Application of Deep Learning methods in robotics is difficult, due to their data and computational requirements. We study methods for visual perception, such as object recognition, detection and segmentation, that can be trained on-line, methods for object pose tracking that are robust to occlusions and can efficiently track moving objects. To support on-line learning, we devise strategies that allow robots to leverage on the interaction with humans and the environment to acquire training examples autonomously.

• Lombardi, M., Maiettini, E., Agnieszka, W., and Natale, L., Gaze estimation learning architecture as support to affective, social and cognitive studies in natural human-robot interaction, ACM Transactions on Human-Robot Interaction, 2025.
• Galliena, T., Apicella, T., Rosa, S., Morerio, P., Del Bue, A., and Natale, L., Embodied Image Captioning: Self-supervised Learning Agents for Spatially Coherent Image Descriptions, in IEEE/CVF International Conference on Computer Vision (ICCV), Honolulu, Hawaii, 2025.
• Taioli, F., Rosa, S., Castellini, A., Natale, L., Del Bue, A., Farinelli, A., Cristani, M., and Wang, Y., Mind the Error! Detection and Localization of Instruction Errors in Vision-and-Language Navigation, in IEEE/RSJ International Conference on Intelligent Robots and Systems, 2024
• Scarpellini, G., Stefano, R., Moreiro, P., Natale, L., and Del Bue, A., Look around and learn: improving object detection by exploration, in European Conference on Computer Vision, 2024
• Rosasco, A., Berti, S., Pasquale, G., Malafronte, D., Sato, S., Segawa, H., Inada, T., and Natale, L., ConCon-Chi: Concept-Context Chimera Benchmark for Personalized Vision-Language Tasks, in IEEE/CVF Conference on Computer Vision and Pattern Recognition 2024, Seattle, Washington, 2024.
• Ceola, F., Maiettini, E., Pasquale, G., Meanti, G., Rosasco, L., and Natale, L., Learn Fast, Segment Well: Fast Object Segmentation Learning on the iCub Robot, 2022, IEEE Transactions on robotics.
• Piga, N., Onyshchuk, Y., Pasquale, G., Pattacini, U., and Natale, L., ROFT: Real-time Optical Flow-aided 6D Object Pose and Velocity Tracking, IEEE Robotics and Automation Letters, vol. 7, no. 1, pp. 159-166, 2022. 
• Maiettini, E., Pasquale, G., Rosasco, L., and Natale, L., On-line Object Detection: a Robotics Challenge, 2020, Autonomous Robots, vol. 44, no. 5, pp. 739–757,
• Pasquale, G., Ciliberto, C., Odone, F., Rosasco, L., and Natale, L., Are we done with object recognition? The iCub robot’s perspective, Robotics and Autonomous Systems, vol. 112, pp. 260-281, 2019.

>> Video Gallery

We propose a computer vision-based system that uses AI methods for controlling the degrees of freedom of a prosthetic arm in a shared control paradigm, in various contributions we explored methods that adapt Computer Vision techniques, robot grasping and, more recently, generative AI method via imitation learning :

• Alessi, C., Vasile, F., Ceola, F., Pasquale, G., Boccardo, N., and Natale, L., HannesImitation: Grasping with the Hannes Prosthetic Hand via Imitation Learning, in IEEE/RSJ International Conference on Intelligent Robots and Systems, Hangzhou, China, 2025
• Vasile, F., Maiettini, E., Pasquale, G., Boccardo, N., and Natale, L., Continuous Wrist Control on the Hannes Prosthesis: A Vision-Based Shared Autonomy Framework, in IEEE-RAS International Conference on Robotics and Automation, Atlanta, GA, USA, 2025. 
• Stracquadanio, G., Vasile, F., Maiettini, E., Boccardo, N., and Natale, L., Bring Your Own Grasp Generator: Leveraging Robot Grasp Generation for Prosthetic Grasping, in IEEE-RAS International Conference on Robotics and Automation, Atlanta, GA, USA, 2025
• Vasile, F., Maiettini, E., Pasquale, G., Florio, A., Boccardo, N., and Natale, L., Grasp Pre-shape Selection by Synthetic Training: Eye-in-hand Shared Control on the Hannes Prosthesis, in IEEE/RSJ International Conference on Intelligent Robots and Systems, Japan, 2022.

Grasping and manipulation allows robots to perform useful tasks but also explore objects to acquire data for learning. We seek to improve grasping capabilities of humanoid robots using multi-modal information (i.e., vision, touch and proprioception). Recent contributions include learning by imitation using generative methods, extraction of tactile features for object pose estimation, force estimation. In past research we have proposed methods for shape completion, object modelling and grasping using superquadrics, and hand detection for precise visual control and methods for benchmarking. 

