A distributed electromechanical spine for bio-inspired robots

Biological mechanisms are embraced in mobile robots to interact with their surroundings. Although current bio-inspired robots perform well, their performance is limited due to the lack of a flexible spine. A spine provides an animal's agility, a wide range of motion, balance, and efficiency. It can be created using motors, which have been widely used for robotic joints. However, this conventional method introduces design complexity, low actuation speed, low efficiency, poor backdrivability, and backlash issue. Moreover, a vertebra in the spine does not fully rotate like a conventional motor. This thesis introduces a distributed and scalable two-dimensional electromechanical spine for bio-inspired robots. It proposes an approach that mimics an actual animal spinal structure and muscles by combining a magnetic core and two coils in a module. Six modules are connected in series to form a spine. A single module and the entire system represent a vertebra and vertebrae, respectively. The proposed actuator utilizes electromagnetic force induced by coil currents to control torque at each module. This actuator has several benefits, including modularity, scalability, distributed actuation, simple structure, and gearless design, as well as better cooling mechanism and compliance. While a motor has a trade-off between torque and speed, the proposed actuator has a trade-off between torque and angular flexibility. Furthermore, the proposed actuator uses normal stress to produce force, while a motor uses shear stress. This approach results in high torque capability without using gears. A distributed air-gap model is proposed to improve force estimation by taking non-uniform air gaps and core saturation into consideration. Core-flux density and coil-current density are considered as design constraints in the design procedure. A torsion spring mechanism is applied to each module to improve the torque capability. Finally, feasibility of the proposed actuation system is verified by both simulation and experimental results.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2020

Author dc:creator

• Ku, Bonhyun

Contributors dc:contributor

• Banerjee, Arijit

• Robotic actuator, robotic spine

Statement dc:rights

• Copyright 2019 Bonhyun Ku

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/106398

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/106398