Nature provides animals and human beings with the ability to efficiently and safely interact with the environment. While vision, auditory and vestibular systems deliver mainly sensory information for navigation and path planning, the somatosensory system with its cutaneous (tactile) and proprioceptive receptors provide feature rich information whenever there is contact with the environment. The somatosensory system allows very fast and save cognition and reaction and is - as soon as the system is in contact with the environment - more reliable than the other parts of the perceptive system: Somatosensory information cannot easily be misinterpreted or occluded. For save, reflex like reactions very fast cognition and responses are needed. Therefore the somatosensory system processes, combines and organizes information already at receptor level and keeps the organization (somatosensory map) up to the highest processing levels in the brain. Also the nervous system conveys and processes information only if there is new information. This asynchronous, event based principle reduces the required transmission bandwidth and processing power and makes the system fast and efficient.
My research focuses on biologically plausible principles for fast and efficient tactile cognition with whole-body multimodal robotic skin and thus will enable robots to new kinds of whole-body interactions and fast reflex-like responses. I let myself inspire from nature while the goal is not to copy nature but to understand the principles nature uses to solve problems that we also have in engineering and robotics. In this way we can benefit from results of the biological optimization process which had millions of years to solve common problems while adjusting the found principles to the special needs of our artificial systems.
I’m currently working on making use of the local distributed processing units of our multimodal self-organizing robotic skin and on implementing efficient event generation in the skin cells themselves for reducing data load and processing power and improving temporal resolution.
tactile sensing, sensors, sensor fusion, electronics, micro-controllers, FPGAs, perception, cognition, the nervous system, the cortex, bio-inspired engineering
- A Robust and Efficient Dynamic Network Protocol for a large-scale artificial robotic skin. 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018, 1600 - 1605 mehr… BibTeX Volltext (mediaTUM)
- Integration of Advanced Robotic Technologies for Rapidly Deployable Industrial Robots. IEEE Transactions on Industrial Informatics (4), 2018 mehr… BibTeX Volltext ( DOI ) Volltext (mediaTUM)
- Efficient Distributed Torque Computation for Large Scale Robot Skin. 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018, 1593 - 1599 mehr… BibTeX Volltext (mediaTUM)
- Enhancing Biped Locomotion on Unknown Terrain Using Tactile Feedback. IEEE-RAS 18th International Conference on Humanoid Robots (Humanoids), 2018, 47 - 52 mehr… BibTeX Volltext ( DOI )
- TOMM: Tactile Omnidirectional Mobile Manipulator. 2017 IEEE International Conference on Robotics and Automation (ICRA), 2017Singapore, May 29 - June 3, 2017, 2441-2447 mehr… BibTeX
- Efficient Event-Driven Reactive Control for Large Scale Robot Skin. 2017 IEEE International Conference on Robotics and Automation (ICRA), 2017, 394 - 400 mehr… BibTeX Volltext (mediaTUM)
- Event-based signaling for large-scale artificial robotic skin - Realization and performance evaluation. 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2016, 4918 - 4924 mehr… BibTeX Volltext (mediaTUM)
- Robotic technologies for fast deployment of industrial robot systems. IEEE Industrial Electronics Conference (IEEE IECON2016), 2016 mehr… BibTeX Volltext (mediaTUM)
- From Multi-modal Tactile Signals to a Compliant Control. IEEE-RAS International Conference on Humanoid Robots 2016, 2016 mehr… BibTeX Volltext (mediaTUM)
- General Recognition Models Capable of Integrating Multiple Sensors for Different Domains. IEEE-RAS International Conference on Humanoid Robots 2016, 2016 mehr… BibTeX Volltext (mediaTUM)
- Extracting general task structures to accelerate the learning of new tasks. IEEE-RAS International Conference on Humanoid Robots 2016 , 2016 mehr… BibTeX Volltext (mediaTUM)