Workshop on Challenges and Novel Approaches for Industry 4.0
By Michael Opieczonek, Innobooster Robotics and Reik Leiterer, Innobooster Databooster
Joint Event by Innovation Booster Robotics and Innovation Booster databooster
March 16th, Biel/Bienne
The event was organized on premises of Switzerland Innovation Park Biel/Bienne. As this is a center of innovation, premises of a smart-factory and cobotics center, the symbolic meaning of this location resonated well with the event. The event started with keynote talks, addressing the topics of smart factory, cobotics, human-machine-interaction and general trends in the field of robotics and data-driven value-added services. After a networking lunch, an interactive, moderated design thinking workshop for identifying challenges and developing ideas and solutions was organized.
In form of impulse presentations. 5 speakers give inspiring insights into their research and application areas as well as highlighting current challenges to solve within their respective fields:
Prof. Dr. Sarah Dégallier Rochat, Lead of Humane Digital Transformation at Bern University of Applied Sciences delivered a presentation on Robots as tools: New approaches to robot integration for SMEs. She highlighted that Swiss SMEs are the makers and can turn workers into makers via the concept of augmented worker. Dr. James Hermu, Postdoctoral Researcher in the Learning Algorithms and Systems (LASA) Laboratory at EPFL, delivered a presentation on Real Time Adaptive Systems for Human Robot Collaboration. He talked about methods to teach robots to perform skills with the level of dexterity displayed by humans in similar tasks. Philipp Schmid, Head Industry 4.0 & Machine Learning at CSEM (Swiss Center for Electronics and Microtechnology), delivered a presentation on Industry 4.0 and Machine Learning. He highlighted the need of how machine learning and robots can automate processes at industrial sites and hence increase future of smart-factories. Dr. Renaud Dubé, CTO and Co-Founder of Sevensense Robotics, delivered a presentation on Visual AI: Empowering a new generation of mobile robots. We learned about robots visual capabilities and challenges: lighting and viewpoints changes and understanding semantics.
Prof. Dr Marc Pollefeys, Professor of Computer Science at ETH Zurich and the Director of the Microsoft Mixed Reality and AI Lab in Zurich, delivered a presentation on spatial computing and the industrial metaverse. He gave interesting examples how metaverse can be used for instructional trainings of workers at industrial settings and how spatial computing is contributing to more sophisticated mapping and localization of robots.
The ideation workshop followed in the afternoon and was moderated by the facilitators Prof. Dr. Patricia Deflorin and Dr. Jürg Meierhofer. The workshop took the format of sequences that build on each other – from identifying and understanding the problem to designing the right solutions. The workshop sessions were closed with the presentations of the devleoped ideas and solutions.
The identified challenges could be roughly clustered into 3 categories.
Cluster 1 relates to the general challenges of integrating automation (both on software and hardware side) into existing processes. On the one hand, this includes the necessary technological knowledge and understanding of WHAT one wants to implement – on the other hand, it also includes the expertise or competence development within the company on HOW it can ultimately be integrated. Decision-makers, employees and customers must all be integrated into this process,
and employee acceptance and training/up-skilling must be ensured – all while considering the short- and long-term cost-benefit relations, ethical and moral issues, and cultural acceptance.
Cluster 2 refers to machine learning systems that react more flexibly/dynamically to process changes. On the one hand, regarding rapidly changing environmental conditions, on the other hand, related to highly dynamic process sequences (small batch manufacturing). This requires not only innovative approaches in human-machine interactions (intuitive, ease-of-use handling, no-code environments etc.) but also standardization in processes and interfaces as well as further developments in modular and self-learning ML systems. In this context, the challenge also arises as to how and whether the individual, experience-based knowledge of experts in a company can be transferred to (semi-)automated processes, e.g., the transformation of human intuition in process understanding to rule-based robot-supported systems.
Cluster 3 concerns the (extended) use of cobots/robots in the field of maintenance. This concerns the large area of logistics/ergonomics, from pick-up, sorting and movement of highly divers component categories, to complex processes in material/surface inspection, automated damage repair/replacement, and assembling and dismantling of large rail vehicles. In these processes, the reliability/accuracy requirements are a major (technical) challenge and addressing them would often involve very high costs.
For each cluster, the workshop participants focused on some of the identified challenges and discussed possible solutions.
Solutions Cluster 1:
- Guideline/framework for the integration of automation processes into existing workflows, considering management, customer, and employee’s perspectives (at the meta-level).
- Framework for integration and regular assessment of compliance with ethical and moral guidelines and legal framework conditions.
- Guideline/framework for the practical implementation of automation processes in the company regarding the involvement of employees: internal acceptance, considering employee’s needs, training/education (up-skilling) and empowering.
- Needs assessment for automation solutions in industry (standardization, interfaces, usability/interactivity).
Solutions Cluster 2:
- Development of automatization solutions that can meet the requirements of low volume/small batch production or highly variable process flows.
- Development of ML systems with improved flexibility in terms of self-learning/self-optimizing components so that they can better adapt to changing environments and high-complex processes.
- Development of monitoring systems to capture unconscious, intuitive human components in the manufacturing process and convert them into a rule-based, machine-executable program (e.g. ViT).
Solutions Cluster 3:
- Development of a tunnel scanning and cleaning system to identify and remove paint from vehicles – a combination of intelligent optical sensing for detection and characterization of paint and non-destructive automatization for cleaning/removal of paint while preserving the underlying paint/coating etc.
- Development of a tunnel scanning system to identify and characterize surface damages/deformations on large vehicles – a combination of intelligent passive and active optical sensing, resulting in a digital 3D-representation and classification of surface damages/deformations.
- System development of an automation solution for different tasks as a mobile implementation which works inside of large vehicles.
- Conceptual development of a holistic system (identifying segments, sub-processes, requirements) to support logistics/ergonomics, both in terms of the potential of autonomous (e.g., for, sorting, transport) and worker assistance systems (e.g., exoskeleton, human-robot collaborations).
If you are interested in the ideas and/or you want to further explore these challenges and ideas, we welcome your submission for proposals during the calls by both boosters:
You can sign up for newsletters on our websites as well as follow us on LinkedIn.