VIDEO SESSION 1
I. Human-Robot Coordination
  1. A Control Strategy for Human-Friendly Robots
  2. Real-Time Stereo Face Tracking System
  3. A Cobot in Automobile Assembly
  4. New Telecommunication using CTerm
  5. Development of a 3,500 Tonne Field Robot
  6. Human-Robot Interactions
  7. Dancing with Juliet
II. Novel Robotic Systems
  1. Micro Inspection Robot for 1-Inch Pipes
  2. 1ms Grasping System Using Visual and Force Feedback
  3. Self-Reconfiguration of 3-Dimensional Homogeneous Modular Structure

VIDEO SESSION 2
III. Mobile Robots
  1. SpaceCat, a Micro-Rover Based on an Innovative Locomotion Concept
  2. Vision Based Navigation System for Autonomous Mobile Robot with Error Recovery
  3. Gait Generation for Multi-Segment Inchworm-Like Robot Locomotion
  4. Sensor Networked Robotics
  5. Reactive Navigation in Outdoor Environments
IV. Medical Robotics
  1. Laparoscopic Telesurgical Workstation
V. Virtual Environments in Robotics
  1. A Virtual Excavator for Operator Training and Controller Evaluation
  2. Interaction with Simulated Environments using a Magnetic Levitation Haptic Interface Device
  3. Interactive Design of a Virtual Factory using Cellular Manufacturing System
  4. Design of Virtual Objects for Exact Collision Detection in Virtual Reality Modeling of Manufacturing Processes

VIDEO SESSION 3
VI. Industrial Automation
  1. Design and Performance Testing of the Sensors with Soft Contact Surfaces Having Eight Sensing Cells
  2. Distributed Control of FTS (Flexible Transfer System) Using Learning Automation
  3. Impedance Based Assembly
  4. Coordination of Multiple Mobile Platforms for Manipulation and Transport
  5. Weightlifting Motions for a Puma 762 Robot
  6. HighLAP: A High Level System for Generating, Representing, and Evaluating Assembly Sequences
  7. 3-D CAD Data Driven Assembly Robot Cell System
  8. Folding Cartons with Fixtures
VII. Robots in Games and Competition
  1. Beach Ball Volley Playing Robot
  2. Autonomous Fire-Fighting Mobile Robot Competition
  3. Robot Soccer System for NaroSot
I. HUMAN-ROBOT COORDINATION
A Control Strategy for Human-Friendly Robots

Authors:  Jochen Heinzmann+, Jon Kieffer*, and Alexander Zelinsky+ (+RSISE, Department of Systems Engineering, *Faculty of Engineering and Information Technology) The Australian National University, Canberra, ACT 0200, Australia
Contact:  Alexander Zelinsky
SISE, Department of Systems Engineering
The Australian National University
Canberra, ACT 0200, Australia
Phone: +61 2 6279 8840
Fax: +61 2 6279 8688
Email: Alex.Zelinsky@anu.edu.au
URL: http://www.syseng.anu.edu.au/rsl/

ABSTRACT:
The aim of this project is to develop a control strategy for human-friendly robots. We consider a robot to be human-friendly if a person can physically interact with the robot to teach it or to guide it while the safety of the person is guaranteed at all times [1]. The video shows the basic behavior of the system: A Zero-Gravity simulation. We are using a Barrett Whole Arm Manipulator (WAM) in our experiments because of its unique characteristics [2]. The WAM is a lightweight robot (low inertia) with 7 DOF driven by cable drives. This provides low friction and no backlash and makes the robot easily backdrivable by a person. The Zero-G module applies the appropriate motor torques to counteract the gravity effects. Thus, the robot is easy to move around for a person and bring it into any desired configuration. The psychological impression is that the robot is completely passive although considerable forces are required to the gravity compensation. It should be noted that the system does not precalculate the robot's trajectory as it would be in zero gravity but only responds to the current configuration of the robot. The robot is slowed down by friction only. In the next stage of the development we are implementing software which guarantees safe robot operation. It encapsulates the Zero-G module so higher-level control strategies can be implemented for a robot without gravitational effects while restricting the robot commands to safe threshold values. In the second part of the video the gravity constant is increased by 60%. This causes the robot to float up into a vertical configuration as if the robot would be mounted upside down.

REFERENCES:
  1. J. Heinzmann, Y. Matsumoto, Jon Kieffer and A. Zelinsky. "Smart interfaces + Safe mechanisms = Human friendly robots." Workshop on Humanoid and Human-Friendly Robots, pp. 25-27, Oct. 98, Tsukuba, Japan.
  2. W.T. Townsend and J.K. Salisbury. "Mechanical design for whole-arm manipulation, robots and biological systems." Toward a New Bionics?, pp. 153-164, 1993.
Go back to top.
Real-Time Stereo Face Tracking System

Authors:  Yoshio Mastumoto and Alex Zelinsky The Australian National University, RSISE, Department of Systems Engineering
Contact:  Alexander Zelinsky
Canberra, ACT 0200, Australia
Phone: +61 2 6279 8040
Fax: +61 2 6279 8688
Email: Alex.Zelinsky@anu.edu.au
URL: http://wwwsyseng.anu.edu.au/rsl/

ABSTRACT:
The aim of the project is to develop a visual human-robot interface, which allows a natural form of interaction between a human operator and the robot. The project is a continuation of our ongoing work in human-friendly robots and human-machine interfaces [1,2]. We consider facial gestures and the operator's gaze point to be important aspects of such a system. The video presents our real-time stereo face tracking system. Our system consists of a standard PC containing a Hitachi IP5000 vision processor. The system takes a stereo NTSC video stream in every frame using a field multiplexing technique [3]. Facial features such as the eyes and mouth are automatically detected using both skin colour and intensity information. Then template images of these facial features are acquired, and a 3D model is built using stereo matching. In the tracking stage, the 3-D measurements from stereo matching are fitted with the previously acquired 3-D model using spring-like connections, where each correlation value is treated as the stiffness in a spring between the model and the measurements. Using this approach, features that are tracking well have strong springs and help to estimate the positions of features that are tracking poorly. This results in a system, which is robust to rotation and translation motions, as well as deformation and partial occlusion of the features. Even when all the features are lost, our system can automatically recover as soon as the features re-appear. By using normalized correlation for feature tracking, our system is tolerant to significant fluctuations in lighting. Since the system runs at video frame rate, we are able to track rapid head motions. Finally the output of the tracking process is visualized by animating an artificial facial model on a Silicon Graphics computer. Not only does our system track facial features, it also can detect the position of the pupils of the eyes. This aspect of our system is also visualize d in the animation.

REFERENCES:
  1. J. Heinzmann, Y. Matsumoto, J. Keiffer, and A. Zelinsky. "Smart interface + Safe mechanisms = Human friendly robots." Proceedings of Workshop on Humanoid and Human-Friendly Robots, Oct. 1998, Tsukuba, Japan.
  2. J. Heinzmann and A. Zelinsky. "3-D facial pose and gaze point estimation using a robust real-time tracking paradigm." Proceedings of International Conference on Automatic Face and Gesture Recognition, 1998, pp. 142-14
  4. Y. Matsutmoto, T. Shibata, K. Sakai, M. Inaba, and H. Inoue. "Real-time color stereo vision system for a mobile robot based on field multiplexing." Proceedings of IEEE International Conference on Robotics and Automation, 1997, pp. 1934-193
Go back to top.
A Cobot in Automobile Assembly

Authors:  Prasad Akella and Nidamaluri Nagesh General Motors Corp. Witaya Wannasuphoprasit, J. Edward Colgate, and Michael Peshkin Northwestern University
Contact:  Michael Peshkin
Northwestern University
2145 Sheridan Road
Evanston, IL 60208, USA
Phone: 847-491-4630
Fax: 847-273-0559
Email: peshkin@nwu.edu

ABSTRACT:
Collaborative robots - cobots - are a new class of robotic device for direct physical interaction with a human operator in a shared workspace. Cobots implement software-defined "virtual surfaces" which can guide human and payload motion. A joint project of General Motors and Northwestern University has brought an alpha prototype cobot into an industrial environment. This cobot guides the removal of an automobile door from a newly painted body prior to assembly. Because of tight tolerances and curved parts, the task requires a specific escape trajectory to prevent collision of the door with the body. The cobot's virtual surfaces provide physical guidance during the critical "escape" phase, while sharing control with the human operator during other task phases. In this application the cobot serves as an example of an Intelligent Assist Device. IADs bring computer control to the traditional field of ergonomic assist devices. IADs, and cobots in particular, promise improved ergonomic performance and higher productivity in materials handling. For other applications of cobots, references, and contact information please start at http://cobot.com.

