• clustering
• computer vision
• Dynamic Movement Primitives
• grasping affordances
• human intention
• learning from demonstration
• robotic manipulation
• Sensory-Motor Contingencies

Industry 5.0 represents an evolution from Industry 4.0, embracing a more human-centric and sustainable approach prioritising collaboration between humans and machines. In this context, enhancing human-robot interaction becomes essential. This thesis introduces a framework that combines the concept of affordances with Learning from Demonstration (LfD), enabling robots to learn, adapt, and generalise behaviours in dynamic environments.
The proposed approach extends the concept of Sensory-Motor Contingencies (SMCs) by integrating affordances, empowering robots to identify potential actions associated with objects and adapt them to specific contexts and operator preferences. Leveraging advanced vision techniques and methods typical of robotic control and machine learning, the framework supports the flexible selection and execution of specific tasks. This work contributes to the development of more intuitive and accessible robotic systems, aiming to bridge the gap between human intent and robotic action.