Companies are seeing their supply chains grow and become increasingly complex as customer demands evolve. In today's market, being able to produce customized products with a short response time provides a key competitive advantage.
Traditional industrial robot arms, that have dominated large-scale manufacturing for the past decades, usually keep performing the same task after they have been assigned to a manufacturing line. This can be costly because programming them is time-consuming and often done by skilled and dedicated experts. Also integrating changes into manufacturing lines to implement a new process incurs additional costs. Consequently, as the demand for flexible manufacturing processes increases, more and more collaborative robots, or cobots, are expected to be deployed.
Cobots are designed to work alongside humans without the need for physical separation and can improve the safety, work quality, and productivity of manufacturing processes by acting as flexible partners to humans. However, by having cobots share their workspace with humans and because of their inherent need for constant reprogramming, several issues arise. Unlike pure automation, collaborative tasks require robots to comprehend their environment and adapt their actions accordingly. Moreover, many commercially available cobots are shipped with specific user-friendly programming environments, sacrificing some of the flexibility that would be characteristic of a system meant to be programmed by an expert.
“Knowledge-driven models and reasoning capabilities could allow robots to gain a better understanding of what they are doing and what people are doing around them, which is a crucial element in collaborative scenarios,” says Alexandre Angleraud.
In his doctoral dissertation, Angleraud investigated ways to offer flexible human-robot collaboration capabilities in manufacturing environments while maintaining user-friendliness. He focused on building a system that would allow operators to model tasks using concepts understandable to both humans and robots. These models are then integrated with planning algorithms, thus serving as instructions to program robots.
“The ability to encode task knowledge and facilitate high-level planning for robotic tasks, thus increasing predictability and transparency in manufacturing processes, is essential for collaborative robots,” Angleraud adds.
Public defence on Wednesday 22 May
The doctoral dissertation of MSc (Tech) Alexandre Angleraud in the field of Automation Science and Engineering titled “Knowledge-Based systems For Human-Robot Collaborative in Manufacturing Environments” will be publicly examined at the Faculty of Engineering and Natural Sciences at Tampere University at 12 o’clock on Wednesday 22 May 2024 at the Hervanta Campus in Festia building, auditorium Pieni Sali 1 (Korkeakoulunkatu 8, Tampere).
The Opponents will be Professor Dimitrios Chrysostomou from Aalborg University, Denmark and Professor David Romero from the Tecnológico de Monterrey, Mexico. The Custos will be Assistant Professor Roel Pieters from the Faculty of Engineering and Natural Sciences at Tampere University.