REFERENCES:
  1. Prasad Akella, Michael Peshkin, Ed Colgate, Wit Wannasuphoprasit, Nidamaluri Nagesh, Jim Wells, Steve Holland, Tom Pearson, and Brian Peacock. "Cobots for the automobile assembly line." 1999 International Conference on Robotics and Automation, Detroit, MI.
  2. W. Wannasuphoprasit, P. Akella, M. Peshkin, E. Colgate. "Cobots: A novel material handling technology." Proceedings of ASME IMECE 1998.
  3. "Intelligent assist. devices: A new generation of ergonomic tools" (Workshop), 1999 International Conference on Robotics and Automation, Detroit, MI.
Go back to top.
New Telecommunication using CTerm

Authors:  Masahiko Mikawa, Mizuki Tanno, Michito Matsumoto, and Junichi Masuda NTT Access Network Systems Laboratories, Nippon Telegraph and Telephone Corporation
Contact:  Masahiko Mikawa
NTT Access Network Systems Laboratories
Nippon Telegraph and Telephone Corporation
Tokai-Mura, Naka-Gun, Ibaraki-Ken 319-1193, Japan
Phone: +81 29 287 7661
Fax: +81 29 287 7294
Email: mikawa@ansl.iecl.ntt.co.jp

ABSTRACT:
Several methods or systems for tele-existence or telecommunication have been proposed [1][2]. However, contact tasks at a remote site are very difficult. This VCR presents a new telecommunication system with which to realize a new type of communication. When a skilled operator at an operation center teaches an unskilled worker at a remote site how to use a device by using a usual telephone, the skilled operator's instructions must be concrete and in full detail, and voice-only communication is sometimes insufficient. Our proposed telecommunication system using our newly developed Communication Terminal (CTerm) allows easy communication between a skilled operator and an unskilled worker without misunderstandings. The operation center and the remote site are connected by this telecommunication system through any public network, such as analog telephone networks, ISDN networks and the Internet. The operator and the worker can speak to each other and transmit images in real time. The core of this telecommunication system is the CTerm, which is set up at the remote site. A CCD camera and a laser pointer, that can be rotated by the skilled operator at the operation center, are attached to the CTerm. The operator can observe the remote site situations through the CCD camera, and make his intentions clear by directing the laser pointer to a particular spot. Moreover, high cost performance is realized by employing servo motors normally used in radio controlled models as the actuators.

REFERENCES:
  1. A. Hiraiwa, H. Tezuka, Y. Katayama, M. Motegi and T. Kakizaki. CYBERSCOPE : A small monorail camera robot system for tele-existence. Proceedings 7th IEEE International Workshop on Robot and Human Communication, pages 154--159, 199
  3. S. Tachi, T. Maeda, Y. Yanagida, M. Koyanagi, and H. Yokoyama. A method of mutual tele-existence in a virtual environment. Proceedings of The 6th International Conference On Artificial Reality and Tele-Existence, pages 9--18, 199
Go back to top.
Development of a 3,500 Tonne Field Robot

Authors:  Jonathan M. Roberts, Peter I. Corke, and Graeme J. Winstanley CSIRO Manufacturing Science & Technology, PO Box 883, Kenmore, Queensland, 4069, Australia
Contact:  Jonathan M. Roberts
CSIRO Manufacturing Science & Technology, CRC for Mining Technology and Equipment
PO Box 883, Kenmore
Queensland, 4069, Australia
URL: http://www.cat.csiro.au/cmst/automation

ABSTRACT:
The mining industry is highly suitable for the application of robotics and automation technology since the work is arduous, dangerous and often repetitive. This video shows the development of a physically large and complex field robotic system -- a 3,500 tonne mining machine called a dragline. Draglines are `walking cranes' used in open-pit coal mining to remove the material covering a coal seam. A very important aspect of this system is that it must work cooperatively with a human operator, seamlessly passing control back and forth in order to achieve the main aim -- increased productivity. The system was designed to operate in all weather conditions, 24 hours a day. A heuristic approach based on the control of drag rope tension and speed was used to control bucket carry angle. A system based on time-of-flight laser scanners was developed to determine the position of the bucket in 3D space. The automation system 'drives' the dragline by physically moving the control joysticks and pedals; somewhat like the cruise control in a car. The automation system has matched human operator swing time for some digging conditions.

REFERENCES:
  1. G. Winstanley, P. Corke, and J. Roberts. "Dragline swing automation." IEEE Conference on Robotics and Automation, 199
  3. J. Roberts, F. Pennerath, P. Corke, and G. Winstanley. "Robust sensing for a 3,500 tonne field robot." Proceedings of IEEE Conf. Robotics and Automation, 199
  5. J. Roberts, P. Corke, R. Kirkham, F. Pennerath, and G. Winstanley. "A real-time software architecture for robotics and automation." Proceedings of IEEE Conference Robotics and Automation, 199

ACKNOWLEDGEMENTS:
This work was funded by the Australian Coal Association Research Program (ACARP) as Project C5003, Rio Tinto, BHP Australia Coal Pty. Ltd., the Cooperative Research Center for Mining Technology and Equipment (CMTE), and CSIRO Manufacturing Science and Technology. Go back to top.
Human-Robot Interactions

Authors:  J.Y.S. Luh and Shuyi Hu Department of Electrical and Computer Engineering, Clemson University, Clemson, SC 29634, USA
Contact:  J.Y.S. Luh Department of Electrical and Computer Engineering Clemson University Clemson, SC 29634, USA
Phone: 864-656-5926
Fax: 864-656-5910
Email: luhj@prism.clemson.edu

ABSTRACT:
In the human-robot cooperative tasks, the robot is required to memorize different trajectories for different assignments, and automatically retrieve a proper one from them in real-time for the robot to follow when any assignment is repeated. Consider the task of carrying a rigid object jointly by a human and a robot (1). Since the human has all the essential sensors, he/she makes most decisions. The robot, which has a sensing wrist, carries the physical load. To start the task, the human leads the robot along a suitable trajectory and thereby achieves the desired goal. For every new task, the human is required to lead the robot. During the process, the trajectories are recorded and stored in memory as "skillful trajectories." The representation of the stored trajectories is position/orientation invariant. For completely or partially repeated tasks, the robot searches for a matching skillful trajectory in storage. If the robot finds one, it travels along t his trajectory without human assistance. Because of the invariance property of the stored trajectories (the matching and skillful ones), the on-going trajectories could be different by a proper translation and rotation. Once the matching process is complete, the robot follows the matched trajectory without human assistance. However, when the human notices that the path of the object is not, or will not be, acceptable, he/she interrupts the motion and applies compliance control (2) to lead the robot until another acceptable, skillful trajectory is identified and followed by the robot. The human then withdraws his/her assistance. However, interruption by the human is repeated intermittently whenever it is needed until the desired goal is accomplished.

REFERENCES:
  1. J.Y.S. Luh and Shuyi Hu. "Interactions and motions in human-robot coordination." Proceedings of 1999 IEEE International Conference on Robotics and Automation, May 10-15, 1999, Detroit, Michigan.
  2. O.M. Al-Jarrah and Y.F. Zheng. "Arm-manipulator coordination for load sharing using compliant control." Proceedings of 1996 IEEE International Conference on Robotics and Automation, April 22-28, 1996, Minneapolis, Minnesota, pp. 1000-1005.
Go back to top.
Dancing with Juliet

Authors:  Oussama Khatib, Kyong-Sok Chang, Oliver Brock, Kazuhito Yokoi, Arancha Casal, and Robert Holmberg Robotics Laboratory, Department of Computer Science, Stanford University
Contact:  Robert Holmberg Robotics Laboratory Department of Computer Science Stanford University Gates Bldg. 1A Stanford, CA 94305, USA
Phone: 650-725-8809
Fax: 650-725-1449
Email: rah@robotics.stanford.edu
URL: http://robotics.stanford.edu/~rah

ABSTRACT:
This video presents experiments in human-robot interaction using the Stanford Mobile Manipulator platforms [1]. Each platform consists of a Puma 560 manipulator mounted on a holonomic mobile base. The experiments shown in this video are the results of the implementation of various methodologies developed for establishing the basic autonomous capabilities needed for robot operations in human environments. The integration of mobility and manipulation is based on a task-oriented control strategy that provides the user with two basic control primitives [1]: end-effector task control and platform self-posture control. The major characteristic of this control structure is that the robot posture behavior has no impact on the end-effector dynamic behavior. In the video this control is applied to collision avoidance. While maintaining a desired task behavior at the end effector, the robot posture is changing in accommodation to the human motion for avoiding collision. This demonstration is a part of the elastic strip framework [2] for the real-time modification of planned motions. A robotic assistant must be capable of interacting and cooperating with a human. Human-guided motions involve tight cooperation performed through compliant motion actions. Several cooperative robots can be used to support a load while being guided by the human. The cooperation between platforms is based on a decentralized control structure that uses the augmented object and virtual linkage models [1] for dealing with the robot/load's dynamics and internal forces.

