Nonparametric regression analysis achieved with topographic mapsdeveloped in combination with projection pursuit learning: anapplication to density estimation and adaptive filtering of grey-scaleimages

A novel approach to nonparametric regression analysis using topographic maps is proposed. The maps are trained with the extended maximum entropy learning rule (eMER) in combination with projection pursuit regression (PPR) learning. Rather than a single map, several maps are developed along optimally chosen projection directions in the input space. In this way, the regression performance improves in the case of sparsely sampled input spaces. We explore two applications of the eMER/PPR combination: (1) probability density estimation from pilot estimates and (2) adaptive filtering of grey-scale images. The first case is used as a testbed for comparing different, both classic and neural network-based, regression techniques. The results show that our eMER/PPR combination yields a superior regression performance for small data sets. In the second case, the regression model is trained on a noisy subimage. The model obtained after training reduces the noise content of the full image by more than 20 dB     

Embedded HIMOS(R) flash memory in 0.35 μm and 0.25 μm CMOStechnologies

In this paper, the performance and reliability characteristics of the 0.35 μm/0.25 μm High Injection MOS (HIMIOS(R)) technology is described in detail. This flash EEPROM technology relies on source-side injection for programming and Fowler-Nordheim tunneling for erasing, and has been successfully implemented in a 1 Mbit memory array embedded in a 0.35 μm CMOS technology, adding only about 30% to the processing cost of digital CMOS. Due to its triple gate structure, the HIMOS(R) cell exhibits a high degree of flexibility and scalability. A fast programming operation (10 μs) at 3.3 V supply voltage is combined with an endurance of well over 100000 program/erase cycles, immunity to all possible disturb effects and a retention time that largely exceeds 100 years at 125°C. Furthermore, the cell has been scaled to a 0.25 μm version, which is a laterally scaled version with the same operating voltages and tunnel oxide thickness. The use of secondary impact ionization is investigated as well and proves to be very promising for future generations when the supply voltage is scaled below 2.5 V     

Mixed analogue/digital phase picking algorithm in oversampling burst-mode clock phase alignment

A novel mixed analogue/digital design of a phase picking algorithm in an oversampling clock phase recovery is presented. The proposed approach results in reduced processing time, improved integrability with analogue front-end and low noise generation. Simulations of a 10 Gbit/s burst-mode clock phase alignment circuit in a 0.25 m SiGe BiCMOS process, show a simulated processing delay of only 280 ps.     

Friction Compensation of an $XY$ Feed Table Using Friction-Model-Based Feedforward and an Inverse-Model-Based Disturbance Observer

Uncompensated friction forces compromise the positioning and tracking accuracy of motion systems. A unique tracking error known as quadrant glitch is the result of complex nonlinear friction behavior at motion reversal or near-zero velocity. Linear-feedback control strategies such as PID, cascade P/PI, or state-feedback control have to be extended with model- and nonmodel-based friction-compensation strategies to acquire sufficiently high path and tracking accuracy. This paper analyzes and validates experimentally three different friction-compensation strategies for a linear motor-based xy feed drive of a high-speed milling machine: (1) friction-model-based feedforward; (2) an inverse-model-based disturbance observer; and (3) the combination of both techniques. The friction models considered are as follows: a simple static-friction model and the recently developed generalized Maxwell-slip (GMS) model. GMS friction-model-based feedforward combined with disturbance observer almost completely eliminates the radial tracking error and quadrant glitches.     

Routing in an AWG-Based Optical Packet Switch

For the next generation of the optical internet, focus is now moving from circuit switched networks, which occupy a wavelength continuously regardless of the demand at that time, towards optical packet/burst switching. By only occupying a wavelength when data is to be transmitted, a more efficient utilization of bandwidth in optical fibers is strived for. As bandwidth in fibers keeps increasing, the bottleneck of the optical network is now moving towards the switching node, since evolution of electronic routers cannot follow the speed of bandwidth increase. Thus a key component in these novel networks is the optical node. Through this node we want to switch traffic very fast and reliably, preferably transparent. Lack of efficient and practically realizable optical buffer, however, makes migration from electronic routers to optical routers a non-straightforward transition. In most optical nodes payload traffic can be switched transparently, whilst control information (e.g., in a header, on a control channel) is still converted to the electronic domain in every node, since optical processing is far from mature. In this paper we present a possible architecture for such a node, array waveguide gratings and all-optical tuneable wavelength converters. The concept of this switch is explained and the node is evaluated in terms of loss rate. We will see that an inherent problem of this switch is its internal blocking. This drawback can be greatly overcome by using an intelligent and efficient wavelength assignment algorithm within the node. Simulation of slotted operation will give some numerical results.     

