Mobile Robot Sensing for Environmental Applications

This paper reports the first application of iterative experimental design methodology for high spatiotemporal resolution characterization of river and lake aquatic systems performed using mobile robot sensing systems. Both applications involve dynamic phenomena spread over large spatial domain: 1) Characterization of contaminant concentration and flow at the confluence of two major rivers displaying dynamics due to flow of the water; and 2) Characterization of rapidly evolving biological processes such as phytoplankton dynamics in a lake system. We describe the development and application of a new general purpose method for mobile robot sensing in such environments - Iterative experiment Design for Environmental Applications (IDEA). IDEA introduces in-field adaptation of mobile robotic sensing system. Analysis of the complex spatial and temporal structures associated with each observed environment is presented. Detailed characterization of the observed environment using IDEA methodology is used as an informed prior to improve the performance of the existing adaptive experimental design approaches for mobile robotic systems - stratified adaptive sampling and hierarchical non-stationary Gaussian Processes.

[1]  Matthias W. Seeger,et al.  Gaussian Processes For Machine Learning , 2004, Int. J. Neural Syst..

[2]  Tamaki Ura,et al.  A new autonomous underwater vehicle designed for lake environment monitoring , 2002, Adv. Robotics.

[3]  Deborah Estrin,et al.  Networked Infomechanical Systems: A Mobile Wireless Sensor Network Platform , 2005 .

[4]  Maxim A. Batalin,et al.  NIMS RD: A Rapidly Deployable Cable Based Robot , 2007, Proceedings 2007 IEEE International Conference on Robotics and Automation.

[5]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[6]  Gaurav S. Sukhatme,et al.  IDEA: Iterative experiment Design for Environmental Applications , 2006 .

[7]  D. Caron,et al.  Networked Aquatic Microbial Observing System , 2006 .

[8]  M. Stealey,et al.  High Resolution River Hydraulic and Water Quality Characterization Using Rapidly Deployable Networked Infomechanical Systems (NIMS RD) , 2007 .

[9]  Gaurav S. Sukhatme,et al.  Adaptive sampling for environmental robotics , 2004, IEEE International Conference on Robotics and Automation, 2004. Proceedings. ICRA '04. 2004.

[10]  Andreas Krause,et al.  Near-Optimal Sensor Placements in Gaussian Processes: Theory, Efficient Algorithms and Empirical Studies , 2008, J. Mach. Learn. Res..

[11]  Paul D. Bates,et al.  The importance of spatial resolution in hydraulic models for floodplain environments , 1999 .

[12]  Joseph M. Wheaton.  Spawning Habitat Rehabilitation , 2003 .

[13]  Joseph M. Wheaton,et al.  Spawning habitat rehabilitation ‐I. Conceptual approach and methods , 2004 .

[14]  Steven W. Effler,et al.  Metabolic rate estimates for a eutrophic lake from diel dissolved oxygen signals , 2002, Hydrobiologia.