H2020 RoBUTCHER Project: A robust, flexible, and scalable cognitive robotics platform
- Type Project
- Status Filled
- Execution 2020 -2023
- Assigned Budget 7.536.303,5 €
- Scope Europeo
- Main source of financing Horizon 2020
- Project website Proyecto RoBUTCHER
The meat sector is one of the most important sectors in EU agriculture, comprising farmers, agricultural cooperatives, and slaughterhouses. One of the biggest challenges is maintaining quality. Therefore, automation is key to complying with EU food safety recommendations. However, advanced robotics and automation systems are expensive, inflexible, and lack scalability and robustness, making them difficult to acquire and implement on a large scale. A pioneering new automation concept developed by scientists in Norway and Denmark, the Meat Factory Cell (MFC), is not autonomous. The EU-funded RoBUTCHER project aims to develop a cognitive FMC with high autonomy using core robotic technologies based on AI and cognition. The project will provide small and medium-sized meat processors with the tools to overcome the technical barriers that impede the adoption of robotic automation systems.
The societal, legislative, and best practice requirements of the meat industry were assessed as expected. From a robotics perspective, current legislation or standards do not prevent the implementation of MFC in an industrial setting. From a meat safety perspective, the project conducted a comprehensive review of legislative texts, the findings of which reflect the need for so-called "functional demands" when developing such texts. It is encouraging that the WHO now uses terminology consistent with these principles. RoBUTCHER has a solid foundation in industrial needs and best practices and included end-users from across Europe as part of a reference group. This has ensured the relevance of the project's work and also ensures a good contribution to studies addressing the societal impacts of automation in meat processing. RoBUTCHER has developed several new vision-based AI approaches: 2D only (trachea location); Combined 2D and 3D data (a method for MFC cutting, as well as limb grasping); 3D and CT (another, more customizable method for MFC cutting); 3D and X-ray (for butchering). The solutions cover a wide range of scenarios, including cutting hot and cold meat, as well as whole and half carcasses, all while taking into account the biological variability this entails. Post-project activities already show actions by the Consortium to extend this learning to other species, such as beef and turkey. Virtual reality (VR) has allowed, for example, a butcher to remotely verify and adjust AI cutting trajectories in a 3D environment, prior to execution. Digital twins have facilitated the planning and testing of offline systems; the Consortium recognizes the need for new R&D activities, where flexible and variable materials are better modeled in such environments. RoBUTCHER identified a deficiency in robot tools for working with meat, resulting in the development of new smart grippers and cutting instruments, including the patent-pending smart knife. R&D will continue after the project. Notably, studies indicate that consumers do not distinguish between robot-cut and human-cut meat; some may even prefer the former. RoBUTCHER successfully developed two industrially relevant prototypes (i.e., at TRL6), for hot and cold cutting. Furthermore, the project has resulted in no fewer than 17 key innovations, ranging from TRL4 to TRL9 at the end of the project. Regarding dissemination, 316 activities were carried out, including workshops, demonstrations, scientific publications, and more. Forty-four additional post-project actions are planned.
The meat sector is one of the most important in the EU. Automation is key to achieving food security and sovereignty in Europe. Traditional robotics and automation systems are too expensive, inflexible, and lack scalability and robustness. A pioneering new automation concept developed by scientists from the RoBUTCHER project, the Meat Factory Cell (MFC), has demonstrated a way to address these problems. This project has developed and successfully tested a highly autonomous MFC prototype using core robotic technologies based on Artificial Intelligence (AI) and cognitive robotics. RoBUTCHER is an innovative approach that offers an alternative to conventional in-line production in slaughterhouses through parallel production in autonomous MFCs. This is particularly relevant for small and medium-sized producers, as it offers the possibility of increasing or decreasing productivity based on demand. This technology has disruptive potential, and the project has developed, produced, and tested this innovation. The results of RoBUTCHER are important for society, as they will allow robots with a higher degree of competence to enter the manufacturing field. Global development has increased attention to flexibility in meat processing. As the meat market becomes increasingly internationalized, product portfolios are expanding to meet new consumer demand. This challenges conventional solutions and emphasizes the need for RoBUTCHER. The inability of small and medium-sized producers to access robotics and automation also exacerbates another time bomb in Europe: labor availability. In addition to economic considerations, robots will be required to overcome labor shortages and contribute to making food processing safer, cleaner, and more sustainable. The recent pandemic and geopolitical events affecting labor movement and the food supply chain have further accelerated acceptance and prepared the sector for the deployment of the novel RoBUTCHER technology. The results may also, over time, impact other sectors (e.g., other species such as beef, other food products, or entirely different materials) where handling materials with high variation and flexibility is important. The overall objectives of RoBUTCHER were: 1. Assessment of societal, legislative, and best practice requirements in the meat industry 2. Development of novel technology modules for autonomous cutting trajectory planning and integration with human-robot interfaces providing remote intervention and training opportunities 3. Adaptation of cutting tools to incorporate cognitive sensing and development of gripping tools for grasping and manipulating muscles, bones, and internal organs 4. Implementation and demonstration of RoBUTCHER MFC in a pilot-scale environment at Technology Readiness Level 6 (industrial prototype). RoBUTCHER met the stated objectives. In conclusion, the outcomes of RoBUTCHER are even more important for society than when the action began. In particular, global events such as the COVID-19 pandemic and the war in Ukraine demonstrated that food safety is a primary concern for society.
