Motor primitive-based control for lower-limb exoskeletons

Authors:

Abstract
Assistive technology forecasts better autonomy for people with lifelong disabilities and for the elderly facing motor decline. As the population of developed countries is becoming greyer, there is thus a high probability of observing a significant increase in the demand for assistive locomotion devices. Designing the controller for such devices is not trivial, and requires both to set up a compliant framework and to manage the cross adaptation between the device and its user (the so-called interface). Taking inspiration from neuromechanical principles governing human locomotion is an approach which is currently investigated by several groups to address these challenges. In this contribution, we develop a bio-inspired controller based on motor primitives for lower-limb assistance. Motor primitives are combined in order to generate muscle stimulations that serve as neural inputs to a virtual musculoskeletal model, which in turn produces reference sagittal joint torques. This controller is also adaptable to different configurations of lower-limb exoskeletons and sensory inputs. This paper further reports the validation of the controller through two different experiments. Results showed the capability of the controller to provide coherent joint torque profiles for different configurations and locomotion tasks. These profiles could be used as assistive torques by scaling them as a function of the level of support being required, and transmitting them through an assistive device like an exoskeleton.

This publication has 24 references indexed in Scilit: