Multi-Input Transcutaneous Neuromuscular Electrical Stimulation for Control of the Lower Limb
| Author(s): R. Nguyen |
Conference/Journal: |
| Abstract: Muscle contractions in the body are caused by electrical impulses that reach the muscle through nerves. These impulses ordinarily originate from within the body, for example, from the brain. Muscle contractions can also be induced artificially by applying electric current near a nerve. It can even be applied non-invasively through a pair of electrodes, the active and reference, placed on the surface of the skin near the targeted muscle, a technique termed transcutaneous neuromuscular electrical stimulation (TNMES). TNMES is even applicable to individuals suffering paralysis so long as nerve innervation to the muscle is intact. This is often still the case after spinal cord injury or stroke. Thus, a major clinical application of TNMES is to induce muscle contractions in such individuals for functional tasks such as grasping and walking. This coordinated use of TNMES requires precise modulation of pulsed electric currents and is termed functional electrical stimulation (FES). It can have both orthotic and therapeutic benefits for individuals. Three major challenges in the widespread use of FES in the lower limb are overcoming muscle fatigue, providing better control of limb motion, and integration of voluntary effort in a safe environment to improve therapeutic outcomes. Each of these challenges is addressed in one of the three parts of this thesis. The first part of this thesis addresses the limitation of the rapid onset of muscle fatigue during FES caused by localized high frequency stimulation of muscle fibers. To mitigate this fatigue, a technique termed spatially distributed sequential stimulation (SDSS) is used. During SDSS, the area of stimulation is sequentially changed, pulse by pulse, using multiple surface electrodes that cover the same area as would be covered by the one active electrode during conventional single electrode stimulation (SES). Three studies showing the effectiveness and elucidating the physiological mechanism of SDSS are presented. The first study in this part demonstrates the effective use of SDSS in paralyzed triceps surae muscles of an individual with complete spinal cord injury. The second study extends this finding to the four major lower limb muscle groups in able-bodied individuals, as well as in individuals with spinal cord injury. The fatigue-reducing ability of SDSS is observed in all muscles except knee flexors of able-bodied individuals, and in all muscles tested in individuals with SCI. The third study elucidates the mechanism underlying SDSS. In particular, electromyography (EMG) is used to show that, as hypothesized, different sets of muscle fibers are indeed being activated alternately by different stimulation electrodes. The second part of this thesis addresses the lack of effective methods to precisely control limb motions during FES. Experiments are carried out to investigate the effectiveness of proportional-integral-derivative (PID) control in controlling ankle dorsiflexion. This controller is limited due to the non-linear nature of the task and thus P-type iterative learning control (ILC) is investigated. ILC is particularly suited for FES during rehabilitation where the task is repetitive. In particular, P-type ILC changes the stimulus intensity during one iteration proportional to the error that occurred in the previous iteration. Successful experimental results are obtained using P-type ILC to control ankle rotation. FES often targets joints with multiple degrees of freedom, however the control of these degrees, using multiple electrodes, is an area of research that has received little attention due to its difficulty. Experiments are conducted to assess the appropriateness of P-type ILC to control both ankle dorsiflexion and inversion. Even though one successful experimental result is demonstrated, consistent results are not obtained. Thus, another method of ILC based on dynamic linearization is investigated in simulation to control knee angle and successful results are obtained. The focus is on the knee, instead of the ankle, since better models exist. Although not investigated further, this method of ILC could be extended and tested in experiments to control the two degrees of freedom of the ankle. The third and final part of this thesis addresses limitations of therapeutic use of FES: muscles often being too weak to provide extended functional use and the requirement of extensive clinician support to assist in movement and provide safety during walking. Robot-assisted gait therapy can provide a safe and easy environment within which FES can be used to aid walking. FES complements well robot-assisted gait therapy where lack of muscle activity can be an issue. A method using P-type ILC to control the knee muscles using FES during robot-assisted gait is investigated and results show that increased muscle activity is possible. It has been shown that the therapeutic effect of FES is improved further when synchronized with voluntary effort. Thus, a method within robot-assisted gait to trigger FES based on user effort, measured through EMG, is developed. This method is used for treatment of three stroke patients, whose improvement is shown through various functional assessments. In summary, this thesis presents three methods to improve FES in the lower limb: the use of multiple electrodes to reduce muscle fatigue, controllers that allow more precise control of joints, and integration of EMG and robot-assisted gait with FES for better rehabilitative outcomes. | Year: 2015 |
| Type of Publication: (03)Ph.D. Thesis | |
| Supervisor: M. Morari | |
% Autogenerated BibTeX entry
@PhDThesis { Xxx:2015:IFA_5350,
author={R. Nguyen},
title={{Multi-Input Transcutaneous Neuromuscular Electrical
Stimulation for Control of the Lower Limb}},
school={},
year={2015},
address={},
month=jul,
url={https://old.control.ee.ethz.ch/index.cgi?page=publications;action=details;id=5350}
} | |