DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited.

Size: px
Start display at page:

Download "DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited."

Transcription

1 AVIA Test Selection through Spatial Variance Bounding Method for Autonomy Under Test By Miles Thompson Senior Research Engineer Aerospace, Transportation, and Advanced Systems Lab DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 1

2 Autonomy Validation, Introspection, and Assessment (AVIA) Rapidly conduct massive numbers of analytic assessments of autonomy system for DARPA's Anti-Submarine Warfare Continuous Trail Unmanned Vessel (ACTUV) in complex, dynamic scenarios Impact Assess capabilities beyond operational envelope: 10x 100x more scenarios Address ACTUV operational concerns Provide > 10x reduction in time, from 42 days to < 1 day DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 2

3 How does AVIA accomplish this? Sample entire test domain intelligently and without bias Currently using Latin hypercube sampling and SVBM Allow perturbations and emergent behaviors to arise Upgradeable: can implement future sampling techniques Parallel scenario execution Demonstrated hour scenarios in < 24 hours (16 hours) Automatic metrics evaluation Metrics stored in MySQL database, easy to visualize with analysis graphical user interface (GUI) Metrics are requirement dependent and include introspection to increase understanding of autonomy logic DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 3

4 How does AVIA accomplish this? Sample entire test domain intelligently and without bias Currently using Latin hypercube sampling and SVBM Allow perturbations and emergent behaviors to arise Upgradeable: can implement future sampling techniques What is Spatial Variance Bounding Method (SVBM)? Test selection method designed for discontinuous performance of complex systems across a large test domain Perfect for autonomous systems and autonomy under test Iteratively selects tests to run based on expected variance Goal is to bound the expectations of performance throughout the test domain including probability of exceeding a requirement value DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 4

5 Requirements for SVBM Assumptions about performance of the Autonomy Under Test (AUT) Performance is a random variable with known or unknown distribution» May still be deterministic, but assumed stochastic or probabilistic Performance is stationary, it is time-invariant» A learning autonomy is not time invariant Performance can be measured with scalar values for all metrics Performance at similar locations in the test domain have similar variances» There is a spatial relationship between test points DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 5

6 How Test Selection Works in SVBM 1. Initial Sampling 2. Iterative Sampling 3. Completion Criteria DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 6

7 Initial Sampling (Seeding) DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 7

8 Requirements for Initial Sampling Defined test domain The set of valid test cases that arise from the set of parameter values May or may not include the system under test (SUT) that the autonomy will inhabit AVIA leverages scenarios Smaller sections of the test domain that contain interesting results Scenarios have certain constants and other parameters are varied Gives the tester/developer more control over the test domain DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 8

9 Initial Sampling Method AVIA uses a random sampling technique to seed the test domain Uses a form of Latin Hypercube Sampling that is not orthogonal under the following conditions: Sample size greater than the number of categorical values in a parameter Sample size greater than the number of integer values in a range of an integer parameter DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 9

10 Example Initial Sample Run # of Transients Visibility (m) Sea State Seed DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 10

11 Running Initial Sample Initial sample is passed to Test Matrix Tool (TMT) Run Manager for simulation and results gathering While the sample is being run, as runs complete, the metrics from those runs are added to a SQL database and available for viewing in the analysis graphical user interface (GUI) DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 11

12 Iterative Sampling DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 12

13 Why Sample Iteratively? Use results to drive selection of desired test locations Know when to stop testing prior to running entire domain Entire test domain likely unfeasible Blind large sample may be costly and not provide desired results SVBM can use a variety of iterative sampling methods as long as those methods provide an estimate of variance DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 13

14 Iterative Sampling Methods Interpolation Estimating the values of test locations that have not been run by combining weighted values from other test locations in the previous sample Genetic Algorithm Using natural selection to guide sampling toward desired test locations based on crossover and mutation of the previous sample Univariate Search Modifying the previous sample by stepping each parameter sequentially to produce a new sample DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 14

15 Background on Kriging Kriging provides a performance estimate at a point in the test domain based on the results of all completed points and their spatial relationship Unlike many interpolation methods, Kriging uses covariance between parameters to estimate metric values Prevents bias from closely located tests on predictive estimates Allows each metric to have its own weights for interpolation based on the spatial relationship between test locations DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 15

16 Applied Kriging Examples Design Optimization of an Engine Joseph, V.R., Hung Y., & Sudjianto, A. (2008). Blind Kriging: A new Method for Developing Models DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 16

17 Applied Kriging Examples System Tradespace Search Fulcoly, D. O., Shaw, N. B., Ross, A. M., & Rhodes, D. H. (2012). Exploiting Multidimensional Design of Experiments and Kriging Methods: An Application to a Satellite Radar System Tradespace and Orbital Transfer Vehicle Tradespace DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 17

18 Kriging in AVIA AVIA follows a very similar path to the tradespace search when it comes to testing Major difference: Tradespace looks for optimal or satisfactory conditions AVIA currently looks for correct/comprehensive performance characterization with regard to requirements AVIA uses Kriging to identify the points in the test domain that the testers know the least about These points have the highest uncertainty highest variance Every iteration of the sample will be the tests with the highest uncertainty AVIA does not bias on passing or failing requirements, which makes it an excellent test tool AVIA could be used as a design tool looking for optimal performance, but currently it is being used for assessing autonomy DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 18

19 Kriging in AVIA With the distance, known as lag, calculated between all existing runs, AVIA can calculate a covariance matrix for each of the metrics being analyzed (see backup slides) It is important to analyze each metric independently because each could have a different variogram, the spatial relationship of lag to variance for that metric. Closest Point of Approach (CPA) DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 19

20 Completion Criteria DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 20

21 Completion Criteria Completion criteria are required so that an iterative sampling mechanism knows when to stop It also should be meaningful to the tester In AVIA, desire to have a comprehensive performance characterization across the test domain Accomplished by getting an accurate representation of the variance per metric Accurate performance variance requires a good sample» AVIA assumes this condition is reached when the metric variances are steady state DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 21

