Food waste is a global problem that, in addition to economic loss, has a negative impact in the environment and society. Millions of the world’s food ends up in the trash every year and it happens at all stages of the food chain, from production to consumption. technology center AINIA It believes that reducing food loss at every stage is crucial to addressing the global problem of hunger and is committed to finding solutions.
FOODCOLLECT is a research project at AINIA Innovative technologies for food processing and preservation, promoting good practices in responsible food consumption.
Through the use of mobile robot systems and artificial vision, this project actively supports the United Nations Sustainable Development Goals (SDGs), a set of international goals and commitments to achieve a more just, equitable, and sustainable future.
Fallen fruits
It has three main goals: responsible production and consumption (Goal 12), eradicating poverty (Goal 2) and improving the technological capacity of industrial sectors (Goal 9). The main goal is to reduce food waste through the independent and selective collection of fallen fruits, that would normally be lost. Intelligent sensors and autonomous robotic systems make this possible.
FOODCOLLECT also uses AI technologies to control and direct the system in real time.. The system is designed to orient itself autonomously across the field using geolocation technologies, allowing for more efficient harvesting.
Three main units
FOODCOLLECT has three main modules: Independent Navigation, Intelligent Perception, and Collaborative Manipulation.
The autonomous navigation unit allows the robot to move around the field in search of the target fruit. A system was developed that tracks the optimal path to pass through the fruit trees in the plot, avoiding hitting obstacles. Experimental tests were conducted in different areas of persimmon and orange to verify the different algorithms and navigation methods.
The Intelligent Perception Unit is responsible for detecting the fruits in the soil around the trees Based on deep learning techniques. A high-resolution 2D image acquisition platform was developed and AI models were trained to detect oranges and persimmons. The neural network was repeatedly tuned to adjust the detection, and the performance was above 90%.
The Collaborative Processing Unit features the KINOVA GEN-3 Collaborative Robotic Arm as a collection tool. The robotic arm is attached to one side of the mobile vehicle and a claw is designed that allows it to adapt to the morphology of persimmons and oranges, and to perform an effective grip, without damaging the already pierced product because it is on the ground. A 3D scan sensor based on time-of-flight (ToF) measurement technology for the 3D characterization of objects was selected and installed on the robot paw, covering light weight and size specifications, but with the ability to acquire spatial point cloud collection in real time.
Valencian Community
The integrated solution developed at FOODCOLLECT II is being validated in the areas of the Valencian Community, which contributes to improving the efficiency and sustainability of the Valencian rural sector. In addition, the project allows Valencia companies in the machinery and capital goods manufacturing sectors to develop state-of-the-art technological applications with great potential for national and international exports.
The project is involved to ANECOOP y ava asjaAnd producing companies and fruit and vegetable centers that help identify the problem of fruit falling and conduct field tests. There are also mobile robot companies, such as robotnikand specialists in agricultural technologies, such as Agrotec Spainwho will participate in validation and clarification sessions.
This activity is part of the research and development line that AINIA is developing with funding from the Valencia Institute for Business Competitiveness (Evas).