AI performance evaluation based on mixed testing environments
Summary
Evaluation of AI embedded in physical devices (e.g. robot with smart camera(s)) in a semi-simulated environment.
S00093
Version: 1.10
- arable farming
- horticulture
- tree crops
- viticulture
- livestock farming
- food processing
- test design
- test setup
- test execution
- collection of test data
- performance evaluation
- physical system
- software or AI model
- Location: in France
- Offered by: LNE
Description
This service propose to assess any devices (with respect to the physical constraints of the bench) integrating AI-based vision processing capabilities in a novel hybrid testing environment. Evaluation on this bench consist of placing devices in the heart of a simulation projected onto a 300° screen while a motion capture system and instrumented conveyors belt measure its movements, if any. These data are continuously sended to 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, in the cases where devices require specific data and/or if the assessment is oriented toward special kind of sensor degradation. Typical products evaluated: mobile robots with visual navigation based on 2D cameras, or "intelligent" cameras (with AI functionalities). Other sensors like : 3D camera, GPS, Lidar, sonar, etc., are supported through data injection from the simulator.