How does robotic handling affect soft fruit quality?
18 December 2024
In the project ‘Autonomous robots for agrifood processes’, a team of Vision+Robotics experts is studying the impact of robotic handling on the quality of soft fruits. What impact does this have and can soft grippers (really) shed a different light on this?
Senior scientist and expertise leader computer vision and robotics Aneesh Chauhan is one of the Vision+Robotics experts of Wageningen University & Research studying how robotic handling impacts different types of soft and mostly also fragile fruit. “The main motivation for the research in the context of the project “Autonomous robots for agrifood processes” is the strong need food processing companies have to automate, to robotise repetitive tasks and processes in food production factories. While the arguments are clear – labour shortages, repetitive tasks, efficiency and hygiene – you cannot blindly introduce robots in jobs where they need to be in direct contact with food products. Indirect contact where robots handle boxes or trays containing food products or fruits and vegetables, is fine and widely in use. This way, the product is not affected by the handling. But as soon as there is a direct contact between a robot or robotic gripper and a product and/or fruit, the product quality can be affected. In a visible way, such as bruising, but also in ways not immediately visible to the human eye and/or to computer vision. Until now, these effects were very limitedly researched and we want to change that with the significant amount of work on quality measurement we do in the context of this project.”
Different fruits, different challenges
The idea and purpose of any post-harvest action and handling is to keep the product alive and in good quality until it is consumed. “We are dealing with biologically active products and some of these are very easily bruised during handling. I mean, there are successful robotic solutions to sort and pack apples, melons and pineapples for instance, but handling bunches of grapes, raspberries and especially soft fruits like red currants and strawberries is much more challenging from a product quality perspective. Introducing robotic end effectors like soft grippers to handle these, does not mean you don’t damage the product. Some damages occur internally and don’t appear until later. During logistics, last mile delivery, during shelf life or just prior to consumption.”
Chauhan has noticed that technological advances in recent years have moved to soft materials because these can lead to more universally applicable solutions. “They are mechanically compliant and can change shape in response to external or internal actuators as well as to their interactions with objects. Such as octopus’ tentacles, gecko feet and coffee bean based jamming grippers. Many end effectors however are not suitable for fragile fruits because these bruise so easily. End effectors can also transfer diseases and pathogens like fungi from one fruit to another without being noticed by quality control procedures. We can introduce gecko feet like grippers with an excellent grip, but every ridge on that gripper’s surface introduces a risk and can become a carrier of fungi. Robots can thus become a vector for diseases and that’s not what we want. You can clean or replace the gripper, but when and how often? Can we detect contact with pathogens?”
Another challenge is how to handle different sizes and shapes of fruit. Humans fingers and hands can easily adapt to handling small as well as large strawberries while for robotic grippers this is very challenging. “That is why I argue that soft and compliant robotics are part of the future of fruit handling.”
Understanding the impact
“We want to understand short term effects of a contact between a gripper and a product as well as the long-term effects that might only appear later on. Obvious impacts can be detected and measured visually, while others require near infrared spectroscopy (NIRS) and/or X-ray technology. To do so, we also test in storage and shelf-life conditions to detect and measure changes occurring at a later stage. Changes that were not apparent directly after a product was handled. And we potentially want to set up benchmark facilities that allow robotic gripper and end effector developers to measure the effects of their handling. I believe that Wageningen University & Research can provide the ideal circumstances for this because all the knowledge across every discipline comes together here. My aim is to create a better understanding of the impact of robotic handling that leads to a better handling of biologically active materials.”
Testing various types of robotic grippers on various soft fruits, such as blackberries, raspberries, and strawberries.
Sensing to be studied
While humans know that their fingers are carrying something, there’s currently no sensing in existing grippers we work with, at least not to the sensitivity levels of human fingers, Chauhan says. “We need to study high sensitivity sensing in grippers as most of the time grippers sense too little too late. With every type of product handling, of technology you introduce, whether it is mechanical or robotic, you also introduce a risk of impacting and potentially damaging the product. Sometimes these effects also surprise us as experts in vision and robotics. Until last year a theory was that soft robotic handling solutions can work. But then we had surprising test results with soft grippers where the softest gripper led to unexpected latent damage on raspberries. Why was that? What has changed in the product’s biology? We still have to understand the biological mechanisms resulting from robotic handling.”
Red currant solution coming up
“We are currently using all the knowledge and insights we gain to introduce robotic grippers for packaging and handling red currants in about two years’ time. That will be an AI supported solution to handle red currants in factories that is not causing any damage to the fruit. But for now, the AI-based robot control and positioning is not yet precise enough. I am convinced however that robotics is going to change food processing automation significantly.”
This study is primarily funded by the Ministry of Agriculture, Fisheries, Food Security and Nature (LVVN) in the Netherlands via the knowledge base programme autonomous robotics under grant number KB-38-001-005.
