Forward Propagation Model Research Articles

Synergistic Use of Artificial Neural Network for the Detection of Underground Coal Fires

ABSTRACTThe spontaneous heating of coal is responsible for releasing a number of combustible gases at characteristic temperatures. However, its accurate prediction is essential to plan efficient preventive measures for impending underground coal fires. For this purpose, artificial neural network can be used as a potential tool to precisely envisage the temperature on the basis of the concentration of various gases released during thermal gas analysis of coal samples. Keeping this in view, a three-layer feed forward back propagation model has been developed. The correlation among the gases released during the thermal gas analysis along with temperature has also been established by developing a dendrogram using statistical software. The results of the neural network model have been conformed to the hierarchy of the gas appearance taking place during thermal gas analysis.

Estimating refractivity from propagation loss in turbulent media

AbstractThis paper estimates lower atmospheric refractivity (M‐profile) given an electromagnetic (EM) propagation loss (PL) measurement. Specifically, height‐independent PL measurements over a range of 10–80 km are used to infer information about the existence and potential parameters of atmospheric ducts in the lowest 1 km of the atmosphere. The main improvement made on previous refractivity estimations is inclusion of range‐dependent fluctuations due to turbulence in the forward propagation model. Using this framework, the maximum likelihood (ML) estimate of atmospheric refractivity has good accuracy, and with prior information about ducting the maximum a priori (MAP) refractivity estimate can be found. Monte Carlo methods are used to estimate the mean and covariance of PL, which are fed into a Gaussian likelihood function for evaluation of estimated refractivity probability. Comparisons were made between inversions performed on propagation loss data simulated by a wide angle parabolic equation (PE) propagation model with added homogeneous and inhomogeneous turbulence. It was found that the turbulence models produce significantly different results, suggesting that accurate modeling of turbulence is key.

Posterior inference of two closely spaced wave vectors for inference regarding the state of an acoustic target

A bistatic sonar arrangement is employed to infer the depth, speed, and range of an oncoming submerged acoustic target from a continuous wave transmission. Surface and direct interacting paths associated with closely spaced angles and frequencies from the returned acoustic wave fronts are resolved. A Gibbs sampler is employed to construct the posterior joint density of all parameters taking full advantage of the analytic tractability of the amplitude and phase of the wave fronts while the ordered wave vectors, their angles and frequencies are sampled numerically by two-dimensional quantile function. The inferred density of target depth, range, and speed is accomplished by constructing a numerical inverse-transformation of the forward propagation model. Simulations are used to demonstrate the potential of sub-Rayleigh resolution accuracy at less than -10dB received SNR with an 9 element vertical array.

Optimal scheduling of electrical power in energy-deficient scenarios using artificial neural network and Bootstrap aggregating

In a developing country like Pakistan, where the electrical power demand is more than the generated power, maintaining the power system stability is a big challenge. In such cases it becomes, thus, essential to shed just the right amount of load to keep a power system stable. This paper presents a case study of Pakistan’s power system where the generated power, the load demand, frequency deviation and the load shed during a 24-h duration have been provided. The data have been analyzed using two techniques; the conventional artificial neural network (ANN) by implementing feed forward back propagation model and the Bootstrap aggregating or bagging algorithm. The simulation results reveal the superiority of the Bootstrap aggregating algorithm over a conventional ANN technique using feed forward back propagation model.

Bayes Factor test for the discrimination of a submerged mobile object from a continuous wave transmission

Detection of an underwater mobile object by means of a continuous wave transmission is a challenging problem in part due to the difficulty of making inferences on closely spaced tones that provide clues regarding the objects' depth and speed. Considered here is a bistatic sonar arrangement to detect the presence of an oncoming submerged object. The optimal receiver computes the Bayes factor associated with the hypothesis test. The analytic intractability of the marginalization associated with the composite nature of the hypothesis leads to numerical methods of integration. The prior density on the target present scenario is constructed by an inverse image transformation through the forward propagation model. Computation of the Bayes factor is accomplished by Markov-chain Monte Carlo sampling with a Gibbs sampler. Analytically tractable conditional and marginal densities of the tone amplitudes are exploited while the conditional density of the ordered frequencies are constructed numerically by an inverse quantile sampler. Detection performance of a receiver against a mobile target are illustrated and discussed.

