In this paper, a low sensitive position control system using a DC series servo motor driven by a chopper is presented. The digital simulation is employed for the design of this system in order to take account of non-linearity of the motor. The simulated responses are compared with the experimental ones. The low sensitivity is verified for the inertia load about ten times as large as that of the motor itself. The responses of duty cycle and driving electric energy for the specification are also presented. As a result, it becomes clear that this system is of much practical use, because the driving apparatus of this system is small and light comparing with the DC amplifier.

View full abstract

This paper considers a decentralized stabilization problem for large-scale inter- connected systems. Dynamic controllers for subsystems are designed sequentially using the proper-stable-factorization approach. Sufficient conditions for simultaneous stabilization of the overall system as well as subsystems are obtained. It is also shown that the closed-loop overall system can be further made connectively stable under appropriate minimum phase conditions.

View full abstract

This paper deals with the development of computer-aided design systems for mold designers. Specifically a model of design process for working out mold specifications and the program structure of a decision support system developed on the basis of such a model are considered. These are the nuclei of AI-CAD and design expert systems for mold designing. Major features of the design process model and the program architecture discussed in this paper may be summarized as follows : (1) The design process is presented in a new form different from such a fixed designing process as described in the existing CAD systems for molds. (2) All items in mold specifications which are called “design variables” in the design process are given as a set of attributes or functions. (3) The process for deciding on the solutions of design variables forms a hierarchical model consisting of three phases represented in the so-called “design knowledge”, including design strategy, a graphic representation of correlations, the values of attributes or functions, and design equations. (4) The knowledge base that forms an integral part of this system can be developed by directly rewriting the design knowledge formulated in the hierarchical model in accordance with the production rule. (5) Knowledge engineering and an optimization method are used in deciding on the values of those variables which are called “design solutions” in the design process. Choosing specifications of injection molds for a design task, this paper also discusses the development of knowledge and optimization processing modules for the decision support system for working out such mold specifications, and then proves the validity of these processing modules through case studies.

View full abstract

This paper presents a model reference adaptive control for the retarded time delay system, which is a continuous-time linear system with a time delay in the state. The time delay can be compensated by the linear filter which belongs to the finite Laplace trnsformation. The adaptive controller estimates unknown integral kernels of the filter and unknown parameters of the time delay system. At the same, time the controller makes the output of the time delay system follow that of the prespecified reference model. A convergence proof of the output error between the time delay system and the model is given under a boundness assumption on the all state included in the control system. The result is verified by a simple numerical simulation performed on a 2nd order retarded time delay system.

View full abstract

In this paper we propose a hierarchical structure of controllers for robot manipulators which allows real time computation of the dynamic control. In this method, the computation of the dynamic control is separated into two levels. In the upper level, mainly the inverse dynamics and the desired trajectory are computed at a low sampling frequency. In the lower level, the computation for servo control is done at a high sampling frequency. Since the dynamic control and the servo control are separated clearly and the control structure is very simple, this method is expected to be useful for practical purposes. The computational amount at each level is estimated for n-degree-of-freedom manipulators. Finally, the effectiveness of the proposed method is shown by some simulation and experimental results using a 2-degree-of-freedom planer manipulator.

View full abstract