REFERENCES:
  1. O. Khatib, K. Yokoi, K. Chang, D. Ruspini, R. Holmberg, and A. Casal. "Coordination and decentralized cooperation of multiple mobile manipulators.'' Journal of Robotic Systems, Vol. 13, 1996, pp. 755-76
  3. O. Brock, and O. Khatib. "Elastic strip: Real-time path modification for mobile manipulation." Robotics Research, The Eighth International Symposium, Y. Shirai and S. Hirose, eds. Springer 1998, pp. 5-1
Go back to top.
II. NOVEL ROBOTIC SYSTEMS
Micro Inspection Robot for 1-Inch Pipes

Authors:  Koichi Suzumori, Toyomi Miyagawa, Masanobu Kimura, and Yukihisa Hasegawa Toshiba R&D Center
Contact:  Koichi Suzumori Mechanical Systems Laboratory, R&D Center, Toshiba Corporation 1, Komukai Toshiba-cho, Saiwai-ku, Kawasaki 210-8582, Japan
Phone: +81 44 288 8087
Fax: +81 44 520 2057
Email: suzumori@eml.rdc.toshiba.co.jp

ABSTRACT:
A micro inspection robot for 1-inch pipes has been developed [1]. The robot is 23 mm in diameter and 110 mm in length and is equipped with a high-quality micro CCD camera and a dual-hand for manipulating small objects in pipes. It can travel through both vertical pipes and curved sections, making possible inspections that would be difficult with conventional endoscopes. Its rate of travel is 6 mm/s and it has a load pulling power of 1 N. To realize this microrobot the authors have specially designed and developed several micro devices and micromechanisms: a novel micromechanism called a planetary wheel mechanism for robot drive; a micro electromagnetic motor with a micro planetary reduction gear to drive the planetary wheel mechanism; a micro pneumatic rubber actuator that acts as a hand; a micro CCD camera with high resolution [2]; and a pneumatic wobble motor [3] for rotating the camera and hands. In this video, the design and performance of the robot and micro devices used in it are shown.

REFERENCES:
  1. K. Suzumori, et al. "Micro inspection robot for 1 inch pipes." ASME/IEEE Trans. on Mechatronics, (To be published in 1999).
  2. Segawa, et al. "A micro miniaturized CCD color camera utilizing a newly developed CCD packaging technique." IEEE Transactions on Consumer Electronics, Vol. 41, No. 3, pp. 946-953, June 199
  4. K. Suzumori, K. Hori, and T. Miyagawa. "A direct-drive pneumatic stepping motor for robots: designs for pipe-inspection microrobots and for human-care robots." Proceedings of IEEE International Conference on Robotics and Automation, May 1998, pp. 3047-3055.
Go back to top.
1ms Grasping System Using Visual and Force Feedback

Authors:  Akio Namiki, Yoshihiro Nakabo, Idaku Ishii, and Masatoshi Ishikawa Department of Mathematical Engineering and Information Physics, University of Tokyo
Contact:  Akio Namiki Info. Phys. #1 Lab. Department of Mathematical Engineering University of Tokyo 7-3-1 Hongo, Bunkyo-ku, 113-8656 Tokyo, Japan
Phone: +81-3-3812-2111 (ex 6902)
Fax: +81-3-5802-2973
Email: namik@k2.t.u-tokyo.ac.jp
URL: http://www.k2.t.u-tokyo.ac.jp/~sfoc/

ABSTRACT:
In most conventional manipulation systems, changes in the environment can not be observed in real time because vision sensor is too slow. As a result the system is powerless under dynamic changes or sudden accidents. To solve this problem we have developed a grasping system using high-speed visual and force feedback. This system consists of three parts; a hierarchical parallel processing system by digital signal processors (DSP), an high speed active vision module based on SPE-256 [1] which is a 16x16 SIMD processing array with photo-detectors, and a human-like hand-arm which has a total of 21 degrees of freedom. Here a hand grasps a moving object observed by high-speed visual feedback and compliance control is executed using force feedback after grasping. The most significant feature of the system is the ability to process sensory feedback at high speed, that is, in about 1 msec. As a result grasping has high responsiveness and adaptiveness to the object motion. Then using a grasping algorithm with parallel sensory feedback on this system, various types of grasping are realized; adaptive grasping to object's changing shape, grasping with avoidance to an obstacle, and grasping with tracking to an object.
REFERENCE:
  1. I. Ishii, Y. Nakabo, and M. Ishikawa. "Target tracking algorithm for 1ms visual feedback system using massively parallel processing vision," Proceedings IEEE International Conference on Robotics and Automation, pp. 2309-2314, 1992.
Go back to top.
Self-Reconfiguration of 3-Dimensional Homogeneous Modular Structure

Authors:  Eiichi Yoshida, Satoshi Murata, Haruhisa Kurokawa, Shigeru Kokaji, and Kohji Tomita Mechanical Engineering Laboratory, AIST, MITI
Contact:  Eiichi Yoshida 1-2 Namiki, Tsukuba-shi, Ibaraki, 305-8564, Japan
Phone: +81-298-58-7108
Fax: +81-298-58-7091
Email: eiichi@mel.go.jp
URL: http://www.mel.go.jp/

ABSTRACT:
A 3-dimensional self-reconfigurable structure has been developed which is composed of identical mechanical units. Each unit is designed to build various 3-dimensional structures with other units and has six connecting arms along three orthogonal axes, which can rotate independently. The units change their configuration by using pair-wise motion. By repeating this kind of motion, the modular structures can reconfigure themselves. We have built prototype units. All rotation and connection are driven by single motor through electro-magnetic clutches. Rigid connection is realized by hand mechanism and sufficient torque is delivered through worm gears, so that one unit can lift another against gravity. The feasibility of 3-dimensional reconfigurable structure has been verified by experiments using four units. A variety of 3-dimensional structures can be constructed by using many units. We have developed a distributed self-reconfiguration algorithm that allows many units to self-assemble a given target shape. In the video, a simulation of self-assembly of 56-unit structure is presented to demonstrate the effectiveness of the algorithm. Possible applications include machines used in environments inaccessible to humans, for instance, planetary exploring vehicles or satellite antennae. Transported in a compact folded form, they can expand to the original structure when working, and repair themselves if some part is damaged.

REFERENCES:
  1. S. Murata, et. al. "A 3-D self-reconfigurable structure." Proceedings of 1998 IEEE International Conference on Robotics and Automation, pp. 432-43
  3. E. Yoshida, et. al. "A distributed reconfiguration method for 3-D homogeneous structure." Proceedings of 1998 IEEE/RSJ International Conf. on Intelligent Robots and Systems, pp. 852-859.


III. MOBILE ROBOTS

SpaceCat, a Micro-Rover Based on an Innovative Locomotion Concept

Authors:  Michel Lauria, Francois Conti, and Roland Siegwart Swiss Federal Institute of Technology, Lausanne (EPFL) Pierre-Alain Maeusli MECANEX S.A., Nyon, Switzerland
Contact:  Roland Siegwart Swiss Federal Institute of Technology Lausanne EPFL, DMT-ISR 1015 Lausanne, Switzerland
Phone: +41 21 693 38 50
Fax: +41 21 693 38 66
Email: roland.siegwart@epfl.ch
URL: http://dmtwww.epfl.ch/isr/asl

ABSTRACT:
The European Space Agency (ESA) is currently developing micro-rovers for planetary exploration. Within an interdisciplinary group of companies specialized in space applications and research labs new designs of micro-rovers have been investigated. Two concepts, a simple and robust one and an innovative one, have been selected and functional breadboard models of them have been built [1-3]. Their main features are: - Stowed dimensions (cm): 30 x 20 x 20 - Total mass including payload: + 3 kg - Maximum obstacle height to overcome: 10 cm The video segment concentrates on the locomotion concept of SpaceCat, the more innovative solution selected by the team. The micro-rover consists of 6 independently driven wheels arranged in two triangles. It therefore allows not only for efficient rolling on flat surfaces but also to step on obstacles. Additionally the center of gravity and the instrumentation carrousel are adjustable using some sort of a robot arm. This allows to optimally balance the micro-rover in almost every situation. Semi-autonomous navigate is based on various distance measurements and an on board inclinometer. It allows to easily control the rover in very rough terrain by using two joysticks. Autonomous control based on visual serving is also implemented, but not shown on the video.