22 Steady State Variance Example Iteration 1 Variances x10^-3 Iteration 2 Variances Collisions/Hour CPA Collisions/Hour CPA SIMULATED DATA FOR THE PURPOSE OF PRESENTATION DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 22

23 Convergence/Divergence/Steady Iteration 3 Variances x10^-3 Iteration 4 Variances Collisions/Hour CPA Collisions/Hour CPA SIMULATED DATA FOR THE PURPOSE OF PRESENTATION DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 23

24 Convergence/Divergence/Steady x10^-3 Iteration 5 Variances x10^-3 Iteration 6 Variances Collisions/Hour CPA Collisions/Hour CPA SIMULATED DATA FOR THE PURPOSE OF PRESENTATION DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 24

25 Convergence/Divergence/Steady x10^-3 Iteration 7 Variances x10^-3 Iteration 8 Variances Collisions/Hour CPA Collisions/Hour CPA SIMULATED DATA FOR THE PURPOSE OF PRESENTATION DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 25

26 Guaranteed Steady for Stationary Guaranteed to converge to steady state for stationary (time-invariant) performance: Even if truly random to some distribution Even if no spatial variance relationship (see CPA above) Finite simulation testing of large test domains with completion criteria in AVIA DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 26

27 Advantages/Disadvantages Advantages/Disadvantages of Spatial Variance Bounding Method in AVIA Pros:» Can bound expected performance for any conditions that were not tested (excellent for acquisition evaluation)» Finite Cons:» Scales with the variability of the system» Each metric assessed individually for steady state» Iterative sampling criteria is flexible and can be made to locate min/max» If high variability in system, will require many runs» Requires all metrics to be scalar» Assumes all metric errors to be normally distributed to remain unbiased DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 27

28 Acknowledgement Thanks to DARPA for funding this research and pushing assessment methods for autonomous systems. QUESTIONS? This research was developed with funding from the Defense Advanced Research Projects Agency (DARPA). The views, opinions and/or findings expressed are those of the author and should not be interpreted as representing the official views or policies of the Department of Defense or the U.S. Government. DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 28

29 BACKUP SLIDES DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 29

30 Explanation of Kriging From Bailey and Gatrell Ordinary Kriging The weights are based on the covariance between the locations of the test that have been completed and the location of the desired untested point DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 30

31 Explanation of Kriging Introduce a Lagrange multiplier λ to ensure that the sum of all the weights = 1 to minimize the variance between the actual and predicted response DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 31

32 Explanation of Kriging The interpolation weights provide the estimated response based on already captured responses Predicting the response at an untested location is good, but AVIA needs the estimated variance at that location for test selection DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 32

33 Explanation of Kriging Minimizing the variance of the prediction errors with the Lagrangian, yields the following: Modifying the previous equations for calculating weights: DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 33

34 Explanation of Kriging Traditional Kriging Methods: Simple Kriging The mean performance is known and it is zero (or a constant)» The sum of the weights does not have to equal 1 Ordinary Kriging The mean performance is not known, but it is constant» The sum of the weights equals 1 Universal Kriging (Kriging with a trend) The mean performance is dictated by an unknown trend-line (typically polynomial)» Requires sufficient sample size to estimate the polynomial Modern Kriging Methods: Disjunctive Kriging Nonlinear generalization of Kriging Blind Kriging Special Universal Kriging using Bayesian inference for estimating trend-line DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 34

35 Kriging Lag Calculation in AVIA Finding the distance in AVIA test domain: Kriging relies on covariance to calculate its weights» The covariance requires a distance commonly referred to as lag between test locations AVIA s parameters are rarely in the same units» Requires normalization to ensure dimensions can be computed as lag Run # of Transients Visibility (m) Sea State Target Present False False True DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 35

36 Kriging Lag Calculation in AVIA Run # of Transients Visibility (m) Sea State Target Present False False True AVIA normalizes each parameter individually, then calculates the lag matrices for each parameter. Categorical parameters are treated with simple matching coefficient. # of Transients Visibility Sea State Target Present Run Run Run Run With lags calculated for each parameter, AVIA calculates an overall distance by Euclidean distance calculations. In this way, all distances will be between 0 and n, where n is the number of parameters. In this case, n = 4 so all distances will be [0, 2]. DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 36

37 Kriging Lag Calculation in AVIA Run By the calculations, runs 1 and 3 are the closest to each other and runs 2 and 3 are farthest from each other. A lag of 0 indicates that two runs are identical in parameters. DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 37

38 Kriging in AVIA Variograms in AVIA Calculated by binning the lags and calculating the variance of all runs metric values in that bin» Only bin up to 50 percent of max lag» Expect variance to approach sill after halfway across the test domain Fit a theoretical variogram model to smooth the spatial variance» Currently use Gaussian because likely to have a non-zero y-intercept (nugget variance)» A horizontal variogram indicates very little spatial relationship DISTRIBUTION STATEMENT A Approved for public release: distribution unlimited. 38

Spatial Interpolation & Geostatistics

Spatial Interpolation & Geostatistics (Z i Z j ) 2 / 2 Spatial Interpolation & Geostatistics Lag Lag Mean Distance between pairs of points 1 Tobler s Law All places are related, but nearby places are related more than distant places Corollary:

More information

Spatial Interpolation - Geostatistics 4/3/2018

Spatial Interpolation - Geostatistics 4/3/2018 Spatial Interpolation - Geostatistics 4/3/201 (Z i Z j ) 2 / 2 Spatial Interpolation & Geostatistics Lag Distance between pairs of points Lag Mean Tobler s Law All places are related, but nearby places

More information

Introduction to ANSYS DesignXplorer

Introduction to ANSYS DesignXplorer Lecture 4 14. 5 Release Introduction to ANSYS DesignXplorer 1 2013 ANSYS, Inc. September 27, 2013 s are functions of different nature where the output parameters are described in terms of the input parameters

More information

Spatial Interpolation & Geostatistics

Spatial Interpolation & Geostatistics (Z i Z j ) 2 / 2 Spatial Interpolation & Geostatistics Lag Lag Mean Distance between pairs of points 11/3/2016 GEO327G/386G, UT Austin 1 Tobler s Law All places are related, but nearby places are related

More information

Computer Experiments: Space Filling Design and Gaussian Process Modeling

Computer Experiments: Space Filling Design and Gaussian Process Modeling Computer Experiments: Space Filling Design and Gaussian Process Modeling Best Practice Authored by: Cory Natoli Sarah Burke, Ph.D. 30 March 2018 The goal of the STAT COE is to assist in developing rigorous,

More information

Using Kriging Methods to Estimate Damage Distribution

Using Kriging Methods to Estimate Damage Distribution Lehigh University Lehigh Preserve Theses and Dissertations 2011 Using Kriging Methods to Estimate Damage Distribution Bing Wang Lehigh University Follow this and additional works at: http://preserve.lehigh.edu/etd

More information

Workshop - Model Calibration and Uncertainty Analysis Using PEST

Workshop - Model Calibration and Uncertainty Analysis Using PEST About PEST PEST (Parameter ESTimation) is a general-purpose, model-independent, parameter estimation and model predictive uncertainty analysis package developed by Dr. John Doherty. PEST is the most advanced

More information

FMA901F: Machine Learning Lecture 3: Linear Models for Regression. Cristian Sminchisescu

FMA901F: Machine Learning Lecture 3: Linear Models for Regression. Cristian Sminchisescu FMA901F: Machine Learning Lecture 3: Linear Models for Regression Cristian Sminchisescu Machine Learning: Frequentist vs. Bayesian In the frequentist setting, we seek a fixed parameter (vector), with value(s)

More information

Non Stationary Variograms Based on Continuously Varying Weights

Non Stationary Variograms Based on Continuously Varying Weights Non Stationary Variograms Based on Continuously Varying Weights David F. Machuca-Mory and Clayton V. Deutsch Centre for Computational Geostatistics Department of Civil & Environmental Engineering University

More information

Spatial Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University

Spatial Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University Spatial Analysis and Modeling (GIST 4302/5302) Guofeng Cao Department of Geosciences Texas Tech University 1 Outline of This Week Last topic, we learned: Spatial autocorrelation of areal data Spatial regression

More information

Surface Smoothing Using Kriging

Surface Smoothing Using Kriging 1 AutoCAD Civil 3D has the ability to add data points to a surface based on mathematical criteria. This gives you the ability to strengthen the surface definition in areas where data may be sparse or where

More information

CHAPTER 2 CONVENTIONAL AND NON-CONVENTIONAL TECHNIQUES TO SOLVE ORPD PROBLEM

CHAPTER 2 CONVENTIONAL AND NON-CONVENTIONAL TECHNIQUES TO SOLVE ORPD PROBLEM 20 CHAPTER 2 CONVENTIONAL AND NON-CONVENTIONAL TECHNIQUES TO SOLVE ORPD PROBLEM 2.1 CLASSIFICATION OF CONVENTIONAL TECHNIQUES Classical optimization methods can be classified into two distinct groups:

More information

Linear Methods for Regression and Shrinkage Methods

Linear Methods for Regression and Shrinkage Methods Linear Methods for Regression and Shrinkage Methods Reference: The Elements of Statistical Learning, by T. Hastie, R. Tibshirani, J. Friedman, Springer 1 Linear Regression Models Least Squares Input vectors

More information

Clustering and Visualisation of Data

Clustering and Visualisation of Data Clustering and Visualisation of Data Hiroshi Shimodaira January-March 28 Cluster analysis aims to partition a data set into meaningful or useful groups, based on distances between data points. In some

More information

Workload Characterization Techniques

Workload Characterization Techniques Workload Characterization Techniques Raj Jain Washington University in Saint Louis Saint Louis, MO 63130 Jain@cse.wustl.edu These slides are available on-line at: http://www.cse.wustl.edu/~jain/cse567-08/

More information

Analytical model A structure and process for analyzing a dataset. For example, a decision tree is a model for the classification of a dataset.

Analytical model A structure and process for analyzing a dataset. For example, a decision tree is a model for the classification of a dataset. Glossary of data mining terms: Accuracy Accuracy is an important factor in assessing the success of data mining. When applied to data, accuracy refers to the rate of correct values in the data. When applied

More information

Massive Data Analysis

Massive Data Analysis Professor, Department of Electrical and Computer Engineering Tennessee Technological University February 25, 2015 Big Data This talk is based on the report [1]. The growth of big data is changing that

More information

Machine Learning. Topic 5: Linear Discriminants. Bryan Pardo, EECS 349 Machine Learning, 2013

Machine Learning. Topic 5: Linear Discriminants. Bryan Pardo, EECS 349 Machine Learning, 2013 Machine Learning Topic 5: Linear Discriminants Bryan Pardo, EECS 349 Machine Learning, 2013 Thanks to Mark Cartwright for his extensive contributions to these slides Thanks to Alpaydin, Bishop, and Duda/Hart/Stork

More information

Mixture Models and the EM Algorithm

Mixture Models and the EM Algorithm Mixture Models and the EM Algorithm Padhraic Smyth, Department of Computer Science University of California, Irvine c 2017 1 Finite Mixture Models Say we have a data set D = {x 1,..., x N } where x i is

More information

Geostatistical Reservoir Characterization of McMurray Formation by 2-D Modeling

Geostatistical Reservoir Characterization of McMurray Formation by 2-D Modeling Geostatistical Reservoir Characterization of McMurray Formation by 2-D Modeling Weishan Ren, Oy Leuangthong and Clayton V. Deutsch Department of Civil & Environmental Engineering, University of Alberta

More information

Dijkstra's Algorithm

Dijkstra's Algorithm Shortest Path Algorithm Dijkstra's Algorithm To find the shortest path from the origin node to the destination node No matrix calculation Floyd s Algorithm To find all the shortest paths from the nodes

