An Inclusive Grasp Taxonomy and a Variable Transmission Prosthesis
Keilani Adachi
with the Berkeley Engineering Design Scholars Program, the Embodied Dexterity Group, Dr. Hannah Stuart, PhD Student Michael Abbott, and Gabriela Gutierrez
Hands are such an important instrument through which we interact with the world and with each other. My two projects this summer, a grasp taxonomy for people with spinal cord injury and a variable transmission prosthesis, attempted to both describe and quantify hand usage and also actually prototype a hand prosthesis concept.
A Grasp Taxonomy for People with Spinal Cord Injury
The Problem:
Expansive work studies human grasping, however nearly all prior studies have focused on normative hand function where the participants have full control of their fingers. I aimed to broaden the field’s understanding of grasping by creating a taxonomy of grasps used by people with cervical spinal cord injury (SCI).
The Process:
By observing YouTube videos of people with C6/7 level spinal cord injuries performing daily tasks, I developed a list of grasps which they tend to use. After multiple reorganizations, a final taxonomy began to take shape.
The taxonomy is useful for having a common language to describe otherwise abstract grasp types. Unlike other taxonomies, this taxonomy includes other body parts besides the hand such as the face, trunk, arms, and mouth. This is an important distinction because it acknowledges the differences between how people with SCI grasp and how people with normative hand function grasp.
In Practice:
I utilized the grasp taxonomy to carefully analyze the YouTube video entitled "Quadriplegic Preparing Hot Tea" (embedded above). I annotated the video by tagging when and for how long the person used each body part and grasp.
It's interesting to note that this man used both hands simultaneously almost as often as he used each hand individually, however for a much shorter duration. He had a similar number of touches with his right and left hands, though the right hand was used for longer tasks than the left. He also uses his mouth on occasion, especially when a very firm grasp was necessary like when tearing open the tea bag package.
The most frequent grasp used was the Lateral Pinch, also known as the "key" grip. This grasp is popular because of its versatility in being able to hold both very small and medium sized objects. The Wrap grasp with the thumb around the object was also used for a long duration. Tenodesis grasp assists this grasp significantly, as people with C6/7 spinal cord injury have little to no control over closing their fingers in this way.
In order to compare normative hand function for this task to the video of the man with SCI preparing hot tea, I made a video and analyzed it of myself making tea. I used only the hands and had far less diversity of grasp types with my top grasps being pinching with the finger pads and wraps.
Looking Forward:
In the future, the taxonomy could be used to analyze the effectiveness and impact of assistive devices and in further human subjects studies of grasping.
Variable Transmission Prosthesis
with PhD Student Michael Abbott and Gabriela Gutierrez
The Problem:
Upper limb prostheses have a very high rejection rate due to design, discomfort, and dysfunction. It is estimated that 26-46% of upper limb prostheses are rejected (Biddiss 2007). Body-Powered prostheses, prostheses where the body instead of electronics control the "hand", are popular because of their force feedback and simple design. They work by moving the opposite shoulder or the affected arm to apply tension to the cable which runs from the harness to the end effector, opening or closing it. However, applying tension to the cable can cause bruising, harness irritation, and fatigue because of the high forces necessary to achieve a high grasp force.
The Goal:
Our goal this summer was to prototype a proof of concept for a variable transmission prosthesis. With this device, a person would be able to exert high grasp forces with less shoulder forces on the cable. This would lessen the probability of harness irritation and bruising.
The Design Process:
After creating 17 different concepts, we used a weighted matrix to evaluate a multitude of ideas. My partner Gabriela and I individually explored the cones and wheel continuously variable transmission and linkage continuously variable transmission concepts respectively.
Initially, we were thinking of using a linear actuator to extend and contract the length of one linkage. However, linear actuators so small cannot tolerate a side load of the magnitude users would be applying on the cable. The rack and pinion model was promising, but ultimately we decided to further pursue the cone and wheel design instead of the linkage idea for its ability to be scaled to a small size and proven effectiveness with varying transmission ratios.
The cones and wheel CVT, as sketched below, works by having an input torque on the first cone and that torque is transferred to the output cone via a roller. The wheel can translate up and down the cones, so depending on the location of the roller, the ratio of the diameters of the cones determines the output speed and torque.
To prototype this proof of concept, Gabriela and I designed the variable transmission mechanism in SolidWorks. After ordering custom shafts, bearings, and other hardware online, we went into lab and got to work 3D printing and laser cutting.
One of our biggest challenges was finding a suitable method to attach the 3D printed parts to the shafts, especially because we would need to be transmitting torque. We ended up using heat set threaded inserts and set screws to overcome this issue which can be seen in the third picture below.
Finding a suitable roller wheel was tricky, as it needed to be small and not too hard to be able to transmit torque via friction. I found a wheel in my old Lego collection.
One of the main design aspects of the mechanism is the slots for adjusting the cone position. It was important to be able to fine tune the side to side position of the cone to ensure the proper amount of contact with the wheel to transmit torque.
The crank and the flag turned out to be slightly too large to function well with the large bolt heads so we decided to reprint the two pieces to have a better fit.
Finally, here is our latest video showcasing how the output speed changes depending on the position of the roller.
Looking Forward:
As this is currently a proof of concept and still driven by a crank instead of a cable as in a typical body powered prosthesis, the next steps on this project would be to connect the input cone to a cable and harness and the output cone to a terminal device. This further integrates the variable transmission concept into the context of a body powered prosthesis. Then further work could automate the movement of the roller as opposed to manually shifting its position.
Thank you for checking out my projects!
I'd also like to thank the Berkeley Engineering Design Scholars Program, Dr. Hannah Stuart, PhD Student Michael Abbott, and Gabriela Gutierrez.
