Student Projects
Group M2 – EMG Glove for Muscular Dystrophy Patients
Final Design Project Lab Report
Introduction
Our design project’s goal is to help people with muscular dystrophy. This is a degenerative
disease which severely reduces functionality of the patient because of increase in muscle tissue friability.
Our project is focused on hand movement. The main aim is for the patient to be able to move his or her
arm, and achieve normal gripping ability. For the design a glove with cables attached to the fingertips will
be used. The fingers are closed by an actuator and opened passively by rubber bands. The motor unit is
turned on/off utilizing a lab view program controlled by EMG signals taken from the patients arm.
Experimental Setup
The project consists of three main parts: EMG signal acquisition, LabVIEW processing and
Glove assembly. The EMG signal is taken from the finger extensor muscles on the forearm of the patient.
Those are the muscles that would normally be used for a gripping motion too. The signal is then amplified
and filtered using a band-pass filter with a total gain of 3000 and cut-off frequencies of 18 Hz and 450
Hz. In LabVIEW, the EMG-Signal is further processed and a band-stop filter is implemented from 59 Hz
to 61 Hz to filter out 60 Hz power line noise. The signal DC offset is also removed in the VI. One EMG
signal is used to toggle between the mechanical relays, the other signal outputs to the digital outputs of
the DAQ and is used to activate the actuator by delivering a 5V DC. The digital outputs are used because
with analog inputs the time lag is too big with analog outputs. The signal controls a MOSFET which
delivers the required current to the selected relay. The actuator is thus directly connected to the switches.
When the actuator, which is placed on the forearm, pulls on the cables attached to the fingertips of the
glove the hand is closed in a grabbing motion. The cables are run through tubing attached to the glove, so
that they are not in the way when the person wants to grab something. To reduce the number of cables
running along the hand we connected two fingers to one cable. The cable of the thumb is run over a
pulley in order to change the direction. On the other side of the hand rubber bands are attached to
passively open the hand. The wrist is stabilized with a wrist brace, without which the wrist would move
instead of the fingers. For a better grip another glove could be worn over this one. An overview of the
glove and electronics can be found in appendix A and B respectively.
Results
During the initial testing phase, we observed multiple sources of lag and error. The first of those
was that the LabVIEW software had a delay in updating the LED vis-à-vis the actual value of the local
variables, which was rectified when we switched the output from analog to digital. We also observed that
the DAQ output was insufficient to power the mechanical relays on its own (limited to 5V) and
necessitated the use of two MOSFET based systems, which allowed us to use the digital outputs as
toggles while drawing power from the 12V DC car battery. Relative motion between the actuator and
forearm was also observed and ameliorated as much as possible by the use of duct tape. A thumb brace
was implemented to limit the motion of the thumb to two dimensions.
For the glove validation we used a spring system (see appendix C). The spring constant for each
spring was calculated using the Instron 3366 Load Frame to be 861N/m (see appendix D for linear
regression). Using four springs, all with the same spring constant, we created 6 test cases using a
combination of series and parallel configurations (see appendix E for different configurations). To
perform ‘efficiency’ (
E) measurements we measured the displacement achieved by a healthy subject’s
forearm contraction against the spring system vs. that achieved by the actuated glove without the use of
the subject’s muscles. As expected there were mechanical losses in the glove that created an
E of 85.5%
(see appendix F).
Any reduction in efficiency is mainly due to the motion of the wrist and movement of the actuator relative
to the forearm.
The first of our demonstrations of dexterity involved the transport of a Styrofoam cup filled with
water, which was chosen given the relative fragility of the cup material. The criteria for success included
not crushing the cup and not spilling the water in the process, which were met. The other test for dexterity
was the use of scissors to cut duct tape, which was achieved in two cycles of actuation.
While we have no quantified data with regard to the reliability of our system, we did perform our
dexterity tests and efficiency measurements multiple times over the course of multiple days and observed
the same results with no malfunctioning of the device.
Discussion and Conclusion
Overall we were able to achieve the desired targets. The person with the EMG-Electrodes was
able to open and close the hand of the person wearing the glove. In an actual scenario this would be the
same person, however in the interest of simulating conditions of atrophy a separate person was assigned
glove control. It was also possible to control the actuator with enough sensitivity to achieve the dexterity
required for the aforementioned Styrofoam cup and scissor tests.
Sources of error in this experiment include crosstalk between the two locations where the
electrodes were placed, mechanical losses and software lag. Increasing the distance between the
electrodes to the point where they are placed on different portions of the arm, could reduce crosstalk, but
we decided against this to maintain a greater semblance to normal hand use. By using digital instead of
analogue outputs we were able to reduce the majority of the time delay. Mechanical losses arise due to the
fact that we were using a two-inch actuator, which for normal range of motion in a hand is not enough.
This error was accentuated by a forward movement of the actuator and a bending motion of the wrist
during contraction. To fix these issues, a longer actuator placed more firmly on the arm would suffice.
Application
With regard to the possible future directions the development of our device could take, we would
like to incorporate the independent motion of the individual fingers and 3D motion of the thumb in an
effort to increase manual dexterity. We would also aim to condense the power supply system as well as
modify the actuator to draw less spike current, allowing us to explore alternate sources of power. The
device in question serves as not only a very useful tool for rehabilitation and/or assistive aid; it also has
the potential to augment user strength thereby reducing fatigue for labor intensive tasks. In an overview
sense, the healthcare provider could recommend our EMGlove to enable people with mild to moderate
symptoms of muscular dystrophy to restore day to day functionality, improving the patient’s quality of
life.
Attachment: Original Report