More information

Metaheuristic Development Methodology. Fall 2009 Instructor: Dr. Masoud Yaghini

Metaheuristic Development Methodology. Fall 2009 Instructor: Dr. Masoud Yaghini Metaheuristic Development Methodology Fall 2009 Instructor: Dr. Masoud Yaghini Phases and Steps Phases and Steps Phase 1: Understanding Problem Step 1: State the Problem Step 2: Review of Existing Solution

More information

PATTERN CLASSIFICATION AND SCENE ANALYSIS

PATTERN CLASSIFICATION AND SCENE ANALYSIS PATTERN CLASSIFICATION AND SCENE ANALYSIS RICHARD O. DUDA PETER E. HART Stanford Research Institute, Menlo Park, California A WILEY-INTERSCIENCE PUBLICATION JOHN WILEY & SONS New York Chichester Brisbane

More information

Geostatistics 2D GMS 7.0 TUTORIALS. 1 Introduction. 1.1 Contents

Geostatistics 2D GMS 7.0 TUTORIALS. 1 Introduction. 1.1 Contents GMS 7.0 TUTORIALS 1 Introduction Two-dimensional geostatistics (interpolation) can be performed in GMS using the 2D Scatter Point module. The module is used to interpolate from sets of 2D scatter points

More information

Note Set 4: Finite Mixture Models and the EM Algorithm

Note Set 4: Finite Mixture Models and the EM Algorithm Note Set 4: Finite Mixture Models and the EM Algorithm Padhraic Smyth, Department of Computer Science University of California, Irvine Finite Mixture Models A finite mixture model with K components, for

More information

You ve already read basics of simulation now I will be taking up method of simulation, that is Random Number Generation

You ve already read basics of simulation now I will be taking up method of simulation, that is Random Number Generation Unit 5 SIMULATION THEORY Lesson 39 Learning objective: To learn random number generation. Methods of simulation. Monte Carlo method of simulation You ve already read basics of simulation now I will be

More information

Missing Data Analysis for the Employee Dataset

Missing Data Analysis for the Employee Dataset Missing Data Analysis for the Employee Dataset 67% of the observations have missing values! Modeling Setup For our analysis goals we would like to do: Y X N (X, 2 I) and then interpret the coefficients

More information

The Cross-Entropy Method

The Cross-Entropy Method The Cross-Entropy Method Guy Weichenberg 7 September 2003 Introduction This report is a summary of the theory underlying the Cross-Entropy (CE) method, as discussed in the tutorial by de Boer, Kroese,

More information

INVERSE PROBLEMS IN GROUNDWATER MODELING

INVERSE PROBLEMS IN GROUNDWATER MODELING INVERSE PROBLEMS IN GROUNDWATER MODELING Theory and Applications of Transport in Porous Media Series Editor: Jacob Bear, Technion - Israel Institute of Technology, Haifa, Israel Volume 6 The titles published

More information

New developments in LS-OPT

New developments in LS-OPT 7. LS-DYNA Anwenderforum, Bamberg 2008 Optimierung II New developments in LS-OPT Nielen Stander, Tushar Goel, Willem Roux Livermore Software Technology Corporation, Livermore, CA94551, USA Summary: This

More information

Finite Element Convergence for Time-Dependent PDEs with a Point Source in COMSOL 4.2

Finite Element Convergence for Time-Dependent PDEs with a Point Source in COMSOL 4.2 Finite Element Convergence for Time-Dependent PDEs with a Point Source in COMSOL 4.2 David W. Trott and Matthias K. Gobbert Department of Mathematics and Statistics, University of Maryland, Baltimore County,

More information

7 Fractions. Number Sense and Numeration Measurement Geometry and Spatial Sense Patterning and Algebra Data Management and Probability

7 Fractions. Number Sense and Numeration Measurement Geometry and Spatial Sense Patterning and Algebra Data Management and Probability 7 Fractions GRADE 7 FRACTIONS continue to develop proficiency by using fractions in mental strategies and in selecting and justifying use; develop proficiency in adding and subtracting simple fractions;

More information

Global modelling of air pollution using multiple data sources

Global modelling of air pollution using multiple data sources Global modelling of air pollution using multiple data sources Matthew Thomas SAMBa, University of Bath Email: M.L.Thomas@bath.ac.uk November 11, 015 1/ 3 OUTLINE Motivation Data Sources Existing Approaches

More information

Mobile Robotics. Mathematics, Models, and Methods. HI Cambridge. Alonzo Kelly. Carnegie Mellon University UNIVERSITY PRESS

Mobile Robotics. Mathematics, Models, and Methods. HI Cambridge. Alonzo Kelly. Carnegie Mellon University UNIVERSITY PRESS Mobile Robotics Mathematics, Models, and Methods Alonzo Kelly Carnegie Mellon University HI Cambridge UNIVERSITY PRESS Contents Preface page xiii 1 Introduction 1 1.1 Applications of Mobile Robots 2 1.2

More information

Gridding and Contouring in Geochemistry for ArcGIS

Gridding and Contouring in Geochemistry for ArcGIS Gridding and Contouring in Geochemistry for ArcGIS The Geochemsitry for ArcGIS extension includes three gridding options: Minimum Curvature Gridding, Kriging and a new Inverse Distance Weighting algorithm.

