Joint training of non-negative Tucker decomposition and discrete density hidden Markov models

Non-negative Tucker decomposition (NTD) is applied to unsupervised training of discrete density HMMs for the discovery of sequential patterns in data, for segmenting sequential data into patterns and for recognition of the discovered patterns in unseen data. Structure constraints are imposed on the NTD such that it shares its parameters with the HMM. Two training schemes are proposed: one uses NTD as a regularizer for the Baum–Welch (BW) training of the HMM, the other alternates between initializing the NTD with the BW output and vice versa. On the task of unsupervised spoken pattern discovery from the TIDIGITS database, both training schemes are observed to improve over BW training in terms of pattern purity, accuracy of the segmentation boundaries and accuracy for speech recognition. Furthermore, we experimentally observe that the alternative training of NTD and BW outperforms the NTD regularized BW, BW training and BW training with simulated annealing.     

Trajectory Tracking Control for a Rotary Wing Vehicle Powered by Four Rotors

The present paper presents a trajectory tracking control scheme for an underactuated rotary wing vehicle. The translational controller is designed based on feedback linearization and saturated control. The rotational controller is designed based on the passivity properties of the rotational dynamics. The interconnection between the translational and the rotational controllers is inspired by the backstepping control technique. The final control structure allows to include an integral action that is able to compensate constant disturbances on the rotational dynamics. It is shown that the resulting closed–loop dynamics has a local asymptotic stability property. Numerical simulations show the performance of the proposed controller.     

Modelling gross primary production of a tropical semi-deciduous forest in the southern Amazon Basin

Semi-deciduous forest in the Amazon Basin is sensitive to temporal variation in surface water availability that can limit seasonal rates of leaf and canopy gas exchange. We estimated the seasonal dynamics of gross primary production (GPP) over 3 years (2005–2008) using eddy covariance and assessed canopy spectral reflectance using MODIS imagery for a mature tropical semi-deciduous forest located near Sinop, Mato Grosso, Brazil. A light-use efficiency model, known as the Vegetation Photosynthesis Model (VPM), was used to estimate seasonal and inter-annual variations in GPP as a function of the enhanced vegetation index (EVI), the land surface water index (LSWI), and local meteorology. Our results indicate that the standard VPM was incapable of reproducing the seasonal variation in GPP, primarily because the model overestimated dry-season GPP. In the standard model, the scalar function that alters light-use efficiency (ε_g) as a function of water availability (W_scalar) is calculated as a linear function of the LSWI derived from MODIS; however, the LSWI is negatively correlated with several measures of water availability including precipitation, soil water content, and relative humidity (RH). Thus, during the dry season, when rainfall, soil water content, and RH are low, LSWI, and therefore, W_scalar, are at a seasonal maximum. Using previous research, we derived new functions for W_scalar based on time series of RH and photosynthetic photon flux density (PPFD) that significantly improved the performance of the VPM. Whether these new functions perform equally well in water stressed and unstressed tropical forests needs to be determined, but presumably unstressed ecosystems would have high cloud cover and humidity, which would minimize variations in W_scalar and GPP to spatial and/or temporal variation in water availability.     

Level 2 analysis to estimate LERF risk from a thermally induced rupture of a steam generator tube defect

This paper outlines the Level 2 portion of a methodology for determining the incremental induced steam generator tube rupture large early release fraction caused by an actual through-wall defect. This defect was responsible for the minor steam generator tube leak that occurred in September 2002 at the Comanche Peak Steam Electric Station Unit 1. In order to quantify the performance of the defect over the operating cycle, a range of defect lengths were input to the PROBFAIL computer code [Kenton, M., 2001. PROBFAIL: A Computer Code for Evaluating the Likelihood of Steam Generator Tube Rupture in Severe Nuclear Power Plant Accidents, CREARE TM-2138], using appropriate boundary conditions derived from MAAP4 [Henry, R., et al., May 1994. MAAP4—Modular Accident Analysis Program for LWR Power Plants, Computer Code Manual, EPRI Research Project 3131-02] runs. From the analysis of the calculated times of burst for each assumed defect length, the minimum through-wall defect length necessary for tube burst to occur prior to hot leg or surge line creep rupture was calculated. The probability that the defect would actually have this length was then estimated by determining the fraction of the cycle for which the defect would be at least that long. The methodology development and implementation relied on MAAP4 runs, which are discussed extensively in connection with their role in: (1) guiding the construction of the accident progression event tree, (2) generating relevant information for probability assignments in the various underlying fault trees and (3) obtaining boundary conditions of pressure and temperature for use in PROBFAIL. The overall increment in LERF due to the existence of the defect was calculated to be approximately 4E−08.     

New chemosensors based on thiomacrocycle-containing coumarin-343 fluoroionophor: X-ray structures and previous results on the effect of cation binding on the photophysical properties

Two new emissive chemosensors based on coumarin-343 have been synthesized, and their photophysical studies conducted. L1 contains an aza-thio macrocycle ring as the chelating unit, which has great affinity for soft metal ions, whereas L2 is a parent species without macrocyclic unit. Both compounds were separated by column chromatography and characterized by elemental analysis, ¹H,¹³C NMR, UV–vis and FAB mass spectroscopy. The X-ray structures of L1 and L2 and the supramolecular interactions in the solid state are discussed. Preliminary results on the metal–ion sensing effects of the ligands are presented. Titrations with Ag⁺, Pd²⁺ and Ni²⁺ have been studied by UV–vis absorption and fluorescence spectroscopy.     

The Use of Infrared Thermography as a Novel Approach for Real-Time Validation of PCR Thermocyclers

Validation of PCR thermocycler performance is crucial to obtain reliable results. In this study, infrared (IR) thermography was evaluated as a novel validation tool. After stabilisation, no significant difference in the temperatures recorded using thermography and a reference block-based system was found. By employing IR thermography, information about the length of the time until temperature stabilisation in the sample could be obtained. This study shows the potential of using IR thermography for validation of thermocyclers.