Pełna lista publikacji

Tytuł Data publikacji Autor
Symptoms caused by cerebral palsy or stroke deprive a person partially or even completely of his ability to move. Nowadays we can observe more technologically advanced rehabil- itation devices which incorporate biofeedback into the process of rehabilitation of such people. However, there is still a lack of devices that would analyse, assess, and control (independently or with limited support) specialised movement exercises. In addition one should also take into account the limited capabilities of physiotherapists in diagnosing and arranging an exercise plan and controlling its course. To fill one of the aforemen- tioned gaps, here we propose an idea of an automated exercise evaluation mechanism based on machine learning techniques. Given a set of established exercise routines used in cerebral palsy or stroke rehabilitation, we evaluate their course and check whether the exercise is performed according to recommendations, i.e., if the motion being executed is in line with the exercise plan and if the right speed of movement is maintained. For this, we mainly focus on lower and upper limb exercises. Movement analysis is performed based on electromyographic activity (EMG) of selected muscles responsible for movement of particular limbs. EMG signals during the exercise are processed and then analysed by selected well-established classifiers, such as: support vector machines, decision tree, random forest, and k-nearest neighbours. The output is the evaluation result whether the exercise is being performed well, too slow, too fast, or some parts of the patient’s movement go beyond the accepted plan. While being only a limited case study, the paper presents the first attempt on creation of a fully automated exercise evaluation procedure based mainly on EMG. Further research on the matter should lead to more sophisticated methods that will take into account more physiological parameters and be able to provide more information about the exercises being performed. In view of the problems described in the first paragraph, such research is justified and necessary.
2018
Aleksander Pałkowski,
Grzegorz Redlarski,
Gustaw Rzyman,
Marek Krawczuk
Currently used body surface area (BSA) formulas give satisfactory results only for individuals with typical physique, while for elderly, obese or anorectic people accurate results cannot be expected. Particularly noteworthy are the results for individuals with severe obesity (body-mass index greater than 35 kg/m2), for which BSA estimation errors reached 80%. The main goal of our study is the development of precise BSA models for specific body parts. We have achieved satisfactory results for a wide range of patients. Using regression models, such as: support vector regression, multilayer perceptron regressor, stochastic gradient descent, or ridge regression, a fourfold decrease in errors proportion is achieved. Machine learning algorithms led to reduction from 1.2 to 8 times for mean estimation error.
2018
Gustaw Rzyman,
Grzegorz Redlarski,
Aleksander Pałkowski,
Piotr Tojza,
Marek Krawczuk,
Janusz Siebert
Symptoms caused by cerebral palsy or stroke deprive a person partially or even completely of his ability to move. Nowadays we can observe more technologically advanced rehabilitation devices which incorporate biofeedback into the process of rehabilitation of such people. However, there is still a lack of devices that would analyse, assess, and control (independently or with limited support) specialised movement exercises. Here we propose an idea of an automated exercise evaluation mechanism based on machine learning techniques, such as: support vector machines, decision trees, random forest, and k-nearest neighbours. While being only a preliminary case study, our research showed that with appropriate processing even a 100% accuracy score can be achieved in classifying whether an exercise is executed well or not.
2018
Aleksander Pałkowski,
Grzegorz Redlarski,
Gustaw Rzyman,
Marek Krawczuk
Modern methods of detection and identification of structural damage direct the activities of scientific groups towards the improvement of diagnostic methods using for example the phenomenon of mechanical wave propagation. Damage detection methods that use mechanical wave propagation in structural components are extremely effective. Many different numerical approaches are used to model this phenomenon, but, due to their universal nature, spectral methods are the most commonly used, of which there are several types. This paper reviews recent research efforts in the field to show basic differences and effectiveness of the two most common spectral methods used for modelling the wave propagation problem in terms of damage detection.
2018
Magdalena Palacz
W artykule przedstawiono zasady doboru żył powrotnych w kablach średniego napięcia z uwzględnieniem zwarć jednofazowych i dwufazowych z udziałem ziemi. Określono warunki zmniejszenia obowiązującego przekroju żyły powrotnej 50 mm2, co uzasadniono analizą nagrzewania przewodu w stanach zwarciowych.
2018
Marek Olesz,
Radosław Sawicz
Wyświetlanie 6 - 10 z 218 rezultatów.
z 44
RSS (otwiera nowe okno)