB that includes proprietary features such as:
- Offline-First MANET Reconvergence
- 3D Volumetric Biomass Estimation
- Hardware-Accelerated NPU Pipelines (Hailo-8)
- Complete EUDR API Integration
Licensing Notice
This project utilizes a Dual-License structure:
- Proprietary Components: Core SLAM algorithms, Zero-Trust security handshakes, trained ML models (e.g., custom YOLO weights), and proprietary business logic are strictly protected under Copyright (c) 2026 Corax CoLAB AB.
- Public Interfaces: Documentation and public interface stubs available in this repository are licensed under a permissive MIT License.
Commercial entities, enterprise partners, and investors looking for licensing, deep-tech professional consulting services, or implementation are encouraged to direct commercial inquiries to info@coraxcolab.com.
Our 'Edge-First' decentralized architecture leverages real-time AI inference locally via Hailo-8L NPUs to maintain true offline autonomy.
Hardware Critical Note: To prevent system brownouts during intensive Hailo-8 NPU inference tasks, the Raspberry Pi 5 MUST use a dedicated 5V/5A BEC and have
usb_max_current_enable=1set in/boot/firmware/config.txt.
flowchart TD
subgraph Governance
EUDR
TRACES
end
subgraph GAP_Edge [GAP Edge-First Platform]
Drone[GAPdrone - UAV]
Bot[GAPbot - UGV]
Registry
end
subgraph Safety_Resilience
SAR
Fire[Firefighting Ad-hoc Mesh]
end
Operator((Forest Manager)) -- "Goal-Oriented Tasks" --> Registry
Registry -- "Dynamic Loading" --> Drone & Bot
Drone -- "6-Decimal Polygons" --> EUDR
Bot -- "MAOC Carbon Analysis" --> TRACES
Drone <--> Bot -- "MANET Mesh" --> Drone
The GAP Ecosystem Architecture: Highlighting the data flow from Edge Sensors to Web3 Audit Ledgers.
The React/Vite based Mission Control dashboard displaying live telemetry and 3D digital twin visualization.
The system utilizes sequential, high-speed pipelines. For example, the GAPdrone's internal pipeline flows: Camera β‘οΈ Hailo-8 NPU β‘οΈ ROS 2 Brain Node β‘οΈ MicroXRCE-DDS β‘οΈ Pixhawk 6C Flight Controller
The NLP pipeline translates human intents into structured JSON representations. The system converts these intents into actionable kinematic waypoints for the swarm coordinators.
{
"mission_id": "msn_alpha_092",
"priority": "high",
"agent_targets": ["drone_1", "drone_2", "hexapod_1"],
"objectives": [
{
"type": "scout_area",
"parameters": {
"bounding_box": {
"north_west": {"lat": 59.3293, "lon": 18.0686},
"south_east": {"lat": 59.3280, "lon": 18.0700}
},
"altitude_m": 35.0,
"search_pattern": "lawnmower"
}
}
]
}The GAP system operates in mission-critical environments. To ensure absolute reliability and safety, the system has successfully passed a comprehensive Master System Audit 2026. This audit verifies:
- Strict hardware-software parity across all deployed edge devices.
- Full compliance with the upcoming EU AI Act, ensuring transparent and accountable autonomous operations.
- Implementation of industry best practices for reproducible AI using DVC (Data Version Control).
- Adherence to our Inclusive Design Guidelines for 2026 HRI Standards.
Corax CoLAB is fundamentally committed to ethical innovation and rigorous regulatory compliance, specifically aligned with upcoming 2026 mandates like the EU Deforestation Regulation (EUDR).
- Horizon Europe & Vinnova Standards: We are fully compliant with Horizon Europe and Vinnova standards.
- Gender Equality Plan: View our formal Gender Equality Plan.
- Community Standards: Please review our CODE_OF_CONDUCT.md and CONTRIBUTING.md.
- First-Mile Traceability: Ensuring immutable polygon mapping for plots exceeding 4 hectares via Web3 Audit Ledgers.
Explore the technical showcases available in this public repository:
π gap_zero_trust
Showcases our approach to cyber-secure Edge AI and Zero-Trust architecture, crucial for protecting high-bandwidth sensor data and ensuring operational telemetry in heavy manufacturing (Industry 5.0).
πΊοΈ core_slam
Demonstrates the integration of advanced sensor fusion (3D-LiDAR and thermal imaging) for autonomous navigation and infrastructure inspections in GPS-denied environments.
π gapdrone_edge_ai
Highlights the airborne Edge AI unit used for ecological interventions (e.g., autonomous seed pod deployment), tactical deployment, and continuous swarm coordination over a B.A.T.M.A.N.-adv mesh network.