More information

Missing Data Missing Data Methods in ML Multiple Imputation

Missing Data Missing Data Methods in ML Multiple Imputation Missing Data Missing Data Methods in ML Multiple Imputation PRE 905: Multivariate Analysis Lecture 11: April 22, 2014 PRE 905: Lecture 11 Missing Data Methods Today s Lecture The basics of missing data:

More information

A workflow to account for uncertainty in well-log data in 3D geostatistical reservoir modeling

A workflow to account for uncertainty in well-log data in 3D geostatistical reservoir modeling A workflow to account for uncertainty in well-log data in 3D geostatistical reservoir Jose Akamine and Jef Caers May, 2007 Stanford Center for Reservoir Forecasting Abstract Traditionally well log data

More information

Soft Threshold Estimation for Varying{coecient Models 2 ations of certain basis functions (e.g. wavelets). These functions are assumed to be smooth an

Soft Threshold Estimation for Varying{coecient Models 2 ations of certain basis functions (e.g. wavelets). These functions are assumed to be smooth an Soft Threshold Estimation for Varying{coecient Models Artur Klinger, Universitat Munchen ABSTRACT: An alternative penalized likelihood estimator for varying{coecient regression in generalized linear models

More information

Themes in the Texas CCRS - Mathematics

Themes in the Texas CCRS - Mathematics 1. Compare real numbers. a. Classify numbers as natural, whole, integers, rational, irrational, real, imaginary, &/or complex. b. Use and apply the relative magnitude of real numbers by using inequality

More information

3 Feature Selection & Feature Extraction

3 Feature Selection & Feature Extraction 3 Feature Selection & Feature Extraction Overview: 3.1 Introduction 3.2 Feature Extraction 3.3 Feature Selection 3.3.1 Max-Dependency, Max-Relevance, Min-Redundancy 3.3.2 Relevance Filter 3.3.3 Redundancy

More information

Analytical Approach for Numerical Accuracy Estimation of Fixed-Point Systems Based on Smooth Operations

Analytical Approach for Numerical Accuracy Estimation of Fixed-Point Systems Based on Smooth Operations 2326 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I: REGULAR PAPERS, VOL 59, NO 10, OCTOBER 2012 Analytical Approach for Numerical Accuracy Estimation of Fixed-Point Systems Based on Smooth Operations Romuald

More information

Fast Automated Estimation of Variance in Discrete Quantitative Stochastic Simulation

Fast Automated Estimation of Variance in Discrete Quantitative Stochastic Simulation Fast Automated Estimation of Variance in Discrete Quantitative Stochastic Simulation November 2010 Nelson Shaw njd50@uclive.ac.nz Department of Computer Science and Software Engineering University of Canterbury,

More information

Machine Learning / Jan 27, 2010

Machine Learning / Jan 27, 2010 Revisiting Logistic Regression & Naïve Bayes Aarti Singh Machine Learning 10-701/15-781 Jan 27, 2010 Generative and Discriminative Classifiers Training classifiers involves learning a mapping f: X -> Y,

More information

GAUSSIAN PROCESS RESPONSE SURFACE MODELING AND GLOBAL SENSITIVITY ANALYSIS USING NESSUS

GAUSSIAN PROCESS RESPONSE SURFACE MODELING AND GLOBAL SENSITIVITY ANALYSIS USING NESSUS UNCECOMP 2017 2 nd ECCOMAS Thematic Conference on International Conference on Uncertainty Quantification in Computational Sciences and Engineering M. Papadrakakis, V. Papadopoulos, G. Stefanou (eds.) Rhodes

More information

Local Recoverable Reserves Prediction with Block LU Simulation

Local Recoverable Reserves Prediction with Block LU Simulation Local Recoverable Reserves Prediction with Block LU Simulation Julián Ortiz C. 1 and Clayton V. Deutsch 1 Department of Mining Engineering University of Chile Centre for Computational Geostatistics (CCG)

More information

Multi-Mesh CFD. Chris Roy Chip Jackson (1 st year PhD student) Aerospace and Ocean Engineering Department Virginia Tech

Multi-Mesh CFD. Chris Roy Chip Jackson (1 st year PhD student) Aerospace and Ocean Engineering Department Virginia Tech Multi-Mesh CFD Chris Roy Chip Jackson (1 st year PhD student) Aerospace and Ocean Engineering Department Virginia Tech cjroy@vt.edu May 21, 2014 CCAS Program Review, Columbus, OH 1 Motivation Automated

More information

Chapter II. Linear Programming

Chapter II. Linear Programming 1 Chapter II Linear Programming 1. Introduction 2. Simplex Method 3. Duality Theory 4. Optimality Conditions 5. Applications (QP & SLP) 6. Sensitivity Analysis 7. Interior Point Methods 1 INTRODUCTION

More information

Selected Implementation Issues with Sequential Gaussian Simulation

Selected Implementation Issues with Sequential Gaussian Simulation Selected Implementation Issues with Sequential Gaussian Simulation Abstract Stefan Zanon (szanon@ualberta.ca) and Oy Leuangthong (oy@ualberta.ca) Department of Civil & Environmental Engineering University

More information

Applying Supervised Learning

Applying Supervised Learning Applying Supervised Learning When to Consider Supervised Learning A supervised learning algorithm takes a known set of input data (the training set) and known responses to the data (output), and trains

More information

Frequencies, Unequal Variance Weights, and Sampling Weights: Similarities and Differences in SAS

Frequencies, Unequal Variance Weights, and Sampling Weights: Similarities and Differences in SAS ABSTRACT Paper 1938-2018 Frequencies, Unequal Variance Weights, and Sampling Weights: Similarities and Differences in SAS Robert M. Lucas, Robert M. Lucas Consulting, Fort Collins, CO, USA There is confusion

More information

Gaussian Processes for Robotics. McGill COMP 765 Oct 24 th, 2017

Gaussian Processes for Robotics. McGill COMP 765 Oct 24 th, 2017 Gaussian Processes for Robotics McGill COMP 765 Oct 24 th, 2017 A robot must learn Modeling the environment is sometimes an end goal: Space exploration Disaster recovery Environmental monitoring Other

More information

WHAT TYPE OF NEURAL NETWORK IS IDEAL FOR PREDICTIONS OF SOLAR FLARES?

