Reconfigurable

INRIA has access to a 100 m x 100 m reconfigurable outdoor facility located at AzurArena in Antibes, ideal for conducting experiments involving Autonomous Vehicles (AVs), Autonomous Ground Robots (AGRs), and Unmanned Aerial Vehicles (UAVs). The site can be equipped with a 4G/5G communication network, enabling real-time connectivity and coordination across multiple robotic platforms. Its open and adaptable layout makes it suitable for both individual and collaborative autonomous system testing.

NEF is a multidomain scientific experimentation platform at INRIA, designed to support high-performance computation, storage, and visualisation for advanced research applications. It is built on a heterogeneous parallel computing architecture, combining multiple generations of high-performance servers into a cluster-based system. The platform features high-speed network interconnects and large-capacity storage and is directly connected to the Inria Sophia interactive visualisation platform, enabling seamless data processing and graphical analysis.

At INRIA, the private parking facility spans an area of 85 meters by 13 meters and is naturally enclosed by plants and trees, creating a semi-structured, forest-like environment. This makes it an ideal outdoor space for conducting development, testing, and validation experiments with Autonomous Ground Robots (AGRs) and Unmanned Aerial Vehicles (UAVs). Its dynamic setting bridges the gap between controlled indoor testing and fully unstructured field environments, allowing for safe and effective experimentation under realistic outdoor conditions.

At ACENTAURI, we host a dedicated indoor UAV Arena measuring 5 meters x 6 meters x 7 meters, specifically designed for the development, testing, and demonstration of mobility algorithms for both Unmanned Aerial Vehicles (UAVs) and Autonomous Ground Robots (AGRs). This facility offers a controlled and customisable indoor environment, making it ideal for preliminary experimentation and fine-tuning of robotics systems before deployment in real-world conditions.

PSNC has a network of 20 units of phenological stations, consisting of a PTZ camera with 40x optical zoom and a GSM router. These stations are used to observe phenological phenomena and monitor the occurrence of pests and plant diseases. They are located throughout Poland. Additionally, PSNC is the owner of data transmitted from a network of 550 agrometeorological stations. These stations are precise tools for conducting continuous meteorological measurements.

The GNSS laboratory allows the use of a GNSS simulator, which produces a comprehensive range of emulated RF signals with industry-leading flexibility, fidelity, performance, and reliability. It covers a full set of known constellations like GPS, GLONASS, Galileo, BeiDou, QZSS, NavIC, and SBAS.

This concrete test platform (30 m * 15 m) located on the INRAE Montpellier premises enables application techniques to be tested under controlled conditions: it features rows of artificial vines and a device for controlling wind conditions on the platform (wall of fans).

Typically, an evaluation plan is drawn up, including identification of the various influencing factors, evaluation criteria definition, and test scenarios. Using task-appropriate metrics, the overall product performance is assessed, and underperforming scenarios can be identified. It is also necessary to have various data from the product, which must be qualified and annotated in order to evaluate the product as accurately as possible. This facility is based on MATICS software and high-performance computing resources.

This modular platform covers several fields of robotics. In particular, it features a climate chamber to validate the robot's operation at different temperatures and tracking cameras to measure its movements. The modules designed for movement tests (rough terrain, mazes, etc.) and manipulation tests (handles, valves, 6D positioning) are based on NIST (National Institute of Standards and Technology) standards.

Evaluation on this bench consists of placing a device at the heart of a simulation projected onto a 300° screen while a motion capture system and an instrumented conveyor belt measure its movements, if any. These data are integrated into the simulator in real time so that the device's digital twin follows the same movements and the projected environment is updated accordingly. The simulator also incorporates a physics engine and advanced sensor models, enabling a virtual sensor output to be substituted for the robot's sensors in real time.