Design, Analysis and Control of Soft Wearable Devices Using Twisted String Actuators
Loading...
Authors
Tsabedze, Thulani
Issue Date
2023
Type
Dissertation
Language
Keywords
Closed-loop control , Soft robotics , Spooled motor tendon actuators , Twisted String Actuators , Wearable robotic devices
Alternative Title
Abstract
Physical assistive robotic devices have demonstrated many desirable advantages in rehabilitation and augmentation, allowing for effective performance of activities of daily living (ADLs). Robotic rehabilitation, for instance, allows for accurate and efficient dynamic exercise routines where performance metrics can be easily retrieved. In human augmentation, wearable devices allows human beings to carry heavier loads for longer duration. However, it is challenging to create wearable robotic systems and devices that are inherently safe, compact, and can produce sufficient power and force. Existing robotic devices are often bulky, actuate limited degrees of freedom or produce limited force outputs. The large footprint of the devices usually lead to the devices being tethered either in the lab or in rehabilitation centers, limiting the duration of usage. In addition, active wearable devices that provide resistive capabilities to enable strength training have been difficult to develop. The properties of the wearable devices --- inherent safety, compactness, high force output --- is heavily driven by the choice of actuation mechanism. Twisted string actuators (TSAs) are appealing for wearable robotic applications because they are compliant, energy-efficient, capable of producing large translational force, and exhibit high power density. To utilize the properties of TSA, it is important to understand TSAs' key performance metrics to allow for ubiquitous usage. The performance analysis of TSAs is challenging due to the strong coupling between the TSA model parameters. It is important to compare TSA's performance to motor-based actuators, helping in understanding the trade-offs of TSAs compared to motor-based tendon-driven actuators. This dissertation first provides a theoretical model-based framework to analyze the performance of TSAs focusing on four metrics: contraction range, linear velocity, effective torque input and force output. The performance of the TSA is then compared to the spooled motor-tendon actuator (SMTA). SMTA is a motor based actuation approach that is similar in some regard to TSA. The results of the analysis and comparisons were then used as a basis to select to use TSAs for the actuation approach for the wearable devices in this work. Next, the design, characterization and open loop control of two devices that are driven by TSAs are described in detail. The two devices explored in detail are: 1) wearable glove and 2) wearable wrist orthosis. Specifically, the evolution of the wearable glove is presented in four versions, named a biomimetic robotic assistive glove (BRAG) v1, BRAG v2, Active Wearable Assistive and Resistive Device (AWARD) v1 and AWARD v2. In particular, BRAG v1 used only stiff strings while BRAG v2 and AWARD utilized stiff strings and compliant super coiled polymer strings for position sensing. In addition, the kinematic modeling, dynamic modeling and closed-loop control of AWARD is presented to demonstrate the different control strategies that can be used to achieve the unique capabilities in trajectory tracking. To allow for full state estimation of the finger motions without requiring tethered approaches, inertial measurement units, supercoiled polymer actuators, force sensitive resistors, encoders and motor current sensors were implemented, ensuring that the footprint and weight of the device is not affected drastically. Position control was implemented in the tendon space and also in the encoder space and demonstrated that desired trajectories were attained.