REFERENCES:
  1. R. Siegwart, M. Lauria, P.-A. MŠusli, Van Winnendael. "Design and implementation of an innovative micro-rover." Proc. of Robotics 98, the 3rd Conference and Exposition on Robotics in Challenging Environments, Albuquerque, New Mexico, April 26-30, 199
  3. M. Lauria, Francois Conti, P.-A. MŠusli, M. Van Winnendael, R. Bertrand, R. Siegwart. "Design and control of an innovative micro-rover." Proc. of 5th ESA Workshop on Advanced Space Technologies for Robotics and Automation, ASTRA'98, The Netherlands, December 1-3, 199
  5. http://dmtwww.epfl.ch/isr/asl/projects/rosa.html
Go back to top.
Vision Based Navigation System for Autonomous Mobile Robot with Error Recovery

Authors:  Yasunori Abe, Masaru Shikano, Toshio Fukuda, Fumihito Arai, and Yoshio Tanaka Shinryo Corporation,Wadai 41 300-4247, Tsukuba, Ibaraki, Japan Department of Micro System Engineering, Nagoya University Center for Cooperative Research in Advanced Science & Technology, Nagoya University Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan
Contact:  Toshio Fukuda Department of Micro System Engineering Nagoya University Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan
Fax: +81-52-789-3909
Email: fukuda@mein.nagoya-u.ac.jp

ABSTRACT:
We have developed the robot that does the inspection of the air volume from square air diffusers that are called "anemo" on the ceiling at office buildings. The robot moves automatically, relocating its self-position by recognizing the anemo as the landmark with CCD cameras. However, self -position of the robot cannot be calculated when the robot fails in recognizing the landmark. Then the robot stops working. Therefore, we have proposed the behaviors of error recovery and "Hierarchical Adaptive and Learning Architecture System (HALAS)" that make the robot to be adapted to the environment by learning so that the robot can continue to work. In HALAS, there are 4 states of the robot. The upper state has more behaviors of error recovery than lower state. So, the robot changes to upper state when the error occurs. And, the time when the robot changes to lower state is decided by value function.

REFERENCES:
  1. T. Fukuda et al. "Navigation system based on ceiling landmark recognition for autonomous mobile robot - landmark detection based on fuzzy template matching (FTM)." Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, Vol. 2, pp. 150-155, 199
  3. T. Fukuda et al. "Navigation system based on ceiling landmark recognition for autonomous mobile robot - position/orientation control by landmark recognition with plus and minus primitives." Proceedings of the IEEE International Conference on Robotics and Automation, Vol. 2, pp. 1720-1725, 199
  5. Y. Abe et al. "Vision based navigation system considering error recovery for autonomous mobile robot." Proceedings of the IEEE International Conference on Robotics and Automation, Vol. 3, pp. 1993-1998, 199
Go back to top.
Gait Generation for Multi-Segment Inchworm-Like Robot Locomotion

Authors:  I-Ming Chen, Song Huat Yeo, Hui Leong Ho, Yan Gao School of Mechanical and Production Engineering Nanyang Technological University Nanyang Ave, Singapore 639798
Contact:  I-Ming Chen School of Mechanical and Production Engineering Nanyang Technological University, Nanyang Ave, Singapore 639798
Phone: 65-799-6203
Fax: 65-791-1859
Email: michen@ntu.edu.sg
URL: http://www.ntu.edu.sg/home/michen/index.html

ABSTRACT:
An inchworm-like robot is a mobile robot that imitates the locomotion pattern of a natural inchworm. This type of robots usually consist of a series of interconnected actuating modules that can either deform in the direction of travel (extensors) or produce friction against the environment (grippers). The gait of an inchworm robot is a sequence of actuator actions that will change the internal shape of the robot to generate net body motion. In this video, we demonstrate the computer simulation and experiment of locomotion gait generation for inchworm-like robot. The inchworm robot is modeled as a finite state automaton in which the grippers and the extensors have only simple binary actions (contracted or extended). Thus, the state of the robot can be expressed as a binary string. And the gait becomes a sequence of state transitions that follow kinematic constraints imposed by the environment. The gait generation problem utilizes the auxiliary actuator con cept and is posed as a search problem on the graph described by the finite state automaton that follows the kinematic of robot locomotion. Single-stride and multi-stride gait generations strategies are investigated. In the single-stride gaits, forward and backward motions are demonstrated as well as the fault-tolerant feature. In the multi-stride gaits, double-stride motion and standing wave gaits are demonstrated. Applications of this gait generation algorithm can be applied to the control of robotic endoscopes for medical inspection and pipe inspection systems using the inchworm locomotion principle. Future research will be emphasized on real-time on-line gait generation and gait change, sensor-based gait planning and general gait planning strategies.

REFERENCES:
  1. A.B. Slatkin and J.W. Burdick. "The development of a robotic endoscope." Proceedings IEEE/RSJ International Conference on Intelligent Robots Systems, pp. 162-171, 199
  3. M.C. Carrozza, L. Lencioni, B. Magnani, P. Dario, D. Reynaerts, M.G. Trivella, and A. Pietrabissa. "A microrobot for colonoscopy." 7th Int. Sym. on Micro Machine and Human Science, 199
  5. T. Fukuda, H. Hosokal, and M. Uemura. "Rubber gas actuator driven by hydrogen storage alloy for in-pipe inspection mobile robot with flexible structure." Proceedings IEEE Conference Robotics Automation, pp. 1847-1852, 198
  7. S.D. Kelly and R.M. Murray. "Geometric phases and robotic locomotion. J. robotic systems," Vol. 12, No. 6, pp. 417-431, 199
  9. J.E. Hopcroft and J.D. Ullman. Introduction to Automata Theory, Languages and Computation, Addison-Wesley, 197
Go back to top.
Sensor Networked Robotics

Authors:  Adam Hoover Electrical & Computer Engineering Department, Visual Computing Lab, University of California, San Diego Bent David Olsen, Department of Medical Information & Image Analysis Laboratory of Image Analysis, Aalborg University, Denmark
Contact:  Adam Hoover Electrical & Computer Engineering Department University of California, San Diego 9500 Gilman Drive La Jolla, CA 92093-0407, USA
Phone: 619-822-0176
Fax: 619-534-1004
Email: hoover@vision.ucsd.edu
URL: http://vision.ucsd.edu/~hoover

ABSTRACT:
The sensor networked robotics project focuses on a symbiosis between "blind" mobile robots and a stationary environment-based sensor network. We have developed novel methods to calibrate a video camera network spanning multiple rooms and corridors to a common coordinate system. Using this sensor network, we have invented a novel system to compute and broadcast an occupancy map in real-time. This spatial-temporal map may be received by any "blind" robot working in the area, to be used for path planning, auto-localization, and dynamic obstacle avoidance. In our video, we demonstrate our sensor networked robots in operation.

REFERENCES:
  1. A. Hoover and B. Olsen. "A Real-Time Occupancy Map from Multiple Video Streams." Proc. of IEEE International Conference on Robotics and Automation, Detroit, MI, May 199
  3. B. Olsen. "Robot Navigation Using a Sensor Network." Master's Thesis, Laboratory of Image Analysis, Aalborg University, June 199
Go back to top.
Reactive Navigation in Outdoor Environments

Authors:  Simon Lacroix, Hassan Haddad, and Matthieu Herrb LAAS/CNRS, 7, Ave du Colonel Roche, F-31077 TOULOUSE Cedex 4, France
Contact:  Simon Lacroix LAAS/CNRS 7, Ave du Colonel Roche, F-31077 TOULOUSE Cedex 4, France
Phone: +33 5 61 33 62 66
Fax: +33 5 61 33 64 55
Email: Simon.Lacroix@laas.fr

ABSTRACT:
We present in this video a reflex (or reactive) navigation mode we have developed and experimented with the robot Lama. The principle of the approach is the following: a pair of stereo images is acquired, and a correlation procedure produces a dense 3-D point image which is projected onto a grid. The probability for each grid cell to correspond to an obstacle is estimated by a Bayesian classifier. The goal to reach is provided by a visual tracker, and the elementary motion commands are finally generated using a potential field technique. This sequence is repeated as long as the goal is not reached, at a rate that is determined by the time necessary to process the data: when available, a new local map (or a new goal position) simply replaces the former one. The local maps are not fused into a global model, and the robot position is only defined relatively to the goal by the goal tracker: no dead-reckoning is required. The originality of the approach essentially relies in the way the probabilistic obstacle map is built, and how potential fields are defined upon such a map, in order to produce smooth and safe motions. The video sketches the various software components, and presents an experiment with the robot Lama.

