December 29, 2022
The COALA project, in its 4th newsletter edition, presents two main enablers of the COALA solution, namely didactic concept and COALA on-the-job training assistant. Both enablers have been integrated and recently tested in the textile production use case.
A short introduction of each of these enablers is presented with an explanatory video to give you a quick overview on their potential implementation in an industrial use case. Our end-users, who lead the textile use case, share their feedbacks on the recent first testing of the on-the-job training assistant and their experiences using the developed learning nuggets.
Karl Hribernik – Project Coordinator – BIBA – Bremer Institut für Produktion und Logistik GmbH
The Coala project
The COALA project (Horizon 2020 Research and Innovation Programme) is part of the ICT-38 Cluster (AI-MAN) and works to provide a solution for cognitive assistance that consists of a composition of trustworthy AI components with a voice-enabled digital intelligent assistant as an interface. The solution will support workers that need to use analytics tools and new workers that perform on-the-job training.
1 Smartening workers with AI competencies: Didactic Concept
One of COALA objectives is to improve the competencies of factory workers in managing AI opportunities, challenges and risks in the shop floor. For this, we have developed a process-oriented didactic concept for training and further education for factory workers through using a voice-based COALA Digital Intelligent Assistant.
The didactic concept aims to help workers building their competencies in human-AI collaboration and to strengthen their control and responsibility during the use of the COALA solution. Media materials, exercises, and competencies tests are offered in a form of micro learning units, so called learning nuggets, to train factory workers and evaluate their learning progress.
Besides that, multi-dimensional learning paths are designed to help workers finding the correct learning content they need while searching for the answers to their current questions. The following video gives an example of implementation of COALA didactic concept in the textile production use case.
Jan Naumann – Vocational Training Expert – ITB – University of Bremen
2 Smartening workers with On-the-job Training Assistant
COALA’s On-the-job training is a core service provided by the Cognitive Advisor. The goal is to reduce the time it takes to train a new operator by 50% and support the transfer of (tacit) knowledge. A custom dialog model and the cognitive advisor service allows identification of the worker’s current skill level, and advice how to prepare and operate a machine. Moreover, the on-the-job training assistant can adapt the advising behaviour with progressing skill increases and under consideration of human feedback about the learning experience, for instance it will be more instructional in the early phase and more supervising in later phases.
The on-the-job training assistant can help operators, for example by suggesting relevant learning nuggets and answering frequently asked questions. If a loom operator needs specific instructions, they can ask COALA directly. The assistant can suggest them a learning nugget on a specific topic.
It can also teach operators to troubleshoot the machine in an efficient way and guide them step-by-step through the troubleshooting process, while offering learning materials from the learning nuggets. The following video shows a short demo on how the COALA Assistant helps operators during on-the-job training by suggesting relevant learning nuggets.
Evangelos Niforatos – Assistant Professor of AI-Powered Human Augmentation – TU Delft
AI-assisted on-the-job training for textile workers
The COALA solution will be deployed and validated against three industrial use cases: (1) white goods, (2) textile and (3) detergent productions.
In the last three editions of the COALA newsletter, we introduced the three industrial use cases, including some challenges face by each use case and what our end-users expectations towards COALA solution. In this edition, COALA partners, who are in charge for the textile use case: Eugenio Alessandro Canepa, R&D Manager of Fratelli Piacenza, and Massimo Angelo Curti, Lecturer of Città Studi, share with us their hands-on experiences during the implementation and first testing of the COALA Assistant for the on-the-job training of warping and weaving operators.
The interview with:
Which pilot use case for textile production has been implemented in the project so far?
The first pilot use case in the textile production, which has been implemented so far, is the “Development of AI competencies of manufacturing workers” scenario. This scenario focuses on the development and demonstration of a didactic concept to train new workers and further train skilled workers. It aims to develop the workers competencies in working with AI-based systems such as COALA Digital Intelligent Assistant.
Città Studi offers academy courses to train the new generation of textile workers. Together with COALA Partner from ITB, Città Studi has implemented a new training plan for warping and weaving operators integrating the didactic concept developed in the COALA Project.
The main targets of this use case are to improve performance in the following activities:
– Faster training for students, as future workers, and for existing workers
– Reduction of errors or defects
– Faster restart after a defect
– Faster change-over from one product to a new one for a skilled operator
– Increase safety
Up to this point, the following features of COALA on-the-job training assistant have been implemented and tested with students and lecturers from Città Studi as well as supervisors and technicians in the weaving department of Piacenza:
– Frequently Asked Questions
– Loom troubleshooting support
– Recommend learning nuggets
What are the lessons learned we may take from the first evaluation?
The didactic concept including the first learning nuggets have been evaluated with some students and technicians in the weaving department. The feedback on the structure of the initial learning nuggets is positive and well appreciated. A further learning nuggets dedicated to specific production processes will be developed and tested in the next year.
Besides that, user experience with the first prototype of the COALA on-the-job training assistant was also evaluated. Some received feedback include improvements and considerations on the following aspects such as single user log-in, speech to text function in local language and recording of the tests, as well as a provision of a user guidelines. These features and procedures will be integrated in the next evaluation planned in the beginning of next year.
What are the next steps in the textile use case?
Besides the planned further evaluation on the first pilot use case mentioned above, our focus for next year is to finalise integration of Piacenza production interface with COALA system to allow production of recommendations based on live machine and production data of Piacenza.
We expects that the COALA solution will allow machine operators to request advices, explanations, and other information via the digital assistant running on a mobile device and that this training support will reduce defects created during the manufacturing process due to human errors.
ICT-38-2020 (AI-MAN Cluster) projects:
- AI-PROFICIENT (Artificial Intelligence for improved PROduction efFICIEncy, quality and maintenance – 957391)
- ASSISTANT (leArning and robuSt deciSIon SupporT systems for agile mANufacTuring environments – 101000165)
- COALA (COgnitive Assisted agile manufacturing for a LAbor force supported by trustworthy Artificial Intelligence – 957296)
- EU-Japan.AI (Advancing Collaboration and Exchange of Knowledge Between the EU and Japan for AI-Driven Innovation in Manufacturing – 957339)
- knowlEdge (Towards AI powered manufacturing services, processes, and products in an edge-to-cloud-knowlEdge continuum for humans [in-the-loop] – 957331)
- STAR (Safe and Trusted Human Centric Artificial Intelligence in Future Manufacturing Lines – 956573)
- MAS4AI (Multi-Agent Systems for Pervasive Artificial Intelligence for assisting Humans in Modular Production Environments – 957204)
- TEAMING.AI (Human-AI Teaming Platform for Maintaining and Evolving AI Systems in Manufacturing – 957402)
- XMANAI (Explainable Manufacturing Artificial Intelligence – 957362)
COALA has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 957296