Joanna Zietkiewicz

Constrained predictive control of a levitation system

In this paper a method of an predictive controller design for the control of constrained MIMO levitation system is presented. The proposed control strategy is based on combining feedback linearization and LQ method with modifying reference signal allowing to keep variables inside constrained region.

Non-minimum phase properties and feedback linearization control of nonlinear chemical reaction

Non-minimum phase objects cause problems in designing control systems due to unstable inverse dynamics. In the case of, considered in the paper, Van de Vusse nonlinear chemical reaction, non-minimum phase properties are in connection with the input multiplicity of the process. Control method proposed for this process relies on feedback linearization and linear quadratic control. Input-state linearization provides an exact linear model. Linear quadratic regulator uses obtained model and determines appropriate control law. This approach, providing that weights for the linear quadratic regulator are adjusted properly, delivers control system with good performance of output and state signals, despite the fact that the output of the process is not a state variable of the linear model. The solution also assures stability of the system.

Performance of Feedback Linearization Based Control of Bicycle Robot in Consideration of Model Inaccuracy

This paper considers the influence of model inaccuracy on control performance when feedback linearization is used. For this purpose we use plant of bicycle robot. The problem is analysed in two ways: by simulations with artificially changed parameters and by comparison of simulated data with the results obtained from the real object. The collected data show that, even if the model differs from the real object, control method provides good results. This indicates that feedback linearization, method strongly relying on model, can be successfully used for some real plants.

Particle Filter in State Vector Estimation Problem for Power System

Particle Filter is a tool, which has been used more frequently over the years. Calculations with using Particle Filter methods are very versatile (in comparison to the Kalman Filter), which can be used in high complex and nonlinear problems. Example of such a problem is the power system, where Particle Filter is used to state estimation of network parameters based on measurements. Paper presents theoretical basis regarding Par- ticle Filter and power system state estimation. Results of experi- ment have shown that Particle Filter usually gives better outcome comparing to the Weighted Least Squares method. In extension Multi Probability Density Function Particle Filter is proposed, which improves obtained results so that they are always better than Weighted Least Squares method.

Robust Actuator Fault-Tolerant LQR Control of Unmanned Bicycle Robot. A Feedback Linearization Approach

Robust control must be able to cope with various system behavior subject to mismodeling and able to assure some performance level. In the paper, we propose to use actuator fault-tolerant control law to stabilize a bicycle robot model with inertial wheel to take into account the unmodeled uncertainty introduced by using linearized model in a LQR fashion with feedback linearization. Our proposal is illustrated by signal plots and performance indexes’ values obtained from a set of experiments and is a natural extension of the results presented in the past.

Input and state constrained nonlinear predictive control: Application to a levitation system

The subject of the article concerns a constrained predictive control with feedback linearization (FBL) applied for multiple-input and multiple-output (MIMO) system. It relies on finding a compromise in every step between feasible and optimal linear quadratic (LQ) control by minimization of one variable. Behaviour of model signals in function of minimized variable is investigated, in order to assure the optimality of the solution. LQ control based applications for feedback linearized models do not meet the problem of choosing weights in linear quadratic cost function. That important problem is solved here by comparison of the cost function with that obtained for the linear approximated system in the operating point. That provides satisfactory behaviour and also justifies the simplified approach relied on minimization of only one variable for MIMO system.

The impact of vocabulary size and language model order on the polish whispery speech recognition

The article presents studies on the automatic whispery speech recognition. In the performed research a new corpus with whispery speech has been used. The aim of studies presented in this paper was to check, how the vocabulary size and the language model order influence on the speech recognition quality. It has been concluded that even using recordings with 5,000 different words only it is possible to prepare large vocabulary continuous speech recognition (LVCSR) model. It has been also found that the third order of language model is the best choice. The difference between normal and whispery speech is negligible and is manifested only in higher word error rate index (about 1.5 times higher for whispery speech).