Inferences on target speed, depth, and range from a continuous wave transmission

Tracking an underwater mobile object by means of a continuous wave transmission is challenging in part due to the difficulty of drawing inferences on closely spaced tones associated with target depth and speed. Considered here is a bistatic sonar arrangement employed to infer the depth, speed and range of an oncoming submerged object. Computation of the full posterior probability distribution of the returned amplitudes and frequencies from both the prior distribution and a finite duration window of the received waveform is made by Markov-chain Monte Carlo sampling. A Gibbs sampler is employed to construct the posterior joint density of all parameters by taking full advantage of the analytic tractability of the conditional and marginal densities of the received amplitudes while those of the ordered frequencies are constructed numerically by either an inverse quantile sampling or a Metropolis-Hastings sampler. The inferred density of depth, range, and speed of the target is accomplished by constructing a numerical inverse-transformation of the forward propagation model.

A novel framework for feature extraction in multi-sensor action potential sorting

BackgroundExtracellular recordings of multi-unit neural activity have become indispensable in neuroscience research. The analysis of the recordings begins with the detection of the action potentials (APs), followed by a classification step where each AP is associated with a given neural source. A feature extraction step is required prior to classification in order to reduce the dimensionality of the data and the impact of noise, allowing source clustering algorithms to work more efficiently. New methodIn this paper, we propose a novel framework for multi-sensor AP feature extraction based on the so-called Matched Subspace Detector (MSD), which is shown to be a natural generalization of standard single-sensor algorithms. ResultsClustering using both simulated data and real AP recordings taken in the locust antennal lobe demonstrates that the proposed approach yields features that are discriminatory and lead to promising results. Comparison with existing method(s)Unlike existing methods, the proposed algorithm finds joint spatio-temporal feature vectors that match the dominant subspace observed in the two-dimensional data without needs for a forward propagation model and AP templates. ConclusionsThe proposed MSD approach provides more discriminatory features for unsupervised AP sorting applications.

Using the General Link Transmission Model in a Dynamic Traffic Assignment to Simulate Congestion on Urban Networks

This article presents two new models of Dynamic User Equilibrium that are particularly suited for ITS applications, where the evolution of vehicle flows and travel times must be simulated on large road networks, possibly in real-time. The key feature of the proposed models is the detail representation of the main congestion phenomena occurring at nodes of urban networks, such as vehicle queues and their spillback, as well as flow conflicts in mergins and diversions. Compared to the simple word of static assignment, where only the congestion along the arc is typically reproduced through a separable relation between vehicle flow and travel time, this type of DTA models are much more complex, as the above relation becomes non-separable, both in time and space.Traffic simulation is here attained through a macroscopic flow model, that extends the theory of kinematic waves to urban networks and non-linear fundamental diagrams: the General Link Transmission Model. The sub-models of the GLTM, namely the Node Intersection Model, the Forward Propagation Model of vehicles and the Backward Propagation Model of spaces, can be combined in two different ways to produce arc travel times starting from turn flows. The first approach is to consider short time intervals of a few seconds and process all nodes for each temporal layer in chronological order. The second approach allows to consider long time intervals of a few minutes and for each sub-model requires to process the whole temporal profile of involved variables. The two resulting DTA models are here analyzed and compared with the aim of identifying their possible use cases.A rigorous mathematical formulation is out of the scope of this paper, as well as a detailed explanation of the solution algorithm.The dynamic equilibrium is anyhow sought through a new method based on Gradient Projection, which is capable to solve both proposed models with any desired precision in a reasonable number of iterations. Its fast convergence is essential to show that the two proposed models for network congestion actually converge at equilibrium to nearly identical solutions in terms of arc flows and travel times, despite their two diametrical approaches wrt the dynamic nature of the problem, as shown in the numerical tests presented here.

