Comparative study on classifying human activities with miniature inertial and magnetic sensors

This paper provides a comparative study on the different techniques of classifying human activities that are performed using body-worn miniature inertial and magnetic sensors. The classification techniques implemented and compared in this study are: Bayesian decision making (BDM), a rule-based algorithm (RBA) or decision tree, the least-squares method (LSM), the k-nearest neighbor algorithm (k-NN), dynamic time warping (DTW), support vector machines (SVM), and artificial neural networks (ANN). Human activities are classified using five sensor units worn on the chest, the arms, and the legs. Each sensor unit comprises a tri-axial gyroscope, a tri-axial accelerometer, and a tri-axial magnetometer. A feature set extracted from the raw sensor data using principal component analysis (PCA) is used in the classification process. A performance comparison of the classification techniques is provided in terms of their correct differentiation rates, confusion matrices, and computational cost, as well as their pre-processing, training, and storage requirements. Three different cross-validation techniques are employed to validate the classifiers. The results indicate that in general, BDM results in the highest correct classification rate with relatively small computational cost.     

A new extended LMI-based robust gain scheduled state feedback <Emphasis Type="Italic">H</Emphasis><Subscript>2</Subscript> controller design

This paper proposes an improved robust H ₂ state feedback control synthesis for the Linear Parameter Varying (LPV) systems by attaining the affine quadratic stability. In place of standard H ₂ computation in the literature, a new H ₂ computation based on extended Linear Matrix Inequality (LMI) is improved by means of the slack variable, where it is obtained by separation Lyapunov matrix from system matrix. State feedback H ₂ synthesis is improved for the systems, and is more effective and less conservative than the common ones in the literature. Therefore, the less conservative results are obtained for gain scheduling controller design for LPV systems. The numerical examples are presented to show the superiority of the proposed controller design.     

A combination of Genetic Algorithm,Particle Swarm Optimization and Neural Network for palmprint recognition

In this study, a new approach to the palmprint recognition phase is presented. 2D Gabor filters are used for feature extraction of palmprints. After Gabor filtering, standard deviations are computed in order to generate the palmprint feature vector. Genetic Algorithm-based feature selection is used to select the best feature subset from the palmprint feature set. An Artificial Neural Network (ANN) based on hybrid algorithm combining Particle Swarm Optimization (PSO) algorithm with back-propagation algorithms has been applied to the selected feature vectors for recognition of the persons. Network architecture and connection weights of ANN are evolved by a PSO method, and then, the appropriate network architecture and connection weights are fed into ANN. Recognition rate equal to 96% is obtained by using conjugate gradient descent algorithm.

     

Experimental investigation of steel fiber reinforced concrete box beams under bending

Reduction of dead weight of a reinforced-concrete (RC) structure without too much concession in its load carrying capacity has always been an attractive study subject because it engenders (1) a decrease in dimensions of the members, (2) a decrease in the reinforcement steel, and (3) a decrease in lateral inertia forces during severe earthquakes. In this study, nine RC beams of outer dimensions of 300 × 300 × 2000 mm, six of which are box beams, designed and produced using a C20 class steel fiber concrete, (SFRC) with the commonly used steel fiber type of Dramix-RC-80/0.60-BN at a dosage of 30 kg/m³, are tested under bending. The mechanical behaviours of all these nine beams under bending are recorded from the beginning of the test till the ultimate failure of the tensile reinforcement in a two-point beam-loading setup. The proportions of (1) loss in ultimate load versus reduction in dead weight and (2) (ultimate experimental load)/(ultimate theoretical load) of the SFRC box beams are determined for two different box thicknesses. Dimensionless behaviour relationships of all the SFRC beams are determined, and the experimentally obtained relationship between the ratio of (actual ultimate load)/(theoretical ultimate load) and the ratio of (wall thickness)/(beam height) for the SFRC box beams is expressed diagrammatically.     

Zinc Oxide Nanowire Decorated Single‐Use Electrodes for Electrochemical DNA Detection

The surfaces of pencil graphite electrodes (PGEs) were decorated with zinc oxide nanowires (ZnO NWs) for the electrochemical detection of nucleic acids. ZnO NWs were synthesized through simple hydrothermal method. PGEs decorated with ZnO NWs (ZnO NW/PGEs) were electrochemically characterized through cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) following morphological characterization through transmission (TEM) and scanning electron microscopy (SEM). Enhanced sensor response obtained using ZnO NW/PGEs contrary to the bare PGE (control) samples. Our preliminary results simply reveal the potential of combining ZnO NWs with disposable sensor technology for the electrochemical detection of DNA.     

Effects of limestone replacement ratio on the sulfate resistance of Portland limestone cement mortars exposed to extraordinary high sulfate concentrations

A comparative study has been performed on the sulfate resistance of Portland limestone cement (PLC) mortars exposed to extraordinary high sulfate concentrations (200 g/l). PLCs have been prepared by using two types of clinkers having different C₃S/C₂S ratios and interstitial phase morphologies. Blended cements have been prepared by replacing 5%, 10%, 20% and 40% of clinker with limestone. Cubic (50 × 50 × 50 mm) and prismatic (25 × 25 × 285 mm) cement mortars were prepared. After two months initial water curing, these samples were exposed to three different sulfate solutions (Na₂SO₄ at 20 °C and 5 °C, MgSO₄ at 5 °C). Solutions were not refreshed and pH values of solutions were monitored during the testing stage. The compressive strength and length changes of samples have been monitored for a period of 1 year. Additional microstructural analyses have been conducted by XRD and SEM/EDS studies. Results indicated that in general, limestone replacement ratio and low temperature negatively affect the sulfate resistance of cement mortars. Additionally, clinkers of high C₃S/C₂S ratios with dendritic interstitial phase structure were found to be more prone to sulfate attack in the presence of high amounts of limestone.From the results, it is postulated that in the absence of solution change, extraordinary high sulfate content modified the mechanism of sulfate reactions and formation of related products. At high limestone replacement ratios, XRD and SEM/EDS studies revealed that while ettringite is the main deterioration product for the samples exposed to Na₂SO₄, gypsum and thaumasite formation were dominant products of deterioration in the case of MgSO₄ attack. It can be concluded that, the difference between reaction mechanisms of Na₂SO₄ and MgSO₄ attack to limestone cement mortars strongly depends on the pH change of sulfate solutions.     

Treatment of skewed multi-dimensional training data to facilitate the task of engineering neural models

Successful application of neural network models relies heavily on problem-dependent internal parameters. As the theory does not facilitate the choice of the optimal parameters of neural models, these can solely be obtained through a tedious trial-and-error process. The process requires performing multiple training simulations with various network parameters, until satisfactory performance criteria of a neural model are met. In literature, it has been shown that neural models are not consistently good in prediction under highly skewed data. Consequently, the cost of engineering neural models rises in such circumstance to seek for appropriate internal parameters. In this paper the aim is to show that a recently proposed treatment of highly skewed data eases the task of practitioners in engineering neural network models to meet satisfactory performance criteria. As the applications of neural models grows dramatically in diverse engineering domains, the understanding of the treatment show indispensable practical values.