Adaptive control of system with hysteresis using neural networks

An adaptive control scheme is developed for a class of single-input nonlinear systems preceded by unknown hysteresis, which is a non-differentiable and multi-value mapping nonlinearity. The controller based on the three-layer neural network （NN）, whose weights are derived from Lyapunov stability analysis, guarantees closed-loop semiglobal stability and convergence of the tracking errors to a small residual set. An example is used to confirm the effectiveness of the proposed control scheme.

Sliding Mode Control Design via Reduced Order Model Approach

This paper presents a design of continuous-time sliding mode control for the higher order systems via reduced order model. It is shown that a continuous-time sliding mode control designed for the reduced order model gives similar performance for thc higher order system. The method is illustrated by numerical examples. The paper also introduces a technique for design of a sliding surface such that the system satisfies a cost-optimality condition when on the sliding surface.

A Reliable and Accurate Calculation of Excitation Capacitance Value for an Induction Generator Based on Interval Computation Technique

A squirrel cage induction generator （SCIG） offers many advantages for wind energy conversion systems but suffers from poor voltage regulation under varying operating conditions. The value of excitation capacitance （C exct ） is very crucial for the selfexcitation and voltage build-up as well as voltage regulation in SCIG. Precise calculation of the value of C exct is, therefore, of considerable practical importance. Most of the existing calculation methods make use of the steady-state model of the SCIG in conjunction with some numerical iterative method to determine the minimum value of C exct . But this results in over estimation, leading to poor transient dynamics. This paper presents a novel method, which can precisely calculate the value of C exct by taking into account the behavior of the magnetizing inductance during saturation. Interval analysis has been used to solve the equations. In the proposed method, a range of magnetizing inductance values in the saturation region are included in the calculation of C exct , required for the self-excitation of a 3-φ induction generator. Mathematical analysis to derive the basic equation and application of interval method is presented. The method also yields the magnetizing inductance value in the saturation region which corresponds to an optimum C exct（min） value. The proposed method is experimentally tested for a 1.1 kW induction generator and has shown improved results.

A New Energy Optimal Control Scheme for a Separately Excited DC Motor Based Incremental Motion Drive

This paper considers minimization of resistive and frictional power dissipation in a separately excited DC motor based incremental motion drive （IMD）. The drive is required to displace a given, fixed load through a definite angle in specified time, with minimum energy dissipation in the motor windings and minimum frictional losses. Accordingly, an energy optimal （EO） control strategy is proposed in which the motor is first accelerated to track a specific speed profile for a pre-determined optimal time period. Thereafter, both armature and field power supplies are disconnected, and the motor decelerates and comes to a halt at the desired displacement point in the desired total displacement time. The optimal time period for the initial acceleration phase is computed so that the motor stores just enough energy to decelerate to the final position at the specified displacement time. The parameters, such as the moment of inertia and coefficient of friction, which depend on the load and other external conditions, have been obtained using system identification method. Comparison with earlier control techniques is included. The results show that the proposed EO control strategy results in significant reduction of energy losses compared to the existing ones.

Modelling of ultrasonic motor with dead-zone based on Hammerstein model structure

The ultrasonic motor (USM) possesses heavy nonlinearities which vary with driving conditions and load-dependent characteristics such as the dead-zone. In this paper, an identification method for the rotary travelling-wave type ultrasonic motor (RTWUSM) with dead-zone is proposed based on a modified Hammerstein model structure. The driving voltage contributing effect on the nonlinearities of the RTWUSM was transformed to the change of dynamic parameters against the driving voltage. The dead-zone of the RTWUSM is identified based upon the above transformation. Experiment results showed good agreement be- tween the output of the proposed model and actual measured output.

A New Subdivision Algorithm for the Bernstein Polynomial Approach to Global Optimization

In this paper, an improved algorithm is proposed for unconstrained global optimization to tackle non-convex nonlinear multivariate polynomial programming problems. The proposed algorithm is based on the Bernstein polynomial approach. Novel features of the proposed algorithm are that it uses a new rule for the selection of the subdivision point, modified rules for the selection of the subdivision direction, and a new acceleration device to avoid some unnecessary subdivisions. The performance of the proposed algorithm is numerically tested on a collection of 16 test problems. The results of the tests show the proposed algorithm to be superior to the existing Bernstein algorithm in terms of the chosen performance metrics.

