Cooperative transportation by two four-legged robots with implicit communication

https://doi.org/10.1016/S0921-8890(99)00034-2Get rights and content

Abstract

In this paper, we show cooperative object transportation by two four-legged robots. To realize cooperative transportation with autonomy, we adopt implicit communication based cooperation. Each robot uses only its own sensors to estimate the state of a task. With this method, people can attend to the cooperation system without any changes. We will show an algorithm and two experiments; one is by two legged robots and the other is by one robot and a human.

Introduction

Recently, many researches about cooperation of multiple mobile robots are studied. In this research field, “cooperation” and “autonomy” play very important roles. But in some cases, they conflict with each other. When the design of a multiple mobile robot system is weighted on “cooperation”, the system can not only work efficiently but also is easily fixed and has less applicability. On the other hand, when it is weighted on “autonomy”, the system has high applicability but has less efficiency and less stability.

In this paper we take up a problem that autonomous legged robots transport one object cooperatively. Cooperative transportation of an object requires high cooperativeness. Control of legged robot requires high autonomy since it needs very complex motion. With this subject matter, we aim to construct a strategy that two-legged robots move autonomously but they cooperate to transport an object over a step or rough terrain. For the autonomy, we adopt an implicit communication between the robots.

Section snippets

Related work

There are some studies on cooperative transportation by mobile robots. Stilwell [9], [10], Hashimoto [2], Kosuge [4], [5] co-workers proposed planning methods for cooperative transportation by wheeled mobile robots, which have one leader robot and some followers in a group. Stilwell and Kosuge did not use any explicit communication which uses transmitter, receiver or any other communication equipment. Follower robots use a force sensor and estimate states of a leader and others, and then decide

Cooperation of legged robots

Fig. 1 shows an image of transportation by two four-legged mobile robots. We adopt legged type mobile robots for cooperation. When a legged robot is planned to transport an object, the most serious problem is its payload. Usually, it is almost all it could do to support its own weight. It is hard to bring another object. Because of this problem of payload, legged robots have never been used for practical work even though it is said that they are suited to work on rough terrain. But it is

Cooperative transportation strategy

As a strategy that two legged robots transport one object without explicit communication, we adopt a method that they use their own sensor and move asynchronously by implicit communication which uses sensor of force from the object. We describe this method in the following sections.

Experimental system

The experimental system is shown in Fig. 3, Fig. 4. The four-legged robot “TITAN-VIII” was developed by Arikawa and Hirose [1]. The body has 600 mm in length and each leg has three-degrees-of-freedom driven through wire.

A six-degrees-of-freedom force/torque sensor which has 5 kg/50 kgf range is attached at the end-effector. In this experiment the robot does not have a limit switch to sense over-going of an object. So we replace a limit switch in the strategy with another threshold for Mx.

As a

Conclusion and future works

In this paper, we showed a strategy for two legged robots for cooperative transportation. To keep autonomy of each robot we do not use tight communication with explicit method but adopt a method that robots get information only with implicit communication. Since each robot does not have any communication equipment and decides its motion only by its own sensor, some robots in the cooperation can be relieved by a human. Practically, one robot and one human could transport an object with our

Acknowledgements

This research is supported by the Grant-in-Aid for Scientific Research on Priority Areas “Intelligent Robots” by Ministry of Education, Japan.

Yasumichi Aiyama was born in 1967 in Tokyo, Japan. He received M. Eng. and Dr. Eng. from the Faculty of Engineering, the University of Tokyo in 1994 and in 1997, respectively. Since 1997, he has been a Research Associate at the Faculty of Engineering, University of Tokyo. His research interests are human-like dexterous manipulation and cooperative transportation by multiple mobile robots.

References (11)

  • K. Arikawa, S. Hirose, Development of quadrupled walking robot TITAN–VIII, in: Proceedings 1996 IEEE/RSJ International...
  • M. Hashimoto, F. Oba, T. Eguchi, Dynamic control approach for motion coordination for multiple wheeled mobile robots...
  • O. Khatib, K. Yokoi, K. Chang, D. Ruspini, R. Holmberg, A. Casal, Vehicle/arm coordination and multiple mobile...
  • K. Kosuge, T. Oosumi, Decentralized control of multiple robots handling and object, in: Proceedings 1996 IEEE/RSJ...
  • M.J. Matarić, M. Nilsson, K.T. Simsarian, Cooperative multi-robot box-pushing, in: Proceedings 1995 IEEE/RSJ...
There are more references available in the full text version of this article.

Cited by (34)

View all citing articles on Scopus

Yasumichi Aiyama was born in 1967 in Tokyo, Japan. He received M. Eng. and Dr. Eng. from the Faculty of Engineering, the University of Tokyo in 1994 and in 1997, respectively. Since 1997, he has been a Research Associate at the Faculty of Engineering, University of Tokyo. His research interests are human-like dexterous manipulation and cooperative transportation by multiple mobile robots.

Mitsuhiro Hara was born in 1972 in Kanagawa, Japan. He received M. Eng. from the Faculty of Engineering, the University of Tokyo in 1998. He is a graduate student in the Engineering Course and working for his doctoral thesis. He is interested in intelligent robotics and is now doing research on cooperation of multiple walking robots.

Takashi Yabuki was born in 1974 in Ibaraki, Japan. He joined Hitachi, Ltd. in 1993. In 1996, he was a research worker in the Faculty of Engineering, University of Tokyo. Now he is being engaged in the work of an intellectual property relation.

Jun Ota was born in 1965 in Saitama, Japan. He received M. Eng. and Dr. Eng. from the Faculty of Engineering, University of Tokyo in 1989 and in 1994, respectively. From 1989 to 1991, he worked with Nippon Steel Cooperation. In 1991, he was a Research Associate of the University of Tokyo. Since 1996, he has been an Associate Professor at the Faculty of Engineering, University of Tokyo. His research interests are motion planning, task planning of multiple mobile robots, cooperative control of multiple robots, and landmark design for robots.

Tamio Arai was born in 1947 in Tokyo, Japan. He received his M.S. degree and D.S. degree in Engineering from the University of Tokyo in 1972 and 1977, respectively. He was a visiting researcher in the Department of Artificial Intelligence, Edinburgh University in 1979–1981, He has been a Professor in the Department of Precision Machinery Engineering, University of Tokyo since 1987. He has mainly worked on robotics and manufacturing engineering. Currently his research interests include (1) automatic assembly, (2) planning and control of plural mobile robots, and (3) robot language and protocols in manufacturing.

View full text