Structured neural network modelling of multi-valued functions for wind retrieval from scatterometer measurements

A conventional neural network approach to regression problems approximates the conditional mean of the output vector. For mappings which are multi-valued this approach breaks down, since the average of two solutions is not necessarily a valid solution. In this article mixture density networks, a principled method to model conditional probability density functions, are applied to retrieving Cartesian wind vector components from satellite scatterometer data. A hybrid mixture density network is implemented to incorporate prior knowledge of the predominantly bimodal function branches. An advantage of a fully probabilistic model is that more sophisticated and principled methods can be used to resolve ambiguities.

Structured neural network modelling of multi-valued functions for wind retrieval from scatterometer measurements

A conventional neural network approach to regression problems approximates the conditional mean of the output vector. For mappings which are multi-valued this approach breaks down, since the average of two solutions is not necessarily a valid solution. In this article mixture density networks, a principled method to model conditional probability density functions, are applied to retrieving Cartesian wind vector components from satellite scatterometer data. A hybrid mixture density network is implemented to incorporate prior knowledge of the predominantly bimodal function branches. An advantage of a fully probabilistic model is that more sophisticated and principled methods can be used to resolve ambiguities.

On the robustness of random Boolean formulae

Random Boolean formulae, generated by a growth process of noisy logical gates are analyzed using the generating functional methodology of statistical physics. We study the type of functions generated for different input distributions, their robustness for a given level of gate error and its dependence on the formulae depth and complexity and the gates used. Bounds on their performance, derived in the information theory literature for specific gates, are straightforwardly retrieved, generalized and identified as the corresponding typical-case phase transitions. Results for error-rates, function-depth and sensitivity of the generated functions are obtained for various gate-type and noise models. © 2010 IOP Publishing Ltd.

Calculating the derivative of piecewise functions

Exercises involving the calculation of the derivative of piecewise defined functions are common in calculus, with the aim of consolidating beginners’ knowledge of applying the definition of the derivative. In such exercises, the piecewise function is commonly made up of two smooth pieces joined together at one point. A strategy which avoids using the definition of the derivative is to find the derivative function of each smooth piece and check whether these functions agree at the chosen point. Showing that this strategy works together with investigating discontinuities of the derivative is usually beyond a calculus course. However, we shall show that elementary arguments can be used to clarify the calculation and behaviour of the derivative for piecewise functions.

Visual signs and symptoms of Parkinson's disease

Parkinson's disease (PD) is a common disorder of middle-aged and elderly people, in which there is degeneration of the extra-pyramidal motor system. In some patients, the disease is associated with a range of visual signs and symptoms, including defects in visual acuity, colour vision, the blink reflex, pupil reactivity, saccadic and smooth pursuit movements and visual evoked potentials. In addition, there may be psychophysical changes, disturbances of complex visual functions such as visuospatial orientation and facial recognition, and chronic visual hallucinations. Some of the treatments associated with PD may have adverse ocular reactions. If visual problems are present, they can have an important effect on overall motor function, and quality of life of patients can be improved by accurate diagnosis and correction of such defects. Moreover, visual testing is useful in separating PD from other movement disorders with visual symptoms, such as dementia with Lewy bodies (DLB), multiple system atrophy (MSA) and progressive supranuclear palsy (PSP). Although not central to PD, visual signs and symptoms can be an important though obscure aspect of the disease and should not be overlooked.

Monetary models of exchange rates and sweep programs

Numerous studies find that monetary models of exchange rates cannot beat a random walk model. Such a finding, however, is not surprising given that such models are built upon money demand functions and traditional money demand functions appear to have broken down in many developed countries. In this article, we investigate whether using a more stable underlying money demand function results in improvements in forecasts of monetary models of exchange rates. More specifically, we use a sweep-adjusted measure of US monetary aggregate M1 which has been shown to have a more stable money demand function than the official M1 measure. The results suggest that the monetary models of exchange rates contain information about future movements of exchange rates, but the success of such models depends on the stability of money demand functions and the specifications of the models.

Interocular masking and summation indicate two stages of divisive contrast gain control

