Electronic Thesis and Dissertation Repository

Thesis Format

Monograph

Degree

Master of Engineering Science

Program

Electrical and Computer Engineering

Supervisor

Polushin, Ilia G.

2nd Supervisor

Patel, Rajni V.

Joint Supervisor

Abstract

This thesis deals with development and experimental evaluation of control algorithms for stabilization of robot-environment interaction based on the conic systems formalism and scattering transformation techniques. A framework for stable robot-environment interaction is presented and evaluated on a real physical system. The proposed algorithm fundamentally generalizes the conventional passivity-based approaches to the coupled stability problem. In particular, it allows for stabilization of not necessarily passive robot-environment interaction. The framework is based on the recently developed non-planar conic systems formalism and generalized scattering-based stabilization methods. A comprehensive theoretical background on the scattering transformation techniques, planar and non-planar conic systems is presented. The dynamics of the robot are estimated using data-driven techniques, which allows the equations for the dynamics of a robot to be obtained in an explicit form. The generalized scattering transformation is used in combination with the Lyapunov-based adaptive trajectory tracking control. It is shown that the original interconnected system is not stable due to its non-passive nature; however, the application of the proposed stabilization algorithm allows stability to be achieved, without affecting the robot’s trajectory tracking performance in free space.

Summary for Lay Audience

Since the advent of robotics, considerable effort has been put in the development of appropriate theory, algorithms and their implementation. Nowadays, robotic systems have found applications in various areas including, but not limited to, industrial manufacturing, the services sector, driving and healthcare fields. Further integration of robotics in the economy requires enabling a higher level of physical interaction between a robot and the outside world. In other words, robots should be able to, for example, safely interact with people, other robots and various systems. The robotics research community has put a great amount of effort into investigating methods for interaction control. To date, most commonly applied algorithms for interaction control are limited to passive systems, i.e., systems that do not have internal sources of energy. Simple examples of passive systems are a mass-spring-damper system and a fixed wall. In contrast, both humans and robots come under the concept of non-passive, or active systems, because they have internal sources of energy. Thus, conventional interaction control algorithms typically fail to interact with active environments in a stable manner. In this thesis, a framework for stable robot-environment interaction for passive and non-passive systems is presented and evaluated on a real robot. This work presents a comprehensive overview and theoretical background on control methods and paradigms used to design the framework. All the necessary steps for implementation of the algorithm are described. These include hardware design, a method for estimation of dynamics and cone parameters. It was experimentally verified that the proposed algorithm successfully stabilizes interaction between the robot and the environment. The proposed framework constitutes a fundamental extension of the existing passivity based approaches for the coupled stability problem. The algorithm can be applied for stabilization of interconnections of active systems. For example, possible applications of the framework are bilateral teleoperation with communication delays, robotic surgery on a beating heart, and haptics-based environments for training.

Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS