The MeaTable project has aimed to develop a novel robotic «cell» for use in abattoirs for assisting with the primary cutting of carcasses.
The main objective of the project has been to reduce the work load and strain on human workers. The cell can hold, lift, manipulate and orientate the carcass for a human-operator to perform the necessary cuts quickly and with the minimum of effort - inherently this shall improve the work environment of workers in Norway and beyond. Further opportunities also exist due to the parallel operation of many cells, including the scalability, flexibility and robustness; features that are desirable in meat production but often not facilitated by line-based methods.
The project has been carried out in a systematic way, working with concepts developed by the NFR project «Meat 2.0» and refining them so that they could be physically realised. This has been further enhanced using robotic simulations, a method which has proved an invaluable interface between the engineering and meat expertise within the project.
In 2019, a significant design effort was undertaken to develop a system for handling carcasses, and this was tested in small-scale experiments within industrially relevant environments. A mechanical system, referred to as the Carcass Handling Unit (CHU), was developed for this purpose. A rack was also made, in collaboration with the Ringsaker videregående school, for presentation of the carcass primal cuts. The MeaTable project has been able to realise and confirm that the non-linear concept, first suggested in Meat 2.0, functions with acceptable results in terms of speed and hygiene. This work has been published as a joint scientific article (Animalia and NMBU), and is considered a new method for the slaughter of pigs tailored to the Meat Factory Cell (MFC) approach.
In 2020, based on the success of this new slaughter method, the project established a pilot Meat Factory Cell at NMBU REALTEK (Ås), where development of semi- and fully- automated process takes place. The MeaTable cell utilises one of two large industrial robots manufactured by ABB (supplied by RobotNorge), which are capable of handling the forces and loads involved in processing a carcass of ca. 120kg. In particular, our focus has been on development of Artificial Intelligence systems for recognition and segmentation of carcasses, in addition to tools for cutting and gripping of the material so that they can be transported away from the cell.
In 2021, several trials of an automatic path planning algorithm developed by the PhD candidate (Ian Esper) were performed, with key publications relating to issues such as the detailed approach and calibration to be published in 2022. An important dataset developed during the project was also published, providing public access to RGB-D information from 25 carcasses slaughtered during this project during development of the MFC. Gathering such data is a time-consuming endeavour, and it is therefore hoped that such data being available will be relevant for other researchers in the fields of Artificial Intelligence, Robotics, and perhaps the Veterinary Sciences.
In addition, further cutting and gripping activities have taken place using tools supplied by external parties to the project. Configuration and orientation of the robots to perform these activities has been investigated. Development of the CHU also continued, with both NMBU and Robot Norge contributing solutions to enable automation of several key features in its design. SINTEF have also been working on a method for automatic gripping point recognition, which has included imaging of carcass limbs at NMBU's facility. Several methods have been compared, with the final results to be published in 2022. Other key results are also expected to be published beyond the contractual end of the project, including discussion of the most appropriate use and placement of collaborative robots in the meat sector, and a detailed description of MFC itself, as an engineered system.
The development of a prototype cell in MeaTable has been a significant step forward for innovative approaches to automation in the meat, and food sector more broadly. It enables continuous development and improvement of systems in Robotics, Cognition, and Artificial Intelligence, as well as complementary fields such as Sensor Technology. We engage with potential end-users of the system (e.g., Nortura, a partner in the project) to ensure that the system vision is aligned with the industrial needs, and the results of MeaTable have gained interest in other sectors, such as fish processing.
- Demonstration of the proof-of-concept semi-automatic MFC approach for meat processors, which is relevant for the Norwegian market in regard of its scalability, flexibility and robustness.
- A new approach to slaughter (the process of dissection) is presented, following from its conception in the project "Meat 2.0", which shows the possibility of extracting the gastrointestinal tract in a single piece.
- Influence and evidence for consideration in the context of existing robotics and food legislation, the latter of which stifles innovation in meat processing, particularly in the abattoir.
- A public data set of RGB-D data from 25 carcasses, the likes of which has not existed before in the public domain, facilitating research of others in this and related areas.
- New knowledge and demonstration regarding the use of vacuum gripping systems for large objects in meat processing.
- Review of the state-of-the-art in automation system for abattoirs.
MeaTable will develop a novel robotic work-bench, or cell, for use in abattoirs for assisting with the primary cutting of carcasses. The main objective of the project is to reduce the work load and strain on human workers. The work-bench will hold, lift, manipulate and orientate the carcass to enable a human-operator to perform the necessary cuts quickly and with the minimum of effort and in this way improve the work environment of thousands of workers in Norway and beyond. Furthermore, transportation of parts from the carcass is incorporated to further alleviate labour intensity, as well as reduce cross-contamination risks.
The project aims to increase productivity in an energy-efficient, sustainable way, by introducing automation to the meat slaughter and cutting process that is suitable for the scale of the Norwegian industry. The process shall improve environmental sustainability performance while maintaining or improving economic and social sustainability. Using the novel Cell concept as its starting point, MeaTable will introduce a flexible, scalable, and robust automation solution to the meat industry. Based on human/robot collaboration, that are suitable for the scale of Norwegian slaughterhouses, we develop a system that is specifically well suited for the Norwegian market.