Home
About
The Intelligent Motion Lab studies motion planning and control for intelligent robotic systems and its relationships to motor cognition in biological systems. Our research is particularly interested in computational and information-theoretic issues in complex, high-dimensional motion spaces, and seeks applications to domestic and industrial robotic manipulation, human-robot interaction, robot- and computer-assisted surgery, and legged locomotion. The lab is directed by Prof. Kris Hauser and is part of the Indiana University School of Informatics and Computing.
IML combines high-performance computing and simulation tools with a variety of robot hardware to bridge theory to practice. Our overall goal is to contribute new tools, techniques, and theories that will help machines perform sophisticated physical tasks with high value to society.
Research Areas
For more information, please see the research page.
Versatile, Robust, and Ubiquitous Manipulation
We are studying methods for forming, adapting, and executing sophisticated manipulation skills with applications to domestic and industrial robots. Particular interests include multi-handed, multi-contact reasoning, dealing with impoverished sensors and low-fidelity actuators, and non-grasping manipulation. Our long-term goals include versatile manipulation, which aims to give robots the ability to interact in a wide range of useful ways with a wide range of objects, and ubiquitous manipulation, which aims to provide all robots the intelligence needed to manipulate objects to the best of their capabilities.
Theory and Implementation of Planning Algorithms
As robots are demanded to perform increasly sophisticated tasks, motion planning software will become increasingly complex. To be reliable, capable, and efficient as software complexity grows, planners and their components must be based on sound theoretical principles. We have developed a multi-modal planner, called Random-MMP, for a particular type of hybrid system that arises in the presence of making and breaking contacts. It is provably reliable and scalable, and has been applied to manipulation and legged locomotion problems in configuration spaces with up to 42 dimensions.
Robot- and Computer-Assisted Surgery
To improve the accuracy and consistency of robot-assisted minimally invasive surgery, we are working with collaborators in Johns Hopkins, UNC Chapel Hill, and UC Berkeley on motion planning and control techniques to predict and compensate for disturbances encountered in surgery. Our work is applied to robot-guided needle insertion and tissue manipulation problems. We are also studying how computer simulation and motion planning can aid in clinician training for novel medical devices, such as a recently developed class of steerable needle.
Motion Complexity Theory
The geometric complexity of an robot's motion space strongly affects the difficulty of operating it. Navigation in "hard" spaces will demand greater computational resources, more sophisticated planning and control algorithms, or more engineering effort to devise good heuristic strategies. We seek a fundamental notion of information that applies to both biological and robotic systems, which includes information gathered on-line using computation, sensors, and cognition, as well as information encoded within the engineering or evolution processes that designed the system. Major goals of this effort are to study and design information-optimal algorithms and to study the effects of complexity on biological systems.