This paper proposes a control method for walking robots by manipulating the zero moment point of the motion. The control law is developed based on a simple model, in which motion of a point mass in a sagital plane is governed by the gravitation and reaction force from the ground. Stability of the system under the control is analyzed by using the state plots of the motion trajectory. The objective of the control method is to control the balance of the robot without depending on the strict tracking control, which is required, in conventional methods, in order to prevent the robot from falling down.
We have developed a Virtual Endoscope System (VES) with force sensation to train inexperienced young doctors and simulate operations that require special technical skills. In this paper, we describe the force simulation mechanism employed by our VES and dynamical models of an endo-scope and colon. The force simulation mechanism was developed with the use of four rubber rollers and differential gears. The simple structure of this mechanism enables easy control and stable linear and rotational drive of the endoscope, and we have confirmed its force simulation ability by several experiments. We also developed the dynamical models of an endoscope and colon to calculate reactive force that doctors receive from a colon in realtime, and we confirmed the accuracy of these models by software simulation of endoscopic. This system is adequate for training young doctors in endoscopic insertion.
A method for estimating object pose and position is proposed. This method is based on integration of data obtained using several kinds of sensors, such as spotting rangefinders, scanning rangefinders, and cameras. Integration of data obtained by such different kinds of sensors is generally difficult because there are various measurement methods and they cause difference of the accuracies. Defined here are local shape models for each of the sensors and equations that express the distance between the sensed data and models. Based on the accuracy of each sensor, these equations are integrated as a cost function that is minimized to estimate the optimal object pose and position. The configuration of the sensors is modified according to the environment and the degree of accuracy required, allowing flexible and robust object location. The feasibility of the proposed method is shown by simulation and experiments.
This paper presents a reflective walk of a quadruped robot based on reflections to realize an adaptive walk in a dynamic environment. The combination of reflections, a vision-cued swaying reflection  and a reflective gait, makes the robot walk reflectively, without programming the exact motion of each joint of the legs. An experiment is shown to demonstrate how the proposed method works.
This paper discusses an EMG based control method of a robotic manipulator as an adaptive human supporting system, which consists of an arm control part and a hand and wrist control part. The arm control part controls joint angles of the manipulator's arm according to the position of the operator's wrist joint measured by a 3D position sensor. The hand and wrist control part selects an active joint out of four joint degrees of freedom and controls it according to EMG signals measured from a human operator. A distinctive feature of our method is to use a statistical neural network for EMG pattern classification. This network can acquire stochastic representation of measured EMG patterns through learning based on the log-linearized Gaussian mixture model, so that it can adapt changes of the EMG patterns according to the differences among individuals, different locations of the electrodes, time variation caused by fatigue or sweat, and so on. It is shown from the experiments that the EMG patterns during hand and wrist motions can be classified sufficiently by using the network. It may be useful as an assistive device for handicapped persons.
This paper deals a strategy to estimating contacting positions between transported object and multiple mobile robots with force sensors. One of the most important problems in cooperative transportation by multiple mobile robots is to realize relative positions among contacting points around an object. It is demanded to recognize current position of a robot itself in world coordinates or to avoid some obstacles; but a physical constraint through an object between cooperating robots is dominant in cooperative transportation. For recognizing a motion error of mobile robots, this system estimates positions of robots by a physical constraint as mentioned above. A physical constraint is obtained by force sensor value from each robot. This constraint and a geometrical shape of an object enable the system to estimate an arrangement of contacting points around the transported object. Our numerical simulation and expetiment have verified that contacting positions between robots and an object can be estimated without an external means to estimate contacting positions; such as vision system or anything else.
Accurate object pose measurement is very important for manipulating objects by robotic hands. In this paper, we propose a systematic method to estimate the object pose with a complementary use of camera images and the finger joint information. Statistical characteristics of measurement errors are considered when two disparate sensor measurements are fused. The object pose is estimated not only at each camera sampling but also at inter-camera-sampling, so that the estimation can be used in the manipulation control servo loop. To verify the proposed pose estimation method, experiments are carried out. The proposed pose estimation method was combined with a control scheme for object manipulation by a multi-fingered hand with soft fingertips, which has also been proposed by the authors. Results of the experiment show the effectiveness of the proposed pose estimation method and the overall control system for object manipulation by a soft-fingered hand.
