Compliance analysis and regulation for flexible robotic manipulators

Abstract

Manipulators composed of rigid links, have been widely applied especially in manufacturing industry, where precision, repeatability as well as efficiency are highly demanding. However, they become less competent when working in narrow space or interacting with human and environment, where flexible robotic manipulators become more advantageous due to the high adaptability resulting from their inherent compliance. However, the inherent compliance also brings difficulties in modelling and control of flexible manipulators, causing a dilemma of either sacrificing the control results or efficiency.

In order to simplify the modelling and control of flexible manipulators, investigations on system compliance analysis and regulation have been performed in this thesis, where two representative flexible manipulators have been studied, and one is the cable-driven manipulator with transmission compliance, while the other is the soft continuum manipulator with hybrid compliance, including both transmission and actuation compliance. Both systems were self-developed and have been used for validation of the proposed methods.

Different strategies were applied for tackling compliance of different robotic systems. For cable-driven manipulator, the system compliance was measured by cable elongations, while the system performance was evaluated by motion and force transmission efficiency as well as motion tracking accuracy. Methods from different aspects for compliance regulation have been applied, including parameters adjustments of transmission system, structural enhancement inspired from bracing technique, and compliant control combined with compensation algorithm. In terms of modelling and control of soft continuum manipulators, a method based on workspace analysis and space searching has been proposed, through three steps specifically, including workspace sampling, gridding and searching, and the pivotal factors of each step can affect the control accuracy as well as the computational speed. The system compliance was inferred from its repeatability accuracy, and lower the repeatability is, higher the system compliance should be. It was found that the system compliance although has constrained the system capability also enabled fast space searching and therefore simplified the computational process. Meanwhile, a quantitative relationship between control accuracy and computational speed was also established for further coping with different requirements.

Different from existing control methods for flexible manipulators of pursuing a control result that is as accurate as possible, the concentrate of this thesis is to achieve a balance between the system performance and computational efficiency by regulating the system compliance. For both cable-driven and soft continuum manipulators, the overall relations between the system performance, control complexity as well as compliance have been constructed both in quality and quantitative aspects, through which, basic position-based control can be accomplished, and the motion tracking results have been validated on both platforms. The specific control methods can be also applied for similar robotic systems, and further a new perspective of dealing with compliance in control aspect was proposed as using instead of eliminating it.