Post Recognition Speech Localization

How much does knowledge regarding a certain spoken word or phrase help with its localization? This is a very fundamental question for speech processing, and will be partially addressed in this paper. In particular, this work will utilize prior information regarding the contents of a speech signal in order to improve the artificial localization of it using Time delay of arrival (TDOA) between two microphones. The prior information, which is used to develop a very simple frequency-selective phase transform (FPT), increases the effective SNR by only using a subset of the highest SNR frequencies in the Phase Transform. Simulations in a reverberant environment show that the proposed approach can more robustly and accurately localize speech sources. For 20 ms signal segments, it is shown that using a subset of 45 percent of available speech frequency bins is superior to using 30, 60, or 100, where using 100 corresponds to the standard Phase Transform.     

Estimation of elbow-induced wrist force with EMG signals using fast orthogonal search

In many studies and applications that include direct human involvement-such as human-robot interaction, control of prosthetic arms, and human factor studies-hand force is needed for monitoring or control purposes. The use of inexpensive and easily portable active electromyogram (EMG) electrodes and position sensors would be advantageous in these applications compared to the use of force sensors, which are often very expensive and require bulky frames. Multilayer perceptron artificial neural networks (MLPANN) have been used commonly in the literature to model the relationship between surface EMG signals and muscle or limb forces for different anatomies. This paper investigates the use of fast orthogonal search (FOS), a time-domain method for rapid nonlinear system identification, for elbow-induced wrist force estimation. It further compares the forces estimated using FOS with the forces estimated by MLPANN for the same human anatomy under an ensemble of operational conditions. In this paper, the EMG signal readings from upper arm muscles involved in elbow joint movement and sensed elbow angular position and velocity are utilized as inputs. A single degree-of-freedom robotic experimental testbed has been constructed and used for data collection, training and validation.     

Correction to: Employing singular value decomposition and similarity criteria for alleviating cold start and sparse data in context-aware recommender systems

Electronic Commerce Research -     

Expanding the Scope of Sortase‐Mediated Ligations by Using Sortase Homologues

Sortase‐catalyzed transacylation reactions are widely used for the construction of non‐natural protein derivatives. However, the most commonly used enzyme for these strategies (sortase A from Staphylococcus aureus) is limited by its narrow substrate scope. To expand the range of substrates compatible with sortase‐mediated reactions, we characterized the in vitro substrate preferences of eight sortase A homologues. From these studies, we identified sortase A enzymes that recognize multiple substrates that are unreactive toward sortase A from S. aureus. We further exploited the ability of sortase A from Streptococcus pneumoniae to recognize an LPATS substrate to perform a site‐specific modification of the N‐terminal serine residue in the naturally occurring antimicrobial peptide DCD‐1L. Finally, we unexpectedly observed that certain substrates (LPATXG, X=Nle, Leu, Phe, Tyr) were susceptible to transacylation at alternative sites within the substrate motif, and sortase A from S. pneumoniae was capable of forming oligomers. Overall, this work provides a foundation for the further development of sortase enzymes for use in protein modification.     

A Stable Analytical Solution Method for Car-Like Robot Trajectory Tracking and Optimization

In this paper, the car-like robot kinematic model trajectory tracking and control problem is revisited by exploring an optimal analytical solution which guarantees the global exponential stability of the tracking error. The problem is formulated in the form of tracking error optimization in which the quadratic errors of the position, velocity, and acceleration are minimized subject to the rear-wheel car-like robot kinematic model. The input-output linearization technique is employed to transform the nonlinear problem into a linear formulation. By using the variational approach, the analytical solution is obtained, which is guaranteed to be globally exponentially stable and is also appropriate for real-time applications. The simulation results demonstrate the validity of the proposed mechanism in generating an optimal trajectory and control inputs by evaluating the proposed method in an eight-shape tracking scenario.     

Investigation of the corrosion behavior of cathodic arc evaporated stainless steel coating in 3.5% NaCl

In this paper, two major technical problems (growth defects and chromium content loss) encountering when cathodic arc evaporation physical vapor deposition (CAE-PVD) was used for deposition of stainless steel and their subsequent effects on corrosion behavior of the coating in 3.5 wt % NaCl solution have been investigated. Growth defects in spherical and needle-like shapes were the common defects that resulted from CAE-PVD of a stainless steel and played a major role in determining the corrosion behavior of the coating. The results showed a composition difference between the coating matrix (with ~11 at % Cr) and the growth defects, particularly needle-like ones (with ~15 at % Cr). According to SEM images, it seemed that the needle-like defects were passivated and were susceptible to pitting corrosion while coating matrix was corroded. The results also showed that the corrosion of the coating was influenced by two factors: building up micro-galvanic cells between the needle-like defects (as passivated regions) and both coating matrix and the spherical defects (as active sites). In addition, an intense localized corrosion (as micro-crevice corrosion) was observed around the growth defects.

     

Influence of Friction Stir Processing Parameters on the Microstructure of Aluminum Foams

In the present study, friction stir processing (FSP) has been used to fabricate aluminium foams. The effects of the number of FSP passes, FSP tool rotational speed, foaming time and temperature on the porosity have been investigated. Aluminium foam with porosity up to 40% was successfully fabricated. In the samples foamed at 923 K (650 °C), a few irregular pores were produced as a result of high aluminium matrix stiffness in this temperature. In general with increase in foaming temperature the porosity increased. However, in the samples foamed for 30 or 60 min, lower porosity was detected at higher foaming temperature. Also, in the samples which were produced with more FSP passes, the foaming time decreased and more uniform pore structure was obtained.     

Applications of genetic algorithms in process planning: tool sequence selection for 2.5-axis pocket machining

Tool sequence selection is an important activity in process-planning for milling and has great bearing on the cost of machining. Currently, it is accomplished manually without consideration of cost factors a priori. Typically, a large tool is selected to quickly generate the rough shape and a smaller clearing tool is used to generate the net-shape. In this paper, we present a new systematic method to select the optimal sequence of tool(s), to machine a 2.5-axis pocket given pocket geometry, a database of cutting tools, cutting parameters, and tool holder geometry. Algorithms have been developed to calculate the geometric constructs such as accessible areas, and pocket decomposition, while considering tool holders. A Genetic Algorithm (GA) formulation is used to find the optimal tool sequence. Two types of selection mechanisms namely “Elitist selection” and “Roulette method” are tested. It is found that the Elitist method converges much faster than the Roulette method. The proposed method is compared to a shortest-path graph formulation that was developed previously by the authors. It is found that the GA formulation generates near optimal solutions while reducing computation by up to 30% as compared to the graph formulation.     

Solubility prediction for furosemide in water-cosolvent mixtures using the minimum number of experiments

The mole fraction solubility of a poorly water soluble loop diuretic, furosemide, was determined in aqueous binary mixtures of ethanol, propylene glycol, and glycerol from 0% to 100% cosolvent concentrations at 25 degrees C. Solubility predictions based on the minimum number of experimental data points were performed using the commonly used accurate cosolvency models: the three-suffix excess free energy (3xEFE), the mixture response surface (MRS), the combined nearly ideal binary solvent/Redlich-Kister (CNIBS/R-K), and the general single model (GSM). This prediction method was tested using three sets of solubility data for furosemide generated in this study and 11 data sets collected from the literature. The average percentage deviations (APDs) were 8.4 +/- 3.8, 13.6 +/- 7.3, 7.4 +/- 2.8, and 7.6 +/- 2.9, respectively, for 3xEFE, MRS, CNIBS/R-K, and GSM models. Using 3xEFE, CNIBS/R-K, and GSM models, which are theoretically related, a mean predicted solubility (MPS) approach was also proposed. The APD for this method was 7.3 +/- 2.3. The mean differences between MRS and the others were statistically significant (p < .001). The results showed that one can employ solubility prediction based on a minimum of five experimental data points, and the expected APD is less than 10%.     

The biological seal of the implant–soft tissue interface evaluated in a tissue-engineered oral mucosal model

For dental implants, it is vital that an initial soft tissue seal is achieved as this helps to stabilize and preserve the peri-implant tissues during the restorative stages following placement. The study of the implant–soft tissue interface is usually undertaken in animal models. We have developed an in vitro three-dimensional tissue-engineered oral mucosal model (3D OMM), which lends itself to the study of the implant–soft tissue interface as it has been shown that cells from the three-dimensional OMM attach onto titanium (Ti) surfaces forming a biological seal (BS). This study compares the quality of the BS achieved using the three-dimensional OMM for four types of Ti surfaces: polished, machined, sandblasted and anodized (TiUnite). The BS was evaluated quantitatively by permeability and cell attachment tests. Tritiated water (HTO) was used as the tracing agent for the permeability test. At the end of the permeability test, the Ti discs were removed from the three-dimensional OMM and an Alamar Blue assay was used for the measurement of residual cells attached to the Ti discs. The penetration of the HTO through the BS for the four types of Ti surfaces was not significantly different, and there was no significant difference in the viability of residual cells that attached to the Ti surfaces. The BS of the tissue-engineered oral mucosa around the four types of Ti surface topographies was not significantly different.