Goran Kvascev

Non-destructive characterisation and classification of ceramic artefacts using pEDXRF and statistical pattern recognition

BackgroundPortable energy dispersive X-ray fluorescence (pEDXRF) spectrometry analysis was applied for the characterisation of archaeological ceramic findings from three Neolithic sites in Serbia. Two dimension reduction techniques, principal component analysis (PCA) and scattering matrices-based dimension reduction were used to examine the possible classification of those findings, and to extract the most discriminant features.ResultsA decision-making procedure is proposed, whose goal is to classify unknown ceramic findings based on their elemental compositions derived by pEDXRF spectrometry. As a major part of decision-making procedure, the possibilities of two dimension reduction methods were tested. Scattering matrices-based dimension reduction was found to be the more efficient method for the purpose. Linear classifiers designed based on the desired output allowed for 7 of 8 unknown samples from the test set to be correctly classified.ConclusionsBased on the results, the conclusion is that despite the constraints typical of the applied analytical technique, the elemental composition can be considered as viable information in provenience studies. With a fully-developed procedure, ceramic artefacts can be classified based on their elemental composition and well-known provenance.

Classification of muscle twitch response using ANN: Application in multi-pad electrode optimization

In this paper we present a method for optimization of spatial selectivity of multi-pad electrode during transcutaneous Functional Electrical Stimulation (FES). The presented method is based on measurent of individual muscle twitches using Micro-Electro-Mechanical Systems (MEMS) accelerometers positioned on hand, while stimulating with low frequency electrical stimulation via pads within multi-pad electrode. When elicited, wrist or fingers flexion/extension produce different, characteristic wave shapes of acceleration, by using trained Artificial Neural Network (ANN) we can detect these characteristic signals and detect correlation of each pad and activated muscle beneath. Results presented in this paper show high degree of accurate classification of the elicited movement in inter-subject testing.

Neural networks in feedforward control of a robot arm driven by antagonistically coupled drives

The paper deals with a rapidly growing trend in robotics - anthropomimetics. Following a human paragon, bio-inspired control of the robot arm is presented using artificial neural networks. This work demonstrates results achieved by feedforward control comparing feedforward backpropagation networks and radial bases networks. Use of radial bases network prevails as an efficient tool to evade the exact mathematical modeling and conventional control of the complex mechanical system that is highly nonlinear and includes passive compliance.

Neural network control approach for a two-tank system

This paper describes two different approaches in a two tank-system control using neural networks - the NARMA-L2 Control and the Model Reference Control. Knowing that the process control has had the most satisfying results using a standard PID controller, this paper compares the two approachespsila results one to another but also every one of them with the PID controllerpsilas results. The goal was to increase the system response speed without heavily increasing the two other relevant parameters - the overshoot and the steady state error. All of the experiments, measurements and simulations were conducted in Matlab/RT Simulink.

Hand gesture recognition using neural network based techniques

In this paper, two neural network based methods were implemented for recognition of images showing 10 hand gestures. Images were available from 24 subjects and captured on two different backgrounds and with several space orientations. Firstly, Histogram of Oriented Gradients method was applied for feature extraction and training was performed with multilayer feed forward neural network with back propagation algorithm. Within the second method, Sparse autoencoder with 5 hidden layers and decreasing number of neurons was implemented. For both methods it was examined how number of descriptors influences the accuracy of classification and found relationship was used to determine best performing case. Both classification methods achieved accuracy of about 92.5%, by using the similar number of estimated parameters.

Adaptive estimation of time-varying parameters in AR models with variable forgetting factor

A new method for estimating time-varying parameters of nonstationary AR signal models, based on adaptive recursive least squares with variable forgetting factors, is described. The adaptive estimator differs from the conventional one by the simultaneously estimation of AR model parameters and scale factor of prediction residuals, while the variable forgetting factor values are adapted to the nonstationary signal via a new extended prediction error detection scheme. The method has good adaptability in the non-stationary situations, and gives low bias and low variance at the stationary situations. The feasibility of the approach is demonstrated with simulations.

Combustion distribution control using the extremum seeking algorithm

Quality regulation of the combustion process inside the furnace is the basis of high demands for increasing robustness, safety and efficiency of thermal power plants. The paper considers the possibility of spatial temperature distribution control inside the boiler, based on the correction of distribution of coal over the mills. Such control system ensures the maintenance of the flame focus away from the walls of the boiler, and thus preserves the equipment and reduces the possibility of ash slugging. At the same time, uniform heat dissipation over mills enhances the energy efficiency of the boiler, while reducing the pollution of the system. A constrained multivariable extremum seeking algorithm is proposed as a tool for combustion process optimization with the main objective of centralizing the flame in the furnace. Simulations are conducted on a model corresponding to the 350MW boiler of the Nikola Tesla Power Plant, in Obrenovac, Serbia.

On signal-to-noise ratio estimation

A new simple algorithm for estimating signal-to-noise ratio (SNR) for a signal consisting of one sinusoid in white Gaussian noise is proposed in this paper. Algorithm is based on autocorrelation and modified covariance methods for AR (Autoregressive) spectral estimation. The validity of the algorithm is examined by comparing its SNR estimate with SNR estimate obtained by sinusoid magnitude estimation using Pisarenko harmonic decomposition method and noise variance estimation using modified covariance method. By a large number of simulations this algorithm has proven itself as a comparably precise even in case of significantly noise-contaminated sinusoidal signal.

TOWARD ACOUSTIC NOISE TYPE DETECTION BASED ON QQ PLOT STATISTICS

Fault detection and state estimation using acoustic signals is a procedure highly affected by ambient noise. This is particularly pronounced in an industrial environment where noise pollution is especially strong. In this paper a noise detection algorithm is proposed and implemented. This algorithm can identify the times in which the recorded acoustic signal is influenced by different types of noise in the form of unwanted impulse disturbance or speech contamination. The algorithm compares statistical parameters of the recordings by generating a series of QQ plots and then using an appropriate stochastic signal analysis tools like hypothesis testing. The main purpose of this algorithm is to eliminate noisy signals and to collect a set of noise free recordings which can then be used for state estimation. The application of these techniques in a real industrial environment is extremely complex because sound contamination usually tends to be intense and nonstationary. The solution described in this paper has been tested on a specific problem of acoustic signal isolation and noise detection of a coal grinding fan mill in thermal power plant in the presence of intense contaminating sound disturbances, mainly impulse disturbance and speech contamination.

Application of Structured and Directional Residuals for Fault Detection and Isolation on Permanent‐Magnet DC Motor with Amplifier

Sensors and actuators are physical components often subjected to non-permissible or unexpected deviations from nominal operating conditions. This paper discusses the application of additive fault detection and isolation (FDI) methods developed for linear and stationary systems on a nonlinear non-stationary system consisting of an electronic amplifier with a DC motor. A temperature-dependent viscous friction coefficient, as well as the non-linearity induced by dry friction, makes the system nonlinear. Residuals were designed using two fundamental residual-enhancement approaches: synthesis of structured residuals and synthesis of directional residuals. A comparative analysis of the results was performed applying four different techniques for residual transformation synthesis. The paper proposes suitable filtering and translation of the structured and directional residuals that enhance FDI, performance. A limiting factor in the application of directional residuals, relating to the number of different faults, which may have independent directions during FDI, is illustrated. The entire procedure is demonstrated on a simple model of a permanent-magnet DC motor with a suitable amplifier. This laboratory system is often encountered in electrical engineering laboratories, and accordingly, the results can be used as useful educational material for student training in FDI. Copyright