As robots are leaving safety fences and starting to share workspaces and even living spaces with humans, they need to dynamically adapt to unpredictable interactions with people and guarantee safety at every moment. On the rapidly growing market for collaborative robots, safety is ensured through specific technologies such as force limitations by design or contact detection and stopping relying on force measurements. Humans, however, possess whole-body awareness drawing on dynamic, context-dependent fusion of multimodal sensory information, which makes them adaptive, flexible, and versatile.
In my research, I strive to add important new dimensions to physical human-robot interaction. The first goal is to make the robot aware of its whole body rather than the end-effector only. We use artificial electronic skins to cover the whole robot's body - this forms the basis for its "whole-body awareness". Second, inspired by the way the brain represents the body, we strive to make these representations multimodal. Using vision, or audition, the body space can be extended to a surface around the body (so-called peripersonal space), facilitating collision avoidance and contact anticipation, eventually leading to safer and more natural interaction of the robot with objects, including humans.
HARMONIOUS — Human-Like Reactive Motion Control and Multimodal Perception for Humanoid Robots
What does it take to play a board game with a humanoid robot? We have developed HARMONIOUS, a real-time reactive motion control system for a humanoid which uses two arms to play the game while perceiving and avoiding contact with humans through touch, proximity, and vision.
Rozlivek, J.; Roncone, A.; Pattacini, U. & Hoffmann, M. (2025), 'HARMONIOUS – Human-like reactive motion control and multimodal perception for humanoid robots', IEEE Transactions on Robotics 41, 378 - 393.
Check the dedicated webpage for more details.
Adaptive collision sensitivity to boost co-bot productivity
Hit it hard with a co-bot!
Collaborative robots in the power and force limiting regime are allowed to collide with human operators, provided that they do not hit them too hard. As a consequence, the robots need to move slowly, hampering productivity. Here, with a co-bot covered with sensitive skin, we adaptively adapt the collision thresholds such that the robot does not need to stop when the collision is mild. The robot effective mass concept is employed to set the adaptive collision thresholds on the run - depending on the robot speed and configuration.
Rustler, L.; Misar, M. & Hoffmann, M. (2024), Adaptive Electronic Skin Sensitivity for Safe Human-Robot Interaction, in 'IEEE-RAS International Conference on Humanoid Robots (Humanoids)', IEEE. [pdf-arxiv][youtube-video][code][data]
Effect of active and passive protective soft skins on collision forces in HRI
Svarny, P., Rozlivek, J., Rustler, L., Sramek, M., Deli, Ö., Zillich, M. and Hoffmann, M. (2022), 'Effect of active and passive protective soft skins on collision forces in human–robot collaboration', Robotics and Computer-Integrated Manufacturing 78, 102363. [DOI - sciencedirect][arxiv-pdf][dataset][youtube-video]
Rustler, L.; Misar, M. & Hoffmann, M. (2024), Adaptive Electronic Skin Sensitivity for Safe Human-Robot Interaction, in 'IEEE-RAS International Conference on Humanoid Robots (Humanoids)', IEEE. [pdf-arxiv]
3D Collision-Force-Map - Empirical measurement and data-driven modeling of impact forces warrants safety and boosts performance
Svarny, P.; Rozlivek, J.; Rustler, L. & Hoffmann, M. (2021), '3D Collision-Force-Map for Safe Human-Robot Collaboration', IEEE International Conference on Robotics and Automation (ICRA), 3829-3835. [IEEE Xplore][pdf-arxiv][youtube-video]
Speed and separation monitoring together with power and force limiting (per ISO/TS 15066) regimes in a single collaborative robot scenario
Svarny, P.; Tesar, M.; Behrens, J. K. & Hoffmann, M. (2019), Safe physical HRI: Toward a unified treatment of speed and separation monitoring together with power and force limiting, in 'Intelligent Robots and Systems (IROS), 2019 IEEE/RSJ International Conference on', IEEE, pp. 7574-7581. [IEEE Xplore][arxiv][youtube video]
Robot speed modulation for safe Human-Robot Collaboration
Zardykhan, D.; Svarny, P.; Hoffmann, M.; Shahriari, E. & Haddadin, S. (2019), Collision Preventing Phase-Progress Control for Velocity Adaptation in Human-Robot Collaboration, in 'Humanoid Robots (Humanoids), 2019 IEEE-RAS 18th International Conference on', IEEE, pp. 282-289.[IEEE Xplore] [youtube video]
Work in collaboration with Sami Haddadin's group from Technical University of Munich.
Reaching with anticipatory avoidance of human using skeleton extraction
Work in collaboration with IIT Genoa and Yale University.
Nguyen, P. D.; Hoffmann, M.; Roncone, A.; Pattacini, U. & Metta, G. (2018), Compact real-time avoidance on a humanoid robot for human-robot interaction, in 'HRI ’18: 2018 ACM/IEEE International Conference on Human-Robot Interaction', ACM, New York, NY, USA, pp. 416-424. [ACM digital library][arxiv][youtube video]
Learning peripersonal space representation from visuo-tactile association
The goal is to extend the model of the body itself to the space surrounding it, while preserving the key role of the tactile modality. The first step is thus acquiring a representation of this so-called peripersonal space. In a second step, contacts over the whole body surface will be perceived and can be handled by the robot depending on the task – they can be kept within limits or they can be sought. This will greatly increase the range of configurations available to reach in cluttered spaces, while ensuring safety for the robot itself as well as its environment at all times.
Roncone, A.; Hoffmann, M.; Pattacini, U.; Fadiga, L. & Metta, G. (2016), 'Peripersonal space and margin of safety around the body: learning tactile-visual associations in a humanoid robot with artificial skin', PLoS ONE 11(10), e0163713. [OPEN ACCESS - doi link]
Roncone, A.; Hoffmann, M.; Pattacini, U. & Metta, G. (2015), Learning peripersonal space representation through artificial skin for avoidance and reaching with whole body surface, in 'Intelligent Robots and Systems (IROS), 2015 IEEE/RSJ International Conference on', pp. 3366-3373. [IEEE Xplore] [postprint]