REFERENCES:
  1. R. Chatila and S. Lacroix. "A case study in machine intelligence: Adaptive autonomous space rovers." A. Zelinsky, Editor, International Conference on Field and Service Robotics, Canberra (Australia), Number XII, Spring - Verlag, July 199
  3. H. Haddad, M. Khatib, S. Lacroix, and R. Chatila. "Reactive navigation in outdoor environments using potential fields." International Conference on Robotics and Automation, Leuven (Belgium), pp. 1232-1237, May 199
  5. S. Lacroix, S. Fleury, H. Haddad, M. Khatib, F. Ingrand, G. Bauzil, M. Herrb, C. Lemaire, and R. Chatila. "Reactive navigation in outdoor environments." LAAS Research Report # 98364, Submitted to International Journal of Robotics Research, 199
IV. MEDICAL ROBOTS Go back to top.
Laparoscopic Telesurgical Workstation

Authors:  Murat Cenk Cavusoglu University of California, Berkeley, Department of Electrical Engineering and Computer Sciences Michael Cohn, Endorobotics, Inc. Frank Tendick University of California, San Francisco Department of Surgery S. Shankar Sastry University of California, Berkeley Department of Electrical Engineering and Computer Sciences
Contact:  Murat Cenk Cavusoglu Department of EECS University of California at Berkeley 211-21 Cory Hall Berkeley, CA 94720-1770, USA
Phone: 510-664 0748
Fax: 510-642 1341
Email: mcenk@robotics.eecs.berkeley.edu
URL: http://robotics.eecs.berkeley.edu/~mcenk/

ABSTRACT:
Robotic telesurgery is a promising application of robotics to medicine, aiming to enhance the dexterity and sensation of minimally invasive surgery through millimeter-scale manipulators under control of the surgeon. With appropriate communication links, it would also be possible to perform remote surgery for care in rural areas where specialty care is unavailable, or to provide emergency care en route to a hospital. The UC Berkeley/Endorobotics/UCSF Telesurgical Workstation is a master-slave telerobotic system, with two 6 degree of freedom (DOF) robotic manipulators, designed for laparoscopic surgery. The slave robot has a 2 DOF wrist inside the body to allow high dexterity manipulation in addition to the 4 DOF of motion possible through the entry port, which are actuated by an external gross motion platform. The kinematics and the controller of the system are designed to accommodate the force and movement requirements of complex tasks, including suturing and knot tying. In this video, the telesurgical system will be introduced, followed by a video clip from the in vivo suturing and knot tying tasks performed at the Experimental Surgery Laboratory of University of California at San Francisco, by the resident surgeons.

REFERENCES:
  1. M.C. Cavusoglu. "Control of a telesurgical workstation." UC Berkeley ERL Memo M97/35, May 199
  3. M.C. Cavusoglu, F. Tendick, M. Cohn, and S. S. Sastry. "Control of a telesurgical workstation." Submitted to the IEEE Transactions on Robotics and Automation 199
  5. M.C. Cavusoglu, M. Cohn, F. Tendick, and S. S. Sastry. "Laparoscopic telesurgical workstation." Proceedings of the SPIE International Symposium on Biological Optics (BIOS'98), San Jose, CA, January 24-30, 199
ACKNOWLEDGEMENTS: This work is supported in part by NASA under grant STTR-NAS-1-20288, NSF under grant IRI-95-31837, and ONR under MURI grant N14-196001001200. V. VIRTUAL ENVIRONMENTS IN ROBOTICS Go back to top.
A Virtual Excavator for Operator Training and Controller Evaluation

Authors:  S.P. DiMaio, S.E. Salcudean, C. Reboulet*, Shahram Tafazoli, and Keyvan Hashtrudi-Zaad Department of Electrical and Computer Engineering, University of British Columbia, 2356 Main Mall, Vancouver, BC, Canada V6T 1Z4 (* ONERA-CERT, 2 avenue Edouard Belin, 31055 Toulouse, France)
Contact:  Simon DiMaio Department of Electrical and Computer Engineering University of British Columbia 2356 Main Mall, Vancouver, BC Canada V6T 1Z4 Tel: 604-822-9215
Fax: 604-822-9209
Email: simond@ece.ubc.ca

ABSTRACT:
Many thousands of excavator-based machines are currently employed worldwide in the construction, forestry and mining industries. Improvements in the control algorithms and user interfaces for such machines, as well as increased operator proficiency, could lead to substantial productivity gains. An excavator simulator has been developed in the Robotics and Control Laboratory at the University of British Columbia in order to facilitate the training of human operators and the evaluation of control strategies for heavy-duty hydraulic machines. This video outlines the simulator sub-systems and their operation. The user/operator controls a virtual excavator by means of a joystick, while experiencing visual and force feedback generated by environment dynamic models. The simulator comprises an impedance model of the excavator arm, a model for the bucket-ground interaction forces, a graphically rendered visual environment and a haptic interface.
REFERENCE:
  1. S.P. DiMaio, S.E. Salcudean, C. Reboulet, S. Tafazoli and K. Hashtrudi-Zaad. "A virtual excavator for controller development and evaluation." Proceedings of the IEEE International Conference on Robotics and Automation, pp. 52-58, May 1998; and cited work therein.
Go back to top.
Interaction with Simulated Environments using a Magnetic Levitation Haptic Interface Device

Authors:  P.J. Berkelman and R.L. Hollis The Robotics Institute, Carnegie Mellon University, Pittsburgh, PA 15213, USA
Contact:  Peter Berkelman The Robotics Institute Carnegie Mellon University Pittsburgh, PA 15213, USA
Phone: 412-268-1519
Fax: 412-268-5571
Email: pjb@ri.cmu.edu
URL: http://www.cs.cmu.edu/~pjb

ABSTRACT:
In the Microdynamic Systems Lab of the Robotics Institute at Carnegie Mellon University we have interfaced a high-performance 6-DOF magnetic levitation haptic interface device with various 3-D simulated environments. To use the device, the user grasps the levitated handle to manipulate a virtual object and feels its force and motion response as it contacts and interacts with other objects in the simulated environment. Noncontact actuation forces are generated by currents through wire ribbon coils in magnetic fields and position sensing is provided by planar position sensing photodiodes with lenses. The motion range of the levitated user handle is 25 mm in translation and 15-20 degrees in rotation to accommodate the fingertip motion of the user. The 75 Hz small-motion bandwidth and the 5 to 10 micron position resolution of the device is sufficiently high to display detailed environment interactions to the user's hand. In the first simulation the user manipulates a cube to feel vertex, edge, and face contacts with a cube-shaped environment. The second simulation demonstrates surface friction and texture effects. In the third simulation, the haptic device controller is coupled to an interactive physical simulation on a workstation to enable the user to dynamically manipulate objects in the simulated environment.

REFERENCES:
  1. P.J. Berkelman, Z.J. Butler and R.L. Hollis. "Design of a hemispherical magnetic levitation haptic interface device." ASME IMECE Symposium on Haptic Interfaces, Atlanta, Nov. 17-22, 1996, DSC-Vol. 58, pp. 483-48
  3. D. Baraff. "Interactive simulation of solid rigid bodies." IEEE Computer Graphics and Applications Vol. 15, No.3, 1995, pp. 63-7
  5. J.E. Colgate, M.C. Stanley, and J.M. Brown. "Issues in the haptic display of tool use." International Conference on Intelligent Robots and Systems, Pittsburgh, August 1995, Vol. 1, pp. 140-14
  6. Go back to top.
Interactive Design of a Virtual Factory using Cellular Manufacturing System

Authors:  M. Ernzer and T. Kesavadas Department of Mechanical and Aerospace Engineering State University of New York, Buffalo, NY, 14260, USA
Contact:  Prof. T. Kesavadas Dept. of Mechanical and Aerospace Engineering, 1006 Furnas Hall Buffalo, NY 14260, USA
Phone: 716-645-2593x2229
Fax: 716-645-3668
Email: kesh@eng.buffalo.edu
URL: http://wings.buffalo.edu/academic/department/eng/mae/vrlab/

ABSTRACT:
An interactive virtual factory testbed was developed as a part of a research to explore the applications of Virtual Environments (VE) in the area of manufacturing automation. This fully immersive factory built with special attention to details consists of a set of modular machines that a designer can drag and place in the factory to study issues such as plant layout, clusters and part flow analysis. This video further demonstrates a Virtual Matrix Interface for inputting information into the virtual factory. Cell formation and visualization techniques developed especially for the design of the virtual factory is demonstrated. A Modified Boolean Matrix approach (MBM) algorithm was improved as an example of applying traditional factory design methods in developing virtual factories of the future. Finally a novel interactive part-machine matrix visualization interface was developed to provide a more intuitive view of the mathematical results. The virtual fa ctory was created on a Silicon Graphics ONYX 2 computer can be visualized using stereo head mounted display and stereoglasses.

