Article On the visual deformation servoing of compliant objects: Uncalibrated control methods and experiments The International Journal of Robotics Research 2014, Vol. 33(11) 1462–1480 © The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0278364914529355 ijr.sagepub.com David Navarro-Alarcon 1 , Yun-hui Liu 1 , Jose Guadalupe Romero 2 and Peng Li 1,2 Abstract In this paper, we address the active deformation control of compliant objects by robot manipulators. The control of defor- mations is needed to automate several important tasks, for example, the manipulation of soft tissues, shaping of food materials, or needle insertion. Note that in many of these applications, the object’s deformation properties are not known. To cope with this issue, in this paper we present two new visual servoing approaches to explicitly servo-control elastic deformations. The novelty of our kinematic controllers lies in its uncalibrated behavior; our adaptive methods do not require the prior identification of the object’s deformation model and the camera’s intrinsic/extrinsic parameters. This feature provides a way to automatically control deformations in a model-free manner. The experimental results that we report validate the feasibility of our controllers. Keywords Deformation control, visual servoing, adaptive control, robot manipulators, compliant objects, Lyapunov stability 1. Introduction The robot manipulation of rigid objects has been studied and successfully implemented in several traditional applica- tion fields, e.g. pick-and-place and assembly, for more than four decades now—we refer the interested reader to Murray et al. (1994), Okamura et al. (2000), Hokayem and Spong (2006), and Yoshikawa (2010) for comprehensive works on the topic. In recent years, the robotics research commu- nity has paid considerable attention to the problem of the automatic manipulation and shaping of deformable objects. In the authors’ opinion, this emergent interest in control- ling deformations is mainly driven by the now widespread use of surgical robotics (Mallapragada et al., 2011), and the automation of economically important tasks that require control of interactions with soft bodies, e.g. shaping of food materials (Tokumoto et al., 2001), or assembling flexible objects (Park and Mills, 2005). The objective of the active deformation control is to pro- vide a desired shape or configuration to a compliant object by the interaction of a mechanical system (typically a robot manipulator). One of the principal issues that complicates the automation of these types of tasks is the difficulty of estimating the deformation properties of soft materi- als. From a control design perspective, this information is needed to coordinate the manipulator’s motion with the visual measurements of the object’s shape. In other words, we need to know how displacements of the manipulator are “mapped” to deformations of the body. The deformation properties of a compliant object can be obtained, for example, through traditional offline parame- ter identification techniques (such as least-squares model fitting). An alternative to this offline and fixed-model approach is to equip model-free controllers with some kind of online adaptation to cope with the unknown elastic prop- erties of an object. Our aim in this paper is precisely to develop uncalibrated methods that do not require the identification of the deformation model. 1.1. Related work To the best of our knowledge, one of the earliest works to comprehensively address the robotic manipulation of deformable objects is reported in Henrich and Worn (2000). 1 Department of Mechanical and Automation Engineering, The Chinese University of Hong Kong, Hong Kong SAR, The People’s Republic of China 2 Laboratoire des Signaux et Systemes (L2S), SUPELEC, Gif sur Yvette, France Corresponding author: David Navarro-Alarcon, Department of Mechanical and Automation Engi- neering, The Chinese University of Hong Kong, Rm 112, WMW Mong Engineering Bldg, Shatin NT, Hong Kong SAR, The People’s Republic of China. Email: navarro-alarcon@cuhk.edu.hk at CHINESE UNIV HONG KONG LIB on May 2, 2015 ijr.sagepub.com Downloaded from