Dual-Arm Space Robot Teleoperation

English/Japanese

Space robotics is one of the important technologies in space developments. Especially, it is highly desired to develop a completely autonomous robot which can work without any aid of the astronauts. However, with the present state of technologies, it is not possible to develop a complete autonomous space robot. Therefore, the teleoperation technologies for the robots with high levels of autonomy become very important. Currently, the technologies where an operator teleoperates a space robot from within a spacecraft are already in practical use, like the capture of a satellite with the shuttle arm. However, the number of astronauts in space is limited, and it is not possible to achieve rapid progresses in space developments with the teleoperation from within the spacecraft. For this reason, it has become highly desired to develop the technologies for the teleoperation of space robots from the ground in the future space missions.

Aerospace Robot System for A-ARM (ARS/A) has been constructed about ten years ago and using ARS/A, we have been carrying out many researches and technology developments for a space robot teleoperation from the ground. A 6-DOF slave arm which is named A-ARM (Aoba-ARM) and a force reflecting joystick with same level of freedom have been developed for ARS/A. With only one slave arm, however, it is not possible to realize the complex extra vehicle activities (EVAs) demonstrated by astronauts. Moreover, it is also difficult to execute tasks demanding flexibility with a non-redundant manipulator arm. Due to these limitations, a new system for Dual-Arm Robot Teleoperation in Space (DARTS) is developed which fulfills these two demands [1], [2]. Figure 1 shows the organization of DARTS, while fig. 2 shows physical look of the actual system.

Fig. 1 Organization of DARTS.

Fig. 2 A look of the DARTS setup.

DARTS consists of the space system, the ground system, and the software development system.

Figure 3 shows the construction of the slave system based in space. The part of the space system is composed of a dual-arm manipulator system and two controlling computers. The manipulator system has 7 degree of freedom in each slave arm and is named as Paired-Arm (P-ARM). A 6-axis force/torque sensor is mounted at the tip of each arm, and a BarrettHand is used as a multipurpose hand for each of the two. This hand can grasp objects of many shapes. The image processing system contains an image processor board and a vision tracking board in it. An ordinary PC compatible machined equipped with the real time OS VxWorks controls P-ARM, BarrettHand and the force/torque sensors. Another machine equipped with Linux controls the image processing system. Linux is a free operating system, yet it can run many powerful softwares like Lisp and MATLAB etc.

Fig. 3 A look of the space system.

The ground system is composed of the master arm and two computers. The 6-DOF haptic interface developed in our laboratory is used as the master arm. A PC equipped again with VxWorks controls this haptic interface. While the other machine is a graphics workstation which simulates the 3D graphics model of the space system. This teleoperation environment, which is quite comfortable for an operator, is based on the virtual reality that effectively uses the haptic interface and the 3D graphics.

Fig. 4 3D graphics model of the space system.

The software development system is also composed of two machines. One of them is a Sun Ultra 10 acting a VxWorks host system. This workstation host is used to develop the control programs for both master and slave systems. While the other machine is an SGI OCTANE SSI system used as a terminal for effective whole system integration.

With this DARTS system, until now, the concept of a virtual radar is proposed as an effective teleoperation technology [3]. While the operator executes fine tasks with teleoperation, he/she always concentrates on the slave system end-effector. Therefore, in this situation, though there is a little chance of collisions between the end-effector and the obstacle, yet it is not possible for the operator to avoid the collisions between the slave system links/joints with the obstacles in the dead space. To avoid this situation, the obstacle avoidance information for a whole slave arm, including that of its links and joints, is transposed to and displayed at the end-effector. The experimental graphics design utilizing the concept of virtual radar is shown in Fig. 5. The red object with blue lattice shown at the hand of the arm's graphics is the virtual radar.

Moreover, the concepts of "virtual grip" and "virtual ball" have also been introduced to help the operator teleoperate the slave system comfortable [1]. Along with this, the procedure to teleoperate two slave arm with only one master controller is also proposed.

Fig. 5 Concept of the virtual radar.

Reference

  1. Woo-Keun Yoon, Yuichi Tsumaki, and Masaru Uchiyama: An Experimental System for Dual-Arm Robot Teleoperation in Space with Concepts of Virtual Grip and Ball, Proceedings of the Ninth International Conference on Advanced Robotics ('99 ICAR), Tokyo, Japan, (1999/10/25-27), 225-230.
  2. Woo-Keun Yoon, Yuichi Tsumaki and Masaru Uchiyama: An Experimental Teleoperation System for Dual-Arm Space Robotics, Journal of Robotics and Mechatronics, 12-4, (2000) (to be published).
  3. M. Takahashi, Y. Tsumaki, D. N. Nenchev and M. Uchiyama: A Teleoperation System with Collision Avoidance Capability Based on Virtual Radar, Proceedings 6th IEEE International Workshop on Robot and Human Communication ('97 RO-MAN), Sendai, Japan, (1997/9/29-10/1), 154-159




[Home] [Researches and Projects]