The Arvanos project and team
WHAT HAS ALREADY BEEN BUILT
ARVANOS IS NOT YET A FIELD-READY PRODUCT. BUT ITS TECHNOLOGICAL FOUNDATION ALREADY EXISTS IN INDIVIDUAL MODELS, SOFTWARE MODULES AND VALIDATED METHODS.
WHAT HAS ALREADY BEEN BUILT
CONCEPT AND SYSTEM ARCHITECTURE
The complete Arvanos concept and the interaction logic between Dynas Maya, Dynas Herus and the AI orchestration layer have been defined.
The architecture specifies how the system combines field history, remote observations, directed subsurface measurements and laboratory calibration. It describes how the initial map is refined and how new evidence is converted into the next field tasks.
The central principle is clear: laboratory samples should not be collected across a blanket grid, but where they will add the most information to the map.
WHAT HAS ALREADY BEEN BUILT
AI AND DATA ENGINEERING
An initial AI model and individual algorithmic components have been trained and evaluated on both public and proprietary datasets.
The team has experience with imagery, video streams, telemetry, geospatial data and signals from heterogeneous sensors. We know how to fuse these sources, identify meaningful patterns and turn the results into an interpretable map or the system’s next task.
WHAT HAS ALREADY BEEN BUILT
ONBOARD AUTONOMY AND CENTRAL ORCHESTRATION
The team has experience operating drones under two control architectures. In autonomous edge mode, the onboard system analyses data, adjusts motion and executes the mission locally without a continuous connection to an operator or server.
Under central orchestration, AI fuses data from field agents, generates tasks, updates routes and coordinates subsequent actions. The hybrid architecture combines local autonomy with system-level analysis and mission planning.
WHAT HAS ALREADY BEEN BUILT
COMPUTER VISION AND GEOSPATIAL SYSTEMS
Our practical experience includes real-time video analytics, object detection, tracking, classification and semantic segmentation, anomaly detection in aerial imagery, and visual navigation.
Models can run on onboard edge devices with limited connectivity and computing resources. The team also has experience in geospatial analytics, maps, routes, zones and historical data, mission planning, and decision-support interfaces.
WHAT HAS ALREADY BEEN BUILT
FIELD ROBOTICS: MAYA AND HERUS
A working Dynas Maya prototype has already been built with visible-spectrum and thermal cameras. It enables flight testing, synchronized data acquisition and onboard processing. The team is now integrating a hyperspectral camera.
Dynas Herus is the next step in field robotics: a compact autonomous platform with precision actuators, integrated sensors, and repeatable drilling, measurement and sampling at specified depths. A complete Herus agent has not yet been built, but the team has relevant engineering groundwork and professional connections with Ukrainian engineering and robotics teams.
NEXT STAGE
THESE TECHNOLOGIES MUST NOW BE INTEGRATED INTO A SINGLE FIELD-READY ARVANOS PROTOTYPE.
The models and software modules need to be adapted to agronomic tasks, integrated with Maya and Herus, trained on proprietary field and laboratory data, and validated in pilot farms.
ABOUT THE TEAM
ABOUT THE TEAM
ARVANOS AND REQUESTUM
Arvanos is an independent project with its own product vision, developed in collaboration with Requestum.
Requestum strengthens the project with engineering and AI expertise in computer vision, machine learning, data science, signal processing, geospatial analytics, and software systems that work with real-world sensor data.
For Arvanos, this makes it possible to build the complete technology stack—from field data acquisition and interpretation to soil mapping and the next task assigned to a field agent.
ABOUT THE TEAM
APPLIED EXPERTISE
The team has hands-on experience in computer vision, drone control, video and aerial-image analysis, heterogeneous sensor fusion, edge AI, geospatial analytics, and operator software.
We build not only individual models but the complete software stack around them: data preparation, model training and validation, deployment on onboard edge devices, servers or hybrid architectures, model updates, map and route workflows, and decision-support interfaces.
ABOUT THE TEAM
TECHNOLOGY ECOSYSTEM
Arvanos is a Ukrainian project whose core team was formed in Kharkiv. We have professional ties across Ukraine’s scientific, engineering and agricultural communities—including the Kharkiv agricultural cluster, agricultural machinery manufacturers, and robotics teams.
Ukraine has a strong network of universities, research institutes, technology companies and engineering schools, while Kharkiv has long been—and remains—one of the country’s key technology centres. The available expertise spans unmanned systems, applied mathematics, AI, robotics, automation, agronomy and soil science: every discipline needed to develop Arvanos.
SHARED FOUNDATION