Civil and Environmental Engineering publications and data sets
https://hdl.handle.net/1813/69366
2020-04-02T17:58:16ZCode and data from: Gaussian Process Regression for Estimating EM Ducting Within the Marine Atmospheric Boundary Layer
https://hdl.handle.net/1813/69525
Code and data from: Gaussian Process Regression for Estimating EM Ducting Within the Marine Atmospheric Boundary Layer
Sit, Hilarie; Earls, Christopher J
We show that Gaussian process regression (GPR) can be used to infer the electromagnetic (EM) duct height within the marine atmospheric boundary layer
(MABL) from sparsely sampled propagation factors within the context of bistaticradars. These propagation factors are simulated using PETOOL, developed by
Ozgun et al. 2011, and the datasets for the three cases that correspond to the different sparse sampling techniques can be found in the data folder. We use GPR
to calculate the posterior predictive distribution on the labels (i.e. duct height) from both noise-free and noise-contaminated array of propagation factors. For
duct height inference from noise-contaminated propagation factors, we compare a naive approach, utilizing one random sample from the input distribution (i.e.
disregarding the input noise), with an inverse-variance weighted approach, utilizing a few random samples to estimate the true predictive distribution. The
resulting posterior predictive distributions from these two approaches are compared to a "ground truth" distribution, which is approximated using a large
number of Monte-Carlo samples. We use Python 3.6.4 and scikit-learn 0.20.2. The ability of GPR to yield accurate duct height predictions using few training
examples, along with its inference speed, indicates the suitability of the proposed method for real-time applications. This is the dataset and code that supports this work.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
* Neither the name of the copyright holders nor the name of Cornell University may be used to endorse or promote products derived from this software without specific prior written permission.
Private, research, and institutional usage is without charge. Distribution of modified versions of this soure code is admissible UNDER THE CONDITION THAT THIS SOURCE CODE REMAINS UNDER COPYRIGHT OF THE ORIGINAL DEVELOPERS, BOTH SOURCE AND OBJECT CODE ARE MADE FREELY AVAILABLE WITHOUT CHARGE, AND CLEAR NOTICE IS GIVEN OF THE MODIFICATIONS. Distribution of this code as part of a commercial system is permissible ONLY BY DIRECT ARRANGEMENT WITH THE DEVELOPERS.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
2019-11-26T00:00:00ZCode and data from: Characterizing Evaporation Ducts Within the Marine Atmospheric Boundary Layer Using Artificial Neural Networks
https://hdl.handle.net/1813/69416
Code and data from: Characterizing Evaporation Ducts Within the Marine Atmospheric Boundary Layer Using Artificial Neural Networks
Sit, Hilarie; Earls, Christopher J
Abstract: We apply a multilayer perceptron machine learning (ML) regression approach to infer electromagnetic (EM) duct heights within the marine atmospheric boundary layer (MABL) using sparsely sampled EM propagation data obtained within a bistatic context. EM propagation data is simulated using PETOOL, a MATLAB-based software developed by Ozgun et al. 2011 for solving the split-step parabolic equation approximation of Helmholtz wave equation. Three cases in the data folder correspond to different sparse sampling techniques detailed in our paper. Artificial neural networks are implemented utilizing Tensorflow, and its hyperparameters are selected with grid search. Results for model selection and evaluation can be found in their respective folders. The resulting speed of our ML predictions of EM duct heights, using sparse data measurements within MABL, indicates the suitability of the proposed method for real-time applications.
Copyright (c) 2019 Hilarie Sit, hs764@cornell.edu
Developed by Hilarie Sit, Cornell University
All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
* Neither the name of the copyright holders nor the name of Cornell University may be used to endorse or promote products derived from this software without specific prior written permission.
Private, research, and institutional usage is without charge. Distribution of modified versions of this soure code is admissible UNDER THE CONDITION THAT THIS SOURCE CODE REMAINS UNDER COPYRIGHT OF THE ORIGINAL DEVELOPERS, BOTH SOURCE AND OBJECT CODE ARE MADE FREELY AVAILABLE WITHOUT CHARGE, AND CLEAR NOTICE IS GIVEN OF THE MODIFICATIONS. Distribution of this code as part of a commercial system is permissible ONLY BY DIRECT ARRANGEMENT WITH THE DEVELOPERS.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
2019-10-22T00:00:00ZReflections and the focusing effect from an ideal three-dimensional rough surface
https://hdl.handle.net/1813/2667
Reflections and the focusing effect from an ideal three-dimensional rough surface
Clavano, Wilhelmina R.; Philpot, William D.
