Authors:
Anh Mai
and
Sesh Commuri
Affiliation:
University of Oklahoma, United States
Keyword(s):
Intelligent Control, Prosthetic Foot, Optimization.
Related
Ontology
Subjects/Areas/Topics:
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Neural Networks Based Control Systems
;
Optimization Algorithms
Abstract:
The ability to control the prosthetic ankle joints of below-knee amputees is a challenging problem due to the lack of adequate mathematical models, the variations in the gait in response to the environment, sensor noise, and unknown intent of users. Artificial ankle joints are required to exhibit variable stiffness based on the gait and aid in locomotion as well as stability of the individual. It is desirable for control strategies for such ankle joints to adapt in real-time to any variations in the gait, have robust performance, and optimize specified performance indices relating to efficiency of the gait. In this paper, we investigate the potential of Direct Neural Dynamic Programming (DNDP) method for learning the gait in real-time and in generating control torque for the ankle joint. The residual limb is first represented by a link-segment model and the kinematic patterns for the model are derived from human gait data. Then augmented training rules are proposed to implement the D
NDP-based control to generate torque which drives the prosthetic ankle joint along the designed kinematic patterns. Numerical results show that the DNDP controller is able to maintain stable gait with robust tracking and reduced performance cost in spite of measurement/actuator noises and variations in walking speed.
(More)