W&M ScholarWorks Dissertations, Theses, and Masters Projects Theses, Dissertations, & Master Projects 2014 Use of Pattern Classification Algorithms ot Interpret Passive and Active Data Streams from a Walking-Speed Robotic Sensor Platform Eric Allen Dieckman College of William & Mary - Arts & Sciences Follow this and additional works at: https://scholarworks.wm.edu/etd Part of the Acoustics, Dynamics, and Controls Commons, and the Robotics Commons Recommended Citation Dieckman, Eric Allen, "Use of Pattern Classification Algorithms ot Interpret Passive and Active Data Streams from a Walking-Speed Robotic Sensor Platform" (2014). Dissertations, Theses, and Masters Projects. Paper 1539623643. https://dx.doi.org/doi:10.21220/s2-5cy6-jk36 This Dissertation is brought to you for free and open access by the Theses, Dissertations, & Master Projects at W&M ScholarWorks. It has been accepted for inclusion in Dissertations, Theses, and Masters Projects by an authorized administrator of W&M ScholarWorks. For more information, please contact [email protected]. Use of Pattern Classification Algorithms to Interpret Passive and Active Data Streams from a Walking-Speed Robotic Sensor Platform Eric Allen Dieckman Wellington, Missouri M.S. Architectural Acoustics, Rensselaer Polytechnic Institute, 2009 B.S. Physics, Truman State University, 2008 A Dissertation presented to the Graduate Faculty of the College of William and Mary in Candidacy for the Degree of Doctor of Philosophy Departm ent of Applied Science The College of William and Mary M ay 2014 APPROVAL PAGE This Dissertation is submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Eric A. Dieckman Approved by the Committee, October 2013 Committee Chair Professor Mark Hinders, Applied Science The College of William and Mary ifessor Gunter Luepke, Appliecr Science The College of William and Mary Research Assistant Professor Saskia MordijcK Computer Science The College of William and Mary Dr. William Fehlman II U.S. Army Capabilities Integration Center ABSTRACT In order to perform useful tasks for us, robots must have the ability to notice, recognize, and respond to objects and events in their environment. This requires the acquisition and synthesis of information from a variety of sensors. Here we investigate the performance of a number of sensor modalities in an unstructured outdoor environment, including the Microsoft Kinect, thermal infrared camera, and coffee can radar. Special attention is given to acoustic echolocation measurements of approaching vehicles, where an acoustic parametric array propagates an audible signal to the oncoming target and the Kinect microphone array records the reflected backscattered signal. Although useful information about the target is hidden inside the noisy time domain measurements, the Dynamic Wavelet Fingerprint process (DW FP) is used to create a time-frequency representation of the data. A small-dimensional feature vector is created for each measurement using an intelligent feature selection process for use in statistical pattern classification routines. Using out experimentally measured data from real vehicles at 50 m, this process is able to correctly classify vehicles into one of five classes w ith 94% accuracy. F u lly three-dimensional simulations allow us to study the nonlinear beam propagation and interaction with real-world targets to improve classification results. T a b l e o f C o n t e n t s Acknowledgments iv Dedication v List of Figures vi List of Tables x List of Symbols xii 1 Overview 1 2 Finding information in noisy data 6 2.1 Time-frequency analysis of signals .............................................................. 7 2.1.1 Fourier transform .............................................................................. 7 2.1.2 Short-time Fourier transform ........................................................ 8 2.1.3 Other methods of time-frequency analysis ................................. 10 2.1.4 Wavelets .............................................................................................. 11 2.2 The Dynamic Wavelet Fingerprint (D W F P ) ........................................... 18 2.2.1 Feature creation ................................................................................. 19 2.2.2 Feature extraction ........................................................................... 22 2.3 An application: Ultrasonic detection of flaws in microelectronics . 24 2.3.1 Background........................................................................................ 25 2.3.2 Ultrasonic measurements of microelectronics .............................. 25 2.3.3 Creating flawed microelectronics samples .................................... 30 2.3.4 DWFP analysis................................................................................. 33 2.3.5 Application to analysis of other signals........................................ 38 3 Statistical pattern classification 40 3.1 Statistical pattern classification by example: analysis of glass.............. 41 3.2 Feature extraction and selection ................................................................. 47 3.3 Classification.................................................................................................. 50 3.3.1 Parametric classifiers........................................................................ 51 3.3.2 Nonparametric classifiers................................................................. 53 3.3.3 Improving classifier performance .................................................... 58 i 3.4 Visualizing classification results................................................................. 64 3.5 Some results .................................................................................................... 69 4 rM ary - a walking-speed mobile sensor platform 72 4.1 Towards an autonomous walking-speed robotic platform....................... 74 4.1.1 Sensor modalities for mobile robots ............................................... 75 4.2 Investigation of sensor modalities using r M a r y ....................................... 77 4.2.1 Thermal infrared (IR) ..................................................................... 79 4.2.2 Kinect .................................................................................................. 82 4.2.3 Audio .................................................................................................. 93 4.2.4 Radar .................................................................................................. 98 5 Acoustic echolocation from a mobile robot 102 5.1 Vehicle classification.................................................................................... 102 5.2 Acquiring acoustic echolocation d a ta ....................................................... 104 5.3 Initial data analysis....................................................................................... 110 5.3.1 Detecting a signal reflected from a w a ll ........................................ 110 5.3.2 Data from oblique angles .................................................................. 112 6 Classification of oncoming vehicles using acoustic echolocation 118 6.1 Pattern classification.................................................................................... 118 6.1.1 Compiling d a ta .................................................................................. 119 6.1.2 Aligning reflected signals .................................................................. 120 6.1.3 Feature creation with DWFP ......................................................... 127 6.1.4 Intelligent feature selection ............................................................... 130 6.1.5 Statistical pattern classification..................................................... 135 6.2 Results.............................................................................................................. 136 6.2.1 Proof-of-concept: Acoustic classification of stationary vehicles 137 6.2.2 Acoustic classification of oncoming vehicles.................................. 139 7 Simulations of scattering from a nonlinear acoustic beam 155 7.1 The acoustic parametric array .................................................................... 156 7.1.1 Modeling nonlinear acoustic wave propagation ........................... 160 7.1.2 Numerical solutions of the KZK equation ..................................... 163 7.2 Simulations of acoustic scattering............................................................. 170 7.2.1 AFIT .................................................................................................. 172 7.2.2 Implementation of AFIT on the SciClone computing cluster . 177 7.3 Simulations of acoustic scattering from vehicles .................................... 179 7.3.1 Creating vehicle scatterers ............................................................... 180 7.3.2 The effect of incident pulse length on backscattered reflection 188 7.3.3 Simulated microphone data from the backscattered pressure field203 8 Future directions in sensor fusion for mobile robotics 218 ii 8.1 Sensor modalities .......................................................................................... 218 8.2 Future directions in sensor fu s io n ............................................................. 220 223 Appendices 266 Bibliography 277 V ita iii Acknowledgments I would like to thank my advisor Dr. Mark Hinders for his guidance, assistance, and discussion