Adaptive Bayesian Inference for Discontinuous Inverse Problems, Application to Hyperbolic Conservation Laws

AbstractVarious works from the literature aimed at accelerating Bayesian inference in inverse problems. Stochastic spectral methods have been recently proposed as surrogate approximations of the forward uncertainty propagation model over the support of the prior distribution. These representations are efficient because they allow affordable simulation of a large number of samples from the posterior distribution. Unfortunately, they do not perform well when the forward model exhibits strong nonlinear behavior with respect to its input.In this work, we first relate the fast (exponential) L2-convergence of the forward approximation to the fast (exponential) convergence (in terms of Kullback-Leibler divergence) of the approximate posterior. In particular, we prove that in case the prior distribution is uniform, the posterior is at least twice as fast as the convergence rate of the forwardmodel in those norms. The Bayesian inference strategy is developed in the framework of a stochastic spectral projectionmethod. The predicted convergence rates are then demonstrated for simple nonlinear inverse problems of varying smoothness.We then propose an efficient numerical approach for the Bayesian solution of inverse problems presenting strongly nonlinear or discontinuous systemresponses. This comes with the improvement of the forward model that is adaptively approximated by an iterative generalized Polynomial Chaos-based representation. The numerical approximations and predicted convergence rates of the former approach are compared to the new iterative numericalmethod for nonlinear time-dependent test cases of varying dimension and complexity, which are relevant regarding our hydrodynamics motivations and therefore regarding hyperbolic conservation laws and the apparition of discontinuities in finite time.

RETRIEVING EVAPORATION DUCT HEIGHTS FROM POWER OF GROUND-BASED GPS OCCULTATION SIGNAL

Evaporation duct is a manifest phenomenon that can afiect microwave propagation seriously with low elevation angles near the sea surface. As an important parameter to describe the characteristic of the evaporation duct, duct height can be estimated from radar sea echo using a technique called as \refractivity from clutter. In this study, we proposed a novel approach to estimating the evaporation duct height. The signal power received by a ground- based GPS receiver is used when the GPS satellites rise or set at the local horizon over the sea. A forward propagation model and genetic algorithm are adopted to implement this method. The performance is evaluated via numerical simulation for inferring evaporation duct with difierent height. The results showed that the proposed method is well efiective, especially for the conditions with higher evaporation duct height.

Propagation modeling of ocean-scattered low-elevation GPS signals for maritime tropospheric duct inversion

The maritime tropospheric duct is a low-altitude anomalous refractivity structure over the ocean surface, and it can significantly affect the performance of many shore-based/shipboard radar and communication systems. We propose the idea that maritime tropospheric ducts can be retrieved from ocean forward-scattered low-elevation global positioning system (GPS) signals. Retrieval is accomplished by matching the measured power patterns of the signals to those predicted by the forward propagation model as a function of the modified refractivity profile. On the basis of a parabolic equation method and bistatic radar equation, we develop such a forward model for computing the trapped propagation characteristics of an ocean forward-scattered GPS signal within a tropospheric duct. A new GPS scattering initial field is defined for this model to start the propagation modeling. A preliminary test on the performance of this model is conducted using measured data obtained from a 2009-experiment in the South China Sea. Results demonstrate that this model can predict GPS propagation characteristics within maritime tropospheric ducts and serve as a forward model for duct inversion.

Performance dependence of hybrid x-ray computed tomography/fluorescence molecular tomography on the optical forward problem

Hybrid imaging systems combining x-ray computed tomography (CT) and fluorescence tomography can improve fluorescence imaging performance by incorporating anatomical x-ray CT information into the optical inversion problem. While the use of image priors has been investigated in the past, little is known about the optimal use of forward photon propagation models in hybrid optical systems. In this paper, we explore the impact on reconstruction accuracy of the use of propagation models of varying complexity, specifically in the context of these hybrid imaging systems where significant structural information is known a priori. Our results demonstrate that the use of generically known parameters provides near optimal performance, even when parameter mismatch remains.

