Collaborative Research: Motion -- Models, Algorithms, and Complexity

Stanford University, Stanford CA

Investigators

Abstract

Pankaj Argarwal Duke University Collaborative 0204118/0204486 Motion, like shape, is one of the fundamental modalities to be modeled in order to represent and manipulate the phys-ical world in a computer. As such, motion representations and the algorithms that operate on them are central to all computational disciplines dealing with physical objects: computer graphics, computer vision, robotics, etc. Mod-eling motion is also crucial for other disciplines dealing with temporally varying data, including mobile networks, temporal databases, etc. Motion algorithms require com-putational resources, and frequently sensing and commu-nication resources as well, in order to accomplish their task. Despite the prominent position that motion plays in so many computer disciplines, little has been done to date to provide a clean conceptual framework for representing motion, describing algorithms on moving objects, and an-alyzing their behavior and performance. It is the goal of this proposal to push the frontiers of the computer sci-ence representations and algorithms that deal with motion in its various manifestations. The aim is to create a firm algorithmic theory for the processes of acquiring, model-ing, reasoning about, planning, manipulating, or executing motion and their relevant complexity measures.