Applied estimation for hybrid dynamical systems using perceptional information

This dissertation uses the motivating example of robotic tracking of mobile deep ocean animals to present innovations in robotic perception and estimation for hybrid dynamical systems. An approach to estimation for hybrid systems is presented that utilizes uncertain perceptional information about the system’s mode to improve tracking of its mode and continuous states. This results in significant improvements in situations where previously reported methods of estimation for hybrid systems perform poorly due to poor distinguishability of the modes. Even in applications where the modes are more easily distinguished, the approach presented can improve performance by decreasing the mode estimation delay of the estimator. The specific application that motivates this research is an automatic underwater robotic observation system that follows and films individual deep ocean animals. A first version of such a system has been developed jointly by the Stanford Aerospace Robotics Laboratory and Monterey Bay Aquarium Research Institute (mbari). This robotic observation system is successfully fielded on mbari’s rovs, but agile specimens often evade the system. When a human rov pilot performs this task, one advantage that he has over the robotic observation system in these situations is the ability to use visual perceptional information about the target. The human pilot can immediately recognize any changes in the specimen’s behavior such as whether it is actively swimming or not and react to those mode changes with the necessary forceful measures required to prevent losing sight of the specimen. With the approach of the human pilot in mind, a new version of the robotic observation system is proposed which is extended to (a) derive perceptional information (visual cues) about the behavior mode of the tracked specimen, and (b) merge this dissimilar, discrete and uncertain information with more traditional continuous noisy sensor data by extending existing algorithms for hybrid estimation. These performance enhancements are enabled by integrating a wide range of techniques in hybrid estimation, computer

[1]  Dragomir Anguelov,et al.  A General Algorithm for Approximate Inference and Its Application to Hybrid Bayes Nets , 1999, UAI.

[2]  Robert J. Elliott,et al.  Modal estimation in hybrid systems , 2000 .

[3]  Stephen M. Rock,et al.  Visual tracking of jellyfish in situ , 2001, Proceedings 2001 International Conference on Image Processing (Cat. No.01CH37205).

[4]  Bruce H. Robison,et al.  Deep pelagic biology , 2004 .

[5]  William M. Hamner,et al.  Predatory behavior ofPhacellophora camtschatica and size-selective predation uponAurelia aurita (Scyphozoa: Cnidaria) in Saanich Inlet, British Columbia , 1988 .

[6]  Stephen M. Rock,et al.  Segmentation methods for visual tracking of deep-ocean jellyfish using a conventional camera , 2003 .

[7]  Gaurav S. Sukhatme,et al.  Bias Reduction and Filter Convergence for Long Range Stereo , 2005, ISRR.

[8]  B. Anderson,et al.  Recursive identification of switched ARX hybrid models: exponential convergence and persistence of excitation , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).

[9]  Stephen M. Rock,et al.  A pilot-aid for ROV based tracking of gelatinous animals in the midwater , 2001, MTS/IEEE Oceans 2001. An Ocean Odyssey. Conference Proceedings (IEEE Cat. No.01CH37295).

[10]  Claire J. Tomlin,et al.  Target tracking and Estimated Time of Arrival (ETA) Prediction for Arrival Aircraft , 2006 .

[11]  Aaron M. Plotnik,et al.  A Multi-Sensor Approach to Automatic Tracking of Midwater Targets by an ROV , 2005 .

[12]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[13]  Bruce H. Robison,et al.  Midwater research methods with MBARI'S ROV , 1992 .

[14]  D.D. Sworder,et al.  Data fusion using multiple models , 2000, Conference Record of the Thirty-Fourth Asilomar Conference on Signals, Systems and Computers (Cat. No.00CH37154).

[15]  Stephen M. Rock,et al.  Field experiments in the control of a Jellyfish tracking ROV , 2002, OCEANS '02 MTS/IEEE.

[16]  Feng Zhao,et al.  Estimation of Distributed Hybrid Systems Using Particle Filtering Methods , 2003, HSCC.

[17]  Aaron M. Plotnik,et al.  IMPROVING PERFORMANCE OF A JELLY-TRACKING UNDERWATER VEHICLE USING RECOGNITION OF ANIMAL MOTION MODES , 2003 .

[18]  C. Mills,et al.  In situ observations of the behavior of mesopelagic Solmissus narcomedusae (Cnidaria, Hydrozoa) , 1988 .

[19]  S. Shankar Sastry,et al.  Observability of Linear Hybrid Systems , 2003, HSCC.

[20]  S. Rock,et al.  RELATIVE POSITION SENSING AND AUTOMATIC CONTROL FOR OBSERVATION IN THE MIDWATER BY AN UNDERWATER VEHICLE , 2005 .

[21]  David D. Sworder,et al.  Estimation Problems in Hybrid Systems , 1999 .

[22]  Stephen M. Rock,et al.  Design and Validation of a Robotic Control Law for Observation of Deep-Ocean Jellyfish , 2006, IEEE Transactions on Robotics.

[23]  Stephen M. Rock,et al.  Automated robotic tracking of gelatinous animals in the deep ocean , 2003 .

[24]  Aaron M. Plotnik,et al.  Quantification of cyclic motion of marine animals from computer vision , 2002, OCEANS '02 MTS/IEEE.

[25]  R. J. Larson.  Costs of transport for the scyphomedusa Stomolophus meleagris L. Agassiz , 1987 .

[26]  Sebastian Thrun,et al.  FastSLAM: a factored solution to the simultaneous localization and mapping problem , 2002, AAAI/IAAI.

[27]  Stephen Rock,et al.  Improved Estimation of Target Velocity Using Multiple Model Estimation and a Dynamic Bayesian Network for a Robotic Tracker of Ocean Animals , 2005, ISRR.

[28]  S. Thrun,et al.  Tractable Particle Filters for Robot Fault Diagnosis , 2004 .

[29]  T. Başar,et al.  A New Approach to Linear Filtering and Prediction Problems , 2001 .

[30]  Claudia E. Mills,et al.  In Situ Foraging and Feeding Behaviour of Narcomedusae (Cnidaria: Hydrozoa) , 1989, Journal of the Marine Biological Association of the United Kingdom.

[31]  Bruce H. Robison,et al.  The coevolution of undersea vehicles and deep-sea research , 1999 .

[32]  S.M. Rock,et al.  Visual Servoing of an ROV for Servicing of Tethered Ocean Moorings , 2006, OCEANS 2006.