• Rosasco, A., Ceola, F., Pasquale, G., and Natale, L., KDPE: A Kernel Density Estimation Strategy for Diffusion Policy Trajectory Selection, in Conference on Robot Learning, Seoul, Korea, 2025.
• Puang, E. Y., Ceola, F., Pasquale, G., and Natale, L., PCHands: PCA-based Hand Pose Synergy Representation on Manipulators with N-DoF, in IEEE-RAS Humanoids, Seoul, Korea, 2025.
• Ceola, F., Rosasco, L., and Natale, L., RESPRECT: Speeding-up Multi-fingered Grasping with Residual Reinforcement Learning, IEEE Robotics & Automation Letters, vol. 9, no. 4, 2024.
• Caddeo, G. M., Maracani, A., Alfano, P. D., Piga, N. A., Rosasco, L., and Natale, L., Sim2Surf: a Sim2Real Surface Classifier for Vision-Based Tactile Sensors with a Bilevel Adaptation Pipeline, IEEE Sensors Journal, vol. 25, no. 5, pp. 1558-1748, 2025.
• Shahidzadeh, A., Caddeo, G., Alapati, K., Natale, L., Fermüller, C., and Aloimonos, Y., FeelAnyForce: Estimating Contact Force Feedback from Tactile Sensation Vision-Based Tactile Sensors, in IEEE-RAS International Conference on Robotics and Automation, Atlanta, GA, USA, 2025
• Caddeo, G., Piga, N., A., Bottarel, F., and Natale, L. Collision-aware In-hand 6D Object Pose Estimation using Multiple Vision-based Tactile Sensors, in Proc. IEEE International Conference on Robotics and Automation London, UK, 2023.
• Rosasco, A., Berti, S., Bottarel, F., Colledanchise, M., and Natale, L., Towards Confidence-guided Shape Completion for Robotic Applications, in Proc. IEEE-RAS International Conference on Humanoid Robots, 2022.
• Bottarel, F., Altobelli, A., Pattacini, U., and Natale, L., GRASPA-fying the Panda: Easily Deployable, Fully Reproducible Benchmarking of Grasp Planning Algorithms, IEEE Robotics and Automation Magazine, 2023. 
• Vezzani, G., Pattacini, U., Pasquale, G., and Natale, L., Improving Superquadric Modeling and Grasping with Prior on Object Shapes, in Proc. IEEE-RAS International Conference on Robotics and Automation, Brisbane, Australia, 2018, pp. 6875-6882. 
• Fantacci, C., Vezzani, G., Pattacini, U., Tikhanoff, V., and Natale, L., Markerless visual servoing on unknown objects for humanoid robot platforms, in Proc. IEEE-RAS International Conference on Robotics and Automation, Brisbane, Australia, 2018, pp. 3099-3106.

>> Video Gallery

In robotics, a large body of work has been devoted to the development and study of software architectures and software engineering techniques specifically tailored to robotics systems. We develop YARP (www.yarp.it) an open source middleware that supports concurrent execution of components on a cluster of computers. In recent work we have been looking at BehaviorTrees as a method for designing and implementing reactive and modular robot behaviors, and formal methods for static and runtime verification of robot behaviors.

• Bernagozzi, S., Faraci, S., Ghiorzi, E., Pedemonte, K., Natale, L., and Tacchella, A., Code Generation and Monitoring for Deliberation Components in Autonomous Robots, in IEEE/RSJ International Conference on Intelligent Robots and Systems, Hangzhou, China, 2025
• Bernagozzi, S., Faraci, S., Ghiorzi, E., Pedemonte, K., Ferrando, A., Natale, L., and Tacchella, A., Model-based Verification and Monitoring for Safe and Responsive Reactive Robots, in IEEE International Conference on Simulation, Modeling, and Programming for Autonomous Robots, 2025.
• Ghiorzi, E., Colledanchise, M., Piquet, G., Tacchella, A., and Natale, L., Learning linear temporal properties for autonomous robotic systems, IEEE Robotics and Automation Letters, 2023. 
• Colledanchise, M., and Natale, L., Handling Concurrency in Behavior Trees, IEEE Transactions on robotics, , vol. 38, no. 4, pp. 2557-2576, 2022. 
• Colledanchise, M., and Natale, L., Analysis and Exploitation of Synchronized Parallel Executions in Behavior Trees, in Proc. IEEE/RSJ International Conference on Intelligent Robots and Systems, Macau, China, 2019, pp. 6399-6406. 
• Natale, L., Paikan, A., Randazzo, M., Domenichelli, D.E.,The iCub Software Architecture: Evolution and Lessons Learned, Frontiers in Robotics and AI, vol. 3, no.24, 2016, doi: 10.3389/frobt.2016.00024.

>> Video Gallery