Piotr Suchomski

J -lossless and extended J -lossless factorizations approach for @-domain H ∞ control

The paper addresses problems of numerical conditioning of the discrete-time H X control design based on J -lossless factorizations of the chain-scattering representation of a plant. It is demonstrated that for a sufficiently small sampling period forward-shift operator techniques may become ill-conditioned and numerical robustness and reliability of design procedures can be significantly improved by utilizing a so-called i operator form of the source problem. State space models of all i -domain controllers are given. The solutions are obtained by considering two coupled i -domain algebraic Riccati equations that can be solved by applying a generalized eigenproblem formulation. Numerical conditioning of these solutions can effectively be evaluated by using a suitable relative condition number. An extended J -lossless factorization for generalized plants with zeros on the stability boundary circle is also given. A so-called zero compensator technique adapted to the i -domain models is utilized to cancel such zeros and some attempt is made to simplify the resulting controller. Two numerical examples are given to illustrate the properties of the approach. The first example deals with a simple problem of mixed sensitivity H X design. The second example concerns a method for robust pole placement design for a plant with zero located on the boundary of the stability region.

An intuitive graphical user interface for the Parkinson's Disease patients

In this paper a discussion on the design and development of the graphical user interface (GUI) dedicated to Parkinsons Disease (PD) patients is presented. The interface is intended for a group of PD patients with less severe motor symptoms, who are living at their home independently or with help of a caregiver. The GUI is designed to enable an interaction for the non-computer literate PD patients with a computer-based system. The system will allow for objectively recording the patient diaries, self assessments, taken medication confirmations and other features important for the diagnosis. This will enable physicians to prepare more accurate evaluation and better diagnostic decisions.

Robust H ∞-Optimal Synthesis of FDI Systems

The entirety of fundamental engineering issues (Chen and Patton, 1999a; Frank, 1990; Gertler, 1998; Isermann, 1984; Patton and Chen, 1993; Wilsky, 1976) is identified with the technical diagnostics of dynamic plants and processes. This domain encompasses three basic subdomains of diagnostic procedures known as the detection, isolation and identification of faults. Therefore, they are referred to as FDI (Fault Detection and Isolation).

Numerically Robust Computer Aided Markov-Equivalent CGPC design

Abstract Properly tuned CGPC control algorithms equipped with anticipated filtering constitute numerically efficient designs. The new approach is based on two principles: mean-square low-order Markovian approximation of the plants step response within a properly determined time interval and a rational restricted Markov-equivalent redesign of the model of the plant. The objective of this paper is to robustify the numerical procedure.

An Approach to Determining Tinnitus Acoustical Characteristic

For many treatment methods, accurate estimation of Tinnitus (ringing in ears) concerning sound type, level, and bandwidth or frequency is inevitable. The proposed way of obtaining Tinnitus parameters is described in this paper. The method employs sound synthesis, aimed at obtaining sound which is closest to perceived Tinnitus. The proposed method assumes running a designed application on a multimedia PC provided with a special graphical user interface to facilitate sound generation and identification. Emphasis is put on issues related to the implementation of the proposed diagnostic procedure. The method was verified during preliminary tests in which people suffering from Tinnitus participated. The obtained results are presented and discussed in this paper.

ENTIRELY LEFT EIGENSTRUCTURE-ASSIGNMENT FOR FAULT DIAGNOSIS OBSERVERS

Abstract In this paper a unified geometrically-embedded methodology of synthesis of static decoupled residual generators is presented that leads to a sub-optimal solution to a robust eigenstructure assignment problem based on a predetermined set of eigenvalues. Robustness of this design means that its eigenvector matrix is sufficiently well conditioned. An original idea of a convenient parameterization of corresponding attainable eigensubspaces is developed and new conditions for disturbance decoupling are derived solely in terms of the left eigenvectors of a system observation state-transition matrix. The resolved completely-static disturbance-decoupling design problem can be treated as an effective preliminary-design stage of an integral synthesis of residual generators. A numerical example illustrates the effectiveness of the presented method.

System supporting speech perception in special educational needs schoolchildren

The system supporting speech perception during the classes is presented in the paper. The system is a combination of portable device, which enables real-time speech stretching, with the workstation designed in order to perform hearing tests. System was designed to help children suffering from Central Auditory Processing Disorders.

Designing Diagnostic Observers via Entirely-Left Eigenstructure Assignment

: The synthesis of observers for dynamic plants is a principal tool in designing diagnostic systems. A recently-proposed sub-optimal methodology for synthesizing statically-decoupled residual generators is based on a predetermined set of observation system eigenvalues and embedded in a geometric framework. The idea of parameterization of corresponding attainable eigensubspaces applied there leads to clear conditions for disturbance decoupling. All developments are set solely in terms of the left eigenvectors of the system observation state-transition matrix under consideration. Copyright

DESIGNING DIAGNOSTIC OBSERVERS VIA ENTIRELY-LEFT EIGENSTRUCTURE ASSIGNMENT

Abstract The synthesis of observers for dynamic plants is a principal tool in designing diagnostic systems. A recently-proposed sub-optimal methodology for synthesizing statically-decoupled residual generators is based on a predetermined set of observation system eigenvalues and embedded in a geometric framework. The idea of parameterization of corresponding attainable eigensubspaces applied there leads to clear conditions for disturbance decoupling. All developments are set solely in terms of the left eigenvectors of the system observation state-transition matrix under consideration.

SOME ASPECTS OF RELIABLE ℋ∞ SYNTHESIS OF STATE ESTIMATORS

Abstract The paper deals with a problem of numerically reliable synthesis of ℋ∞-optimal discrete-time state estimators based on a dual J-lossless factorisations of the delta-domain dual chain-scattering models of continuous-time plants. Necessary and sufficient conditions for the existence of strictly proper ℋ∞ discrete-time estimators are given. Both the regular problem concerning models with no zeros on the stability boundary and the extended problem of models with such zeros are discussed. It is shown that in the latter case some structural constraints always appear.