The adoption of advanced robotics and automation in the agri-food sector, specifically in meat processing, has been stifled due to the perceived high cost, as well as a lack of flexibility, robustness, and scalability to accommodate different volumes, particularly smaller ones. This conflicts with recommendations that governments across Europe should encourage greater efficiency and safety in the food system. Pioneers in Norway (NMBU/Animalia) and Denmark (DTI) have been working to address this issue and have created a new automation concept for the meat sector: the Meat Factory Cell (MFC). The current MFC is simple and relies on the intelligence of human experts to complete complex tasks. However, meat processing plants are among the poorest work environments in Europe, making autonomy imperative. RoBUTCHER aims to develop a cognitive MFC, capable of autonomy. To achieve this, RoBUTCHER has the following core objectives: (1) Assessment of the societal, legislative, and best-practice requirements of the meat industry; (2) Development of novel technological modules for autonomous cutting path planning and integration with cooperative human-robot interfaces; (3) Creation of smart tools to evaluate the system, primarily for cutting and handling; and (4) industrial-scale piloting of cognitive MFC. Core robotic technologies are integral to RoBUTCHER, where the main emphasis is on AI and cognition, but there is a natural overlap to include cooperative human-robot interfaces and cognitive mechatronics. Currently, no suitable and available solutions exist. The ambition of RoBUTCHER is to develop a system for TRL6, using the existing MFC infrastructure within the Consortium as a catalyst for research and innovation. Successful implementation of the project will provide the robustness, flexibility, and scalability that small and medium-sized meat processors need to reduce the technical barriers they face when adopting robotic automation, which would improve employment quality and food security in Europe.
Researchers are helping the meat industry overcome obstacles to the adoption of robotic automation systems. The EU is the fourth-largest producer of beef, veal, and chicken, and the second-largest producer of pork in the world, reinforcing the importance of the meat industry in the EU. However, the sector faces several challenges, such as maintaining quality while ensuring a consistent supply of food. Automation can contribute to this goal, as well as helping achieve EU goals for greater efficiency and safety in the food system. “Automation is critical to the sustainability of meat production, where the heavy reliance on human labor is a growing challenge. Without automation, the sector could struggle to meet the increasing demands for efficiency, productivity, and environmental responsibility needed to remain competitive in the future,” notes Alex Mason, RoBUTCHER project coordinator. However, for many industry professionals, advanced robotics and automation systems are inaccessible, as they are expensive and lack flexibility, scalability, and robustness. An automation concept developed by scientists in Norway and Denmark, the Meat Factory Cell (MFC), aims to address these issues. In this context, the EU-funded RoBUTCHER project worked on developing a novel MFC platform, which offers an opportunity for unconventional automation in pork processing, particularly in slaughterhouses. “Rather than following production-line-based approaches, which are the main option today, it uses robotics and artificial intelligence (AI) to perform complex cutting and handling operations on whole, unchilled pig carcasses, taking into account biological variation and deformation,” confirms Mason. Disrupting the meat sector One of the key outcomes of the project was the successful demonstration of two prototype robotic cells for hot and cold meat cutting. One consisted of a prototype for the complete disassembly of unchilled whole pig carcasses, while the other implemented a more conventional approach based on what is known as chilled half-carcass trisection. "We also developed new software tools, including AI systems for meat cutting, interactive virtual reality approaches for large-scale robot butchering, and solutions for smart system integration," Mason reports. This is in addition to new tools to facilitate interaction with meat products and robotic systems, including knives and gripping tools. The project's work has also led to a better understanding of the legislative frameworks and societal impact associated with innovative robotic solutions for highly manual industrial manufacturing environments. "In total, the RoBUTCHER Consortium identified 17 key innovations, including the two demonstrated prototype cutting systems, resulting from the project. In addition, the consortium has been very active in publishing the results, with around 30 scientific papers published to date and many more in various stages of preparation," Mason highlights. Moving Forward: New Developments for the Meat Industry The consortium has identified new initiatives to continue the work of RoBUTCHER and plans to develop new R&D ideas within Horizon Europe over the next 12 months. “At the national level, several partners are already working on projects. They are using RoBUTCHER results to lay the foundations for new developments in smart tools, collaborative robotics, and AI,” concludes Mason. The “cell” concept has sparked interest from stakeholders globally, as highlighted at the recent International Meat Automation Congress in Spain, where the RoBUTCHER results were presented. More information on the project progress and results can be found on their website.
New AI methodologies for meat cutting, capable of recognizing and adapting to biological variability. Prototypes of automatic systems for meat cutting, applying new robotic methodologies. New gripping and cutting tools, including sensor enablement, to provide robotic feedback. New implementation of integration through software platforms, including Node-Red and Ocellus. Introduction of collaborative robotics concepts for meat cutting. Teleoperation of industrial robots through VR. RoBUTCHER has advanced research in core technologies, such as cognition and human-robot interfaces, and has produced economic impact with at least two key innovations already reaching the market. Overall, the level of innovation in the agri-food sector has increased, particularly for meat processors. Leading meat processors and suppliers are aware of RoBUTCHER's novel approach. A post-project goal of the Consortium is to broaden the applicability of RoBUTCHER results to sectors with similar functional demands. In the long term, RoBUTCHER lays the foundation for more scalable (both physically and economically), robust, and flexible automation than that currently available. According to end users, implementing the project's results would reduce or eliminate the risk of injury to workers in traditional meat factories, in addition to improving working conditions. Furthermore, it is likely to offer improvements in gender equality and workforce diversification, while also making employment in the sector more attractive to young people.
- NORGES MILJO-OG BIOVITENSKAPLIGE UNIVERSITET (NMBU)
Related projects