REFERENCE:
  1. M. Ernzer and T. Kesavadas (1998). "Interactive design of a virtual factory using cellular manufacturing system." To be presented at the IEEE International Conference of Automation and Robotics to be held in Detroit, MI, May 199
  2. Go back to top.
Design of Virtual Objects for Exact Collision Detection in Virtual Reality Modeling of Manufacturing Processes

Authors:  Rade Tesic and Pat Banerjee Department of Mechanical Engineering, University of Illinois at Chicago
Contact:  Pat Banerjee Department of Mechanical Engineering University of Illinois (M/C 251) 842 W. Taylor, 2039 ERF Chicago, IL 60607, USA
Phone: 312-996-5599
Fax: 312-413-0447
Email: banerjee@uic.edu
URL: http://www_ivri.me.uic.edu/

ABSTRACT:
Collision detection becomes a key issue when we want to model interactions between general, nonconvex objects in virtual reality applications which arise in manufacturing process domain. Despite significant progress, which has been made in developing efficient, exact collision detection algorithms for convex objects, limited and slow progress has been reported in developing collision detection algorithms for general, nonconvex objects at real-time interactive speeds. To narrow this gap we introduce a concept of virtual objects, which extends applicability of exact collision detection algorithms to nonconvex objects. This paper presents a methodology to encapsulate into virtual objects the surface patches of interest for collision detection as well as the automatic procedures for creation of virtual objects and for partitioning them into convex pieces. The collision detection technique described in this paper is best suited for interactive simulation and animation applications where high accuracy of object contact modeling is required. Examples include virtual assembly; mobile robot simulation; and simulation of manufacturing processes where accurate modeling of near-miss detection is essential, e.g. robotic painting, robotic welding, and NC machining operations.

REFERENCES:
  1. S. Gottschalk, M.C. Lin, and D. Manocha. "OBB-Tree: A hierarchical structure for rapid interference detection." Proc. SIGGRAPH, 199
  3. R. Tesic. "Collision detection and motion generation for virtual manufacturing simulator." Ph.D. Thesis, Department of Mechanical Engineering, University of Illinois at Chicago, 199
  4. VI. INDUSTRIAL AUTOMATION Go back to top.
Design and Performance Testing of the Sensors with Soft Contact Surfaces Having Eight Sensing Cells

Authors:  Branislav Borovac, Laszlo Nagy, Edvard Begovic, Milan Nikolic, Slobodan Dudic Faculty of Technical Sciences, University of Novi Sad, 21000-Novi Sad, Trg D. Obradovica 6, Yugoslavia
Contact:  Branislav Borovac Faculty of Technical Sciences, University of Novi Sad 21000-Novi Sad Trg D. Obradovica 6, Yugoslavia
Phone: +381-21-55 011
Fax: +381-21-59 536
Email: borovac@uns.ns.ac.yu

ABSTRACT:
A newly-designed force sensor with soft contact surfaces (SwSCS) consisting of eight sensing cells [1] is presented. By arranging the cells in an appropriate layout it is possible to obtain a sensor of desired complexity, capable of measuring all necessary component of the force acting on the object, held by the robot fingers. A main feature of the sensor is its capability to ensure simultaneously the object passive compliance caused by local surface irregularities and preserve contact between the object and the environment during the task realization. The sensor capabilities are verified and demonstrated experimentally. Its excellent sensitivity is shown by sensing the force produced by human breath. In the first experiment the robot holds a plastic straw and follows by it elastic band keeping permanently the sensor surface tangential to the band contour. In the second one, the contour surface is formed by two human fingers having very sharp singularity and, additionally, the surface quality changing from hard to soft. In the last experiment, using cylindrical probe the robot follows the curved slot with variable inclination. The robot has to change simultaneously the probe inclination relative to the slot plane caused by probe jamming, while tracking the slot profile.

REFERENCES:
  1. B. Borovac, L. Nagy, E. Begovic, M. Nikolic, and S. Dudic. "Performance testing of the sensors with soft contact surfaces having eight sensing cells." Fourth ECPD International Conference on Advanced Robotics, Intelligent Automation and Active Systems, August 1998, Moscow, pp. 294-29
  3. B. Borovac, L. Nagy, E. Begovic, and M. Sabli. "Use of sensors with soft contact surfaces for handling flat objects." Video Proceedings of the 1998 IEEE International Conference on Robotics and Automation, Leuven, Belgium, May 199
  4. Go back to top.
Distributed Control of FTS (Flexible Transfer System) Using Learning Automation

Authors:  Toshio Fukuda (1), Isao Takagawa (2), Kosuke Sekiyama (3), Yoshiaki Hasebe (4), Susumu Shibata (5), Hironobu Yamamoto (6), Yuji Inada (7)
Contact:  Isao Takagawa Department of Mechano-Infomatics and Systems School of Engineering, Nagoya University Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan
Phone: +81-52-789-4481 (Students' room of Fukuda laboratory)
Fax: +81-52-789-3909 (Office of Department of Micro Systems)
Email: fukuda@mein.nagoya-u.ac.jp takagawa@robo.mein.nagoya-u.ac.jp
URL: http://www.mein.nagoya-u.ac.jp/

ABSTRACT:
For many years, limitation of centralized control such as difficulty of reconfiguration has been strongly pointed out. Therefore, Distributed Autonomous Robotic System (DARS) has been attracting research interest as a flexible system that can be an alternative to a centralized system. As one of the DARS type production system, we have proposed Flexible Transfer System (FTS). The FTS is composed of a number of modules each of that equips actuators on the surface and a palette put on it can be transferred to arbitrary point in surface plane by passing it from module to module. Each module has its controller and controls itself without any central controller. Modules are connected each other via Ethernet and can exchange information. We adopted Learning Automata and reinforcement learning to generate a flexible path. A vector field is gradually formed over processes of real-time learning and finally a feasible path can be generated based only on local inform ation, without global information. A remarkable advantage of our approach is the adaptability against faults such as module break down. In our video, we show the experimental system of 3x4 modules and show that the adaptive transfer path can be formed based on the proposed method.

REFERENCES:
  1. N. Kubota, T. Fukuda, S. Shibata, H. Yamamoto, and Y. Inada. "Intelligent manufacturing system by the evolutionary computation - Schedule and planning of the flexible transportation." Proceedings of the 23rd Int'l Conference on Industrial Electronics, Control, and Instrumentation (IECON), Vol. 3 of 4, pp. 1136-1141 (1997).
  2. K. Najim and A.S. Pozinyak. "Learning automata theory and applications." PERGAMON, (1993).
  3. T. Fukuda, K. Sekiyama, Y. Hasebe, Y. Hasegawa, S. Shibata, H. Yamamoto, and Y. Inada. "Distributed control of flexible transfer system (FTS) Using Learning Automata." Proceedings of ICRA '99 (To appear). ACKNOWLEDGMENT
  4. Center for Cooperative Research in Advanced Science and Technology, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan.
  5. Department of Mechano-Infomatics and Systems, School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan.
  6. Department of Management and System Science, Science University of Tokyo Suwa College, Japan.
  7. Department of Mechano-Infomatics and Systems, School of Engineering, Nagoya University, Furo-ho, Chikusa-ku, Nagoya 464-8603, Japan.
  8. Toyo Engineering Corporation.
  9. Toyo Engineering Corporation.
  10. Toyo Engineering Corporation. Go back to top.
Impedance Based Assembly

Authors:  Wyatt S. Newman, Michael S. Branicky, Siddharth Chhatpar, Ling Huang, and Hao Zhang EECS Department, Case Western Reserve University, Cleveland, OH 44106, USA
Contact:  Wyatt Newman EECS Department Case Western Reserve University Cleveland, OH 44106, USA
Phone: 216-368-6432
Fax: 216-368-6039
Email: wsn@po.cwru.edu

ABSTRACT:
A large class of assemblies is still performed manually because conventional automation is unsuccessful. Under position control, automation will fail if the position uncertainty exceeds the tolerance for successful assembly. To emulate the success of manual assembly, robots must behave compliantly. The experiments shown in this video were conducted on an AdeptOne robot, retrofit with an open-architecture controller. Our control implementation is a flavor of impedance control that we call Natural Admittance Control [1,2]. In this technique, a model of a robot, incorporating estimates of actual inertias, is simulated in real time to respond to sensed end-effector forces, plus virtual forces defined with respect to a computed attractor. The physical robot is served to track the idealized simulation. Using this technique, we were successful in assembling automatic transmission components, including a forward clutch and a planetary gear-set. We demonstrated that a passive remote center of compliance was unsuccessful for these tasks, even in combination with implicit force control. With natural admittance control, assembly was gentle and reliable. In addition, the virtual attractor technique was found to be effective in bridging lower-level impedance behaviors with higher-level planning.

