What did you find out?

We have developed a unified framework for modeling a teleoperation system. This consists of an upper limb exoskeleton and the service robot GARMI. The dynamics of this tendon-driven exoskeleton are modeled via two subsystems: Via the remote actuation system and via the exoskeleton with its rigid limbs. In addition, we have derived the dynamics of GARMI as a “mobile bimanipulator platform”. That means we simplify this system as a mobile base with two arms mounted on it for manipulation. To ensure the safe operation of GARMI, we use controllers that avoid self-collisions and know the limits of joint mobility.

What challenges did you have to face during the project?

One of the biggest challenges was modeling the friction of the exoskeleton's Bowden cable, which is influenced by the pretension and curvature of the cable. The combination of compliance and high friction within the cable complicates the dynamics and modeling of the remote control. Another challenge was to achieve stable linearized dynamics of GARMI. Overall, the high number of degrees of freedom in the system requires an efficient calculation for the real-time controller.

What practical use do your research results have?

We have now created the prerequisites for a controller design for teleoperation. In doing so, we can incorporate the dynamics of the exoskeleton. The result: interaction forces between the human arm and the exoskeleton are minimized, thus improving the mechanical transparency and haptic experience of the exoskeleton.

Publication on the IROS:

A Tactile Lightweight Exoskeleton for Teleoperation: Design and Control Performance; 
Moein Forouhar, Hamid Sadeghian, Daniel Pérez-Suay, Abdeldjallil Naceri; IROS 2024 

Interview: Sandra von Löbbecke