This paper proposes a simple and high performance motion control method of robot manipulators in task-oriented coordinates. In the proposed method, it is necessary that each joint generates the specified acceleration by the controller of each joint. To extend the result to multi-degrees-of-freedom robot manipulators, the acceleration tracing orientation method (ATOM) is obtained. The ATOM is a controller design method in task-oriented coordinates. High performance position controller and hybrid position/force controller are easily realized by the ATOM. From the analyses of the controllers, it is known that the precise acceleration controller is the key to improve the performance of the ATOM. The realization of the strict acceleration controller, however, is almost impossible by using the conventional disturbance compensation methods. For example, the inverse dynamics has complicating calculation and is weak against parameter variations. A resent research about disturbance observer enables the realization of the acceleration controller. Contrary to the inverse dynamics, the observer is simple and robust against parameter variations. High performance acceleration controller is realized by the disturbance observer. The effectivness of the ATOM is shown by some experiments of a three-degrees-of-freedom robot.
This research is concerned with master slave manipulation. A new master slave system in teleexistence mode, i. e., impedance controlled master slave system, is presented. Impedance control of the manipulators is applied to this system. The system regulates the impedances of the master and the slave so that they coincide, and it generates the impedance models of the environment. Four basic control methods are proposed as the extension of conventional bilateral systems and it is shown that most of them can be transformed to each other by calculating force information even with less force information. They are also applied to the system with time delay. These methods are verified in the experiments with a two degrees of freedom direct drive manipulator.
This research is concerned with master slave manipulation. A new master slave system in teleexistence mode, i. e., the basic part of impedance controlled master slave system, has been presented in the part I. In the part II, the system is expanded so that it controls both impedance models of the environment and the manipulators, and modifies the force sensation of the master and slave manipulators as if an operator were in the different physical environment, and the impedances of the master and slave manipulators are determined by using similarity transformation in physical modelling when the manipulators have different scales, force and/or velocity conditions. These methods are verified in the experiments with the two degrees of freedom direct drive manipulator.
This paper describes a basic design of cooperative control system for multiple robot arms. In the system two robot arms and a vision sub-system have their own node controllers connected by a computer network. Cooperative control software developed is embedded in each node controller. Node software is written by object oriented programming technique. System model is constructed as a basis to represent and interpret cooperative task. To provide uniform measure to get external information of modeled objects in the real world, virtual information source-sensorlink-is designed. Real information sources distributed among node locations are transparently accessed by sensorlink. Task requirements imposed on the system are resolved to inner regulative laws. When motion trajectory of the task constrains two arms relatively, the laws leave degrees of freedom for actions of arms to give an alternative way of fulfillment. The example task executed is to push a button on the box held by one arm with the other's hand tip. Two arms move hands to suffice the laws using visual information through sensorlink. Then local node trouble in keeping the law-joint stop caused by given workspace-can be avoided using unconstrained degrees of freedom cooperatively.
In the real world, there are many kinds of uncertainties in motions of robots. Moreover, robots cannot always get sufficient knowledge about tasks in advance. Therefore, intelligent robots must posses both problem solving ability to decide on proper motions with insufficient knowledge and learning ability to acquire knowledge from experiences. Effective integration of the two abilities is also important. In this paper, we describe an experimental system named ARPEX-L (Advanced Robot Planning and Execution System with Learning Ability) . We applied ARPEX-L to two kinds of robot tasks, pick and place task and pushing block task. The former is an simple example of automatic robot programming. The latter is an attempt to make an actual robot learn by a symbolic method. Current defects of the system and future works are also described.
In this paper, a force control of a robotic manipulator based on a neural network model is proposed with consideration of the dynamics of both the force sensor and objects. This proposed system consists of the standard PID controller, the gains of which are augmented and adjusted depending on objects through a process of learning. The authors proposed a similar method previously for the force control of the robotic manipulator with consideration of dynamics of objects, but without consideration of dynamics of the force sensor, showing only simulation results. This paper shows the similar structure of the controller via the neural network model applicable to the cases with consideration of both effects and demonstrates that the proposed method shows the better performance than the conventional PID type of controller, yielding to the wider range of applications, consequently. Therefore, this method can be applied to the force/compliance control problems. The effects of the number of neurons and hidden layers of the neural network model are also discussed through the simulation and experimental results as well as the stability of the control system.
Many studies concerning remote manipulator systems have been published recently. Most of these studies are concerned about using end effectors with limited degrees of freedom. In order to carry out complicated tasks, it is necessary to develop a bilateral master/slave hand system. We developed a configuration differing bilateral master/slave hand. A control experiment using force feedback is carried out, and the bilateral servo system for master/slave control is investigated. The necessity of slip sensation feedback in a master/slave hand system is pointed out, and a slip sensation feedback method is suggested and validated through experiments.
We are researching ways to use neurocomputers that have highly parallel data processing and learning functions for robot control. There are three requirements for the robots: The robot must be easy to control, but the neural network must be sophisticated enough to handle multiple sensor input. Second, the robot must be able to learn easily. Third, the robot must be able to adjust its own actions. We developed a new mobile mechanism, created a network model, and increased the network learning speed. Sensor signals from the robot are input to the neural network. The network outputs a certain reaction pattern in response to the sensor input. Then the reaction is refined to an ideal one using training patterns. A robot can change its reaction pattern by changing the training pattern. We created two robots with different action patterns: one chases other robots, the other runs away from other robots. We confirmed that a neurocomputer can effectively control robot actions.