On-board Adaptive Informative Sampling for AUVs: a Feasibility Study

Autonomous underwater vehicles (AUVs) can be used for sampling our lakes and oceans to measure temperature, Chlorophyll abundance, or other physical properties of the water. A commonly proposed method for collecting informative data in robotics is informative sampling or informative path planning. In informative sampling, information-theoretic metrics are used on top of a created model to decide where to deploy sensors or robots. If a model of the environment is already available, this can be done off-line. Alternatively a model can be learned and optimized for on-line, in adaptive informative sampling (AIS). In this paper we report on field trials where we assess the feasibility of running AIS on board a commercial-off-the-shelf AUV. We propose a method of quantifying AIS field performance by using bathymetric data. We show that it is possible to create a Gaussian Process model on board a small commercial-off-the-shelf AUV, and to use this model to determine waypoints for sampling. Our field results show similar performance curves between full and half duration runs, resulting in good models of bathymetry and Chlorophyll fields.

[1]  P. Holgate.  Lognormal Distributions: Theory and Applications , 1989 .

[2]  Lehel Csató,et al.  Sparse On-Line Gaussian Processes , 2002, Neural Computation.

[3]  Andreas Krause,et al.  Efficient Planning of Informative Paths for Multiple Robots , 2006, IJCAI.

[4]  Andreas Krause,et al.  Nonmyopic active learning of Gaussian processes: an exploration-exploitation approach , 2007, ICML '07.

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

[6]  K. H. Low,et al.  Multi-robot adaptive exploration and mapping for environmental sensing applications , 2009 .

[7]  John J. Leonard,et al.  An Overview of MOOS-IvP and a Brief Users Guide to the IvP Helm Autonomy Software , 2009 .

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

[9]  David R. Thompson,et al.  Autonomous science during large‐scale robotic survey , 2011, J. Field Robotics.

[10]  Kian Hsiang Low,et al.  Decentralized active robotic exploration and mapping for probabilistic field classification in environmental sensing , 2012, AAMAS.

[11]  Gaurav S. Sukhatme,et al.  Optimizing waypoints for monitoring spatiotemporal phenomena , 2013, Int. J. Robotics Res..

[12]  Geoffrey A. Hollinger,et al.  Sampling-based robotic information gathering algorithms , 2014, Int. J. Robotics Res..

[13]  Alkis Gotovos,et al.  Fully autonomous focused exploration for robotic environmental monitoring , 2014, 2014 IEEE International Conference on Robotics and Automation (ICRA).

[14]  Kian Hsiang Low,et al.  Multi-robot active sensing of non-stationary gaussian process-based environmental phenomena , 2014, AAMAS.

[15]  Gaurav S. Sukhatme,et al.  Adaptive informative sampling with autonomous underwater vehicles: Acoustic versus surface communications , 2016, OCEANS 2016 MTS/IEEE Monterey.

[16]  Mac Schwager,et al.  Robust Adaptive Coverage Control for Robotic Sensor Networks , 2017, IEEE Transactions on Control of Network Systems.

[17]  Roland Siegwart,et al.  Adaptive continuous‐space informative path planning for online environmental monitoring , 2017, J. Field Robotics.

[18]  Gaurav S. Sukhatme,et al.  Pilot Surveys for Adaptive Sampling , 2018, ICRA 2018.

[19]  Gaurav S. Sukhatme,et al.  Data‐driven learning and planning for environmental sampling , 2017, J. Field Robotics.