Parallel wire driven robots have some advantages such as high speed, heavy load and so on. In the previous works in the parallel wire driven robots, the object or the end-effector is controlled through each wire length. However, each actuator position in the base coordinates and each measured length of wires include errors. Therefore, even if fine positioning in wire length coordinates is realized, the end-effector position may deviate from a desired position in task oriented coordinates. To overcome such difficulty, we propose a new motion control scheme for the parallel wire driven robots in this paper. In the proposed scheme, a pseudo inverse matrix which shows the relation between a wire tensions vector and a force vector of the end-effector is utilized to operate the object in the task oriented coodinates. When the proposed control scheme based on the force relation is used, it is easy to attain more precise positioning by using external sensors such as cameras, and to apply it to force control. In this paper, we prove the motion convergence to desired points and discuss its robustness based on Lyapunov stability analysis. Finally the usefulness of the proposed control scheme is demonstrated through some experimental results.
A snake is able to attain high terrain adaptability and versatile locomotion even though it has an extremely simple onedimensional configuration. In order to utilize these functions for robotics, we have adapted the basic biological machine elements of the snake into the Active Cord Mechanism (ACM) . And we have discussed about the creeping dynamics and applications to manipulation. In this paper, we developed a new experimental model named“ACM-R1”with a selfcontained system, which realizes higher mobility and terrain adaptability compared with a past model. Next, gliding experiments on ice were carried out in order to demonstrate that the creeping motion is the same as the principle of skating. Finally, a new terrain adaptive control method for sloping surfaces is proposed, and we verified the effectiveness by slope climbing experiments.
In this paper, we propose a control scheme of parallel manipulators focusing on the accuracy of acceleration on the endplate, which is an important factor when parallel manipulators are used as acceleration displays. We use two controllers—dynamic controller to achieve accuracy of position and to stabilize the system, and H∞ controller to feedback the acceleration measured on the endplate. The main problem of dynamic control is computational com-plexity. In order to reduce computation time for inverse dynamics, parallel processing method called multi-thread programming is applied. H∞ controller is added outside the closed loop of dynamic control to remove the vibration due to the elasticity of the links and the influence of modeling errors in the dynamic controller.
In this paper, we propose a new formulation of the dynamics of human figures, which are characterized by their frequent contact with the environment and underactuation due to the free joint of the body link. The formulation is shown to have wide variety of applications from dynamics simulation to generation of physically consistent motions of underactuated systems including human figures, meaning that the proposed formulation gives a common basis for the problems concerning motions of human figures. The computational method of simulating collisions and contacts itself has advantages compared to previous methods: efficiency and stiffness of computation are achieved by two factors, that is, (1) to apply rigid contact model instead of spring-damper model in which extremely large collision and contact forces are applied to make the computation instable, and (2) to avoid the inequality conditions of contact forces used in conventional rigid contact model approaches, leading to quite simple and efficient solution.
This paper presents the space test of the ARH (Advanced Robotic Hand System), which is the world's first precise extravehicular robot aboard the satellite“Hikoboshi”. The telerobotic system has features of dexterity, autonomy and flexible operability, using a three-finger multisensory hand at a work site in space and a computer-graphics-based desktop interface at an operation site on the ground. The concept of sensor-fused telerobotics utilizing multisensory information is introduced, and the system is implemented to perform high precision tasks under the barrier of inter-satellite space communication. The robot system was launched into low earth orbit, and the capability of sensor-fused telerobotics was successfully demonstrated in precise in-orbit servicing.
A mobile robot is difficult to move in a pipe in general, since the locomotional plane in a curved surface and the gravity direction operating on the robot are changeable. However, the wheel-type robot is able to locomote even in the curved pipe so that it can avoid an obstacle when an appropriate steering control is found. Also, the robot may locomote automatically by adjusting itself without the assist of an operator. This paper describes the analysis of the wheel-type robot's locomotion trajectories in pipe by using Ackermann-geometry expanded in three dimensions. At first the robot and the pipe are modeled for the analysis. Then the conventional Ackermann-geometry is expanded to become useful in three dimensions, and the method for calculating and estimating the trajectories is discussed. Later the locomotion in the straight pipe is simulated with the slimmed model of the wheel. Furthermore, the experiment to get actual locomotion trajectories is performed and the validity of the analysis is verified by comparing the analytical data with the experimental results.