System identification

About: System identification is a research topic. Over the lifetime, 21291 publications have been published within this topic receiving 439142 citations.

A wavelet-based approach for model and parameter identification of non-linear systems

Abstract: A procedure is presented for identifying the mechanical parameters of zero-memory non-linear discrete structural systems. The procedure allows both the parameter estimation of a priori known dynamical models as well as the identification of classes of suitable non-linear models based on input–output data. The method relies on a wavelet-based discretization of the non-linear governing differential equation of motion. Orthogonal Daubechies scaling functions are used in the analysis. The scaling functions localization properties permit the tracking of fast variations of the state of the dynamical system which may be associated with unmodeled dynamics of measurement noise. The method is based on the knowledge of measured state variables and excitations and applies to single and multi-degree-of-freedom systems under either free or forced vibrations.

Agnostic System Identification for Model-Based Reinforcement Learning

Abstract: A fundamental problem in control is to learn a model of a system from observations that is useful for controller synthesis. To provide good performance guarantees, existing methods must assume that the real system is in the class of models considered during learning. We present an iterative method with strong guarantees even in the agnostic case where the system is not in the class. In particular, we show that any no-regret online learning algorithm can be used to obtain a near-optimal policy, provided some model achieves low training error and access to a good exploration distribution. Our approach applies to both discrete and continuous domains. We demonstrate its efficacy and scalability on a challenging helicopter domain from the literature.

Low-order black-box models for control system design in large power systems

Abstract: The paper studies two multi-input multi-output (MIMO) procedures for the identification of low-order state space models of power systems, by probing the network in open loop with low-energy pulses or random signals. Although such data may result from actual measurements, the development assumes simulated responses from a transient stability program, hence benefiting from the existing large base of stability models. While pulse data is processed using the eigensystem realization algorithm, the analysis of random responses is done by means of subspace identification methods. On a prototype Hydro-Quebec power system, including SVCs, DC power lines, series compensation, and more than 1100 buses, it is verified that the two approaches are equivalent only when strict requirements are imposed on the pulse length and magnitude. The 10th-order equivalent models derived by random-signal probing allow for effective tuning of decentralized power system stabilizers (PSSs) able to damp both local and very slow inter-area modes.

Distributed parameter systems: theory and applications

Abstract: Part 1 Mathematical theory: some basic results in the theory of partial differential equations - Bellman-Gronwall inequality, Sobolev spaces, Green's formula stochastic partial differential equations - radon measures, cylindrical probability, Gaussian cylindrical probability, nuclear and Hilbert-Schmidt operators, conditional expectation, Hilbert- space-valued Wiener processes optimal control of deterministic distributed parameter systems - elliptic systems, the Dirichlet problem, the Neumann problem, parabolic systems, Riccati equation, Hamilton-Jacobi equation, hyperbolic systems controllability and observability linear estimation theory - finite-dimensional estimation theory, estimation for random linear functionals optimal filter for distributed parameter systems - the filtering problems, Wiener filter, Kalman-Bucy filter, recursive formula for the optimal filter, innovation theory, duality between estimation and control, optimal filter for hyperbolic systems stochastic optimal control of distributed parameter systems formulation of the model, the stochastic optimal control problem, necessary and sufficient conditions for optimality, the separation principle identification of distributed parameter systems - the basic concept of system identification, modal approximation for identification, regularization. Part 2 Engineering Applications: formal approach to optimal filtering and control of distributed parameter systems - Wiener-Hopf theorem, the optimal filter, predictor and smoothing estimator, various approaches to linear estimation problems stochastic optimal control problems optimal sensor and actuator location problems - optimal sensor location problems, optimal actuator locations computational techniques for identification of distributed parameter systems - stochastic approximation, least squares identification, the Galerkin finite-element model, discrete regularization and minimization.