Anytime Sensing, Planning and Action: A Practical Model for Robot Control

Abstract

Anytime algorithms, whose quality of results improves gradually as computation time increases, provide useful performance components for timecritical planning and control of robotic systems. In earlier work, we introduced a compilation scheme for optimal composition of anytime algorithms. In this paper we present an implementation of a navigation system in which an off-line compilation process and a run-time monitoring component guarantee the optimal allocation of time to the anytime modules. The crucial meta-level knowledge is kept in the anytime library in the form of conditional performance profiles. We also extend the notion of gradual improvement to sensing and plan execution.