WHAT TYPE OF NEURAL NETWORK IS IDEAL FOR PREDICTIONS OF SOLAR FLARES? WHAT TYPE OF NEURAL NETWORK IS IDEAL FOR PREDICTIONS OF SOLAR FLARES? Initially considered for this model was a feed forward neural network. Essentially, this means connections between units do not form

More information

Knowledge Discovery and Data Mining

Knowledge Discovery and Data Mining Knowledge Discovery and Data Mining Basis Functions Tom Kelsey School of Computer Science University of St Andrews http://www.cs.st-andrews.ac.uk/~tom/ tom@cs.st-andrews.ac.uk Tom Kelsey ID5059-02-BF 2015-02-04

More information

Homework. Gaussian, Bishop 2.3 Non-parametric, Bishop 2.5 Linear regression Pod-cast lecture on-line. Next lectures:

Homework. Gaussian, Bishop 2.3 Non-parametric, Bishop 2.5 Linear regression Pod-cast lecture on-line. Next lectures: Homework Gaussian, Bishop 2.3 Non-parametric, Bishop 2.5 Linear regression 3.0-3.2 Pod-cast lecture on-line Next lectures: I posted a rough plan. It is flexible though so please come with suggestions Bayes

More information

Petrel TIPS&TRICKS from SCM

Petrel TIPS&TRICKS from SCM Petrel TIPS&TRICKS from SCM Knowledge Worth Sharing Petrel 2D Grid Algorithms and the Data Types they are Commonly Used With Specific algorithms in Petrel s Make/Edit Surfaces process are more appropriate

More information

1. Assumptions. 1. Introduction. 2. Terminology

1. Assumptions. 1. Introduction. 2. Terminology 4. Process Modeling 4. Process Modeling The goal for this chapter is to present the background and specific analysis techniques needed to construct a statistical model that describes a particular scientific

More information

3 Nonlinear Regression

3 Nonlinear Regression CSC 4 / CSC D / CSC C 3 Sometimes linear models are not sufficient to capture the real-world phenomena, and thus nonlinear models are necessary. In regression, all such models will have the same basic

More information

An Automated Method for Hot-to-Cold Geometry Mapping

An Automated Method for Hot-to-Cold Geometry Mapping Brigham Young University BYU ScholarsArchive All Theses and Dissertations 2015-05-01 An Automated Method for Hot-to-Cold Geometry Mapping Brandon Levi Doolin Brigham Young University - Provo Follow this

More information

Feature Selection. CE-725: Statistical Pattern Recognition Sharif University of Technology Spring Soleymani

Feature Selection. CE-725: Statistical Pattern Recognition Sharif University of Technology Spring Soleymani Feature Selection CE-725: Statistical Pattern Recognition Sharif University of Technology Spring 2013 Soleymani Outline Dimensionality reduction Feature selection vs. feature extraction Filter univariate

More information

INFRARED AUTONOMOUS ACQUISITION AND TRACKING

INFRARED AUTONOMOUS ACQUISITION AND TRACKING INFRARED AUTONOMOUS ACQUISITION AND TRACKING Teresa L.P. Olson and Harry C. Lee Teresa.Lolson@lmco.com (407) 356-7109 Harrv.c.lee@lmco.com (407) 356-6997 Lockheed Martin Missiles and Fire Control - Orlando

More information

Expectation-Maximization Methods in Population Analysis. Robert J. Bauer, Ph.D. ICON plc.

Expectation-Maximization Methods in Population Analysis. Robert J. Bauer, Ph.D. ICON plc. Expectation-Maximization Methods in Population Analysis Robert J. Bauer, Ph.D. ICON plc. 1 Objective The objective of this tutorial is to briefly describe the statistical basis of Expectation-Maximization

More information

Just the Facts Small-Sliding Contact in ANSYS Mechanical

Just the Facts Small-Sliding Contact in ANSYS Mechanical Just the Facts Small-Sliding Contact in ANSYS Mechanical ANSYS, Inc. 2600 ANSYS Drive Canonsburg, PA 15317 29 March 2018 Although this document provides information that customers may find useful, it is

More information

3 Nonlinear Regression

3 Nonlinear Regression 3 Linear models are often insufficient to capture the real-world phenomena. That is, the relation between the inputs and the outputs we want to be able to predict are not linear. As a consequence, nonlinear

More information

Uncertainty analysis in spatial environmental modelling. Kasia Sawicka Athens, 29 Mar 2017

Uncertainty analysis in spatial environmental modelling. Kasia Sawicka Athens, 29 Mar 2017 Uncertainty analysis in spatial environmental modelling Kasia Sawicka Athens, 29 Mar 2017 Uncertainty propagation overview Input ± U Model (± U) Output ± U Input data Model parameters Model structure Solution

More information

USE OF STOCHASTIC ANALYSIS FOR FMVSS210 SIMULATION READINESS FOR CORRELATION TO HARDWARE TESTING

USE OF STOCHASTIC ANALYSIS FOR FMVSS210 SIMULATION READINESS FOR CORRELATION TO HARDWARE TESTING 7 th International LS-DYNA Users Conference Simulation Technology (3) USE OF STOCHASTIC ANALYSIS FOR FMVSS21 SIMULATION READINESS FOR CORRELATION TO HARDWARE TESTING Amit Sharma Ashok Deshpande Raviraj

More information

A Comparison of Spatial Prediction Techniques Using Both Hard and Soft Data

A Comparison of Spatial Prediction Techniques Using Both Hard and Soft Data University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln Dissertations and Theses in Statistics Statistics, Department of 5-2011 A Comparison of Spatial Prediction Techniques Using

More information

HOW GEOVIA GEMS DEFINES BLOCK DISCRETIZATION AND BLOCK VARIANCE:

HOW GEOVIA GEMS DEFINES BLOCK DISCRETIZATION AND BLOCK VARIANCE: HOW GEOVIA GEMS DEFINES BLOCK DISCRETIZATION AND BLOCK VARIANCE: EXECUTIVE SUMMARY For confidence and clarity when discussing interpolation results from GEOVIA GEMS, it is important to understand how GEMS

More information

Modeling Uncertainty in the Earth Sciences Jef Caers Stanford University

Modeling Uncertainty in the Earth Sciences Jef Caers Stanford University Modeling spatial continuity Modeling Uncertainty in the Earth Sciences Jef Caers Stanford University Motivation uncertain uncertain certain or uncertain uncertain Spatial Input parameters Spatial Stochastic

More information

Watershed Sciences 4930 & 6920 GEOGRAPHIC INFORMATION SYSTEMS

Watershed Sciences 4930 & 6920 GEOGRAPHIC INFORMATION SYSTEMS HOUSEKEEPING Watershed Sciences 4930 & 6920 GEOGRAPHIC INFORMATION SYSTEMS Quizzes Lab 8? WEEK EIGHT Lecture INTERPOLATION & SPATIAL ESTIMATION Joe Wheaton READING FOR TODAY WHAT CAN WE COLLECT AT POINTS?

