Accession Number : ADA296903

Title :   Learning Maneuvers Using Neural Network Models.

Descriptive Note : Final rept. 1 Apr 93-31 Mar 94,

Corporate Author : MASSACHUSETTS INST OF TECH CAMBRIDGE ARTIFICIAL INTELLIGENCE LAB

Personal Author(s) : Atkeson, Christopher

PDF Url : ADA296903

Report Date : 31 MAR 1994

Pagination or Media Count : 128

Abstract : These researchers explored issued involved in implementing robot learning for a challenging dynamic task, using a case study from robot juggling. They used a memory based local modeling approach (locally weighted regression) to represent a learned model of the task to be performed. Statistical tests are given to examine the uncertainty of a model, to optimize its prediction quality, and to deal with noisy and corrupted data. They developed an exploration algorithm that explicitly deals with prediction accuracy requirement during exploration. Using all these ingredients in combination with methods from optimal control, the robot achievers fast real-time learning of the task within 40 to 100 trials.(KAR) P. 1

Descriptors :   *MATHEMATICAL MODELS, *UNCERTAINTY, *NEURAL NETS, *STATISTICAL TESTS, ALGORITHMS, CONTROL, REQUIREMENTS, OPTIMIZATION, PREDICTIONS, REAL TIME, DYNAMICS, ROBOTS, ACCURACY, REGRESSION ANALYSIS, WEIGHTING FUNCTIONS, CASE STUDIES, QUALITY, MANEUVERS, LEARNING.

Subject Categories : Statistics and Probability

Distribution Statement : APPROVED FOR PUBLIC RELEASE