Our understanding of early spatial vision owes much to contrast masking and summation paradigms. In particular, the deep region of facilitation at low mask contrasts is thought to indicate a rapidly accelerating contrast transducer (eg a square-law or greater). In experiment 1, we tapped an early stage of this process by measuring monocular and binocular thresholds for patches of 1 cycle deg-1 sine-wave grating. Threshold ratios were around 1.7, implying a nearly linear transducer with an exponent around 1.3. With this form of transducer, two previous models (Legge, 1984 Vision Research 24 385 - 394; Meese et al, 2004 Perception 33 Supplement, 41) failed to fit the monocular, binocular, and dichoptic masking functions measured in experiment 2. However, a new model with two-stages of divisive gain control fits the data very well. Stage 1 incorporates nearly linear monocular transducers (to account for the high level of binocular summation and slight dichoptic facilitation), and monocular and interocular suppression (to fit the profound 42 Oral presentations: Spatial vision Thursday dichoptic masking). Stage 2 incorporates steeply accelerating transduction (to fit the deep regions of monocular and binocular facilitation), and binocular summation and suppression (to fit the monocular and binocular masking). With all model parameters fixed from the discrimination thresholds, we examined the slopes of the psychometric functions. The monocular and binocular slopes were steep (Weibull ߘ3-4) at very low mask contrasts and shallow (ߘ1.2) at all higher contrasts, as predicted by all three models. The dichoptic slopes were steep (ߘ3-4) at very low contrasts, and very steep (ß>5.5) at high contrasts (confirming Meese et al, loco cit.). A crucial new result was that intermediate dichoptic mask contrasts produced shallow slopes (ߘ2). Only the two-stage model predicted the observed pattern of slope variation, so providing good empirical support for a two-stage process of binocular contrast transduction. [Supported by EPSRC GR/S74515/01]

Neurophysiology of CSWS-associated cognitive dysfunction

The phenomenon of continuous spikes and waves during slow-wave sleep (CSWS) is associated with a number of epileptic syndromes, which share a behavioral phenotype characterized by deterioration of cognitive, behavioral, or sensorimotor functions. Available evidence seems to suggest that spike-wave activity is a result of a complex interaction between cortical and subcortical inhibitory networks and can "per se" produce a transient loss of underlying cortical functions. Syndromes like Landau-Kleffner syndrome, CSWS, and phenomena such as negative myoclonus could share in common--at least at the neurophysiological level--some similarities. Differences in behavioral phenotypes could be explained in term of maturational and genetic differences, as well as by the functional specificity of the involved areas.

Experimental and theoretical characterisation of tunable fibre gratings

This thesis presents details on both theoretical and experimental aspects of UV written fibre gratings. The main body of the thesis deals with the design, fabrication and testing of telecommunication optical fibre grating devices, but also an accurate theoretical analysis of intra-core fibre gratings is presented. Since more than a decade, fibre gratings have been extensively used in the telecommunication field (as filters, dispersion compensators, and add/drop multiplexers for instance). Gratings for telecommunication should conform to very high fabrication standards as the presence of any imperfection raises the noise level in the transmission system compromising its ability of transmitting intelligible sequence of bits to the receiver. Strong side lobes suppression and high and sharp reflection profile are then necessary characteristics. A fundamental part of the theoretical and experimental work reported in this thesis is about apodisation. The physical principle of apodisation is introduced and a number of apodisation techniques, experimental results and numerical optimisation of the shading functions and all the practical parameters involved in the fabrication are detailed. The measurement of chromatic dispersion in fibres and FBGs is detailed and an estimation of its accuracy is given. An overview on the possible methods that can be implemented for the fabrication of tunable fibre gratings is given before detailing a new dispersion compensator device based on the action of a distributed strain onto a linearly chirped FBG. It is shown that tuning of second and third order dispersion of the grating can be obtained by the use of a specially designed multipoint bending rig. Experiments on the recompression of optical pulses travelling long distances are detailed for 10 Gb/s and 40 Gb/s. The characterisation of a new kind of double section LPG fabricated on a metal-clad coated fibre is reported. The fabrication of the device is made easier by directly writing the grating through the metal coating. This device may be used to overcome the recoating problems associated with standard LPGs written in step-index fibre. Also, it can be used as a sensor for simultaneous measurements of temperature and surrounding medium refractive index.

Simulation and control of a stepping motor

Open-loop operatlon of the stepping motor exploits the inherent advantages of the machine. For near optimum operation: in this mode, however, an accurate system model is required to facilitate controller design. Such a model must be comprehensive and take account of the non-linearities inherent in the system. The result is a complex formulation which can be made manageable with a computational aid. A digital simulation of a hybrid type stepping motor and its associated drive circuit is proposed. The simulation is based upon a block diagram model which includes reasonable approximations to the major non-linearities. The simulation is shown to yield accurate performance predictions. The determination of the transfer functions is based upon the consideration of the physical processes involved rather than upon direct input-outout measurements. The effects of eddy currents, saturation, hysteresis, drive circuit characteristics and non-linear torque displacement characteristics are considered and methods of determining transfer functions, which take account of these effects, are offered. The static torque displacement characteristic is considered in detail and a model is proposed which predicts static torque for any combination of phase currents and shaft position. Methods of predicting the characteristic directly from machine geometry are investigated. Drive circuit design for high efficiency operation is considered and a model of a bipolar, bilevel circuit is proposed. The transfers between stator voltage and stator current and between stator current and air gap flux are complicated by the effects of eddy currents, saturation and hysteresis. Frequency response methods, combined with average inductance measurements, are shown to yield reasonable transfer functions. The modelling procedure and subsequent digital simulation is concluded to be a powerful method of non-linear analysis.

