Brief bio

Lantao Liu is an Associate Professor in the Department of Intelligent Systems Engineering at Indiana University-Bloomington. His main research interest lies in Autonomy that integrates real physical robotic systems with data driven methods. He has been working on various autonomous systems (air, ground, aquatic) involving single or multiple robots,  and his various unmanned vehicles have been deployed to the real world with “field trials” in those complex and unstructured environments. Before joining Indiana University, he was a Postdoctoral Research Associate in the Department of Computer Science at the University of Southern California during 2015 - 2017. He also worked as a Postdoctoral Fellow in the Robotics Institute at Carnegie Mellon University during 2013 - 2015. He received a Ph.D. from the Department of Computer Science and Engineering at Texas A&M University in 2013.

TALK TITLE

Information-driven path planning for ocean sampling and mapping

ABSTRACT

In this talk, I will present two of our recent endeavors focused on the sampling and mapping of oceanic environments.  Our primary goal is to tackle the challenge of efficiently mapping an initially unknown aquatic environment by deploying a surface robot equipped with sensors that provide sparse sensing measurements. A practical example of this is constructing a pollution distribution map, where environmental sensors can only collect pollutant concentration data at specific points, resulting in sparse measurements along the robot's trajectory.  To optimize the robot's sensing path, we often rely on predictions from an underlying probabilistic model, leveraging its uncertainty estimates to pinpoint critical areas for gathering informative data.  I will first introduce a new kernel design that enhances the sampling robot's ability to characterize the map with greater precision, improving both model accuracy and uncertainty quantification. Following this, I will discuss a path planning method based on tree-search algorithms. This method enables the sampling robot to handle potentially conflicting objectives, such as optimizing exploration of a large environment while simultaneously exploiting points of interest within limited time constraints.