Tsuyoshi Okada

Robust control system with observer

This paper proposes a design method for a robust controller including a state observer. Though the optimal regulator has often been used as a desirable controller, it is not necessarily robust if it includes a state observer whose input uses the plant input as well as the plant output. Some attempts to recover robustness have been reported, but perfect robustness has not been achieved. Our method is based on the principle that the return difference is made to satisfy the circle condition for the system to remain robust, and the response properties are improved by the extended perfect model-following system. Design examples are given, and the responses of the designed systems are simulated to exemplify the effectiveness of our method.

Robust Control System Design Synthesis with Observers

A method of designing a robust controller including an observer for a multi-input/multi-output controlled system is presented. The well-known robustness of the optimal regulator is valid only when the regulator is constructed by full-state feedback, and this robustness is not guaranteed if an observer is introduced. In this study, we add a new output feedback loop to recover the robustness of the system with an observer and use a precompensator to improve the response characteristics. Singular-value analysis is applied to improve the robustness of the synthesized system. An example for an aircraft is calculated, and good simulated results are demonstrated.

Sensitivity Reduction by Double Perfect Model Following

We propose a double perfect model following (DPMF) system which uses two models: one is used to define the desirable response and the other is used to reduce sensitivity to parameter perturbation. The concept is shown and methods of model selection and controller synthesis are discussed, followed by numerical examples and an aircraft control example which demonstrate performance of the DPMF concept. It is shown that the DPMF system can be less sensitive than a nominally equivalent single perfect model following (SPMF) system and that sensitivity reduction is enhanced by proper selection of the model. It is also shown that the sensitivity of multiinput systems can be reduced componentwise or blockwise by using a decoupled type model.

Reduced conservative singular value analysis for robustness

In recent years, many studies have been reported on methods of examining the problems of the robustness of multi-input and multi-output systems in the frequency domain using the singular value technique. However, since the conventional criteria have been derived as sufficient conditions for stability or sensitivity reduction, they are more than sufficient (or so-called conservative conditions) and therefore they have the disadvantage that they cannot be used for the evaluation of stability and/or sensitivity reduction. In this paper we propose new criteria for robust stability and sensitivity reduction with reduced conservatism which utilize equivalent systems with the same transfer function as that of the plant. In order to illustrate the effectiveness of these criteria, examples of the longitudinal and lateral motion of aircraft are analysed by both the conventional and the proposed methods, and the results are compared to show the advantage of the proposed method.

Robust model following systems

This paper proposes a robust model following (RMF) system which realizes good properties such as asymptotic stability, disturbance rejection and model following with reduced sensitivity for plant parameter variation. For a given plant, the RMF system is constructed as a double perfect model following system which uses two models, one used to give a good response to be followed and the other to provide the robustness as well as to reduce the sensitivity. It is shown that RMF system can be generally designed for the plant whose output has the same dimension as the input. It is also shown that when some conditions about the forms of the system and parameter variations are satisfied, the plant output which has the larger dimension than the input can be made robust. Furthermore, the robustness for variations of parameters other than the plant is discussed. A design procedure for the RMF system is given. As an example, the flight control system of a helicopter is designed and its characteristics are examined by...

Rigorous Treatment of the Hunting of Relay Servomechanism Operating on Sampled Data

The signals of radar, digital computer and many other systems are sampled signals, In this paper the hunting of on-off control system operating on such a sampled signal is treated. The analysis in this paper is a rigorous one, solving piece-wise linear differential equations with conditions given at the point of discontinuity. The relays considered in this analysis are two-position relays with and without hysteresis. The graph obtained by the analysis can also be used in synthesis. The experiment was carried out using a servo system, and the result of the analysis agrees very well with that of the experiment.

A Method of Decreasing the Amplitude of Hunting of Relay Servomechanism Operating on Sampled Data

In the preceding papers, the method of analyzing the hunting or on-off control system operating on sampled signals were treated. In this paper, we will show how to decrease the amplitude of hunting and design a desired control system. A device of obtaining differentials will also be explained. Especially for the controller designed by the authors, the decrease of amplitude is analytically calculted and the result is compared with that of the experiment. Experiments were also carried out with continuous and sampling relay control systems, with and without hysteresis and the results were compared with each other.

Construction and Test of an Analog-Digital Converter

Summery ; This paper reports on the test results, their data and also circuit elements of an Analog-Digital converter, constructed in our laboratory. This converter is a device which converts the d-c voltage to the binary code signal. The pulse counting method is applied and the converting process is as follows. (1) A sawtooth voltage is generated in the device. (2) The d-c voltage is compared with the sawtooth voltage. (3) The elapsed time between the starting point of the sawtooth and the point where it coincides with the d-c voltage is measured by chronometric pulses. (4) These pulses are digitized by the binary counter. This converter is of an all electronic type and we obtained rather satisfactory results.