Date on Master's Thesis/Doctoral Dissertation

5-2024

Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department

Mechanical Engineering

Degree Program

Mechanical Engineering, PhD

Committee Chair

Harnett, Cindy

Committee Co-Chair (if applicable)

Popa, Dan

Committee Member

Popa, Dan

Committee Member

Murphy, Kevin

Committee Member

Kate, Kunal

Committee Member

Chitalia,Yahs

Author's Keywords

Tactile sensor; soft robotics; robot manipulation; biomimetics; control; optoelectronics

Abstract

Robotic manipulation is one of the main types of automated labor that humans want to achieve in order to free ourselves from mundane and hazardous tasks. The range of realizations could be very broad, from massive industrial robots to highly advanced medical devices for human healthcare. Given the critical nature of these application spaces, they require robust, safe, and most of all versatile designs. In that direction, one of the most interesting topics is the anthropomorphic robotic manipulator. Since it is inspired by human hands’ function and structure, it is able to achieve the aforementioned system requirements. However, the perfect realization of the human-inspired design will take time, because there is still room for advancement in mechanical design, material, control strategy, computation performance, and more. Therefore, in this dissertation, we cover a comprehensive overview in the development of anthropomorphic robotic hand starting from subcomponents of skin, bones, tendons and nerves, and look into biological inspirations that can be implemented with the current state of technology. Another interesting topic regarding robotic manipulation is the sensing ability of robots, which is critical to processing information in situations involving human interaction. Safety is especially vital when it comes to Human-Robot Interaction (HRI). Thus, many recent studies prioritize soft materials and highly conformable systems. Soft designs with bio-inspired aspects have intrinsic advantages in safety as well as robustness. Therefore, integrating both manipulation and sensing within a soft design is a promising path to safe and capable automated labor. In this work, we have developed a new type of soft tactile sensor, notably one that can detect both normal and lateral forces having 0 - 5 N range, 3 Hz sensitivity. Beyond that sensor development, an adaptive control method is demonstrated that adjusts the robotic manipulator’s grasping force to counteract slipping forces on the object being handled in the real-time control manner. In contrast to most soft sensor research involving resistive or capacitive electronic sensors, here a soft optical fiber is utilized to measure the force within a soft silicone housing. Ultimately, we developed a human-inspired robotic finger manipulator incorporating soft tactile sensing. The finger manipulator is not only sensed, but driven by a soft optical fiber tendon which is able to detect finger motion and applied force according to the physical deformation of the tendon fiber. Maximum loading force measured in this study is 3 N on the tactile tip. The final design of the manipulator has 3-joint, 4- link structure so that makes human finger-like motion(70◦/70◦/80◦, DIP/PIP/MCP). Through this novel design of the manipulator and sensor, we have overcome numerous technical hurdles including spatial limitation, cost-efficiency, and complexity of control, coming one step closer to realizing human-like functional robotic manipulation. Therefore, in each section the main contributions of this work can be itemized as below. • Create a new soft tactile sensor design capable of detecting both normal force and lateral motion, including stick-slip phase. • Introduce an adaptive grasping control system using the developed soft sensor and standard robotic grasper to prevent slipping. This controller ensures stability and robustness through analysis and implementation. • Explore human hand anatomy extensively, focusing on structure and sensing features, to inspire the development of robotic hands. • Propose a human finger-inspired robotic finger design with anthropomorphic features and an optical fiber tendon. • Utilize the optical fiber tendon for both force transmission and finger posture measurement in a single unit. • Leverage the optical fiber to estimate finger posture and detect contact force on the fingertip by analyzing its physical deformation.

Download

Share

COinS