REFERENCES:
  1. W.S. Newman. "Stability and performance limits of interaction controllers." ASME J. Dynamics Systems, Measurement and Control, Vol. 114, No. 4, pp. 563-570, 199
  3. W.S. Newman, and Y. Zhang. "Stable interaction control and coulomb friction compensation using natural admittance control." Journal of Robotic Systems, Vol. 11, No. 1, 199
  4. Go back to top.
Coordination of Multiple Mobile Platforms for Manipulation and Transport

Authors:  Tom Sugar and Vijay Kumar University of Pennsylvania, 3401 Walnut Street, Suite 300C, Philadelphia, PA 19104-6228, USA
Contact:  Tom Sugar University of Pennsylvania 3401 Walnut Street, Suite 300C Philadelphia, PA 19104-6228, USA
Phone: 215-898-0352
Fax: 215-573-2048
Email: tgs@grip.cis.upenn.edu
URL: http://www.seas.upenn.edu/~tgs/web

ABSTRACT:
We have developed a small team of robots, which cooperatively grasp objects and transport them in a partially structured or unstructured environment. The robots are heterogeneous: a TRC Labmate equipped with a passive arm; a TRC Labmate equipped with an active force-controlled arm; and a Nomad XR4000 equipped with a stiff, position controlled fork-lift. Robots are used to move objects of different geometry and mass and can perform turns to avoid obstacles. With three platforms our system is also capable of carrying flexible objects such as a board. A critical aspect of the system is the compliant, force-controlled arm. The rear platform is equipped with a novel forklift-like arm, which possesses intrinsic mechanical compliance. Because the arm is compliant, a mobile robot can manipulate or interact with stiff objects not precisely positioned in the environment. Another important feature of this system is the ability to transport objects without special purpose fixtures. We use palm like end effectors to grasp the object and maintain force closure by coordinated control. Each robot is controlled by an independent controller. The robots are organized into a team with one or more leaders. The leaders plan paths based on sensory information avoiding obstacles. The plan is broadcast via a wireless peer to peer network to the one or more follower robots. The follower platforms are position controlled to precisely follow the leaders while the grasp forces are controlled by the active arm. Because the control is decentralized, the robots can be organized into different teams or function independently.

REFERENCES:
  1. T. Sugar and V. Kumar. "Design and control of a compliant parallel manipulator for a mobile platform." Proceedings of the 1998 ASME Design Engineering and Technical Conferences, Atlanta, Georgia, September 199
  3. T. Sugar and V. Kumar. "Decentralized control of cooperating manipulators." Proceedings of the IEEE International Conference on Robotics and Automation, Leuven, Belgium, May 199
  4. Go back to top.
Weightlifting Motions for a Puma 762 Robot

Authors:  Professor James Bobrow Graduate Students: Chia-Yu Wang and Wojciech Timoszyk Department of Mechanical Engineering, University of California, Irvine, CA 92697, USA
Contact:  Professor James Bobrow Department of Mechanical Engineering University of California Irvine, CA 92697, USA

ABSTRACT:
The goal of the project is to extend the payload of a Puma 762 robot beyond the 20-kilogram limit set by the manufacturer. We assume that the initial and final pose of the robot is given and that the motion between them is unknown. We then transform the kinodynamic motion planning problem into an optimal control problem. A direct method is then used to solve it. The joint trajectories are represented by B-spline polynomials, and the polynomial coefficients are varied in a gradient-based parameter optimization. Besides maximizing the payload, our solutions also minimize joint torques for a given load, thereby reducing wear on the robot. This work extends our previous results, Martin and Bobrow 1996, where effort was minimized, to payload maximization on a real robot. In addition, we added a time scale factor as a parameter in order to obtain the solution to the free final time problem. In our numerical solutions, we have found that the optimal motions rout inely swing through singular configurations. This is interesting in itself since most researchers try to avoid singular configurations. Just as is often done by human weightlifters, our solutions also often used a swinging motion to assist with the lift. Our software enabled us to increase the Puma's payload to over three times the manufacturer's specifications.

REFERENCES:
  1. C-Y. E. Wang, W.K. Timoszyk, and J.E. Bobrow. "Weightlifting motion planning for a puma 762 robot." To appear in the Proceedings of the 1999 IEEE International Conference on Robotics and Automation, Detroit, MI.
  2. B.J. Martin and J.E. Bobrow. "Minimum effort motions for open chain manipulators with task-dependent end-effector constraints.'' Proceedings of the 1997 IEEE International Conference on Robotics and Automation, Albuquerque, NM, April 1997, pp. 2044-204
  3. Also to appear in the International Journal of Robotics Research.
  4. F.C. Park, J.E. Bobrow and S.R. Ploen. "A lie group formulation of robot dynamics." International Journal of Robotics Research, Vol. 14, No. 6, pp. 609-618, 199

  5. ACKNOWLEDGMENTS:
    The research was supported by the National Science Foundation under grant number IRI-9711782.     Go back to top.
HighLAP: A High Level System for Generating, Representing, and Evaluating Assembly Sequences

Authors:  H. Mosemann, F. Rohrdanz, H. Ochs-Vietinghoff, S. Vietinghoff, T. Bierwirth, and F. Wahl Institute for Robotics and Process Control, Technical University of Braunschweig Hamburger Strasse 267, D-38114 Braunschweig, Germany
Contact:  Heiko Mosemann
Institute for Robotics & Process Control
Technical University of Braunschweig
Hamburger Strasse 267
D-38114 Braunschweig, Germany
Phone: +49 5313917463
Fax: +49 5313915696
Email: H.Mosemann@tu-bs.de
URL: http://www.cs.tu-bs.de/rob/

ABSTRACT:
High-level assembly planning systems generate plans for the automated assembly of mechanical products by robots. The sequences to be generated underlie several physical and geometrical constraints and in addition have to be efficient to increase productivity. The challenges still facing the field are to develop efficient and robust analysis tools and to develop planners capable of finding optimal sequences rather than just feasible sequences. The presented high level assembly planning system HighLAP automatically considers physical and geometrical constraints to generate and to evaluate stable assembly sequences. In this video we propose a relational assembly model including a CAD description and the specification of features and relations of the assembly components. We use an optional specification of an arbitrary hierarchy of assemblies to speed up and guide the generation of sequences. HighLAP evaluates all feasible assembly sequences considering sever al criteria like separability and manipulability of the generated (sub) assemblies. Furthermore, the necessity of reorientation for a mating operation and parallelism during plan execution is considered. Another important criterion is the stability of the generated (sub) assemblies. Therefore, we introduce a stability metric and an algorithm to calculate all stable orientations of an assembly considering friction. Experimental results are presented to demonstrate the efficiency of our assembly planning system.

REFERENCES:
  1. H. Mosemann, F. Rohrdanz, and F.M. Wahl. "Stability analysis of assemblies considering friction." IEEE Transactions on Robotics and Automation, 13(6):805-813, December 199
  3. H. Mosemann, F. Rohrdanz, and F.M. Wahl. "Assembly stability as constraint for assembly sequence planning." IEEE International Conference on Robotics and Automation, pp. 233-238, Leuven, Belgium, May 199
  5. F. Rohrdanz, H. Mosemann, and F.M. Wahl. "Generating and evaluating stable assembly sequences." Journal of Advanced Robotics, 11(2):97-126, 199
  6. Go back to top.
3-D CAD Data Driven Assembly Robot Cell System
Author &
Contact:  Satori Kojima
2nd Manufacturing System Research Department
Manufacturing Technology R&D Center
Ricoh Company, Ltd.
1005 Shimo-Ogino, Atsugi City, Kanagawa-Pref.
243-0298, Japan
Email: koji@ptrc.ricoh.co.jp

ABSTRACT:
A 3D-CAD-data driven Assembly Robot Cell (ARC) system is proposed in order to cope with the frequent change on production design and amount for the customers' demand, promptly and flexibly. Focusing on the information generated by designers during the production development process, the necessary information is extracted and integrated into Product Model as Assembly Model Data (AMD). Using the ARC, which is supplied the CAD data and AMD corresponding to each part, it can plan assembly process and assembly sequence, and generate assembly robot task program in the virtual environment automatically. Also the system is able to show the designers visually the difficult parts of an assembly, and to assemble actual products via an autonomous robot cell with a function of 3-D object recognition. The products' CAD data and AMD are supplied to each ARC through a network. The utility of AMD and the best format for these data are investigated in the assembly process. The data format was evaluated through loading actual AMD in the ARC system.

