I received the Ph.D. degree in Biomedical, Electrical, and System Engineering from the University of Bologna in 2021 and the M.Sc. degree with honors in Control Systems Engineering from the University of Padova in 2016.
From 2021 to 2023, I worked as a postdoc at the University of Bologna. Currently, I am a postdoc at the Robotics and Mechatronics (RaM) group at the University of Twente.
As a robotics researcher, I focus on developing effective solutions for physical robot interaction in unstructured environments by integrating machine learning and optimization with physical knowledge, such as energy flows and geometric structures.
My current research interests include: 1) advancing control frameworks that combine performance optimization with energy-aware safety for human- and environment-robot interactions; and 2) addressing the complexity and unpredictability of robotic systems through robust perception and intelligent automation pipelines.
Follow the link to see my Curriculum Vitae
Organisaties
My research focuses on developing effective solutions for physical robotic interaction in unstructured environments. These include highly unpredictable scenarios such as the manipulation of deformable objects or direct interaction with humans. In such cases, the complexity and uncertainty often cannot be fully captured by analytical models alone, while relying solely on data-driven approaches may not be sufficient to ensure safe and effective physical interaction. To address these challenges, my work integrates physical structure and system-theoretic principles with optimization and learning-based algorithms.
Machine learning is increasingly seen as a promising approach for designing controllers capable of handling complex tasks in unknown environments. However, its practical use in robotic interaction is challenged by limitations including low reliability, high data demands, and limited interpretability. To overcome these challenges, it is crucial to embed structure into the control architecture by incorporating prior physical knowledge, thereby enhancing both reliability and data efficiency. One promising approach to achieving this integration is through energy-based methods. Rooted in fundamental physical principles, these methods provide a modular and interpretable framework for modeling and regulating energy flow and balance across subsystems. This structured approach is particularly valuable for physical interaction scenarios. Moreover, the inherent flexibility of these methods makes them highly compatible with learning/optimization techniques. This perspective shapes a core element of my research strategy. My work has long focused on learning-based methods for robot sensing and manipulation of deformable objects, where complex dynamics make explicit modeling overly simplistic or computationally impractical, often causing model-based control to fail in real-world scenarios. Motivated by the goal of finding new ways to incorporate prior physical knowledge into learning-based controllers for robotic interaction, I developed a strong interest in energy-based methods. Consequently, my research has increasingly focused on combining energy-based approaches with learning and optimization techniques to enhance control efficiency, flexibility, reliability, safety, and interpretability in robotic tasks. To achieve this, I follow two complementary directions: on one hand, I integrate learning-based control algorithms with energy-awareness; on the other, I extend traditional energy-based control frameworks to enable seamless fusion with optimization techniques and data-driven strategies.
Complementing this control-oriented work and to achieve my overarching objective, I dedicate part of my research to developing robust and efficient perception systems, which are critical in any real-world robotic application operating in unstructured and dynamic environments. I develop methods to extract meaningful and reliable features that simplify and manage the complexity inherent in the environment and interaction dynamics, enabling more effective decision-making, planning and control.
Onderzoeksprofielen
Adres
Universiteit Twente
Carré (gebouwnr. 15), kamer C3530
Hallenweg 23
7522 NH Enschede
Universiteit Twente
Carré C3530
Postbus 217
7500 AE Enschede
Scan de QR-code of