Da-Wei Ding

Nonlinear adaptive control using multiple models and dynamic neural networks

Abstract The adaptive control of nonlinear dynamic system using multiple models has emerged as a mathematically attractive and practically available method. It can be used to improve the transient response of adaptive control. In this paper, parallel dynamic neural networks (PDNNs) are used to identify the unknown dynamic system. Three kinds of combinations of adaptive model and fixed model are established to set up multiple model adaptive control (MMAC) by using an effective switching scheme. The stability and convergence of MMAC using multiple PDNNs are discussed by means of Lyapunov-like analysis. Considering the unmodeled dynamics, the new learning law ensures that the identification error converges to a bounded zone, and the proposed MMAC with PDNNs can improve the control property greatly. By simulation studies, the results of different combinations of fixed and adaptive models are compared, and the effectiveness of the proposed method can be seen.

Adaptive Control of Nonlinear Discrete-Time Systems by Using OS-ELM Neural Networks

As a kind of novel feedforward neural network with single hidden layer, ELM (extreme learning machine) neural networks are studied for the identification and control of nonlinear dynamic systems. The property of simple structure and fast convergence of ELM can be shown clearly. In this paper, we are interested in adaptive control of nonlinear dynamic plants by using OS-ELM (online sequential extreme learning machine) neural networks. Based on data scope division, the problem that training process of ELM neural network is sensitive to the initial training data is also solved. According to the output range of the controlled plant, the data corresponding to this range will be used to initialize ELM. Furthermore, due to the drawback of conventional adaptive control, when the OS-ELM neural network is used for adaptive control of the system with jumping parameters, the topological structure of the neural network can be adjusted dynamically by using multiple model switching strategy, and an MMAC (multiple model adaptive control) will be used to improve the control performance. Simulation results are included to complement the theoretical results.

Nonfragile filtering for discrete-time linear systems in finite-frequency domain

This article investigates the problem of nonfragile filter design for discrete-time linear systems subject to noises with known frequency ranges. Additive interval uncertainty reflecting imprecision in filter implementation is considered. By the aid of generalised KYP lemma, both deterministic and randomised filtering algorithms are proposed to deal with noises in low-, middle- and high-frequency domain, respectively. The proposed nonfragile finite-frequency filters can get a better noise attenuation performance when frequency ranges of noises are known beforehand. An example about F-18 aircraft model is given to illustrate the effectiveness of the proposed algorithms.

Attitude estimation for UAV using extended Kalman filter

A novel attitude estimation algorithm is proposed for unmanned aerial vehicles(UAV) in this paper. It uses attitude quaternion to represent the attitude of UAV, and uses extended Kalman filter(EKF) to fuse the merits of magnetic, angular rate, and gravity(MARG) sensors. First, attitude quaternion and drift bias of gyroscope are selected to construct the state vector, and the state equation is established based on the kinematics model of gyroscope. Then, an orthogonalization method is utilized to obtain the unit attitude quaternion from the outputs of accelerometer and magnetometer, it makes the magnetic field vector perpendicular to the measured gravity vector, which avoids the geomagnetic disturbance. And the unit attitude quaternion is used for the measurements for the EKF. Finally, the EKF update equation is used to determine the attitude of UAV. Experiments are provided on a real-world data set and the results show that the algorithm can precisely represents the orientation of UAV in both static and dynamic situation.

Finite-frequency model reduction of discrete-time T-S fuzzy state-delay systems

This paper is concerned with the problem of model reduction for discrete-time Takagi-Sugeno (T-S) fuzzy state-delay systems with finite-frequency input signals. A new finite-frequency model reduction algorithm is proposed, which can get a better approximation performance than the existing full-frequency methods. Firstly, a finite-frequency H ∞ performance index is defined in the frequency domain. Then, a stability condition and a finite-frequency H ∞ performance analysis condition are developed by the aid of Jensens inequality and Parsevals theorem, respectively. Based on these conditions, sufficient model reduction conditions are derived for discrete-time T-S fuzzy state-delay systems. An optimization algorithm is proposed to obtain a stable reduced-order model satisfying the finite-frequency performance specification. Finally, the effectiveness of the proposed method is illustrated by a numerical example.

Adaptive control of nonlinear system using online error minimum neural networks

In this paper, a new learning algorithm named OEM-ELM (Online Error Minimized-ELM) is proposed based on ELM (Extreme Learning Machine) neural network algorithm and the spreading of its main structure. The core idea of this OEM-ELM algorithm is: online learning, evaluation of network performance, and increasing of the number of hidden nodes. It combines the advantages of OS-ELM and EM-ELM, which can improve the capability of identification and avoid the redundancy of networks. The adaptive control based on the proposed algorithm OEM-ELM is set up which has stronger adaptive capability to the change of environment. The adaptive control of chemical process Continuous Stirred Tank Reactor (CSTR) is also given for application. The simulation results show that the proposed algorithm with respect to the traditional ELM algorithm can avoid network redundancy and improve the control performance greatly.

Application of fuzzy neural network in burden surface clustering

Blast furnace charging distribution plays an important role in the steel production. The radar data containing the information of present burden surface situation and the cross thermometer data reflecting the trend of burden surface are taken as the training data of fuzzy neural network. The trained network will be used for the future clustering of the data. Considering the demand of energy-saving and consumption-decreasing, the method sets up the basic for the distribution control in the next step. The simulation results show the effectiveness of the proposed method.

Adaptive control using multiple parallel dynamic neural networks

The control problem of an unknown nonlinear dynamic system which contains the abrupt changes of parameters is concerned. Multiple models based on dynamic neural networks are used to approximate the dynamic character of unknown system. Different controllers based on these models and an effectively switching mechanism are applied to an unknown system to trace a reference trajectory. Further, we propose different switching and turning schemes for adaptive control which combine fixed and adaptive models. From the simulation, it can be shown that the multiple model adaptive control method proposed in this paper can improve the control performance greatly compared with the conventional adaptive control.

Fault detection for dissipative nonlinear systems: An energy balance method

This paper addresses the problem of fault detection for dissipative nonlinear systems. A new method based on energy balance is proposed for dissipative nonlinear systems. The nonlinear system is described by the T-S fuzzy model. Based on the dissipation of the T-S fuzzy model, the energy balance is constructed. Then faults can be detected by judging whether the energy balance of system construction is established. The effectiveness of the proposed scheme is illustrated by a numerical example.

Nonfragile finite-frequency filtering for discrete-time linear systems

The paper investigates the problem of nonfragile filter design for discrete-time linear systems subject to noises with known frequency ranges. Additive interval uncertainty reflecting imprecision in filter implementation is also considered. By the aid of generalized KYP lemma, both deterministic and randomized filtering algorithms are proposed to deal with noises in low-, middle-, and high-frequency domain, respectively. The proposed nonfragile finite-frequency filters can get a better noise attenuation performance when frequency ranges of noises are known beforehand. An example is given to illustrate the effectiveness of the proposed method.