This paper treats the positioning control problem of the endtip of space manipulator without con-trolling the motion of the spacecraft which is the manipulator's base. The conventional local feedback at each joint is not applicable for such a case, because the position and orientation of the endtip cannot be functions only of the joint displacements. We propose a sensory feedback scheme for space manipulators based on the artificial potential defined in a sensor coordinate frame. In this scheme, information measured with a visual sensor on the base is fed back to the joints in term of control torques. Since the base's motion changes the view of the visual sensor, the manipulator must track the target seeming to move. However, we can prove that the endtip asymptotically approaches thee target by using the Liapunov's method. Finally the effectiveness of the proposed scheme is shown by results of computer simulation for a 6-link space manipulator model.

View full abstract

Walking according to change of various environments is called an adaptive walking. We define the lower level adaptive walking as a walk consisted of only trajectory control and walk-patern generating. The purpose of this study is to realize the lower level adaptive walking in sagittal plane by a biped locomotion robot in indoor space.
In this paper, we propose a simplified procedure of walk-pattern generation as follows that walk-patterns for various environments are generated by adjusting and combining the basic walk-patterns which are previously given for typical environments. Based on this idea, walking experiments are carried out for various environments. As a result, dynamic walking in flat plane (about 1.5 second per a step with 0.3 m), change of step width in walking and walk in various combined environments of flat plane, an obstacle as pipe and a stair (up, down) are realized.

View full abstract

This research deals with an active three-dimensional position sensor with laser beam. The sensor has a unique function of differentiation processing of a linear image to extract only the points to be detected at which luminous intensity changes. Since it is composed of a linear image sensor and a one-dimensional PSD, both with high resolution, it is designated to be of high speed and accuracy.
In the present paper, the structure of the sensor and the sensing mechanism with image processing are described and a signal processing circuit to improve the accuracy of position detection is considered. Finally, the sensor is installed as a vision to a manipulator, to detect points on a stair-shaped object.

View full abstract

This paper discusses input-dependent stability on torque servo system of robot hands. With an insufficient space to mount actuators inside the finger joints, a finger joint is normally actuated by a tendon-sheath driving system. Although this driving system enables us to construct a finger system with simple mechanism and large flexibility, friction and compliance existing in tendon-sheath driving system generally bring a nonlinear characteristic involving a hysteresis into the dependence of joint torque output from actuator displacement. While this transmission characteristic is close to the well-known backlash behaviour appearing for the gears system, the author newly found input-dependent characteristics existing in the backlash region of the transmission system through the precise experiments. Namely, the characteristic is that the tendon-sheath transmission system behaves as if there were virtual springs in the backlash region and the strength of virtual spring depending on the input. Furthermore, it was confirmed that there are close relationships between the input-dependent backlash characteristics and the input-dependent stability. Based on the experiments, the transmission characteristics are modelled in simple equations, and the system stability is explored by using the techniques of sinusoidal-input-describing-function (SIDF) . Resultantly, we obtained non-dimensional stability-criterion-maps explaining the experimental results successfully.

View full abstract

A unified method of the dynamic analysis for general spatial mechanisms and over-constrained mechanisms has been developed. When a screw space representing motion of a mechanism is a mobile space, the velocity and the acceleration motors of the mechanism belong to the mobile space which is invariant with the motion of the mechanism. All the motors in the equations of motion of the mechanism can be expressed by the elements in accord with the basis of the screw space and this make it possible to derive equations which can be applied to mechanisms with several constraint conditions. These equations can be utilized in the unified computer program for the dynamic analysis of serial mechanisms and closed loop mechanisms having arbitrary mobile screw space. They can also be applied to the robot simulator by combining them with the dynamics of servo-mechanisms of actuators.

View full abstract

This paper proposes a new actuator using magnetic fluid. The new actuator uses electromagnetic force driven by a solenoid coil, which is controlled by electric current of the coil. Theoretical and experimental results clarify the possibility of the actuator. Theoretical results also clarify that the power feedback can linearize the control system using a force control servosystem.

View full abstract

PD-Tyre Two-Stage Robust Tracking Control System for robot manipulators was proposed by Osuka et. al. in 1988. This controller has the following properties.
(a) The desired trajectory can be specified.
(b) The control accuracy can be specified.
(c) The convergence characteristics of initial error can be specified.
(d) The robustness to modelling errors and to measurement noises is ensured.
(e) The structure of the controller is simple.
In this paper, we applied the controller to DAIKIN Miniature Robot, and showed the effectiveness of the structure of Two-Stage Control Scheme with some experimental results. However, since the miniature robot has an undesirable property such as a magnitude of measurement noises in angular velocity is so big, we had to solve a few technical problems in order to design the controller for even such a special robot. Therefore, in the final part of this paper, we considered a way to overcome those problems.

View full abstract

In this report, we propose a new trajectory control law of robot arm using sliding mode control and show several experimental results and related problems. Many researchers have proposed several sliding mode algorisms for the trajectory control of robot arm. however, there is no explicit algorism which need no inverse inertia matrix calculation. Furthermore, there are few reports treating the experimental results.
Therefore this report should be useful to the engineers studying this fields.

View full abstract

While knowledge about the working environment expands the autonomy of a telerobot, human support is necessary to complement its autonomy. Especially, human intervention at robot motion level is required to support the flexible task execution intuitively. Since the motion level intervention changes the real world directly without modifying the robot knowledge, it causes the inconsistency between them.
This paper presents a method for the world model management using the motion understanding system which recognizes the meaning of robot motions commanded by the motion level intervention.
The system is designed to have a hierarchical structure composed of symbolizers and a rule based motion interpreter. The symbolizer extracts events in the task execution from such signals as the robot hand pose, finger width, grasping force and external force, and sends them to the interpreter. The event driven production system realizes efficient interpretation and ensures the readability of the rules and the extensibility of the system.
The experiment proves that the motion understanding system can recognize a pick and place task even from the fluctuating operator's task execution using a master-slave manipulator. It also shows that the consistency of the world model can be maintained automatically.

View full abstract

This paper presents an experimental study on a control problem of a space manipulator installed on a free-flying telerobot. The authors' main goal is to develop a control method for dexterous operation of telerobot, with considering the dynamical interaction between the manipulator arm and the base satellite in space micro-gravity environment. In order to control space free-flying manipulators, the generalized Jacobian matrix is introduced. By substituting the conventional Jacobian for this matrix, conventional control methods for ground-fixed manipulators is directly applicable for space manipulators. In order to simulate the free-flying behavior of mechanical links in micro-gravity environment, a laboratory model of space telerobot supported on air bearings is developed. An on-line resolved motion-rate control scheme with vision feedback is proposed for experimenting target capture operation, by means of the generalized Jacobian matrix. The telerobot model can properly chase and capture both a standing target and a moving target in spite of complex satellite/manipulator dynamical interaction. The experimental results confirm the validity of the generalized Jacobian concept and proposed control method.

View full abstract