Description

The ability of robots to interpret human signals remains a key challenge in the development of efficient and adaptive autonomous systems. Building on previous studies in philosophy and sociology on the definition of human intention, how these intentions are transformed into actions, and their application to robotics, this project will follow the Perception-Intention-Action model for Robot-Human interactions. This project proposes an innovative approach that combines computer vision, natural language processing, and deep learning to achieve a more intuitive and fluid interaction between humans and robots. The aim is to develop artificial intelligence-based learning models that allow robots not only to recognize but also to anticipate human intentions and act proactively in dynamic environments. This is expected to improve human-robot interaction in applications such as industrial automation and social robotics.

Amb el suport del programa d'ajuts predoctorals FI-STEP (número d'expedient: 2025 STEP 00188) del Departament de Recerca i Universitats de la Generalitat de Catalunya i el cofinançament pel Fons Social Europeu Plus

The work is under the scope of the following projects:

• LENA: Lifelong navigation learning using human-robot interaction (web)
• TORNADO: foundaTion mOdels for Robots that haNdle smAll, soft and Deformable Objects (web)