More information

Markov Bayes Simulation for Structural Uncertainty Estimation

Markov Bayes Simulation for Structural Uncertainty Estimation P - 200 Markov Bayes Simulation for Structural Uncertainty Estimation Samik Sil*, Sanjay Srinivasan and Mrinal K Sen. University of Texas at Austin, samiksil@gmail.com Summary Reservoir models are built

More information

Multidisciplinary Analysis and Optimization

Multidisciplinary Analysis and Optimization OptiY Multidisciplinary Analysis and Optimization Process Integration OptiY is an open and multidisciplinary design environment, which provides direct and generic interfaces to many CAD/CAE-systems and

More information

Oceanic Domain Awareness- A Concept Rear Admiral Dr. S Kulshrestha, retd

Oceanic Domain Awareness- A Concept Rear Admiral Dr. S Kulshrestha, retd Oceanic Domain Awareness- A Concept Rear Admiral Dr. S Kulshrestha, retd He knows the course of the stars and can always orient himself. He knows the value of signs, both regular, accidental and abnormal,

More information

STA 4273H: Statistical Machine Learning

STA 4273H: Statistical Machine Learning STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Statistics! rsalakhu@utstat.toronto.edu! http://www.utstat.utoronto.ca/~rsalakhu/ Sidney Smith Hall, Room 6002 Lecture 12 Combining

More information

Performance Characterization in Computer Vision

Performance Characterization in Computer Vision Performance Characterization in Computer Vision Robert M. Haralick University of Washington Seattle WA 98195 Abstract Computer vision algorithms axe composed of different sub-algorithms often applied in

More information

Variogram Inversion and Uncertainty Using Dynamic Data. Simultaneouos Inversion with Variogram Updating

Variogram Inversion and Uncertainty Using Dynamic Data. Simultaneouos Inversion with Variogram Updating Variogram Inversion and Uncertainty Using Dynamic Data Z. A. Reza (zreza@ualberta.ca) and C. V. Deutsch (cdeutsch@civil.ualberta.ca) Department of Civil & Environmental Engineering, University of Alberta

More information

Non-Linearity of Scorecard Log-Odds

Non-Linearity of Scorecard Log-Odds Non-Linearity of Scorecard Log-Odds Ross McDonald, Keith Smith, Matthew Sturgess, Edward Huang Retail Decision Science, Lloyds Banking Group Edinburgh Credit Scoring Conference 6 th August 9 Lloyds Banking

More information

Missing Data. Where did it go?

Missing Data. Where did it go? Missing Data Where did it go? 1 Learning Objectives High-level discussion of some techniques Identify type of missingness Single vs Multiple Imputation My favourite technique 2 Problem Uh data are missing

More information

Software Tutorial Session Universal Kriging

Software Tutorial Session Universal Kriging Software Tutorial Session Universal Kriging The example session with PG2000 which is described in this and Part 1 is intended as an example run to familiarise the user with the package. This documented

More information

MODFLOW Stochastic Modeling, PEST Null Space Monte Carlo I. Use PEST to create multiple calibrated MODFLOW simulations

MODFLOW Stochastic Modeling, PEST Null Space Monte Carlo I. Use PEST to create multiple calibrated MODFLOW simulations v. 10.1 GMS 10.1 Tutorial MODFLOW Stochastic Modeling, PEST Null Space Monte Carlo I Use PEST to create multiple calibrated MODFLOW simulations Objectives The PEST Null Space Monte Carlo modeling option

More information

Theoretical Concepts of Machine Learning

Theoretical Concepts of Machine Learning Theoretical Concepts of Machine Learning Part 2 Institute of Bioinformatics Johannes Kepler University, Linz, Austria Outline 1 Introduction 2 Generalization Error 3 Maximum Likelihood 4 Noise Models 5

More information

Influence of the tape number on the optimized structural performance of locally reinforced composite structures

Influence of the tape number on the optimized structural performance of locally reinforced composite structures Proceedings of the 7th GACM Colloquium on Computational Mechanics for Young Scientists from Academia and Industry October 11-13, 2017 in Stuttgart, Germany Influence of the tape number on the optimized

More information

CS 450 Numerical Analysis. Chapter 7: Interpolation

CS 450 Numerical Analysis. Chapter 7: Interpolation Lecture slides based on the textbook Scientific Computing: An Introductory Survey by Michael T. Heath, copyright c 2018 by the Society for Industrial and Applied Mathematics. http://www.siam.org/books/cl80

More information

THE preceding chapters were all devoted to the analysis of images and signals which

THE preceding chapters were all devoted to the analysis of images and signals which Chapter 5 Segmentation of Color, Texture, and Orientation Images THE preceding chapters were all devoted to the analysis of images and signals which take values in IR. It is often necessary, however, to

More information

Nonparametric Methods Recap

Nonparametric Methods Recap Nonparametric Methods Recap Aarti Singh Machine Learning 10-701/15-781 Oct 4, 2010 Nonparametric Methods Kernel Density estimate (also Histogram) Weighted frequency Classification - K-NN Classifier Majority

More information

Overcompressing JPEG images with Evolution Algorithms

Overcompressing JPEG images with Evolution Algorithms Author manuscript, published in "EvoIASP2007, Valencia : Spain (2007)" Overcompressing JPEG images with Evolution Algorithms Jacques Lévy Véhel 1, Franklin Mendivil 2 and Evelyne Lutton 1 1 Inria, Complex

