Session: 15-01-01: ASME International Undergraduate Research and Design Exposition

Paper Number: 149553

149553 - A Testbed Demonstration of a Statically Adjustable Continuous Variable Transmission Mechanism 

Current designs for body-powered (BP) prosthetic graspers rely on simple levers to link input and output forces at the end-effector. However, the high forces experienced by users using graspers with a fixed lever arm reduce grasp force control and increase fatigue. Developing a body-powered prosthesis that incorporates variable mechanical leverage to enable users to augment their input force may lead to a prosthetic design that offers haptic feedback for accurate, quick control of the device while providing a range of low to high output forces; this reduces fatigue while enabling users to more efficiently exert forces and use cable excursion compared to traditional fixed transmissions.

Furthermore, most current CVT mechanisms require the input and output mechanisms to move during a transmission ratio (TR) change. Users must expend a portion of their limited range of motion to adjust a TR—this may be cumbersome. For instance, the conical Evans Variable Speed countershaft has too much static friction at rest to adjust the belts to another TR while the system is at rest. The user must expend some of their range of motion while moving the belt to change the ratio. Another CVT design uses elastomers, which have a short fatigue life. The need for a durable CVT that grants users a greater range of motion motivates our search for a CVT that enables TR change while the input/output remains fixed with minimal damage to the system. 

We built a novel CVT design that uses a friction drive system to transmit torque with two perpendicular discs. One rigid circular plate is the "drive" wheel output, and an omni wheel is the input side. Rolling wheels around the omni wheel enable it to slide axially with less friction. We further proposed that a rubber-coated drive wheel and elastic SLA printed material would improve surface adhesion and reduce slippage. 

We kept the input force constant using a linear actuator to pull a cable attached to a drum linked to the omni wheel to obtain preliminary results about our CVT design. In contrast, friction alone held the contact point of the omni wheel at a near-constant distance from the drive wheel center point. We manually pulled on a cable like a drum connected to the drive wheel and applied a roughly constant output tension by feel. We installed tension sensors in line with the cables at the input and output sides of the mechanisms to measure tension—this helped us characterize the efficiency of force transmission through the CVT. Furthermore, we compared the force required to axially move an omni wheel versus a regular rubber wheel installed at the same spot to determine if the omni wheel would more easily enable TR changes compared to a regular wheel while we fixed the drive wheel. 

Preliminary findings show that the omni wheel reduced the amount of force to axially move the input wheel by a mean of 22.1% to 54.7% (with a standard deviation of 7.72% and 9.33% respectively) compared to a rubber wheel sliding against the drive wheel. The higher force reduction correlates with the longer roller on the omni wheel, while the shorter correlates with the short roller. Comparing different TRs shows that the mechanism produces a TR (output force over input force) range of ~1x to 1.5x. Furthermore, we found a mean transmission efficiency of 48.53% with a standard deviation of 8.23%. 

These results verify that our CVT and its novel inclusion of an omni wheel generates a meaningful range of TRs and efficiently transfers force that may help users accomplish a broader scope of grasping tasks more comfortably. We plan to develop an analytic model of the mechanism and install an actuator to produce constant resistive force. This actuator will provide more consistent trials than applying manual load to the mechanism. A motorized control linked to the omni wheel system will also provide more reliable data about the transmission ratio.

Presenting Author: Franklin Ho University of California, Berkeley

Presenting Author Biography: Franklin is a fourth-year undergraduate student at UC Berkeley pursuing a B.S. in Mechanical Engineering and a minor in Aerospace Engineering. He has a broad interest in physics-based computational models that directly influence the design and optimization of robotic and aerospace systems—especially those that have applications in assistive and consumer technology. He also enjoys exploring novel materials and manufacturing methods, and is teaching a student-led course in additive manufacturing in addition to researching electrostatic treatments that strengthen composites.

Authors:

Franklin Ho University of California, Berkeley
Hannah Stuart University of California, Berkeley
Michael Abbott Santa Clara University