In this talk, I will present our recent advancements in autonomous robotic systems, spanning unmanned ground, aerial, and aquatic vehicles. I will first discuss a data-driven adaptive sensing framework designed for rapid modeling and mapping of unknown environments, such as expansive ocean floors or spatiotemporal air/water pollution.
Central to this framework is the development of a novel family of nonstationary kernels, called Attentive Kernels, used in Gaussian Process Regression. These kernels enable efficient learning of underlying functions, such as capturing critical environmental variations in 3D maps, and have been successfully validated in field experiments using unmanned surface vehicles.
Lantao Liu,
Indiana University-Bloomington
Next, I will introduce a unified framework that addresses multiple key objectives of robotic autonomy: identifying free space for navigation, constructing metric-topological maps, and optimizing spatial coverage for exploration. Unlike traditional approaches that handle these tasks separately, our framework integrates them using a sparse variant of Gaussian Processes, allowing seamless navigation, mapping, and exploration in unknown environments like wild forests, validated by outdoor field results.
Lastly, I will share our progress in stochastic motion planning and control, where the value functions are represented using kernel methods. This approach enables autonomous vehicles to maneuver through cluttered, complex environments under motion disturbances. Field results from unmanned ground vehicles reveal the system’s effectiveness in real-world challenging scenarios.
Lantao Liu is an associate professor in the Departments of Intelligent Systems Engineering and Computer Science at Indiana University-Bloomington, specializing in autonomy, robotics, and AI. His research focuses on developing autonomous systems for diverse environments, with a particular interest in unmanned aerial, ground, and aquatic vehicles.
His robotic systems have been tested in real-world field trials in challenging environments such as construction sites, emergency zones, farmlands, outdoor aquatic areas, and, recently, motor speedways. Dr. Liu’s work has garnered multiple best paper nominations and awards at important robotics conferences like Robotics Science and Systems (RSS), the International Conference on Intelligent Robots and Systems (IROS), and the International Symposium on Distributed Autonomous Robotic Systems (DARS).
Before his role at Indiana University, he was a postdoctoral research associate at the University of Southern California’s Department of Computer Science from 2015 to 2017 and a Postdoctoral Fellow at Carnegie Mellon University’s Robotics Institute from 2013 to 2015. He earned his Ph.D. in computer science and engineering from Texas A&M University in 2013.