Papers by Keyword: Multivariable System

Abstract: System identification is the basis for control system design. For linear time-invariant systems have a variety of identification methods, identification methods for nonlinear dynamic system is still in the exploratory stage. Nonlinear identification method based on neural network is a simple and effective general method that does not require too much priori experience about the system to be identified. Through training and learning, the network weights are corrected to achieve the purpose of system identification. The paper is about the identification of multivariable nonlinear dynamic system based on PID neural network. The structure and algorithm of PID neural network are introduced and the properties and characteristics are analyzed. The system identification is completed and the results are fast convergence.

Abstract: The discrete-time adaptive sliding mode controller for spinning rockets in presence of parameter error is proposed. Considering the nonlinear characteristics for the system, input-output feedback linearization is utilized to transform the system model into two standard form subsystems. Then a discrete-time controller for guided rockets is designed based on discrete-time sliding mode control principle. In order to diminish the switch width of the discrete-time sliding mode system corresponding to parameter error, a dead-zone parameter adaptive law is designed. The stability of the uncertain closed-loop system is proved by Lyapunov theory, which make the controller have high robustness. Simulation result indicates that the proposed controller is robust with respect to large aerodynamic parametric uncertainty, and has excellent dynamic tracking performance.

Abstract: This paper studies the output-feedback model reference robust control for MIMOlinear systems with generalized relative degree one. A new robust controlscheme is proposed within the framework of model reference control. Under theassumption that the high-frequency gain matrix of the plant can be transformedto a glass of main diagonal dominant matrix via full rank transformation, it isshown that all signals of the closed-loop system are globally uniformly boundedand meanwhile, the tracking errors converge to a residual set that can be madearbitrarily small by properly choosing some design parameters. Simulationresults are presented to illustrate the effectiveness of the proposed scheme.

Abstract: Multivariable system with time delay and coupling widely exist in industrial which may destroy the normal work of control system. An unconventional internal model controller design method will be introduced in this paper. The closed loop system can be decouple by calculate the inverse of transfer function matrix and the optimal diagonal decomposition matrix. Finally, this method was applied in a multivariable system with different time delays, the simulation results can show the effectiveness of this method.

Abstract: The multivariable PID neural network (MPIDNN) control system is introduced in this paper. MPIDNN is used to perform both the control and the decouple at the same time and to get better performance. It is difficult to control multivariable system by conventional controller because the strong coupling properties of the system. Generally, the decoupling system should be designed first and the multivariable object would be divided into several single variable objects. Then, several simple controller would achieve the control of those objects. The decoupling system and the controller exist in theory but the design process is very difficult actually because the transfer function of the object is difficult to get. Especially, if the number of the object inputs is not equal to that of the object outputs, which is called unsymmetry object, the conventional decoupling is impossible. A actual example is discussed in the paper in order to prove the function of the MPIDNN, in which an un-symmetry multivariable system which has 3 inputs and 2 outputs is controlled by a MPIDNN and the perfect control property is obtained by self-learning process.

Abstract: An IMC-PID based on V-norm decoupling for multivariable stable process with time delay was proposed in this paper. Previous multivariable IMC is usually designed after the decoupling of the process, however the industrial process modeling is difficult to be precise, so the controller design is often not ideal after decoupling. The method in this paper designs both the decoupling and controller at the same time, the IMC is not only a controller, but also a decoupling compensation. A filter parameter optimization method based on a new performance function was adopted. The controller can be realized by time delay compensation when there is an advance part. The IMC-PID designed by this method has strong robustness when the model is mismatch and excellent performance both in decoupling and control. The simulation results show that the method is very effective and has a wide range of application in actual industry area.

Abstract: For multiple-axis stages, it is required to operate the axes simultaneously, such that the resulting trajectory of platform follows a given contour. For most stage systems, friction acts as the major disturbance which degrades the precision of system motion and its effect should be compensated. In this paper, contouring control of a two-dimensional stage system subjected to friction is investigated. A systematic contour controller design based on task coordinate frame is proposed and its effectiveness is studied through theoretical analysis and numerical simulations.