Control characteristics of a control object including a specimen (10), a hydraulic cylinder (12), and a servo valve (16) are expressed in a model, and stored in a model storage section (34). In modeling, dynamical characteristic values (a mass M, a damping coefficient C, a spring constant K) of a specimen (10) are expressed is variables. After substituting the actual characteristic values of a specimen (10) to be tested in the variables of the model, a simulation is run using the model to obtain an optimal feedback gain. When the optimal feedback gain is determined, the pole position of the corresponding system to be controlled is specified. Thereafter, when different specimens (10) having different characteristics are tested, optimal feedback gains for the respective systems are determined by calculating a value which ensures the respective systems the same pole position as the above.

The present invention relates to an arrangement for the discrete-time adaptive on-off switching control of a continuous-time process with a binary switching actuator, which uses for the determination of the on-off actuating signal a prediction of a process output sequence over several future sampling intervals as reaction to a possible process input sequence that is applied to a discrete-time linear process model and which estimates and updates in every sampling interval the parameters of the process model by means of a parameter estimation device in order to adapt them to the process to be controlled, even when the process behavior changes, and which has a device for the input and change of the setpoint, the limits of the process output and the sampling time as well as a measuring device for the periodical measurement of the process input. The arrangement further comprises two alternatively working control devices one of which is active in the stationary phase and the other of which is active in the start-up phase of the process or after setpoint changes, and a switching device which activates one or the other said control device in dependence of the result of the measurement of the process output and of input current setpoint, wherein the actuator is served by the active control device.

A servomotor controller provides high speed servomotor positioning be means of an on line adaptive state-feedback control algorithm. Servomotor position and velocity are the feedback states and each has an associated gain selected by means of a servomotor characterization scheme. The mechanical break frequency is used to select optimum gains and a high speed estimation technique is used to determine the mechanical break frequency. Position step change may also be used to select optimum gains. Gain tables are produced by an off-line simulation technique.

A method and apparatus for integral control of a dependent variable in a system having at least two independent variables which influence the dependent variable is disclosed. A difference signal indicative of the difference between the sensed dependent variable and a reference is integrated and provided to a hysteresis transfer function. The output of the transfer function is subtracted from the integrated difference signal and the resulting signal is used to control an independent variable which thereby influences the dependent variable only for small magnitude signal variations in the integrated signal. The integrated signal is used to control another independent variable for large magnitude signal variations in the integrated signal. The other independent variable is typically only capable of influencing the dependent variable for large magnitude signal variations.

Apparatus and method for controlling the output of a dynamic system which is susceptible to changing dynamic characteristics. The desired present and future outputs of the system are applied to a predictor which determines the inputs to a model reference adaptive control subsystem from which the actual outputs are produced. The predictor uses an impulse model of the subsystem to simulate and predict future outputs. The adaptive control subsystem includes adjustable gain feedback or controlloops which are adjusted to make the dynamic system appear to have constant characteristics even when its dynamic characteristics are changing. A reference model of the dynamic system is used as the basis for the gain adjustments. The equation weights for the mathematical impulse model used by the input predictor are derived from the reference model of the adaptive control subsystem, and remain constant throughout the operation of the controller.