Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
TEMPLATE DESIGN © 2009
www.PosterPresentations.com
A Two-Bit Binary Control System for an Orthotic, Hand-
Assistive ExoskeletonDonald J. Bucci, Sana F. Fathima, Brett F. BuSha
The College of New Jersey: Department of Electrical and Computer Engineering2000 Pennington Road, Ewing, NJ, 08628
Abstract
157
A combined 450,000 Americans suffer from multiple sclerosis, chronic carpal tunnel
syndrome, and muscular dystrophy, debilitative disorders that result in a loss of muscle
strength and dexterity within the afflicted individual. A controlled orthotic exoskeleton
presents a solution by amplifying a user’s residual strength to restore precision pinch
and/or power grasp motions. In order to optimize system response time and control
accuracy, a two-bit binary state control algorithm was designed using LabVIEW v8.5. Thestates of the control system were refined using positional and object-sensing information
supplied via negative feedback from Hall effect angular sensors and force sensing
resistors. Forearm EMG was recorded and used to characterize hand strength with and
without the mechanical assistance generated from the hand-assistive exoskeleton. The
control system was tested using simulated sensor feedback data and has proven to be a
robust design. The controller will be integrated with a glove-like, hand-based exoskeleton.
The complete system was designed to restore hand functionality through the amplification
of precision pinch and/or power grasp.
Background
Problem:
A combined 450,000 Americans suffer from debilitative disorders, such as multiple
sclerosis and muscular dystrophy, resulting in the loss of muscle strength and dexterity
in the hand.
Solution:Orthotic hand exoskeletons may be used to amplify the remaining muscular control in
the hand and facilitate pinching and grasping movements. The restoration of dexterityand the achievement of synergistic motion lie primarily within the device’s control
algorithm.
Current Designs and Limitations:•Binary Control Algorithms -->The controller reads the input signal, compares it
against a predefined threshold, and sets the output state accordingly. Binary control
usually exhibits faster object interactions through increased system response times,
but at a cost of a decrease in control precision [1].
•Variable Control Algorithms --> System output signals are related proportionally
and linearly to input control signals. Saturation of the output occurs if the input
signal falls outside of a predefined range. Variable control systems exhibit arelatively greater level of control precision for more delicate objects; however this
was contrasted with a slower response time when compared to other binary
algorithms [1].
Design Objective:
Our objective was to design a dynamic digital control system with the
benefits of binary and variable control architectures: one that would exhibit
the necessary balance of control accuracy and response that would more
accurately match hand physiology.
Figure 3. Proposed dynamic state control system flowchart. Flowchart represents the logic for a
single motor driven output (capability for the independent control of four motors).
Proposed Control System Design
Input Signal Quantification:
Based on the notion that a binary control system acts as a state controller, a two-bit binary
control algorithm was realized (four total states).The input control signals were divided into
four areas based off of the maximum expected value of the incoming signal (Fig. 2). Each
area was associated with a different state for each motor driven output.
Figure 2. Two-Bit binary control example given a generic input control signal (red). Based on where the
instantaneous value of the signal is located with respect to the ranges (blue, green, and yellow), a specific
output state will be selected.
Dynamic State System:
• Joint angles --> Continuously measured by the control system. If a single joint angle
existed outside of its predefined range, then all system output states would be forced
to zero, preventing any output that could result in phalangeal hyperextension.
• Palmar Resistance --> Quantified to detect the presence of an object. These
resistance measurements were grouped based on their applicability to either the
precision pinch or the power grasp. If a resistance measurement surpassed a pre-
defined threshold, then the flexion output states would be reduced accordingly to
allow for more accurate control and delicate motions.
Figure 1. An orthotic hand exoskeleton for the amplification of a precision pinch and power grasp.
Hardware and Feedback
Digital Control System:• Designed in LabVIEW v8.5 on a laptop computer with an external Data Acquistion
Unit controlled via USB (Fig. 4).
Instrumentation:• Force sensing resistors (FSRs) used to obtain flexion and extension data, in addition
to palmar resistance measurements (Fig. 5)
• Hall Effect sensors used to obtain joint angles (Fig. 5)
Figure 4. Front panel of designed digital control system. Developed in LabVIEW v8.5.
Figure 5. Instrumentation placement diagram. Force Sensing Resistors for flexion/extension (green circles),
Force Sensing Resistors for palmar resistance feedback (light green circles), and Hall Effect sensors (green
squares) shown respectively.
Conclusion
We designed and implemented a dynamic state control system for an orthotic assistivehand exoskeleton that is able to more closely mimic physiological movement. Futureevaluations of control system functionality include:
• Quantification of grasping strength improvement
• Quantification of dexterity and control precision
References and Acknowledgments[1] Lucas, Lenny. Yoky Matsuoka, and Matthew DiCicco. "An EMG-Controlled Hand Exoskeleton for Natural
Pinching." Journal of Robotics and Mechatronics. 2004. 482-8.
We would like to thank The College of New Jersey School of Engineering and Dean Steven Schreiner for funding
support. We would also like to acknowledge the contribution of Dr. Marvin Kurland towards the design of the control
algorithm.
Resulting Control System Flowchart: