Doctoral Thesis Proposal - Arjun Laksmipathy

— 1:45pm

Location:
In Person and Virtual - ET - Gates Hillman 8102 and Zoom

Speaker:
ARJUN LAKSMIPATHY, Ph.D. Student, Computer Science Department, Carnegie Mellon University
https://www.linkedin.com/in/arjun-lakshmipathy-a48b8535

Humans use their hands to effortlessly manipulate objects of arbitrarily complex geometries and physical properties every day; however, adapting such manipulations to dexterous robots and virtual characters is an extremely difficult task. Understanding the ways in which humans exploit contact to perform these manipulations has the potential to greatly advance progress towards this goal. 

Unsurprisingly, research efforts have analyzed contact in the context of dexterous manipulation for decades. We now have numerous metrics for evaluating grasp quality in terms of contacts, efficient means of computing contact in physical simulation, and countless strategies exploiting contact correspondences between hands and objects to synthesize grasps and manipulations. But the majority of existing works fundamentally characterize contact in the same way: as points, lines, or planes of interaction. 

But contacts in the real world are much more complicated. Instead, real bodies interface with one another via areas of contact which greatly vary with the geometries of contacting surfaces. If we wish to model the complexities of manipulations as they actually occur, then we must progress beyond such simplifying assumptions and deal with the messy nature of reality. This thesis aims to do so by presenting frameworks and algorithms for the modeling, capture, mutation, and exploitation of contact areas. Our intention is to establish the foundations necessary to elevate contact areas to first-class primitives and demonstrate their inherent value across a range of practical applications in dexterous manipulation and adjacent domains.    

First, we introduce three novel models of contact areas alongside operations supported by each model which are fundamentally designed to run on real geometries rather than simple primitive shapes. Second, we introduce a low cost approach for capturing human contacts from the real world. Then, using these models, we introduce: a new set of intuitive artist tools for drafting high quality grasps, a novel motion retargeting pipeline for dexterous manipulations, a novel contact-driven control framework for dexterous robot hands, and two practical extensions of our work. The contributions in this thesis are not intended to be the last words, but rather important first steps designed to promote future research in contact area modeling.

Thesis Committee

Nancy S. Pollard (Chair)
Jessica K. Hodgins
Keenan Crane
C. Karen Liu (Stanford University)

Additional Information

In Person and Zoom Participation.  See announcement.

Event Website:
https://csd.cmu.edu/calendar/doctoral-thesis-proposal-arjun-laksmipathy


Add event to Google
Add event to iCal