Learning Contact-Rich Tasks for the Industry 5.0

With industry 4.0, manufacturing transitioned from mass production to mass customization, thus requiring robotic systems to constantly adapt to new requirements. In this context, traditional programming techniques might not be sufficient to cope with contact-rich tasks, especially when the surface to be treated is complex. With this work, we review advancements in Learning from Demonstration techniques that tackle these issues by injecting human knowledge in the robotic systems.