Telerobotic Mini-Golf: System Design for Enhanced Teleoperator Performance Sudath R. Munasinghe, Ju-Jang Lee Dept. of Electrical Engineering and Computer Science Korea Advanced Institute of Science and Technology 373-1 Guseong-dong, Yuseong-gu Dejeon 305-701, Korea Tel: 82-42-8698032 Fax: 82-42-8698506 Email: rohan@odyssey.kaist.ac.kr Tatsumi Usui, Masatoshi Nakamura Dept. of Advanced Systems Control Engineering Saga University 1 Honjomachi, Saga 840-8502, Japan Tel: 81-952-288643 Fax: 81-952-288666 Email: usui@cntl.ee.saga-u.ac.jp Naruto Egashira Dept. of Control and Information Systems Engineering, Kurume National College of Technology 1-1-1 Kumorino Kurume-City Fukuoka 830-8555, Japan Tel/Fax: 81-942-359323 Email: naruto@kurume-nct.ac.jp Abstract— This paper discusses the safety assurance, capability features, and user-friendliness of Internet-based telerobotics, in the sense that how these features could be incorporated in order to enhance teleoperator performance. A comprehensive set of safety features are incorporated to the interface to screen and accept only the safe commands from the remote operator while discarding risky commands. A laser pointer is used to help remote operator in perceiving self-location and navigation, whereas orientation control has been completely automated and synchronized to the position commands of the teleoperator. The effectiveness of these features has been demonstrated by playing telerobotic mini-golf. I. I NTRODUCTION Internet-based telerobotics is just about a decade old [1], yet by today, there are a wide range of devices being teleoperated on the Internet [2]. Telerobotics and teleoperations will be more appealing and applicable in the future particularly in space systems, hazardous tasks such as fire-fighting, war- front operations, and search and rescue operations, which have turned out to be the current needs of the world. Teleoperations that span long distances over transmission networks deal with substantially long time-delays, which could cause some of the high frequency signals to undergo positive feedback in the control loop. As a result, the control system could become unstable [3]. Alternatively, move-and-wait technique [4] has been adapted for teleoperations in its very early development stages, and still in the main stream of teleoperation techniques. Supervisory control [5] has been the most popular strategy for stable teleoperations in that the tasks are decomposed into simple subtasks, each of which can be planned and executed by a local controller. Teleprogramming or model-based preplanning is another pop- ular technique for stable real-world teleoperations [6]. This technique is appealing in static and structured work-sites, how- ever, the actual telerobotic performance may be deteriorated due to modeling errors and system uncertainties. On the other hand, most real-world telerobotic applications such as live power line maintenance [7], the tasks to be performed cannot be pre-planned. Instead, the teleoperator has to constantly interact with the visual feedback and teleoperate the manipu- lator by reacting to constantly emerging unpredictable work- site scenarios. Therefore, more direct teleoperation capabilities are required. However, the key issue here is the significant incapability of the teleoperator to conceive the geometry and motions of the work-site. Even though the teleoperator is supplemented with sophisticated graphics-based systems [8], a slight difference from actual physical perception could significantly deteriorate teleoperator performance [9], [10]. Therefore, lot of capabilities are required to be automated, and supplimented to the teleoperator commands. Furthermore, the tasks that cannot be pre-planned, and those that require more constant human interaction, and also in unstructured workspaces, more direct teleoperation is very much a need. In this view, the major attention of this paper is to devel- oped and integrate automated capabilities to the teleoperator commands so that to realize more direct teleoperation. Our strategy starts with supervisory control, and reduces the super- visory action to a point-to-point motion as shorter as 1[mm]. With sufficient end-effector speed as high as 20[cm/s]. This technique has been successfully implemented on the recently built telerobotic test-bed between Korea Advanced Institute of Science and Technology (KAIST) and Saga University, Japan, in its telerobotic mini-golf mode, which allows a remote operator to play golf by teleoperating a 5 DOF industrial robot manipulator. The system is versatile in terms of its manipu- lation capabilities, safety features, and user friendliness. The system demonstrated its capabilities that allow the teleoperator to interact with constantly emerging scenarios in the work site, and thereby to perform complex tasks that cannot be pre- planned or predicted. II. SYSTEM DESIGN A. Telerobotic Testbed and Telerobot A complete account of the KAIST-Saga University teler- obotic testbed can be found in [11], and also available online at http://medic.kaist.ac.kr/rohan/career/projects.html. The telerobotic ma- nipulator is shown in Fig.1, and it has 5 DOF in R-R-R-R-P Proceedings of the 2004 IEEE International Conference on Robotics & Automation New Orleans, LA • April 2004 0-7803-8232-3/04/$17.00 ©2004 IEEE 1688