The wavenumber algorithm for full-matrix imaging using an ultrasonic array

Ultrasonic imaging using full-matrix capture, e.g., via the total focusing method (TFM), has been shown to increase angular inspection coverage and improve sensitivity to small defects in nondestructive evaluation. In this paper, we develop a Fourier-domain approach to full-matrix imaging based on the wavenumber algorithm used in synthetic aperture radar and sonar. The extension to the wavenumber algorithm for full-matrix data is described and the performance of the new algorithm compared with the TFM, which we use as a representative benchmark for the time-domain algorithms. The wavenumber algorithm provides a mathematically rigorous solution to the inverse problem for the assumed forward wave propagation model, whereas the TFM employs heuristic delay-and-sum beamforming. Consequently, the wavenumber algorithm has an improved point-spread function and provides better imagery. However, the major advantage of the wavenumber algorithm is its superior computational performance. For large arrays and images, the wavenumber algorithm is several orders of magnitude faster than the TFM. On the other hand, the key advantage of the TFM is its flexibility. The wavenumber algorithm requires a regularly sampled linear array, while the TFM can handle arbitrary imaging geometries. The TFM and the wavenumber algorithm are compared using simulated and experimental data.

Geoacoustic inversion of short range source data using a plane wave reflection coefficient approach

Acoustic time series data were collected in a shallow, hard bottom lake environment located in central Texas using both short range (2 m) implosive data, obtained with the source and a single hydrophone located near mid-depth in the waveguide, along with longer range implosive and explosive data from a near surface source to a bottom mounted hydrophone. Matched field inversions using simulated annealing were performed with a ray trace plus complex plane wave reflection coefficient forward propagation model that was validated in previous work. Isolating bottom interacting paths to perform the inversions is shown to be essential to reduce parameter uncertainties in the hard bottom environment and enables a systematic approach to the inversions which establishes the number of layers needed to represent the lake environment. Measured transmission loss data from a towed source were compared through a RMS error analysis to modeled transmission loss, constructed with the parameters from inversions of data from several source types, to further establish the validity of the inversion approach for this environment. Geoacoustic parameters obtained by inversions of short range, low frequency impulsive data are used to predict transmission loss at longer ranges and higher frequencies. The range dependence of the global minimum is discussed.

Determination of the optical properties of tissue-simulating phantoms from interstitial frequency domain measurements of relative fluence and phase difference

We estimated the absorption and reduced scattering coefficients of tissue-simulating phantoms from interstitial measurements of the phase difference and relative amplitude signals at two distances from a sinusoidally modulated isotropic source. It was found that absorption and reduced scattering coefficients can be recovered within 10% and slightly over 10% respectively, using either the data collected by two detectors 3mm apart or by two detectors 5mm apart with light collected by one detector attenuated by a neutral density filter. This accuracy was achieved over a wide range of optical properties, mu(a)=0.008 to 0.17mm(-1) and mu(s)'=0.3 to 1.8mm(-1). Additional factors affecting accuracy including source anisotropy, uncertainty in fiber placement, phase amplitude crosstalk, and the forward light propagation model (the combined isotropic similarity model and standard diffusion approximation versus the modified spherical harmonics method) were studied by Monte Carlo simulations (first two factors) and experiments (last two factors).

Probability density function predictions of acoustic uncertainty in an underwater sound channel

The utility of forward acoustic propagation models is often limited by incomplete knowledge of environmental parameters. For this reason, information about the resulting uncertainty of acoustic field calculations is a valuable addition to the solutions generated by current propagation routines. However, extensive Monte Carlo simulations are presently the only means for quantifying this uncertainty. This presentation explores alternative methods for obtaining a complete description of the uncertainty of a time-harmonic acoustic field solution in the form of a probability density function (PDF). The isobaric contours of the field solution, including the uncertain parameters as independent variables, produce the relationship between acoustic pressure and the uncertain variables. In an ocean waveguide, the isobaric contours can be obtained approximately from a modal sum Green’s function using techniques similar to those that yield the waveguide invariant. Thus, information from a single field calculation can be used to transform PDFs for the uncertain environmental parameters into an uncertainty distribution for the calculated field solution. Results are shown for an ideal underwater sound channel and compared to results from Monte Carlo simulations. [Work sponsored by the ONR, Code 321.]

