Joseph C Musto

Entropy-based reliability analysis for intelligent machines

A new metric for performance assessment of intelligent machines has been developed. The method fuses concepts from the Theory of Intelligent Machines proposed by Saridis (1988) with traditional reliability analysis in the development of a measure which reflects both the uncertainty inherent in the intelligent machine and the uncertainty allowed by the task description. The metric is entropy based, and is shown to be analogous to a measure of system reliability.

Task reliability assessment in robotic systems

A new method for assessing the performance of a robotic system performing a high-precision task has been developed. The method utilizes an entropy-based formulation to provide a measure of the reliability of a robotic system with respect to a task specification. The derivation of the proposed technique is presented, including two alternative formulations that differ in terms of their complexity and level of approximation. Methods for combining these measures for complex robotic tasks are detailed. The new method is demonstrated in a case study involving a robotic assembly system, in which the entropy formulation is used to select an optimal system design from a set of alternatives. (c) 1995 John Wiley & Sons, Inc.

Reliable Plan Selection by Intelligent Machines

From the Publisher: This book derives techniques which allow reliable plans to be automatically selected by intelligent machines. It concentrates on the uncertainty analysis of candidate plans so that a highly reliable candidate may be identified and used. For robotic components, such as a particular vision algorithm for pose estimation or a joint controller, methods are explained for directly calculating the reliability. However, these methods become excessively complex when several components are used together to complete a plan. Consequently, entropy minimization techniques are used to estimate which complex tasks will be performed reliably. The book first develops tools for directly calculating the reliability of sub-systems, and methods of using entropy minimization to greatly facilitate the analysis are explained. Since these sub-systems are used together to accomplish complex tasks, the book then explains how complex tasks can be efficiently evaluated.