Mixing matrix estimation using discriminative clustering for blind source separation

Mixing matrix estimation in instantaneous blind source separation (BSS) can be performed by exploiting the sparsity and disjoint orthogonality of source signals. As a result, approaches for estimating the unknown mixing process typically employ clustering algorithms on the mixtures in a parametric domain, where the signals can be sparsely represented. In this paper, we propose two algorithms to perform discriminative clustering of the mixture signals for estimating the mixing matrix. For the case of overdetermined BSS, we develop an algorithm to perform linear discriminant analysis based on similarity measures and combine it with K-hyperline clustering. Furthermore, we propose to perform discriminative clustering in a high-dimensional feature space obtained by an implicit mapping, using the kernel trick, for the case of underdetermined source separation. Using simulations on synthetic data, we demonstrate the improvements in mixing matrix estimation performance obtained using the proposed algorithms in comparison to other clustering methods. Finally we perform mixing matrix estimation from speech mixtures, by clustering single source points in the time-frequency domain, and show that the proposed algorithms achieve higher signal to interference ratio when compared to other baseline algorithms.     

Hydrogen production from used lubricating oils

Used lubricating oils (lube oils) are generated throughout the year and collected in central locations in many communities. Disposing lube oil in an improper manner contaminates environment to a great degree. Used lube oil can be valuable as a re-refined lubricant or as an energy source. Lube oil is a complex mixture of aliphatic and polycyclic hydrocarbons formulated to withstand high service temperatures in internal combustion engines. Both synthetic and mineral oils contain a high concentration of hydrogen (about 13–14 wt%). At the Florida Solar Energy Center, we have developed a process that converts lube oils to hydrogen and other valuable low molecular weight hydrocarbons. The lube oil reformation experiments were carried out using several commercially available dehydrogenation catalysts at a range of reactor temperatures and pressures, residence times and steam to carbon ratios. In this paper, the data obtained to date and the results are presented and discussed.     

Preparation of YBCO Films on CeO2-Buffered (001) YSZ Substrates by a Non-Fluorine MOD Method

YBa₂Cu₃O_7−δ (YBCO) films were fabricated via a fluorine-free metal organic deposition (MOD) method followed by high-temperature, low oxygen partial pressure annealing. Trimethyl acetate salts of copper, yttrium, and barium hydroxide were used as the precursors, which were dissolved in proponic acid- and amine-based solvents. After spin-coating and burnout, samples were annealed at 740°C in 180 ppm oxygen partial pressure and exposed to humid atmosphere for different times. A critical transition temperature, T _c( R =0) of 90.2 K and a transport critical current density ( J _c) of 0.55 MA/cm² (77 K and self-field) were obtained for 0.2 μm YBCO films on CeO₂-buffered yttria-stabilized zirconia (YSZ) substrates. X-ray studies shows that the YBCO films have sharp in-plane and out-of-plane texture for all samples; however, the porosity of the YBCO film varies with the time of exposure to the humid atmosphere. A reaction between YBCO and CeO₂ during the high-temperature anneals and formation of the reaction product BaCeO₃ was confirmed by X-ray diffraction (XRD) studies. The XRD and transmission electron microscopy analysis indicated that the epitaxial relations in the film were YBCO (00 l )//CeO₂ (00 l )//YSZ (00 l ) and YBCO [100]//CeO₂ [110]//YSZ [110].     

Investigation on Maximizing Packet Delivery Rate in WSN Using Cluster Approach

Wireless Personal Communications - Wireless Sensor Network plays a vital role and part of real time communication applications. Location of unknown node is difficult to find in the presence of...     

Structural,electrical and electrochemical properties of polyacrylonitrile-ammonium hexaflurophosphate polymer electrolyte system

Polyacrylonitrile (PAN) based polymer electrolyte membranes complexed with Ammonium hexafluorophosphate (NH₄PF₆) with different molar concentration are prepared by solution casting method. Increase in the amorphous nature by the addition of Ammonium salt and the formation of polymer-salt complex are confirmed by X ray diffraction studies and infrared spectroscopy respectively. The glass transition temperature is measured for all membranes and it showed a lowest value for the PAN complexed with 20 mol% of NH₄PF₆. Electrical properties are studied by AC impedance spectroscopy. An ionic conductivity of the order of 10^?3 Scm^?1 is obtained for the 80 PAN / 20 NH₄PF₆ polymer electrolyte. Conductivity, dielectric and modulus spectra from the impedance data are analysed to understand the ionic transport mechanism. Transference number measurement is done to study the ionic contribution to the charge transport. A proton battery with the configuration, Zn+ ZnSO₄. 7H₂O /80 PAN / 20 NH₄PF₆ / PbO₂ +V₂O₅ has been constructed and its discharge characteristics are studied.     

A.c. conductivity studies on the silver molybdo-arsenate glassy system

A new quaternary fast-ion conducting silver molybdo-arsenate [Agl-Ag₂O-(MoO₃ + As₂O₅)] (SMA) glassy system has been prepared using the melt-quenching technique for various dopant salt (Agl) concentrations by fixing the formers (MoO₃ + As₂O₅) composition and the modifier (Ag₂O) to formers (M/F) ratio. The prepared compounds were characterized by X-ray diffraction. The impedance measurements were made on different Agl compositions of the SMA glasses as a function of frequency (6.5 Hz–65 kHz) and temperature (303–343 K), using the Solatron frequency-response analyser (model 1250). The bulk conductivity and the appropriate physical model (equivalent circuit) of the SMA glass were obtained from the impedance analysis. The a.c. conductivity was calculated for different Agl compositions of SMA glasses at various temperatures and the obtained a.c. conductivity results were analysed using Jonscher's Universal Law. The conduction mechanism for the highest conducting SMA glassy compound has been explained using the diffusion path model.     

Optimization of Machining Characteristics for Al／SiCp Composites using ANN／GA

The present work is focused on optimization of machining characteristics of Al/SiCp composites.The machining characteristics such as specific energy,tool wear and surface roughness were studied.The parameters such as volume fraction of SiC,cutting speed and feed rate were considered.Artificial neural networks(NN) was used to train and simulate the experimental data.Genetic algorithms(GA) was interfaced with ANN to optimize the machining conditions for the desired machining characteristics .Validation of optimized results was also performed by confirmation experiments.     

Plasma-Sprayed Ceramic Reaction Couples

In preparing specimens for studying reaction kinetics, it is essential to ensure proper contact between couple end-members. In the present work, reaction couples were prepared by spraying layers of the end-member materials over one another. This method has many advantages, including precise control over the properties of the end-members.