Exploring human – machine collaborations for flexible manufacturing
The Philips factory in Drachten, the Netherlands, is an advanced factory for the mass manufacturing of consumer goods (e.g. shavers, OneBlade, baby bottles, soothers) where over 2100 enthusiastic colleagues from all over the world work together to develop and produce products that improve people’s lives. By putting focus on sustainability through innovation, we are always striving to work on the edge of current technical development. However, pushing for innovation is a process that is hard to do alone. Therefore, we are privileged to team up with, among others, RTO’s, technology providers, and other manufacturers in the STAR project, developing knowledge on key enabling technologies and solutions for manufacturing challenges of today and the future.
One of these challenges revolves around the introduction of flexible manufacturing. Within the production of shavers in Drachten there is a strong emphasis on standardization, automation, and short cycle-times in its production lines. These production lines are often specifically tailored for the mass production of one product series in the most efficient way. However, due to a shift in customer demand, production batches become smaller and personalized. To comply with this shift in customer demand, production lines need to be reconfigured more often to be able to produce different products. In the current production setup, these reconfigurations are expensive and time-consuming. During the STAR project, AI-based solutions will be developed to facilitate more flexible production approaches.
As a global leader in the manufacturing of mass-produced consumer products, Philips Drachten supplies a use-case revolving around ‘HUMAN – COBOT Collaboration improving robust Quality Inspections by Vision’ as a pilot demonstrator in a relevant production environment. This pilot offers the opportunity to demonstrate results developed during the STAR project and gives insight into how these results can be leveraged to improve flexibility of today’s as well as future production systems.
To keep a clear vision of the goals, the pilot line within the factory in Drachten explores three different topics regarding the creation of such a flexible production system.
1) Easy reconfiguration for automated part handling,
2) Human supervised learning for visual quality inspections,
3) Safe collaboration between human and cobot.
During the first topic, flexibility from a part handling perspective is explored. This means that research is done into the design of a system that is useable in many different scenarios where an incoming part can be detected, identified, and handled based on the identification of the part.
Next to flexible part handling, also flexible visual quality inspection is critical to ensure the delivery of correct products. Normally, visual quality inspection systems are trained based on extensive datasets and can be easily optimized due to the mass-production of products. However, in low-volume production, these extensive datasets are often not available. Therefore, flexible visual quality inspection systems that can be trained based on small, incomplete datasets, and human input are developed and tested to be able to guarantee the required quality without the need of an extensive dataset.
Finally, the pilot will demonstrate optimal collaboration between humans and machines to achieve flexibility and to make that succeed, safety is critical. Therefore, the third focus point revolves around the creation of a safe production system where human and machine can work in a shared space were unsafe situations are preemptively identified and dealt with. Next to the physical safety of such a collaborative system, also the digital safety and mental well-being of the human in the loop is explored.
By: Jelle Keizer, Philips