More information

Inclusion of Aleatory and Epistemic Uncertainty in Design Optimization

Inclusion of Aleatory and Epistemic Uncertainty in Design Optimization 10 th World Congress on Structural and Multidisciplinary Optimization May 19-24, 2013, Orlando, Florida, USA Inclusion of Aleatory and Epistemic Uncertainty in Design Optimization Sirisha Rangavajhala

More information

Analysis and optimization methods of graph based meta-models for data flow simulation

Analysis and optimization methods of graph based meta-models for data flow simulation Rochester Institute of Technology RIT Scholar Works Theses Thesis/Dissertation Collections 8-1-2010 Analysis and optimization methods of graph based meta-models for data flow simulation Jeffrey Harrison

More information

Module 1 Lecture Notes 2. Optimization Problem and Model Formulation

Module 1 Lecture Notes 2. Optimization Problem and Model Formulation Optimization Methods: Introduction and Basic concepts 1 Module 1 Lecture Notes 2 Optimization Problem and Model Formulation Introduction In the previous lecture we studied the evolution of optimization

More information

1.2 Numerical Solutions of Flow Problems

1.2 Numerical Solutions of Flow Problems 1.2 Numerical Solutions of Flow Problems DIFFERENTIAL EQUATIONS OF MOTION FOR A SIMPLIFIED FLOW PROBLEM Continuity equation for incompressible flow: 0 Momentum (Navier-Stokes) equations for a Newtonian

More information

USE OF DIRECT AND INDIRECT DISTRIBUTIONS OF SELECTIVE MINING UNITS FOR ESTIMATION OF RECOVERABLE RESOURCE/RESERVES FOR NEW MINING PROJECTS

USE OF DIRECT AND INDIRECT DISTRIBUTIONS OF SELECTIVE MINING UNITS FOR ESTIMATION OF RECOVERABLE RESOURCE/RESERVES FOR NEW MINING PROJECTS USE OF DIRECT AND INDIRECT DISTRIBUTIONS OF SELECTIVE MINING UNITS FOR ESTIMATION OF RECOVERABLE RESOURCE/RESERVES FOR NEW MINING PROJECTS W Assibey-Bonsu 1 and D G Krige 2 1 Gold Fields Limited, South

More information

Tensor Based Approaches for LVA Field Inference

Tensor Based Approaches for LVA Field Inference Tensor Based Approaches for LVA Field Inference Maksuda Lillah and Jeff Boisvert The importance of locally varying anisotropy (LVA) in model construction can be significant; however, it is often ignored

More information

Detection and Mitigation of Cyber-Attacks using Game Theory

Detection and Mitigation of Cyber-Attacks using Game Theory Detection and Mitigation of Cyber-Attacks using Game Theory João P. Hespanha Kyriakos G. Vamvoudakis Correlation Engine COAs Data Data Data Data Cyber Situation Awareness Framework Mission Cyber-Assets

More information

PRIMAL-DUAL INTERIOR POINT METHOD FOR LINEAR PROGRAMMING. 1. Introduction

PRIMAL-DUAL INTERIOR POINT METHOD FOR LINEAR PROGRAMMING. 1. Introduction PRIMAL-DUAL INTERIOR POINT METHOD FOR LINEAR PROGRAMMING KELLER VANDEBOGERT AND CHARLES LANNING 1. Introduction Interior point methods are, put simply, a technique of optimization where, given a problem

More information

Time Series Analysis by State Space Methods

Time Series Analysis by State Space Methods Time Series Analysis by State Space Methods Second Edition J. Durbin London School of Economics and Political Science and University College London S. J. Koopman Vrije Universiteit Amsterdam OXFORD UNIVERSITY

More information

w KLUWER ACADEMIC PUBLISHERS Global Optimization with Non-Convex Constraints Sequential and Parallel Algorithms Roman G. Strongin Yaroslav D.

w KLUWER ACADEMIC PUBLISHERS Global Optimization with Non-Convex Constraints Sequential and Parallel Algorithms Roman G. Strongin Yaroslav D. Global Optimization with Non-Convex Constraints Sequential and Parallel Algorithms by Roman G. Strongin Nizhni Novgorod State University, Nizhni Novgorod, Russia and Yaroslav D. Sergeyev Institute of Systems

More information

Data Mining Chapter 8: Search and Optimization Methods Fall 2011 Ming Li Department of Computer Science and Technology Nanjing University

Data Mining Chapter 8: Search and Optimization Methods Fall 2011 Ming Li Department of Computer Science and Technology Nanjing University Data Mining Chapter 8: Search and Optimization Methods Fall 2011 Ming Li Department of Computer Science and Technology Nanjing University Search & Optimization Search and Optimization method deals with

More information

Image Transformation Techniques Dr. Rajeev Srivastava Dept. of Computer Engineering, ITBHU, Varanasi

Image Transformation Techniques Dr. Rajeev Srivastava Dept. of Computer Engineering, ITBHU, Varanasi Image Transformation Techniques Dr. Rajeev Srivastava Dept. of Computer Engineering, ITBHU, Varanasi 1. Introduction The choice of a particular transform in a given application depends on the amount of

More information

Integration. Volume Estimation

Integration. Volume Estimation Monte Carlo Integration Lab Objective: Many important integrals cannot be evaluated symbolically because the integrand has no antiderivative. Traditional numerical integration techniques like Newton-Cotes

More information

Package acebayes. R topics documented: November 21, Type Package

Package acebayes. R topics documented: November 21, Type Package Type Package Package acebayes November 21, 2018 Title Optimal Bayesian Experimental Design using the ACE Algorithm Version 1.5.2 Date 2018-11-21 Author Antony M. Overstall, David C. Woods & Maria Adamou

More information

NUMERICAL METHODS, NM (4776) AS

NUMERICAL METHODS, NM (4776) AS NUMERICAL METHODS, NM (4776) AS Objectives To provide students with an understanding that many mathematical problems cannot be solved analytically but require numerical methods. To develop a repertoire

More information