Session: Rising Stars of Mechanical Engineering Celebration & Showcase

Paper Number: 147952

147952 - Career: Exploring Robust Robot Manipulation Through Compliance- and Motion-Based Manipulation Funnels 

Given the rapidly increasing interests in highly dynamic manipulation in unstructured environments, the capability of dealing with uncertainties has become an unprecedentedly critical consideration, since any small uncertainty can potentially result in large differences in manipulation outcomes as the errors rapidly accumulate through the process. The proposed work seeks to improve the robustness of manipulation systems through the lens of “manipulation funnels”. By broadening and generalizing the funnel concept, we are interested in investigating novel manipulation formulations, beyond its geometric models, that guarantee to map system states from a large input space to a strictly smaller output space – a natural mechanism to significantly reduce the uncertainties. As enabled by the recent fast development in compliant robot hardware and high-efficiency algorithms, we will study funnels that can be dynamically identified during a manipulation process, or computationally formulated in any task-relevant spaces. In particular, we aim to enable robots to dexterously manipulate through: 1) compliance-based funnels that can smoothly steer itself to complete tasks; 2) motion-based funnels that can actively establish “uncertainty-tolerating” manipulation actions; and 3) composed and transferred funnels to generalize the manipulation skills in novel and complex tasks.

Presenting Author: Kaiyu Hang Rice University

Presenting Author Biography: Kaiyu Hang is an Assistant Professor of Computer Science at Rice University, where he leads the Robotics and Physical Interactions Lab. He is broadly interested in robotic systems that can physically interact with other robots, people, and the world. By developing algorithms in optimization, planning, learning, estimation, and control, his research is focused on efficient, robust, and generalizable manipulation systems, addressing problems that range from small scale grasping and in-hand manipulation, to large scale dual-arm manipulation, mobile manipulation, and multi-robot manipulation.

Authors: