Statistically correlated multi-task learning for autonomous driving

Neural Computing and Applications - Autonomous driving research is an emerging area in the machine learning domain. Most existing methods perform single-task learning, while multi-task learning...     

Modeling guided wave propagation with application to the long-range defect detection in railroad tracks

The aim of this work is to demonstrate the use of a commercial finite element package, ABAQUS EXPLICIT, to model ultrasonic guided waves in structural components. The particular application of interest is modeling the interaction of a broadband vertical bending mode with transverse-type defects in railroad tracks. This topic is part of a broader project on high-speed defect detection in rails by long-range ultrasonic inspections. Reflection coefficient spectra in the 20–45 kHz range are obtained for four different sizes and three different orientations of transverse head flaws. A preliminary study of Lamb waves in a free plate helps drawing modeling guidelines for the rail.     

Random-effects regression analysis of correlated grouped-time survival data

Random-effects regression modelling is proposed for analysis of correlated grouped-time survival data. Two analysis approaches are considered. The first treats survival time as an ordinal outcome, which is either right-censored or not. The second approach treats survival time as a set of dichotomous indicators of whether the event occurred for time periods up to the period of the event or censor. For either approach both proportional hazards and proportional odds versions of the random-effects model are developed, while partial proportional hazards and odds generalizations are described for the latter approach. For estimation, a full-information maximum marginal likelihood solution is implemented using numerical quadrature to integrate over the distribution of multiple random effects. The quadrature solution allows some flexibility in the choice of distributions for the random effects; both normal and rectangular distributions are considered in this article. An analysis of a dataset where students are clustered within schools is used to illustrate features of random-effects analysis of clustered grouped-time survival data.     

Alkaline fuel cells: contemporary advancement and limitations

ZeTek Power recently introduced mass manufacturable and cost effective alkaline fuel cells on the market. Today's research is focused on further improvement both in terms of performance increase and cost reduction. This research is classically performed using small (4 cm²) experimental electrodes in the half-cell configuration. This allows the primary electrochemical losses in an anode or cathode to be determined independently. Additional performance losses occur when one integrates large electrodes into a module of 24 cells and in a stack comprised of many modules. By comparing the performance of half-cell experiments to that of modules, these losses can be distinguished and addressed. The information thus obtained, both for the small electrodes and in up-scaling is vital if one is to identify the key areas in which improvement is possible and where to focus future research. Furthermore, the identification of the losses in a module and system allows us to predict the final performance from half-cell measurements of a new laboratory scale experimental electrode.     

Infrared spectroscopic investigation of tannins

Infrared spectroscopy is suggested as a diagnostic method for the characterisation and qualitative estimation of the two classes of tannins. Gallic acid, tannic acid and chebulinic acid have been taken as model compounds for the hydrolysable and catechin for the condensed tannins. The former class is marked by the presence of strong absorption maxima at 1710 – 35 cm^?1. The two classes have characteristic pattern of absorption, from which it is possible to characterise the particular type of tannin.     

Semi- and fully interpenetrating polymer networks based on polyurethane–polyacrylate systems. VI. Polyurethane–poly(methyl acrylate-co-ethyl acrylate) semi-1-interpenetrating polymer networks

Dynamic mechanical and longitudinal sonic velocity measurements have been made on a series of semi-1-IPNs in which the network component is a polyurethane and the linear constituent a copolymer of methyl acrylate and ethyl acrylate. Dynamic mechanical analysis reveals phase separation. The shifting of the polyurethane glass transition in both the tan δ? and the E″–temperature plots indicates that some mixing occurs. The longitudinal sonic velocity results indicate that the polyurethane is present as a continuous phase in all the materials investigated.     

An image composition algorithm for handling global visual effects

Image composition is widely used in television and film industry to create synthetic visual effects. It requires seamless integration of different parts of two or more images into a single image. Existing image composition techniques only change the local contents of the resulting image while in many cases local changes may also require some global effects as well. For example, if the image of sun from one image is transferred to another image, the global brightness pattern should also be transferred. Unfortunately existing techniques cannot handle global effects of local content manipulations. This paper describes a novel image composition technique which captures global effects associated with a specific local content from one image and incorporates in the second image. In our proposed technique, all images are transformed to the frequency domain. The composite image is created in frequency domain by mixing different frequencies from multiple images and then transformed back to the spatial domain. We have experimented the proposed technique to shift the image of sun along with its global brightness pattern, the global effects of rain and also for transferring global texture pattern from one image to the other. In most of the cases the results produced by our algorithm appear far close to real images than state of the art existing image composition techniques.     

Direct analysis of annealing of nodular crystals in isotactic polypropylene by atomic force microscopy, and its correlation with calorimetric data

The process of isothermal annealing of nodular monoclinic crystals of isotactic polypropylene (iPP) was analyzed by atomic force microscopy (AFM) and temperature-modulated differential scanning calorimetry (TMDSC). Initially nodular and mesomorphic domains were obtained by controlled melt-crystallization at high cooling rate. Subsequent heating triggers transition from mesomorphic to monoclinic structure, and melting of unstable nodules. Annealing allows re-crystallization, which is recognized by enlargement of domains from initially about 20 nm to about 35 and 55 nm after annealing at 393 and 433 K, respectively. Furthermore, the re-crystallization process is connected with a slight change of the aspect ratio of crystals. The isothermal re-crystallization of the liquid is superimposed by aggregation of crystals, to yield blocky, and string-like objects. The direct analysis of structure on isothermal annealing by AFM is for the first time compared with the isothermal decrease of the apparent specific heat capacity, or change of enthalpy, monitored by TMDSC. The apparent specific heat capacity decreases during annealing with an identical non-linear time dependence as the directly observed growth of the crystal size. Analysis of the annealing processes at different temperatures yields proportionality between the increase of the crystal size and the reduction of the apparent specific heat capacity.     

Investigations on the adhesion and interfacial properties of polyurethane foam/thermoplastic materials

The adhesion and interfacial properties of polyurethane (PU) foams with thermoplastic (TP) materials were investigated using different techniques. The adhesion performance of PU foam with TP materials was evaluated using the peel test method, and the adhesion durability was checked after different climate treatments. X‐ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM), and contact angle measurements were used to study the surface and interface morphology of PU foam and TP material system. Three types of PU foam samples which differ in their composition and also five commercially available TP blends systems, based on poly(carbonate), poly(styrene‐co‐maleic anhydride), poly(acrylonitrile‐butadiene‐styrene), and silicone acrylate rubber have been used. The slow reacting foam shows the best adhesion properties with all the TP materials. The climate treatments strongly effected the PU foam adhesion durability with poly(carbonate) containing TP materials (70–80% loss in adhesion), but nearly no effect with poly(styrene‐co‐maleic anhydride). The samples with lowered adhesion could be separated by peeling without visible foam residues on the TP surface. AFM, XPS, and surface tension studies have shown that the surface properties of the TP material are still governed by the PU foam. © 2007 Wiley Periodicals, Inc. J Appl Polym Sci 104: 479–488, 2007     

Transfer learning in deep neural network-based receiver coil sensitivity map estimation

Magnetic Resonance Materials in Physics, Biology and Medicine - The success of parallel Magnetic Resonance Imaging algorithms like SENSitivity Encoding (SENSE) depends on an accurate estimation of...