Volume 8 (2008): Issue 3 (January 2008)

The empirical Bayes Gaussian rule, which in the normal case yields good values of the probability of total error, may yield high values of the maximum probability error. From this point of view the presented modified version of the classification rule of Broffitt, Randles and Hogg appears to be superior. The modification included in this paper is termed as a WR method, and the choice of its weights is discussed. The mentioned methods are also compared with the K nearest neighbours classification rule.

In this work, classification of spirometric pulmonary function test data performed using two artificial neural network methods is compared and reported. The pulmonary function data (N=150) were obtained from volunteers, using commercially available Spirometer, and recorded by standard data acquisition protocol. The data were then used to train (N=100) as well as to test (N=50) the neural networks. The classification was carried out using back propagation and radial basis function neural networks. The results confirm that the artificial neural network methods are useful for the classification of spirometric pulmonary function data. Further, it appears that the Radial basis function neural network is more sensitive when compared to back propagation neural networks. In this paper, the methodology, data collection procedure and neural network based analysis are described in details.

The purpose of this paper was to determine the coefficient of viscosity of a magneto-rheological fluid for different values of the magnetic field and to determine parameters at which the flow of the fluid through a capillary is stopped. To determine the coefficient of viscosity, a method of indirect measurement was implemented using a reference fluid with the known properties. A test stand with a capillary viscometer was constructed. The measurements showed that the viscosity of the magneto-rheological fluid was linearly dependent in a wide range of values of the magnetic induction.

A microcomputer-controlled measuring instrument for capacitance and inductance measurement is described. It is based on an oscillator circuit with the oscillation frequency dependent on a measured element. An analysis of the oscillator used is also given. Equations for the oscillation frequency and its deviation from the resonance frequency of a frequency controlling resonance circuit are derived. The measured results can be transferred into a personal computer (PC) which can process and display these results and control the instrument via RS-232 serial interface.

Some applications of Wireless Sensor Networks (WSNs) to the automobile are identified, and the use of Crossbow MICAz motes operating at 2.4 GHz is considered together with TinyOS support. These WSNs are conceived in order to measure, process and supply to the user diverse types of information during an automobile journey. Examples are acceleration and fuel consumption, identification of incorrect tire pressure, verification of illumination, and evaluation of the vital signals of the driver. A brief survey on WSNs concepts is presented, as well as the way the wireless sensor network itself was developed. Calibration curves were produced which allowed for obtaining luminous intensity and temperature values in the appropriate units. Aspects of the definition of the architecture and the choice/implementation of the protocols are identified. Security aspects are also addressed.

Ensuring the safety and high operation reliability of nuclear reactors takes 100% inspection of geometrical parameters of fuel assemblies, which include the grid spacers fabricated as cellular structure with fuel elements. The required grid spacers' geometry of assembly in the transverse and longitudinal cross sections is extremely important for maintaining the necessary heat regime. A universal method for 3D grid spacer inspection using the diffractive optical element, which generates, as the structured light, a multiple-ring pattern on the inner surface of a grid spacer cell is investigated. The experimental measurement error for cell centers position deviation is ±7 μm, and the error for overall dimensions is ±11 μm.