Identification of nonlinear systems by correlation using pseudorandom signals

This thesis is concerned with the measurement of the characteristics of nonlinear systems by crosscorrelation, using pseudorandom input signals based on m sequences. The systems are characterised by Volterra series, and analytical expressions relating the rth order Volterra kernel to r-dimensional crosscorrelation measurements are derived. It is shown that the two-dimensional crosscorrelation measurements are related to the corresponding second order kernel values by a set of equations which may be structured into a number of independent subsets. The m sequence properties determine how the maximum order of the subsets for off-diagonal values is related to the upper bound of the arguments for nonzero kernel values. The upper bound of the arguments is used as a performance index, and the performance of antisymmetric pseudorandom binary, ternary and quinary signals is investigated. The performance indices obtained above are small in relation to the periods of the corresponding signals. To achieve higher performance with ternary signals, a method is proposed for combining the estimates of the second order kernel values so that the effects of some of the undesirable nonzero values in the fourth order autocorrelation function of the input signal are removed. The identification of the dynamics of two-input, single-output systems with multiplicative nonlinearity is investigated. It is shown that the characteristics of such a system may be determined by crosscorrelation experiments using phase-shifted versions of a common signal as inputs. The effects of nonlinearities on the estimates of system weighting functions obtained by crosscorrelation are also investigated. Results obtained by correlation testing of an industrial process are presented, and the differences between theoretical and experimental results discussed for this case;

Inference and optimization of real edges on sparse graphs:A statistical physics perspective

Inference and optimization of real-value edge variables in sparse graphs are studied using the Bethe approximation and replica method of statistical physics. Equilibrium states of general energy functions involving a large set of real edge variables that interact at the network nodes are obtained in various cases. When applied to the representative problem of network resource allocation, efficient distributed algorithms are also devised. Scaling properties with respect to the network connectivity and the resource availability are found, and links to probabilistic Bayesian approximation methods are established. Different cost measures are considered and algorithmic solutions in the various cases are devised and examined numerically. Simulation results are in full agreement with the theory. © 2007 The American Physical Society.

Resonance studies of artificial earth satellites

Orbit determination from artificial satellite observations is a key process in obtaining information about the Earth and its environment. A study of the perturbations experienced by these satellites enables knowledge to be gained of the upper atmosphere, the gravity field, ocean tides, solid-Earth tides and solar radiation. The gravity field is expressed as a double infinite series of associated Legendre functions (tesseral harmonics). In contemporary global gravity field models the overall geoid is well determined. An independent check on these gravity field harmonics of a particular order may be made by analysis of satellites that pass through resonance of that order. For such satellites the perturbations of the orbital elements close to resonance are analysed to derive lumped harmonic coefficients. The orbital parameters of 1984-106A have been determined at 43 epochs, during which time the satellite was close to 14th order resonance. Analysis of the inclination and eccentricity yielded 6 lumped harmonic coefficients of order 14 whilst analysis of the mean motion yielded additional pairs of lumped harmonics of orders 14, 28 and 42, with the 14th order harmonics superseding those obtained from analysis of the inclination. This thesis concentrates in detail on the theoretical changes of a near-circular satellite orbit perturbed by the Earth's gravity field under the influence of minimal air-drag whilst in resonance with the Earth. The satellite 1984-106A experienced the interesting property of being temporarily trapped with respect to a secondary resonance parameter due to the low air-drag in 1987. This prompted the theoretical investigation of such a phenomenon. Expressions obtained for the resonance parameter led to the determination of 8 lumped harmonic coefficients, coincidental to those already obtained. All the derived lumped harmonic values arc used to test the accuracy of contemporary gravity field models and the underlying theory in this thesis.

Applications of Reissner's principle to structural dynamics

The analysis and prediction of the dynamic behaviour of s7ructural components plays an important role in modern engineering design. :n this work, the so-called "mixed" finite element models based on Reissnen's variational principle are applied to the solution of free and forced vibration problems, for beam and :late structures. The mixed beam models are obtained by using elements of various shape functions ranging from simple linear to complex cubic and quadratic functions. The elements were in general capable of predicting the natural frequencies and dynamic responses with good accuracy. An isoparametric quadrilateral element with 8-nodes was developed for application to thin plate problems. The element has 32 degrees of freedom (one deflection, two bending and one twisting moment per node) which is suitable for discretization of plates with arbitrary geometry. A linear isoparametric element and two non-conforming displacement elements (4-node and 8-node quadrilateral) were extended to the solution of dynamic problems. An auto-mesh generation program was used to facilitate the preparation of input data required by the 8-node quadrilateral elements of mixed and displacement type. Numerical examples were solved using both the mixed beam and plate elements for predicting a structure's natural frequencies and dynamic response to a variety of forcing functions. The solutions were compared with the available analytical and displacement model solutions. The mixed elements developed have been found to have significant advantages over the conventional displacement elements in the solution of plate type problems. A dramatic saving in computational time is possible without any loss in solution accuracy. With beam type problems, there appears to be no significant advantages in using mixed models.