Parameter Error Identification Research Articles

Minimization with few violated constraints and its application in set-membership identification

Abstract We study the problem of finding an optimal solution satisfying all but k of the given n constraints. A solution is obtained along with an algorithm of the complexity min{ O(n. k d ), O(n. d k+1 ) }, where d is the dimension of the problem. We then use the results to solve successfully the problem of robust estimation in the presence of outliers in the setting of bounded error parameter identification. It is shown that the obtained estimate converges to the true but unknown parameter even in the presence of unknown outliers.

A Consideration of the Upper Bound of Parameter Identification Error with Unstructured Modelling Uncertainty and Disturbance

When a control system is designed, it is important to make a model which can explain the input-output relation precisely. To identify the model’s parameter as precisely as possible, it is important to make the experimental signal less affected by the disturbance or the unstructured model uncertainty and sufficiently informative. In this report, the upper bound of the estimated parameter error is proposed when the unstructured model uncertainty exists. This upper bound can be used to decide the optimal identification input.

Parameter Estimation with Interacting Bad Data in State Estimation

This paper proposes a new parameter error identification and estimation method with taking into account interacting bad data. The problem is solved in two steps. First, it uses an adaptive sensitivity approach to pick out the suspicious branches, and detects the bad data by replacing the impedance with power flow variables for those branches. Second, it identifies the parameter errors and estimates the correct network parameters by using the estimated power flows of the suspicious branches. The performance of the method is evaluated by using IEEE-118 bus system and a 250 bus real system.

An experimental study on model reference adaptive control for robot manipulators

Model reference adaptive control based on Popov's hyperstability theory is applied to the trajectory tracking control of a two-link SCARA-type manipulator, and its effectiveness is examined experimentally. A linearized state equation with an optimal state feedback is used as a reference model. The similarity of the structure of the model and a real system like this is useful for simplifying the design of the adaptive control system. Further, by introducing a new adaptation law which consists of a combination of linear and non-linear control, it is possible to reduce high frequency chattering in the control inputs. The results of a simulation and experiment show that by adding a rather simple adaptation mechanism to a conventional state feedback, the effects of Coulomb frictions and parameter identification errors are removed and the accuracy of trajectory tracking is greatly improved.

Experiments Using Lattice Filters to Identify the Dynamics of a Flexible Beam

An approach for identifying the dynamics of large space structures is applied to a free-free beam. In this approach the system’s order is determined on-line, along with mode shapes, using recursive lattice filters which provide a linear least square estimate of the measurement data. The mode shapes determined are orthonormal in the space of the measurements and, hence, are not the natural modes of the structure. To determine the natural modes of the structure, a method based on the fast Fourier transform is used on the outputs of the lattice filter. These natural modes are used to obtain the modal amplitude time series from the measurements. The modal time series provides the input data for an output error parameter identification scheme that identifies the autoregressive moving average (ARMA) parameters of the difference equation model of the modes. Results using this approach for experimentally identifying the dynamics of a flexible beam hardware are presented.

Sensitivity and application of difference equation models to real-time urban runoff forecasting

This paper examines the problems of parameter identification and forecast error analysis under the specification of erroneous model structures based on synthetically generated rainfall-runoff sequences. Sensitivity analysis is conducted by using a variety of recursive parameter identification algorithms, including equation-error and output-error algorithms. Results of the analysis indicate that overestimation of the structural model parameters na and/or nb is not as serious as underestimation of those parameters when measured in terms of the RMSE in the simulated flows. The forecast models are applied to a combined sewer catchment for lead times ranging from 5 to 60 min. The EEM algorithm was found to provide slightly better forecasts than either the RELS or ROLS algorithms for lead times of up to 30 min.

Error bounds for model-plant mismatch in identifiers and adaptive observers

Abstract : The stability properties of six different adaptive schemes with respect to model error are analyzed. Bounds on parameter identification and state errors are established. All adaptive schemes considered are robust in the sense that the error is of order of the speed ratio micron between the modeled slow phenomena vs. the neglected fact. The dependence of the error on the input signal is shown to be crucial. The bounds obtained indicate possibilities for reducing the error by a proper choice of the input signal. (Author)