GreenPatrol faces its final stretch with the satisfaction of having achieved all the technical objectives set at the beginning of this European project. The development of a precise positioning solution based on GNSS for semi-indoor environments, the system for early detection of pests affecting crops, the technology implemented to facilitate autonomous navigation of this device and validation in a real scenario have been achieved within the parameters established by the specifications of this research.The last meeting of the international consortium led by the CTC Technology Centre has reviewed and validated the work carried out by the six partners of the initiative. Representatives of all the CTC, the Tekniker Foundation, Inkoa Sistemas SL, the Czech university Mendelova Univerzita V Brne, the English company Nottingham Scientific Ltd and the Dutch company Aerovision Bv, as well as the Project Officer of the GSA and the reviewers of this initiative, attended this last meeting.
During the course of the meeting, the circumstances that conditioned the final part of the project were explained. The outbreak of COVID-19 forced to reduce as much as possible the tests carried out in real environments to minimise the risk of contagion. To alleviate this situation and complete all the tests to thoroughly test the robotic solution, the researchers created two virtual greenhouses faithful to the real test environment. They replicated the crop rows and the width of the corridors to perfection. The same number of plants as in a real greenhouse were introduced and randomly positioned to replicate the unpredictability of crops and possible obstacles. The environment also included the simulation of different levels of positioning error, including areas of GNSS signal degradation and even areas of complete obstruction to simulate the effect of plant growth and the metal structure of the coating.
Thus, even though the validation tests in real greenhouses were reduced to a single week of testing, the extensive battery of tests carried out in simulation allowed many of the system’s functionalities to be validated in a real environment, such as the accuracy of the positioning solution; the autonomous navigation skills, the detection of pests using vision and Deep Learning techniques; or the robot’s ability to apply pesticides locally.
Also, the dissemination activities carried out during the last period of this project linked to the H2020 Framework Programme, which has a budget of 2.4 million euros, from the GSA (European Global Navigation Satellite System Agency) through the European H2020 Programme, were reviewed during the meeting.