Pawel Piatek

Six-minute walk test distance and resting oxygen saturations but not functional class predict outcome in adult patients with Eisenmenger syndrome

BACKGROUND Eisenmenger syndrome (ES) represents the extreme manifestation of pulmonary arterial hypertension in patients with congenital heart disease, associated with significant exercise intolerance and mortality. Even though of six-minute-walk-test (6MWT) is routinely used in these patients, little is known about its prognostic value in comparison to functional class. METHODS AND RESULTS We included 210 adult patients with ES who underwent a total of 822 6MWTs. Median walking distance (6MWD) was 330 m [IQR 260-395], oxygen saturation (SO2) at baseline 86% [IQR 82-91%] and SO2 at peak-exercise 69% [IQR 60-80%]. In patients commenced on advanced therapy for pulmonary hypertension, but not in the reminder, there was a significant improvement in walking distance (297±97 m vs. 325±87 m,P=0.0019), SO2 at rest (84.9±7.1 vs. 86.8±5.9%,P=0.003), SO2 at peak exercise (69.1±12.7 vs. 72.3±12.2%,P=0.04) and NYHA functional class (P=0.0047). During a follow up of 3.3 years, 29 patients died. On time-dependent Cox analysis, 6MWD (HR 0.94 per 10 m, 95%CI: 0.91-0.97,P<0.001) and baseline SO2 (HR 0.90, 95%CI:0.86-0.94,P<0.0001) were predictors of death. In contrast, age, functional class, peak-exercise SO2 and SO2 change were not related to mortality. A three-fold increased risk of death was identified in patients not reaching a 6MWD of 350 m or with baseline SO2 below 85%. CONCLUSIONS The 6MWD and resting SO2, but not functional class were predictive of outcome in this contemporary cohort of Eisenmenger patients and should be incorporated in both risk stratification and management algorithms for these patients.

Impulse response approximation method for bi-fractional filter

In this paper method of impulse response approximation with orthogonal Laguerre functions is used for the approximation of bi-fractional filter (BFF). This filter is a new development in the field of non-integer order signal processing. Filter with three parameters allows advanced shaping of frequency response, especially the phase characteristic. Method of impulse response approximation is a very promising one because of its great conditioning and sound theoretical basis.

Robust non-integer order controller for air heater

Robust non-integer control is an emerging field in control theory. Multiple works are focused on designing controllers with flatphase and robust stability margins. In this paper design of non-integer order with robust properties is considered. This controller is designed for the model of air heating process trainer system belonging to the Department of Automatics and Biomedical Engineering of AGH University and Science and Technology and verified experimentally with that system.

Time-domain Oustaloup approximation

Paper presents a new method for implementing Oustaloup approximation of non-integer order systems in digital environment. Classical method and its limitations are presented and discussed. Method of realization is formulated and schemes for discretization are presented. Operation of method is illustrated with two examples. Work is ended with discussion of method limitations and conclusions.

Bi-fractional filters, part 2: Right half-plane case

In this paper new non-integer order filter is proposed. Considered filter is given by a transfer function \(\cfrac{c}{s^{2\alpha}+ b s^{\alpha}+c}\), with parameters b and c chosen in a way, that locates the eigenvalues of the system in left open complex half plane. Dependence of frequency characteristic of the system on parameters α, b and c is investigated. Also method for realisation in the form of non-integer order differential equations is analysed.

Implementation of an inexpensive EEG headset for the pattern recognition purpose

There are many types of bio-signals with various control application prospects. In this work possible control application domain of electroencephalographic signal obtained from an easily available, inexpensive EEG headset - Emotiv EPOC was presented. This work also involved application of an embedded system platform. That solution caused limits in choosing an appropriate signal processing method, as embedded platforms characterise with a little efficiency and low computing power. Potential implementation of the embedded platform enables to extend the possible future application of the proposed BCI. It also gives more flexibility, as the platform is able to simulate various environments. In this work traditional, statistical methods were neither used nor described.

Nonlinear dynamical feedback for motion control of magnetic levitation system

For fast, nonlinear and unstable dynamical systems, such as magnetic levitation, high quality of control is essential. Such control can be generated by different feedback structures, most of which require measurements of velocities or derivatives of the system output. In case of magnetic levitation the velocity of levitating object (a ferromagnetic sphere) is not measured. That is the reason for application of state observers which along with different controllers will create a nonlinear dynamical feedback structure. In this paper we will consider two such structures - PID controller with a nonlinear feed-forward and a cascade linearising feedback using two nonlinear state observers: high-gain observer (HGO) and an observer with reduced, linear error dynamics (abbreviated as RO). Operation of these control algorithms will be compared with a classical PID structure, derivative part of which will be realised either as a finite difference or with an observer.

Discretisation of different non-integer order system approximations

The goal of this paper is to compare popular methods of approximation on non-integer order systems and propose strategies of increasing their robustness to numerical errors of their coefficients. It is shown, that popular methods during discretisation lead to instability.

On the practical implementation of non-integer order filters

Non-integer order filters allow much greater flexibility in shaping of frequency responses than classical methods. In this paper authors present the methodology of digital filter design using non-integer order filter as a prototype. Implementation on the embedded testbed was performed to validate the presented method. Experiments with measuring of frequency response show consistency of implementation with desired characteristics. The presented approach is promising and can be successfully used in multiple applications.

Parametric optimization of PD α controller using Laguerre function approximation

In this paper the method for optimisation of PDα controllers for linear systems was investigated. A new method for approximation of performance index with infinite horizon was presented and used. Optimisation was undertaken on approximated performance index with satisfactory results for multiple test systems. Results are promising, convergence of approximant and performance index was observed and numerically illustrated. Method is easily extendable and shows promise as a general method of non-integer controller tuning.