The research centres on the integration of knowledge, methods and systems to design and implement all the processes involved in the optimal resolution of complex problems. For example, in the case of Precision Agriculture, a Site Specific Weed Management system is conceived as a three-stage intelligent and complex artificial system which includes:
1) Perception or detection and identification of natural structures such as crop rows and weeds.
2) Decision-making, or the development of an action plan that must consider elements such as the amount of herbicide to be applied, information supplied by the perception stage, additional information available (type of crop and weeds, field history, farmer experience, etc.), and the targets set for the treatment.
3) Action or implementation of the treatment plan; specifically, the generation of the appropriate signals to be sent to the treatment equipment, for example a sprayer.
In the area of outdoor mobile robotics, the GPA is dedicated to the development of various small mobile platforms that can be used for both field scouting and treatment application. The aim is for robots to move autonomously guided by the available natural structures, such as crop rows. The first prototypes also integrate a hybrid architecture (deliberative and reactive) for behaviour generation.
Since 2010 the GPA participates in the European RHEA project – Robot Fleets for Highly Effective Agriculture and Forestry Management” (http://www.rhea-project.eu). The RHEA project is a FP7-NMP-2009-LARGE-3 project that involves 19 groups from 8 countries. The head of the GPA is also part of the RHEA coordination team. She is in charge of supervising the scientific and technical aspects of the project. The project has a total budget of approximately 7 million euros and is a great opportunity for Europe to lead the development of agricultural technologies. RHEA provides a set of complex real problems and specific objectives for the development of perception-decision-action systems that will enhance scientific progress in research areas such as perception, coordination of agents (physical), decision making, mobile robotics, software architectures for the generation of behaviours, etc.
In addition, the group is coordinating the Spanish research project GroW (Ground Inspection System in Autonomous Vehicles and their Effective Application for Weed Detection and Control – AGL2011-30442-C02-02) since January of 2012.
In summary, the research tasks of the GPA focus on:
– Perception Systems: Merging sensory and contextual information.
– Structures for representing knowledge and information.
– Knowledge induction methods.
– Planning and decision-making mechanisms.
– Distributed processes architectures: coordination and interaction of agents and behaviour generation.
– Artificial vision systems.
– Discrimination in real time of natural structures.
– Autonomous visual navigation of mobile robots.
– Sensor nodes in wireless networks.
– Solutions based on evolutionary computation.
