On interval methods applied to robot reliability quantification

Abstract Interval methods have recently been successfully applied to obtain significantly improved robot reliability estimates via fault trees for the case of uncertain and time-varying input reliability data. These initial studies generated output distributions of failure probabilities by extending standard interval arithmetic with new abstractions called interval grids which can be parameterized to control the complexity and accuracy of the estimation process. In this paper different parameterization strategies are evaluated in order to gain a more complete understanding of the potential benefits of the approach. A canonical example of a robot manipulator system is used to show that an appropriate selection of parameters is a key issue for the successful application of such novel interval-based methodologies.