Teleo-Reactive Programs for Agent Control

摘要

A formalism is presented for computing and organizing actions for autonomousagents in dynamic environments. We introduce the notion of teleo-reactive (T-R)programs whose execution entails the construction of circuitry for thecontinuous computation of the parameters and conditions on which agent actionis based. In addition to continuous feedback, T-R programs support parameterbinding and recursion. A primary difference between T-R programs and many othercircuit-based systems is that the circuitry of T-R programs is more compact; itis constructed at run time and thus does not have to anticipate all thecontingencies that might arise over all possible runs. In addition, T-Rprograms are intuitive and easy to write and are written in a form that iscompatible with automatic planning and learning methods. We briefly describesome experimental applications of T-R programs in the control of simulated andactual mobile robots.