Title: Unfolding the Potential of Point-Based Correspondences for Cloth Manipulation

Abstract:

Robotic cloth manipulation is an active area of research with numerous applications in domestic and industrial environments. However, prior work in this field have limitations that restrict their applicability in real-world scenarios, such as handling unaligned configurations and requiring subgoals for long-horizon tasks. The objective of this thesis is to develop an approach for cloth manipulation that is efficient, robust, and capable of long-horizon planning. To achieve this, we propose the utilization of point-based correspondences, which capture geometric relationships and deformations in cloth surfaces. Leveraging these correspondences can enhance the accuracy and reliability of cloth manipulation algorithms. In this work, we present a general framework for cloth manipulation that incorporates the use of point-based correspondences. We focus on fundamental cloth manipulation tasks such as folding, smoothing, and alignment. Through extensive experiments and evaluations, we demonstrate the effectiveness of our proposed approach, surpassing state-of-the-art methods. Experiments comparing our methods against existing techniques highlight the advantages of using point-based correspondences to achieve efficient, robust, long-horizon cloth manipulation.

Committee:

Prof. David Held (advisor)

Prof. Oliver Kroemer

Prof. Shubham Tulsiani

Alex LaGrassa

Zoom link: https://cmu.zoom.us/j/93763495998?pwd=QU42Zm9PTGxXdDd1dzg5ekRUT1JPdz09

Meeting ID: 937 6349 5998
Passcode: 304538