COMING SOON! PQDT Open is getting a new home!

ProQuest Open Access Dissertations & Theses will remain freely available as part of a new and enhanced search experience at

Questions? Please refer to this FAQ.

Dissertation/Thesis Abstract

FLIHI: Fuzzy Logic Implemented HIll-based muscle model
by O'Brien, Amy Joy, D.Sc., The George Washington University, 2006, 202; 3304088
Abstract (Summary)

FLIHI, the Fuzzy Logic Implemented HIll-based muscle model, is a model of skeletal muscle that is implemented in fuzzy logic rather than high-order dynamic differential state equations. Archibald V. Hill's empirical muscle model proposed in 1938 comprises a passive series elastic element (SE, e.g., tendon), a passive parallel elastic element (PE, e.g., ligament), and an active contractile element (CE) that produces force in response to neural drive. All three elements are nonlinear. FLIHI implements a Hill-based model in fuzzy logic, which allows computers to handle human language terms (e.g., hot/cold) and partial set membership (e.g., shades of gray plus black or white).

Original contributions made in this dissertation include FLIHI and the Generalized Fuzzy Inference Engine (GFIE) that runs in MATLAB and was developed for FLIHI. GFIE has been field tested in a graduate biomedical engineering course. GFIE demonstrated that a very simple fuzzy inference system can reproduce a 1/(τs+1) filter.

FLIHI was evaluated for three classic single-muscle tasks using theoretical, physiologically appropriate electromyogram (EMG), muscle length, and/or external load inputs. FLIHI was compared to Virtual Muscle, a mature MATLAB-Simulink Hill-style model, for three validation tasks. For a sudden perturbation task, FLIHI estimated force with 10.7% RMS error compared to Virtual Muscle; the corresponding RMS EMG yielded 46% RMS error. A linear approximation of FLIHI was proven stable for a sample quick-release task. The robustness of FLIHI is indicated by its insensitivity to almost all of its fuzzy modeling parameters. The FLIHI baseline proves that this modeling approach is feasible. FLIHI can be made neurofuzzy, thus combining neural networks' ability to learn with fuzzy logic's ability to express the model in human language form.

Future work for FLIHI includes addressing limitations by improving performance during submaximal EMG; defining rest, or optimal, length dynamically; replacing the simple, single-stage linear low-pass filters used to simulate CE and SE damping with biologically, physiologically appropriate nonlinear CE and SE damping; and testing/refining behavior for lengthening muscle. Future work also includes using FLIHI with biological signals and muscle/joint tasks (e.g., in a clinical setting).

FLIHI has potential applications in research, medicine, and academia.

Keywords. fuzzy logic, Hill muscle model, GFIE, FLIHI

Indexing (document details)
Advisor: Carroll, Robert L., Szu, Harold H.
Commitee: De Witte, Joseph, Loew, Murray H., Myklebust, Barbara M., Zara, Jason M.
School: The George Washington University
Department: Electrical Engineering
School Location: United States -- District of Columbia
Source: DAI-B 69/03, Dissertation Abstracts International
Subjects: Biomedical research, Electrical engineering, Systems design
Keywords: FLIHI, Fuzzy logic, GFIE, Hill muscle model
Publication Number: 3304088
ISBN: 978-0-549-52133-4
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy