Adaptive Identification Algorithm Research Articles

Application of Adaptive Identification Algorithms to Closed-Loop Adaptive Control Systems with an Identifier

The paper is concerned with the identifiability of a plant in the closed-loop adaptive system of control with an identifier (ASI), The r.m.s. errors of monitoring parameters of a nonstationary plant are obtained. Consideration is given to algorithms used for identification of plant parameters along both, disturbance and control, channels. Stabilization errors of the nonstationary plant output in ASI have been estimated. The closed-loop ASI are analyzed for the case when the input disturbance vector has statistically independent components whereas the unknown parameters of the plant vary in a random way.

Implementation of Adaptive Filters Using Residue Number Systems

Implementation of adaptive infinite impulse response (IIR) digital filters using residue number is presented. Earlier, Jenkins and his co-workers have considered implementation of IIR filters with fixed coefficients, and Soderstrand has implemented an adaptive finite impulse response (FIR) filter. Here, we extend the earlier work to the adaptive IIR filter. An output and equation error adaptive identification algorithm which has been shown by Johnson to be convergent at low convergence rates is used. Residue arithmetic decomposes the arithmetic operations which are well suited for high speed ROM table look-up, for computations on parallel microcomputers, or for special computer architectures involving only shift registers and adders.

Exponential convergence of adaptive identification and control algorithms

Output and equation error adaptive identification algorithms are shown to be exponentially convergent under a deterministic or stochastic persistently exciting (or spanning) condition on the system inputs together with several other standard conditions. An adaptive control algorithm is shown to be exponentially convergent under a deterministic or stochastic persistently exciting condition on the reference trajectory together with some standard conditions.

Exponential Convergence of Adaptive Identification and Control Algorithms

Output and equation error adaptive identification algorithms are shown to be exponentially convergent under a persistently exciting (or spanning) condition on the system inputs together with several other standard conditions. An adaptive control algorithm is shown to be exponentially convergent under a persistently exciting condition on the reference trajectory together with some standard conditions.

A Modified Single Step Identification Algorithm and its Application in Machining Control Systems

Publisher Summary This chapter discusses a modified single-step identification algorithm and its application in machining control systems. Adaptive control systems with an identifier in the feedback loop (ASI) are multipurpose and may thus be used in design of physically different control systems. Such systems usually employ adaptive identification algorithms. Because ASI realize direct digital control, the basic requirements to adaptive algorithms of identification is a high convergence rate, fault tolerance and implementation simplicity. This chapter presents results of the study of a modified single-step algorithm obtained from that currently used in ASI Rajbman, 1976. The modification implied averaging of several observations which allowed the increase of the rate of convergence in process control systems in real-life conditions when input and output variables were measured with noise.

ON SYSTEM IDENTIFICATION WITH AND WITHOUT CERTAINTY

Abstract Information-theoretic concepts are utilized to develop a procedure for identifying a parameter of a stochastic linear discrete time dynamic scalar system based on noisy linear measurements of the system's state. After various simplifying approximations, the derived error entropy identification algorithm reduces to an on-line adaptive identification algorithm that is similar in many respects to well-established identification techniques. Conditions under which the developed on-line adaptive algorithm identifies the system with certainty are presented. Using an error entropy estimation lower bound, which is independent of any estimation procedure, conditions for which identification cannot be made with certainty are also presented. Examples involving non-Gaussian statistics are used to illustrate the efficiency of the error entropy adaptive identification algorithm as well as to compare it with several other identification procedures.

Design of Nonlinear Adaptation and Adaptive Identification Algorithms by Using Nonsearching Self-Adjusting Systems (NSAS) with a Reference Model

The paper is concerned with the adaptive control of a continuous linear nonstationary multidimensional plant all of whose states are observable. The coefficients of the controller are tuned by parameter adjustment loops in an open (adaptive identification) or closed (USAS with a reference model) cycle. The adaptation and identification algorithms are designed by the direct Lyapunov method with Lyapunov function which is the sum of the quadratic form of generalized adaptation errors and the quadratic forms of a special types of parametric misalignments between plant and model. New classes of adaptation and identificational algorithms are suggested which are nonlinear vector functions of bilinear forms of the system states and coordinate adaptation errors.

Comparison of Continuous and Discrete Adaptive Identification Algorithms

Discretization of a popular continuous-time control algorithm is effected and an equivalent discrete-time identification law developed and compared to a published discrete identification algorithm developed from Lyapunov Theory. Results are compared as regards asymptotic stability as insured using Lyapunov theory. Some analysis and design guidelines are proposed as regards implementation and practical utility.

A condition for decentralization in model reference adaptive systems

It is shown that when the interconnection pattern within a system conforms to certain prescribed symmetry requirements, decentralized adaptive algorithms for identification and control can be developed within a decomposition-decentralization framework. This is useful in the presence of constraints on the available information structure of the system.