Bionics (2006–2016)

The abilities of humans provide both a goal and inspiration, for the development of robotic systems. We try to understand these abilities to meld them into technical systems and to assist the human in various settings.

In order to understand the human “model”, we acquire bio-data like muscle activity, exerted forces, and performed motions by recording different measures, e.g., electromyography (EMG) and ultrasound. This data can expose very different sets of information, including the behavior of the subject, his intentions, or abilities. Our goal is to process the data in a manner that lets us extract the information of interest. Once the information is extracted we can use it for different purposes.

On the one hand, we want to build and optimize technical systems that—under a specific focus—show performance similar to the human, an example being bi-articular muscles allowing an enriched capability to control arm position, force and stiffness.
On the other hand, we want to use the decoded information as control commands for robotic systems. With this approach we focus on the control of prosthetic hands, assistive robotic arms for people with limited upper limb function, and the control of robotic devices for post-stroke rehabilitation.

Bio-data exploitation for supportive robotics

Assistive robotics, human-machine interfaces for prosthetics, imaging EMG (iEMG) and rehabilitation robotics

Intent detection via biological signals

Online, real-time examination and interpretation of signals gathered from the body (electromyography, surface pressure information, gaze tracking, gesture recognition, etc.) with the aim of detecting a human subject's intention: when to grasp, how to grasp, what body parts to move and how (position, speed, force, torque), and possibly also for the entire upper limb - wrist, elbow, shoulder.