Academic Commons Search Results
http://academiccommons.columbia.edu/catalog.rss?f%5Bauthor_facet%5D%5B%5D=Roberts%2C+Kenneth+S.&f%5Bgenre_facet%5D%5B%5D=Technical+reports&q=&rows=500&sort=record_creation_date+desc
Academic Commons Search Resultsen-usLearning the Probability Distribution for Probabilistic Expert Systems
http://academiccommons.columbia.edu/catalog/ac:143336
Baker, Michelle; Roberts, Kenneth S.http://hdl.handle.net/10022/AC:P:12218Tue, 17 Jan 2012 00:00:00 +0000This paper describes a method for learning the joint probability distribution of a set of variables from a sample of instances from the domain. The method is based on a straightforward application of Bayes Law to the problem of estimating individual probabilities from a probability distribution. We use a maximum entropy distribution as an initial estimate and show how this estimate can be easily updated each time an additional example is observed. Although developed for the purpose of estimating the conditional probabilities required for Bayesian inference networks, this method can be adopted to simplify knowledge acquisition in any expert system that uses knowledge in the form of probabilities.Computer scienceComputer ScienceTechnical reportsRobot Active Touch Exploration: Constraints and Strategies
http://academiccommons.columbia.edu/catalog/ac:143204
Roberts, Kenneth S.http://hdl.handle.net/10022/AC:P:12188Wed, 11 Jan 2012 00:00:00 +0000We investigate the problem of using active touch ("haptic") exploration to recognize a 3D object taken from a known set of models. "That is new is that we combine two approaches: (1) using geometric constraints between components to eliminate interpretations, and interpretation tree methods for choosing the best active sensing move; (2) exploratory moves made by tracing continually along the surface of the object (and not through free space). We restrict ourselves to polyhedral, and give a set of geometric constraints tailored for matching components acquired from haptic exploration against components in the models. We present a new constraint using pairs of line segments. We then give a set of active sensing moves, each with an associated cost measure, and our strategies for choosing the next move.Computer scienceComputer ScienceTechnical reportsCoordinating a Robot Arm and: Multi-finger Hand Using the Quaternion Representation
http://academiccommons.columbia.edu/catalog/ac:143207
Roberts, Kenneth S.http://hdl.handle.net/10022/AC:P:12189Wed, 11 Jan 2012 00:00:00 +0000An autonomous hand-arm system must be able to move its multiple fingertips to specified 3-D world locations. We present an algorithm which, given a desired position and normal for each fingertip, computes all the joint angles for the fingers and arm. Our primary method for handling this under-constrained problem is to optimize a cost function. We also give methods for generating good candidates to be optimized. We present several new techniques for using the quaternion form to optimize rotation. We give experimental results from using the algorithm to apply complicated grasps with a Utah/MIT hand-arm system.Computer scienceComputer ScienceTechnical reportsAn Integrated System for Dextrous Manipulation
http://academiccommons.columbia.edu/catalog/ac:142841
Allen, Peter K.; Michelman, Paul; Roberts, Kenneth S.http://hdl.handle.net/10022/AC:P:12076Wed, 21 Dec 2011 00:00:00 +0000This paper describes an integrated system for dextrous manipulation using a Utah-MIT hand that allows one to look at the higher levels of control in a number of grasping and manipulation tasks. The system consists of a number of low-level system primitives for grasping, integrated hand and robotic arm movement, tactile sensors mounted on the fingertips, sensing primitives to utilize joint position, tendon force and tactile array feedback, and a high-level programming environment that allows task level scripts to be created for grasping and manipulation tasks. A number of grasping and manipulation tasks are described that have been implemented with this system.Computer science, Roboticspka1Computer ScienceTechnical reportsHaptic Object Recognition Using A Multi-Fingered Dextrous Hand
http://academiccommons.columbia.edu/catalog/ac:142735
Allen, Peter K.; Roberts, Kenneth S.http://hdl.handle.net/10022/AC:P:12042Mon, 19 Dec 2011 00:00:00 +0000It is becoming increasingly clear that robotic systems need to have capabilities similar to the human haptic system in order to perform complex grasping, manipulation and object recognition tasks using dextrous hands. This paper is an exploration of using a dextrous, multi-fingered hand for high-level object recognition tasks. The paradigm is model-base recognition in which the objects are modeled and recovered as superquadrics, which are shown to have a number of important attributes that make them well suited for such a task. Experiments have been performed to recover the shape of objects using sparse contact point data from the hand with promising results. We also present our approach to using tactile data in conjunction with the dextrous hand to build a library of grasping and exploration primitives that can be used in recognizing and grasping more complex multi-part object.Computer sciencepka1Computer ScienceTechnical reportsEquivalent Descriptions of Generalized Cylinders
http://academiccommons.columbia.edu/catalog/ac:141271
Roberts, Kenneth S.http://hdl.handle.net/10022/AC:P:11745Mon, 07 Nov 2011 00:00:00 +0000The "equivalence" problem for shape descriptions is that a single three-dimensional shape may have several different descriptions. The Slant Theorem (Shafer) for equivalent generalized cylinder descriptions was proven under the restrictions that the same radius function and the same axis be used for all the descriptions. A proof is given that the theorem still holds when the "same radius function" condition is removed. It does not hold when the "same axis" condition is removed. The ellipsoid is a. counter-example.Computer scienceComputer ScienceTechnical reports