Diller_cmu_0041E_10332.pdf (5.47 MB)
Design, Characterization, and Implementation of Lightwieght and Energy-Efficient Electroadhesive Clutches for Robotics
Despite decades of academic and industry effort, achieving efficient and dynamic movement
of robots remains a significant challenge. Many robots, particularly humaniod robots
and wearable robots such as exoskeletons and prostheses, are quite limited in their versatility
and usefulness because of the force and speed limitations of actuators. Weight and
power consumption are particularly important factors in determining the operating range
and effectiveness of these devices. Geared electric motors are most commonly used in these
applications, but often result in slow, stiff, and halting operation. Other options include hydraulic
actuators, pneumatic actuators, electroactive polymer actuators, and shape memory
materials, but none of these are able to achieve the combination of high power output, high
efficiency, and low weight that would enable dynamic movement of untethered robots.
Many proposed solutions to this problem involve using clutches to improve the efficiency
and capability of actuation systems. However, conventional clutches such as electromagnetic
and magnetorheological clutches are themselves too heavy and power-hungry to be
practical. This thesis presents an electroadhesive clutch that has 10£ lower weight and
1000£ lower power consumption than conventional clutches. To inform a variety of possible
implementations, I extensively characterized the effects of design choices on the holding
force, responsiveness, and power consumption of the electroadhesive clutch. Next, I investigated
the use of the clutch in a walking assistance exoskeleton, demonstrating the reliability
and advantages of the electroadhesive clutch in a challenging robotics application. Finally, I
studied the use of electroadhesive clutches to harvest, store, and return mechanical energy
with rubber springs, and used multiple of these units in parallel to create a force controllable
energy recycling actuator. My aspiration is that the work in this dissertation will
lead to improved robotic hardware that enables exciting new capabilities in next generation
robotics.
of robots remains a significant challenge. Many robots, particularly humaniod robots
and wearable robots such as exoskeletons and prostheses, are quite limited in their versatility
and usefulness because of the force and speed limitations of actuators. Weight and
power consumption are particularly important factors in determining the operating range
and effectiveness of these devices. Geared electric motors are most commonly used in these
applications, but often result in slow, stiff, and halting operation. Other options include hydraulic
actuators, pneumatic actuators, electroactive polymer actuators, and shape memory
materials, but none of these are able to achieve the combination of high power output, high
efficiency, and low weight that would enable dynamic movement of untethered robots.
Many proposed solutions to this problem involve using clutches to improve the efficiency
and capability of actuation systems. However, conventional clutches such as electromagnetic
and magnetorheological clutches are themselves too heavy and power-hungry to be
practical. This thesis presents an electroadhesive clutch that has 10£ lower weight and
1000£ lower power consumption than conventional clutches. To inform a variety of possible
implementations, I extensively characterized the effects of design choices on the holding
force, responsiveness, and power consumption of the electroadhesive clutch. Next, I investigated
the use of the clutch in a walking assistance exoskeleton, demonstrating the reliability
and advantages of the electroadhesive clutch in a challenging robotics application. Finally, I
studied the use of electroadhesive clutches to harvest, store, and return mechanical energy
with rubber springs, and used multiple of these units in parallel to create a force controllable
energy recycling actuator. My aspiration is that the work in this dissertation will
lead to improved robotic hardware that enables exciting new capabilities in next generation
robotics.
History
Date
2018-12-01Degree Type
- Dissertation
Department
- Mechanical Engineering
Degree Name
- Doctor of Philosophy (PhD)