REFERENCES:
  1. S. Kojima, T. Hayashi, P. Kerites, and H. Hashimoto. "Research for 3-D-CAD data driven autonomous assembly robot cell systems." JSME Proc. JSME Annual Conference on Robotics and Mechatronics (ROBOMEC'98), June 1998 (in Japanese).
  2. S. Kojima, P. Kerites, T. Hayashi, and H. Hashimoto. "Assembly model data in robot cell systems." IEEE Proc. IEEE/RSJ Int. Conference on Intelligent Robots and Systems, Vol. 3, pp. 1037-1038, Oct. 199
  4. A. Fuchigami, S. Kojima, H. Hashimoto. "Object pose estimation from range image detected by laser range finder." SICE Proceedings of the 5th Symposium on Robot Sensors, pp. 115-118, Apr. 1996 (in Japanese). Go back to top.
Folding Cartons with Fixtures

Authors:  Liang Lu and Srinivas Akella Beckman Institute, University of Illinois at Urbana-Champaign, 405 North Mathews Avenue, Urbana, IL 61801, USA
Contact:  Dr. Srinivas Akella
Beckman Institute
University of Illinois at Urbana-Champaign,
405 North Mathews Avenue
Urbana, IL 61801, USA
Phone: 217-244-1622
Fax: 217-244-8371
Email: sakella@uiuc.edu
URL: http://www.uiuc.edu/ph/www/sakella

ABSTRACT:
Packaging products such as telephones and two-way radios after assembly is a common manufacturing task. Carton folding is a packaging operation typically performed by human operators or with fixed automation. In this video, we demonstrate a flexible method to fold cardboard cartons using fixtures; a carton blank is folded by moving it through a fixture with a robot. This method enables rapid changeovers between product models by interchanging fixtures, where each fixture is designed for a carton and a selected folding sequence. We have developed a motion planning algorithm that generates all folding sequences for a carton by modeling it as a manipulator arm with revolute joints and branching links. A fixture constrains the carton to move along paths consisting of line segments in its configuration space, and the motion planner generates these paths. To illustrate the method, we selected a folding sequence for an example carton, designed a fixture, and dem onstrated folding of the carton from blanks with an AdeptOne robot.

REFERENCES:
  1. Liang Lu and Srinivas Akella. "Folding cartons with fixtures: A motion planning approach." Proceedings of the 1999 IEEE International Conference on Robotics and Automation, Detroit, MI, May 1999.
  2. Tomas Lozano-Perez. "A simple motion-planning algorithm for general robot manipulators." IEEE Journal of Robotics and Automation, RA-3(3):224-238, June 1987.


VII. ROBOTS IN GAMES AND COMPETITION

Go back to top.
Beach Ball Volley Playing Robot

Authors:  Ozaki Fumio, Sato Hirokazu, Hashimoto Hideaki, Oaki Junji, Ogawa Hideki, Uenohara Michihiro, Yoshimi Takashi, Asari Yukio, Maeda Katsuhiro, Tatsuno Kyoichi, Nakai Hiroaki, Taniguchi Yasuhiro, Tanaka Eiji, Yamaguchi Osamu, Tachimori Mitsuyoshi, Okada Satoshi, and Bamba Hiroyuki R&D Center, Toshiba Corporation, 4-1 Ukishima-cho, Kawasaki-ku, Kawasaki 210-0862, Japan
Contact:  Ozaki Fumio
R&D Center
Toshiba Corporation
4-1 Ukishima-cho, Kawasaki-ku, Kawasaki 210-0862, Japan
Email: fumio.ozaki@toshiba.co.jp

ABSTRACT:
Future robots are going to work in your houses and hospitals with humans. Toshiba has developed a beach ball volley playing robot as a demonstration of such a human friendly robot technology. We believe that it is essential to interact with robot using everyday words, such as "Lets play volleyball." For the everyday word commands to work, the robot needs 1)to measure the target relative position with respect to the robot position, 2)to know the mechanics and procedures of the tasks, and 3)to have a good database for the environment and the target. Here we have realized two of them: we have 1)developed a controller which can calculate a target relative position at a rate of 1/60 seconds with newly developed field rate image labeling boards and 2)described the procedures and mechanics of the following three tasks. We have demonstrated a robot which can play beach ball volley with an opponent using the ball position data feedback to calculate the hitting po sition and time, pick up a ball from the floor using visual servo control, and shake hands with a human using the mechanics describing hand motion, by only one respective everyday word command.

REFERENCES:
  1. H. Nakai, et al. "A volleyball playing robot." Proceedings of the 1998 IEEE International Conference on Robotics & Automation, Leuven, Belgium, pp.1083-1089, 1998.
  2. F. Ozaki, et al. "Beach ball volley playing robot." Proceedings of the IARP 1st International Workshop on Humanoid and Human Friendly Robotics, Tsukuba, Japan, VI-1, 199 T. Hashimoto, T. Kimoto, T. Ebine, and H. Kimura, "Manipulator control with image-based visual servo," 1991 IEEE International Conference on Robotics and Automation ICRA '91, Sacramento, California, pp. 2267-2272, April 1991. Go back to top.
Autonomous Fire-Fighting Mobile Robot Competition

Authors:  David J. Ahlgren and Jacob E. Mendelssohn Department of Engineering, Trinity College, 300 Summit Street, Hartford, CT 06106, USA
Contact:  David J.Ahlgren
Department of Engineering
Trinity College
300 Summit Street
Hartford, CT 06106, USA
Phone: 860-297-25884
Fax: 860-297-3531
Email: dahlgren@trincoll.edu

ABSTRACT:
This video shows fire-fighting mobile robots in competition in the 1998 Trinity College Fire fighting Home Robot Contest, the largest robot contest in the U.S. open to all ages, affiliations, and ability levels. In this contest, fully autonomous mobile robots navigate through a 8 ft. by 8 ft. maze, find a candle, and extinguish it in minimum time. The maze includes four rooms with connecting hallways that are 18" wide. Each robot makes three runs. The score is the sum of the two best run times with reduction factors for reliability, obstacle avoidance, return to the starting point, and ability to trigger on a smoke alarm. Deductions are given for non dead-reckoning operation. The contest presents a challenging interdisciplinary design problem requiring development of sensors, motor drives, and reliable, fault-tolerant real-time software [1], [2]. Successful robots have employed various sensing technologies including IR, ultrasonic, and UV devices to enable navigation, obstacle avoidance, and flame detection. Word sizes of on-board computers have ranged from four bits to 32 bits, and codes have been developed in C, BASIC, and assembly language. Robots that compete in this contest are scaled-down versions of full-scale fire-fighting robots that will provide fire security in single-story homes of the future.

REFERENCES:
  1. R. Avanzato. "Collaborative mobile robot design in an introductory programming course for engineers," Presented at the 1998 ASEE Annual Conference, Seattle, July 1, 1998.
  2. D.J. Pack, A.M. Mankowski, and G.J. Freeman. "A fire-fighting robot and its impact on educational outcomes," Presented at the 1998 ASEE Annual Conference, Seattle, July 1, 1998. ACKNOWLEDGEMENT: The authors gratefully acknowledge the support of Motorola, Inc. Go back to top.
Robot Soccer System for NaroSot

Authors:  Jong-Hwan Kim, Byung-Kook Kim, Kui-Hong Park, Heung-Soo Kim, Sung-Ho Kim, and Jong-Suk Choi Department of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), 373-1 Kusong-dong, Yusong-gu, Taejon 305-701, Korea
Contact:  Jong-Hwan Kim
Department of Electrical Engineering
Korea Advanced Institute of Science and Technology (KAIST)
373-1 Kusong-dong, Yusong-gu, Taejon 305-701, Korea

Phone: +82-42-869-8048/3448
Fax: +82-42-869-8010
Email: johkim@vivaldi.kaist.ac.kr
URL: http://www.fira.net

ABSTRACT:
The FIRA Robot World Cup France'98 in Paris witnessed the smallest robots ever to play the game of soccer between the Y2K2 and the BEST teams from Korea. It was the NaroSot (Nano-Robot World Cup Soccer Tournament) category with robots sized 4cm x 4cm x
  • 5cm. In the NaroSot category, only the vision information is used as a feedback, and there are no internal sensors such as encoders within the robots. Due to this, precise and fast robot control is difficult. Cooperation and coordination among robots are very important in NaroSot, as there are ten robots on the playground. This video shows two nano-robot soccer systems, the BEST and the Y2K
  • The BEST used a three-layered controller, which is hierarchical in nature. Role allocation and libero strategies are used by this team. The Y2K2 is designed with an action selection layer and it used a fixed role allocation scheme. The NaroSot system is a good benchmark for multi-robotic system research.

    REFERENCES:
    1. J.-H. Kim, H.-S. Shim, H.-S. Kim, M.-J. Jung, I.-H. Choi, and K.-O. Kim. "A cooperative multi-agent system and its real time application to robot soccer." Proc. IEEE International Conf. on Robotics and Automation, pp. 638-643, 1996.
    2. H.-S. Shim, H.-S. Kim, M.-J. Jung, I.-H. Choi, J.-H. Kim, and J.-O. Kim. "Designing distributed control architecture for cooperative multi-agent system and its real-time application to soccer robot." Journal of Robotics and Autonomous Systems, Ed. J.-H. Kim, Vol. 21, No. 2, pp. 149-165, Sept. 1997.