Adjoint modeling for acoustic inversion.

The use of adjoint modeling for acoustic inversion is investigated. An adjoint model is derived from a linearized forward propagation model to propagate data-model misfit at the observation points back through the medium to the medium perturbations not being accounted for in the model. This adjoint model can be used to aid in inverting for these unaccounted medium perturbations. Adjoint methods are being applied to a variety of inversion problems, but have not drawn much attention from the underwater acoustic community. This paper presents an application of adjoint methods to acoustic inversion. Inversions are demonstrated in simulation for both range-independent and range-dependent sound speed profiles using the adjoint of a parabolic equation model. Sensitivity and error analyses are discussed showing how the adjoint model enables calculations to be performed in the space of observations, rather than the often much larger space of model parameters. Using an adjoint model enables directions of steepest descent in the model parameters (what we invert for) to be calculated using far fewer modeling runs than if a forward model only were used.

A neural network controller for hydronic heating systems of solar buildings

An artificial neural network (ANN)-based controller for hydronic heating plants of buildings is presented. The controller has forecasting capabilities: it includes a meteorological module, forecasting the ambient temperature and solar irradiance, an indoor temperature predictor module, a supply temperature predictor module and an optimizing module for the water supply temperature. All ANN modules are based on the Feed Forward Back Propagation (FFBP) model. The operation of the controller has been tested experimentally, on a real-scale office building during real operating conditions. The operation results were compared to those of a conventional controller. The performance was also assessed via numerical simulation. The detailed thermal simulation tool for solar systems and buildings TRNSYS was used. Both experimental and numerical results showed that the expected percentage of energy savings with respect to a conventional controller is of about 15% under North European weather conditions.

Rapid geoacoustic inversion with a curved horizontal array

Real-time geoacoustic inversions with a towed array are of interest for rapidly characterizing the sediment properties over changing regions, and require an efficient and accurate forward propagation model. A wavenumber integration solution using the computationally efficient method of undetermined coefficients and the Fast-Field Program is implemented with a simulated annealing optimization method and a coordinate rotation technique. The source function of interest may be broadband ship self-noise or a controlled source. The parameters sought in the inversion include, but are not limited to, geoacoustic sediment properties and receiver array geometry; the water column is assumed known. The forward propagation method is efficient and accurate for very short ranges. Results are presented of simulations with a controlled source, as well as preliminary MAPEX2000 ship self-noise. [Work supported by ONR.]

Broadband sound propagation in shallow water and geoacoustic inversion.

Part of an experiment to test a measurement package in a shallow water region in the Gulf of Mexico was designed to gather broadband acoustic data suitable for inversion to estimate seabed geoacoustic parameters. Continuous wave tow acoustic signals at multiple frequencies and broadband impulsive source signals were recorded on a horizontal line array in a high-noise environment. Simulated annealing with a normal mode forward propagation model is utilized to invert for a geoacoustic representation of the seabed. Several inversions are made from different data samples of two light bulb implosions, the measured sound speed profiles at the HLA and at the positions of the light bulb deployments, and for two different cost functions. The different cost functions, measured sound speed profiles, and measured time series result in different inverted geoacoustic profiles from which transmission loss is generated for comparison with measurements. On the basis of physical consistency and from the comparison of the transmission loss and time series, a best estimate geoacoustic profile is selected and compared to those obtained from previously reported inversions. Uncertainties in the sound speed profile are shown to affect the uncertainties of the estimated seabed parameters.