[home] [Personal Program] [Help]

---

Session: Prostheses
Session starts: Thursday 24 January, 13:30
Presentation starts: 13:45
Room: Lecture room 558

---

Letian Wang ()
Edwin van Asseldonk ()
Herman van der Kooij ()

Abstract:
ABSTRACT Assisted by lower extremity exoskeletons, walking disabled people can stand up again and regain their walking abilities. It is a wise design to underactuate certain joints of the exoskeleton in order to achieve lighter weight, lower power consumption and as a consequence, better mobility. A reasonable choice is to leave the ankle joints passive[1, 2]. In this way, the gait can still be realized by actuating the hip and knee joints while less weight is put at the end of the leg, where has the most effect on the inertia of the leg. However, the sacrifice of underactuated ankle is the performance on the balance control especially during stance. We believe, human subject inside such exoskeleton with unimpaired upper body abilities , can assist the exoskeleton to enhance the balance using the upper body movement given properly designed exoskeleton controllers. In this way, less assistive devices, e.g. crutches, are needed for assisting the balance. Experiments are designed to test the ability of human subjects to assist the controller to balance a virtual human-exoskeleton system under random perturbation (push or pull in sagittal and lateral plane) during stance. The virtual exoskeleton has three degrees of freedom (DoFs) at hip, one DoF at knee and one DoF at ankle. Hip endo-exo and ankle dorsiflexion are passive while the rest joints are actively controlled by PD controllers and model predictive controllers[3]. The human subject is given visual feedback of the posture of the virtual human-exoskeleton. A mapping from the trunk of the human subject to the trunk of the virtual human is provided so that the human subject is able to manipulate the centre of mass of the virtual human-exoskeleton. Results show that the virtual human-exoskeleton can withstand larger perturbation under the balance assistance of the human subject compared the case of no balance assistance. The results also suggest the performance of the balance assistance by human subjects can be increased by training and learning.