Test campaign realised by the CEOL robot, in collaboration with Agreenculture and INRAE.
The aim of GEOSUR project is to ensure the integrity performance of GNSS terminals dedicated to autonomous navigation in agricultural environments.
The 3 main activities of GUIDE are to:
- Assess the GNSS terminals accuracy performance that will be used during the project;
- Perform test scenarios based on the corrupted GNSS measurements to realistically recreate feared events
- Develop certification processes to establish the various conformities of GNSS terminals tested with client requirements.
Testing in constrained agricultural environments
With AGREENCULTURE, a secure GNSS terminals manufacturer, and the National Institute for Agriculture, Food, and the Environment (INRAE), responsible for agronomy research, GUIDE has validated the operational capabilities of the GNSS terminal named the AGCBox.
Performance requirements were established to assess the use of this product through fruit tree plantations and a solar panel field. The local environment was considered relatively constrained. Physical and natural obstacles can cause position errors by masking or altering satellite signals, or by causing parasitic reflections such as multipaths or diffractions.
GUIDE performed the field tests, in particular to determine the ground truth with measurement uncertainties and to analyzed the collected data. The results were published in a test report and the key measurements in a test certificate.
GNSS data corruption
To estimate the vulnerability of GNSS terminals to feared events, laboratories need to put them to the test by recreating realistic test conditions with all the hybrid measurements leading to position calculations.
These test conditions can be difficult to reproduce on site or with a GNSS simulation, particularly when the feared events are the result of a combination of erroneous measurements.
In this context, selected real-world measurements are reused after data contextualization operations. GNSS measurements are time-stamped to synchronize them with the test scenarios. Above all, the feared errors are merged into the scenarios by corrupting the original measurements.
To this end, the laboratory has developed methods based on automatic learning techniques. GUIDE uses large volumes of data collected by its test engineers in real-life situations to train its models. The results provide a digital simulation of GNSS measurements (RINEX or RTCM) that is perfectly representative of reality at a lower cost.
The development of a certification bench for agriculture
The aim of the GEOSUR project is to develop a test bench for assessing the integrity of GNSS terminal position calculations used for agricultural and autonomous robotics.
The certification process will enable to achieve the following objectives:
- Identify the potential feared events and analyze the failure modes, their effects and their criticality,
- Design representative test scenarios of the feared events to be overcome,
- Develop the sets of measurements, with or without errors, necessary for the specified test scenarios,
- Validate the test bench used for replaying the data sets dedicated to the integrity performance,
- Document the results and establish their conformity with the integrity requirements.
The main objective of GEOSUR project is to contribute to the enhancement the safety of positioning functions on agricultural robots. This condition is crucial for the autonomous mode of these machines, which are intended to be operated without any operator.
The results of this initiative will help to determine the degree of confidence associated with their use, in a carefully defined context.
Ce projet a été financé par le gouvernement dans le cadre de France 2030.
Ce projet a été financé par l’Union européenne – Next Generation EU dans le cadre du plan France Relance.