Method R estimates of heritability for milk, fat, and protein yields of United States dairy cattle.

Heritabilities for milk, fat, and protein yields were estimated from first lactation data used for USDA-Dairy Herd Improvement Association (DHIA) genetic evaluations. Contemporary group assignments and standard deviations within herd-year were determined with the procedure used for national evaluations. Pedigree data were included for animals born since 1970; yield data were included for cows born since 1980. Lactation records were divided into four mutually exclusive data sets based on standard deviations. Ranges for standard deviations were chosen so that data sets were approximately equal in size. Method R was used to estimate heritability with 25 different random samples of half of the data for each data set. Because of the large number of Holstein observations, estimates of heritability for Holsteins were based on random subsets of the complete data file; each subset included approximately 5% of the data. Mean heritability estimates increased with standard deviations, and estimates ranged from 0.18 to 0.51 across breeds. Repeatability estimates for milk yield of Holsteins were approximately 0.50 and did not change with standard deviation. These heritability estimates were higher than those previously used in the USDA-DHIA genetic evaluation. Heritability used in the USDA-DHIA genetic evaluation have been increased based on these results.     

Introducing optimal experimental design in predictive modeling: a motivating example

Predictive microbiology emerges more and more as a rational quantitative framework for predicting and understanding microbial evolution in food products. During the mathematical modeling of microbial growth and/or inactivation, great, but not always efficient, effort is spent on the determination of the model parameters from experimental data. In order to optimize experimental conditions with respect to parameter estimation, experimental design has been extensively studied since the 1980s in the field of bioreactor engineering. The so-called methodology of optimal experimental design established in this research area enabled the reliable estimation of model parameters from data collected in well-designed fed-batch reactor experiments. In this paper, we introduce the optimal experimental design methodology for parameter estimation in the field of predictive microbiology. This study points out that optimal design of dynamic input signals is necessary to maximize the information content contained within the resulting experimental data. It is shown that from few dynamic experiments, more pertinent information can be extracted than from the classical static experiments. By introducing optimal experimental design into the field of predictive microbiology, a new promising frame for maximization of the information content of experimental data with respect to parameter estimation is provided. As a case study, the design of an optimal temperature profile for estimation of the parameters D(ref) and z of an Arrhenius-type model for the maximum inactivation rate kmax as a function of the temperature, T, was considered. Microbial inactivation by heating is described using the model of Geeraerd et al. (1999). The need for dynamic temperature profiles in experiments aimed at the simultaneous estimation of the model parameters from measurements of the microbial population density is clearly illustrated by analytical elaboration of the mathematical expressions involved on the one hand, and by numerical simulations on the other.     

Restoration of signal polarity in a set of inversion recovery NMR images

A method is described which enables unambiguous retrieval of sign information in a set of magnetic resonance magnitude images of the inversion recovery type. The proposed method starts from the observation that the inversion recovery curve S is a monotonically increasing function of the inversion time TI, and comes down to finding the zero-crossing time TI0 of this curve for each pixel within the image. Absolute Value(S)and S are then related by S(TI) = -Absolute Value(S(TI)) for TI =/< TI(0) and S(TI) = + Absolute Value(S(TI)) for TI =/> TI0. The method, which does not require additional knowledge with respect to any of the NMR parameters involved, is shown to be effective when at least four inversion recovery images with different inversion times of a particular object slice are available. The efficacy of sign retrieval is demonstrated by imaging experiments on phantoms and human subjects. The validity of the polarity restoration method is established by viewing its results against the results of conventional methods, i.e., NMR spectroscopy.     

A Cross-Layer Approach to Enhance QoS for Multimedia Applications Over Satellite

The need for on-demand QoS support for communications over satellite is of primary importance for distributed multimedia applications. This is particularly true for the return link which is often a bottleneck due to the large set of end-users accessing a very limited uplink resource. Facing this need, Demand Assignment Multiple Access (DAMA) is a classical technique that allows satellite operators to offer various types of services, while managing the resources of the satellite system efficiently. Tackling the quality degradation and delay accumulation issues that can result from the use of these techniques, this paper proposes an instantiation of the Application Layer Framing (ALF) approach, using a cross-layer interpreter (xQoS-Interpreter). The information provided by this interpreter is used to manage the resource provided to a terminal by the satellite system in order to improve the quality of multimedia presentations from the end user’s point of view. Several experiments are carried out for different loads on the return link. Their impact on QoS is measured through different application as well as network level metrics.