An analytical expression for higher-order reflectances from a shallow-water homogeneous ocean bottom modeled as an egg-carton surface is presented. Roughness of this ideal surface is expressed as the amplitude-to-length ratio of its basic sinusoidal function. Any real surface that can be approximated by an egg-carton function will effectively have a comparable roughness metric. Incidence and reflection directions are considered in full azimuthal variation. The detector is located just below the water surface so that only in-water reflections are considered and there are no air-water transmission effects. Furthermore, this setup allows for an understanding of reflections that occur in media with any index of refraction or absorption coefficient. Fixing the detector footprint but adjusting its field-of-view enables the observation of the same bottom surface area as the depth varies while keeping the roughness and the number of waveforms viewed constant.
First-order reflectance decreases as the roughness increases, as was shown in the two-dimensional case. This is true as the roughness varies, regardless of the bottom reference level chosen. Focusing effects are expected from (but are not limited to) second-order reflectance and are due to parts of the bottom whose angles maximize both incoming light and the reflections toward the detector. Along a plane about the vertical axis, the roughness ratio for a fixed-length waveform that returns the highest reflectance can be found. In three dimensions, this phenomenon is complicated by reflections from all hemispherical directions. Shadowing and obscuration behave similarly as in the two-dimensional case although shadowed areas will have an increased potential to reflect light from other directions (than the plane defined by the source incidence and the vertical directions). This is expected to cause higher order reflections to increase as the roughness increases.
2006-03-04T15:33:50ZThe off-specular peak and polarisation effects of an undulating underwater suface
https://hdl.handle.net/1813/2666
The off-specular peak and polarisation effects of an undulating underwater suface
Clavano, Wilhelmina R.; Philpot, William D.
Periodic undulations are used to describe underwater bottom roughness. An expression of the bi-directional reflectance distribution function (BRDF) is given that is dependent on the given roughness metric. Highlights include an off-specular peak and polarisation effects. For an undulating underwater surface, we have shown through geometric optics that reflectance from a rough diffuse surface increases as the viewing direction approaches the backward direction even in the absence of shadowing and/or self-shading (Clavano & Philpot (2003), see also Cox & Munk (1956)). The effects of shadowing and self-shading are equivalent to applying a geometrical attenuation factor to specular reflectance, which is similar to an analysis of morphological effects using triangular waves by Zaneveld & Boss (2003). We show that a reflectance peak displaced away from the specular direction occurs at large angles of incidence (relative to the global normal) as the surface gets rougher (part of work in Clavano & Philpot (2004)). Similar results have been shown for oil films on ocean surfaces using Monte Carlo methods by Otremba & Piskozub (2004) and Otremba (2004). As a general result, an expression of the full bi-directional reflectance distribution function (BRDF) is given. While geometrical effects play a significant role in the reflectance distribution, we consider polarisation effects (as in Mullamaa (1962, 1964)) to gain more insight into real-world reflectances and compare with empirical distributions described by Cox & Munk (1956).
http://www.hydrooptics.spb.ru/onw2005/index.php
2006-03-03T18:45:20ZBackscattering anisotropy near $180^{\circ}$: an indication of particle size and shape
https://hdl.handle.net/1813/2656
Backscattering anisotropy near $180^{\circ}$: an indication of particle size and shape
Clavano, Wilhelmina R.; Boss, Emmanuel; Agrawal, Yogesh C.
By modelling the single scattering of particles in the exact backward direction ($180^{\circ}$) and $5^{\circ}$ around, the field of view of an instrument measuring backscattering is simulated. Calculations of the scattering Mueller matrix $M_{ij}$ using a development of the extended boundary condition method [1] are made for spheroidal particles with sizes ($D$ in $\mu$), shapes (defined by spheroidal aspect ratio $\frac{s}{t}$) and refractive indices similar to ($m = 1.05 + 0.01i$) marine particles found in the natural environment.
Results show that information about size and shape can be gathered from the intensity patterns of the backscattering for particles within the anomalous diffraction region. Comparison between the polarised scattering intensity patterns ($I_{\parallel}$ and $I_{\perp}$) produced by these non-spheres and their volume-equivalent spheres provides insight into the information available from backscattering polarimetry on the effects of size and shape in light scattering by differently shaped particles.
2006-03-01T18:45:55Z