Model-based Tracking for Agent-based Control Systems in the Case of Sensor Failures

The study on artificial intelligence(AI) methods for tuning of particle accelerators has been reported in many literatures.This paper presents tuning method for agent-based control systems of transport lines in the case of sensor/actuator failures.The method uses model-based tracking concept to relax the demand on sensor data.The condition for successful operation of the stated scheme is derived,and the concept is demonstrated through simulation by applying it to the model of microtron,transport line-1 and booster of indus accelerator.The results show that this approach is very effective in transport line control during sensor/actuator failures.

Construction of Bode Envelopes Using REP Based Range Finding Algorithms

The frequency domain analysis of systems is an important topic in control theory. Powerful graphical tools exist in classic control, such as the Nyquist plot, Bode plots, and Nichols chart. These methods have been widely used to evaluate the frequency domain behavior of system. A literature survey shows that various approaches are available for the computation of the frequency response of control systems under different types of parametric dependencies, such as affine, multi-linear, polynomial, etc. However, there is a lack of tools in the literature to construct the Bode envelopes for the general nonlinear type of parametric dependencies. In this paper, we address the problem of computation of the envelope of Bode frequency response of a non-rational transfer function with nonlinear parametric uncertainties varying over a box. We propose two techniques to compute the Bode envelopes：first, based on the natural interval extensions （NIE） combined with uniform subdivision and second, based on the existing Taylor model combined with subdivision strategy. We also propose the algorithms to further speed up both methods through extrapolation techniques.

Adaptive cubature Kalman filter with the estimation of correlation between multiplicative noise and additive measurement noise

Mobile robots are often subject to multiplicative noise in the target tracking tasks,where the multiplicative measurement noise is correlated with additive measurement noise.In this paper,first,a correlation multiplicative measurement noise model is established.It is able to more accurately represent the measurement error caused by the distance sensor dependence state.Then,the estimated performance mismatch problem of Cubature Kalman Filter(CKF)under multiplicative noise is analyzed.An improved Gaussian filter algorithm is introduced to help obtain the CKF algorithm with correlated multiplicative noise.In practice,the model parameters are unknown or inaccurate,especially the correlation of noise is difficult to obtain,which can lead to a decrease in filtering accuracy or even divergence.To address this,an adaptive CKF algorithm is further provided to achieve reliable state estimation for the unknown noise correlation coefficient and thus the application of the CKF algorithm is extended.Finally,the estimated performance is analyzed theoretically,and the simulation study is conducted to validate the effectiveness of the proposed algorithm.

QFT preflter design for multivariable systems using interval constraint satisfaction technique

In this paper, a computationally efficient method is proposed for automated design of the prefilters for multivariable systems. In quantitative feedback theory （QFT） method, proposed by Horowitz, the prefilter is designed to achieve the desired tracking specifications. In the proposed approach, we pose the prefilter design problem as an interval constraint satisfaction problem and solve it using the well-established interval constraint satisfaction techniques. The proposed method finds optimal values of the parameters of fixed structure prefilter within the initial search domain. An approach based on prefilter synthesis for single-input single-output is already developed. The purpose of this paper is to extend this approach to QFT prefilter design for general multivariable systems. To validate the above design approach, we applied the method to a laboratory setup of magnetic levitation system.

Dynamic Fault Diagnosis Using the Improved Linear Evidence Updating Strategy

The majority of the existing fault diagnosis methods using Dempster-Shafer(DS) evidence theory(DST) all provide the "static" fused results by combining several pieces of diagnosis evidence, which only reflect the current running status of monitored equipment. This paper presents the dynamic diagnosis strategy by using recursively the improved linear evidence updating rule. Its updated result can synthesize the diagnosis evidence collected at historical, current and future time steps by dynamically adjusting the proposed smoothing linear combination weights. The diagnosis examples of machine rotor show that the proposed method can provide more reliable and accurate results than the diagnosis methods based on the classical updating strategies.