>
About the Project
AIGreenBots establishes an advanced doctoral training program
spanning multiple institutions, aimed at creating innovative agricultural robotic platforms, sensor-fusion
systems, reliable AI and machine learning models, and decision-making mechanisms for agricultural automation.
Additionally, the project addresses crucial regulatory and safety challenges related to the field deployment
of these technologies. The consortium, comprising both academic and industrial partners, integrates hands-on
training, secondments, and collaborations with non-academic sectors to ensure the development of practical,
market-ready solutions for agricultural robots.
The AIGreenBots project tackles the challenges addressed in each technical WP within the context
of agricultural robotics through a comprehensive and interdisciplinary approach. Among beneficiaries and
associated partners, the network is comprised of 5 universities, 3 research centers, and 5 industry partners
from various countries, bringing together top-class expertise in robotics, AI, machine learning, sensor fusion,
perception systems, and agricultural technology. This consortium ensures inter-sectoral collaboration, with
industrial partners actively supporting through facility access, real case studies, practical training, and
participation in project-wide events.
The AIGreenBots Doctoral Candidates (DCs) will be trained to become highly skilled scientists and
engineers, prepared for careers in field robotics, AI development, and related sectors. They will gain valuable
experience in a range of environments, ensuring their expertise aligns with the evolving demands of sustainable
precision agriculture and related challenging domains.
Coordinate the AIGreenBots project to ensure deliverables are met within the budget and
timeline. It includes organising meetings, managing recruitment, setting up management structures,
and monitoring project progress. The work package also oversees legal and financial compliance,
intellectual property management, and the dissemination of knowledge to maximise the project's impact.
Provide interdisciplinary research training to Doctoral Candidates (DCs) to prepare them for
careers in academia or industry. It includes creating an Open Science and Innovation platform,
implementing a career development plan, and providing technical and project management training.
The work package also promotes public engagement and diversity in STEM fields through outreach
activities.
Design user-centred robotic platforms for agriculture, capable of safely operating in
real-world terrains. It adopts industrial-grade software, the Robot Operating System (ROS), and best
software engineering practices. Key tasks include developing a 3D simulation environment, a
self-diagnosis toolkit, task-planning systems, and a machine learning framework for terrain
analysis, with solutions evaluated through user feedback and field testing.
Solve perception challenges in agricultural robotics by developing multi-sensor data fusion
techniques. It focuses on integrating data from various sensors, addressing communication protocols,
and data storage. The main goal is to design an architecture that enhances perception accuracy for
complex tasks. Key tasks include creating a probabilistic perception model, defining sensor
specifications, implementing time-based data fusion, and developing a distributed robotic
architecture for efficient data sharing.
Develop reasoning and task planning for agricultural robots in uncertain environments. It
focuses on adaptive, long-term planning using data from WP4 and machine learning to update the
robot's world model. The key tasks involve adaptive task planning, mission management, and an
ontology-based inference system to enhance robot decision-making. Techniques such as Bayesian
inference and Graph Neural Networks are used to improve action selection and planning efficiency.
Create a decision-making platform using eXtended Reality (XR) for human-robot collaboration
in agriculture. It allows users to control robots via XR glasses, using sensor and machine learning
data for safe and accurate operations. Key tasks include developing an XR-integrated environment,
standardising sensor interfaces, building a decision-learning platform, and demonstrating robotic
arm use cases.
Equips researchers with knowledge of legal, operational, and safety issues in agricultural
robotics. It involves reviewing international standards, developing safety systems for robot
coordination, and creating a sensor-based control system for farm integration. The work includes a
workshop for hands-on farming experience and field trials.