Ryozaburo Tagawa

Evaluation of robust model matching for the control of a DC servo motor

The evaluation of a new robust control design method called Robust Model Matching (RMM) is treated by applying it to the design of electrical position servos. RMM is a practical method for the control of plants with wide parameter variations, however, the effect of RMM has not yet been sufficiently examined in experiments and simulations. RMM is applied to the design of electrical position servos, and is evaluated with simulations and experiments. The results of these simulations and/or experiments show that the electrical position servos designed by RMM can eliminate the undesirable influences of: (a) load variations; (b) torque disturbances; (c) non-linear friction; (d) dead zone in a power amplifier; and (e) uncertainty of a velocity sensors output signal; upon controlled variables. It is also shown that electrical position servos can be robustly stabilized simultaneously by RMM controllers for plants with arbitrary bounded parameter variations.

Characteristic transfer function matrix-based linear feedback control system analysis and synthesis

In this article, the concept of control system synthesis via a method using the characteristic transfer function matrix (CTFM), is explained. Then, the formal definition of CTFM and the associated concept of connectivity are presented. Next, CTFM-based control system analysis and synthesis are discussed, together with the derivation of the necessary and sufficient conditions for the realisation of a proper controller. A summary of the design procedure is presented and three design examples are introduced, in order to illustrate this procedure. Finally, the main conclusions to the work are presented.

Design of a robust electrical position servo based on dual model matching

This paper deals with the robust control of electrical servomotors with parametric uncertainties and some non-linear factors. A powerful two-degrees-of-freedom (TDOF) approach via sensitivity and complimentary sensitivity functions has been used previously to obtain results, subject to unstructured uncertainties. This approach has the advantages that any stable linear control systems can be realized and also that the reference input response and disturbance supression can be specified independently. In this paper a procedure is proposed for the design of robust electrical position servos based on a novel control system design method called dual model matching (DMM). The DMM approach also possesses the above properties but with additional advantages subject to parametric uncertainties. A comparison of the usual TDOF approach with the DMM approach is discussed. The efficacy of the proposed servos is confirmed with computer simulations and laboratory experiments.

Low-Order Robust Model Matching Controller Design for SISO Plants

Abstract Two design methods of low-order robust model matching controllers for SISO plants with parameter and order variations are presented. It is assumed that the relative degree and the sign of the gain constant of the transfer function of the plant are fixed, the zeros of the plant are asymptotically stable and the coefficients and the degrees of the denominator and numerator polynomials of the plant transfer function vary in step-wise fashion in bounded and known ranges. A low-order robust model matching controller consists of an ordinary model matching controller and a robust compensator. The model matching controller is designed bythe usual method for a reduced-order “reference plant” constructed suitably. The nominal and actual plants are considered as those the reference plant changed into. The robust compensator is designed to reduce the influence of the plant characteristic variations from the reference plant. It is shown that, by suitably designing the robust compensator, one can ensure that the closed-loop system with a low-order robust model matching controller has robust stability, robust steady-state and robust transient properties simultaneously. For given reference model and reference input, the order of a controller designed by these methods is determined by the relative degree of the plant, in spite of the order of the plant.

A Robust Control of Fuel Injection Servo System

Abstract In this paper, Robusust Model Matching (RMM) is applied to spill position control of fuel injection pump of diesel engine car. The parameters uf this plant vary owing to the change of temperature. It is desirable that, even if parameters of plant the change, the spill position tracks the reference input rapidly and correctly. KMM is a design method of realizing model matching from reference inputs to controlled variables and assuring high robustness for the parameter uncertainties by the robust compensator simultaneously. Since this plant is a non-minimum phase and single input single output system, the robust compensator is designed as follows: 11) To achieve robust steady-state property, the robust compensator is designed so as to minimize H 2 norm of R W dy within the desired bounded frequency range. (2) To maintain stability, the robust compensator is designed such that the transfer function R W U*V from the conceptual input to the control variable has suitable margin of gain and phase. Simulation and experiment results show that the method proposed in this paper is effective the position control of fuel injection pump.

A ROBUST CONTROL OF FUEL INJECTION SERVO SYSTEM

In this paper, Robust Model Matching (RMM) is applied to spill position control of fuel injection pump of diesel engine car. The parameters of this plant vary owing to the change of temperature. It is desirable that, even if parameters of plant the change, the spill position tracks the reference input rapidly and correctly. RMM is a design method of realizing model matching from reference inputs to controlled variables and assuring high robustness for the parameter uncertainties by the robust compensator simultaneously. Since this plant is a non-minimum phase and single input single output system, the robust compensator is designed as follows: (1) To achieve robust steady-state property, the robust compensator is designed so as to minimize H2 norm of RWdy within the desired bounded frequency range. (2) To maintain stability, the robust compensator is designed such that the transfer function RWu*v from the conceptual input to the control variable has suitable margin of gain and phase. Simulation and experiment results show that the method proposed in this paper is effective the position control of fuel injection pump.

A computer aided technique to derive the class of realizable transfer function matrices of a control system for a prescribed order controller

Although it has been considered that a transfer function based method is difficult to use to treat multi-input, multi-output (MIMO) systems, the MIMO control system can be treated in a comparatively easy way using a characteristic transfer function matrix (CTFM) concept. In this paper, based on properties of the CTFM, a computer aided technique is presented for deriving the class of realizable transfer function matrices of a control system for a prescribed order controller. Then the possibility as a practical design tool is discussed.

Control of a Magnetic Levitation System by Robust Model Matching

Abstract Magnetic levitation systems possess a lot of advantages as actuators, however, model uncertainties lose its excellent performance. Robust Model Matching (RMM) is a practical design method for the control of minimum-phase plants with wide parameter variations. As for digital control, the conventional RMM can not be applied to magnetic levitation systems because of unstable limiting zeros. In this paper, we show a novel design method of a digital controller within the RMM framework using inter-sample output signal to avoid the problem occurred by unstable zeros, and apply it to a magnetic levitation system.

Dual Model Matching

Abstract In control systems, when two types of characteristic transfer function matrices between the reference inputs and the outputs, and between the conceptual inputs and the outputs respectively are considered, there exist one-to-one corresponding relationships between these characteristic transfer function matrices and those of the controllers. Therefore, all the properties of the control systems are determined by these two types of characteristic transfer function matrices. In this paper, new design methods of control systems are proposed based on the ideas, i.e. dual model matching, that for the given plants, appropriate controllers are derived by assigning the model (i.e. dual model) of the characteristic transfer function matrices of the two types stated above

Design of Electrical Position Servo Based on Robust Model Matching

Abstract The applications of the electrical position servo are extending to industrial robot. NC machine tool and computer peripheral unit, etc. In these fields, the requirements for high accuracy and high speed control have been increasing more and more. But the following problems make it difficult to meet these requirements. (1) Load variations (2) Torque disturbance (3) Nonlinear friction. Recently, one of the authors proposed a new design method of robust control system and named it ‘Robust Model Matching’. This method deals with the problem of wide parameter variations of controlled object. When the robust model matching is applied to the electrical servo, it is expected that this method is effective not only for the parameter variations of controlled object but also for the torque disturbance and the nonlinear friction. The present paper proves by simulations and experiments that electrical position servo designed by robust model matching method can overcomes the problems mentioned above.