Extending the Teleo-Reactive paradigm for robotic agent task control using Zadehan (fuzzy) logic

摘要

The Teleo-Reactive (T-R) paradigm for computing and organizing actions for autonomous agents in dynamic environments can be viewed as having a foundation triplet of durative actions, circuit semantics, and Boolean logic. This foundation is intended to enable the construction of semantic circuitry for the continuous computation of parameters and conditions on which agent action is based. Drawing on the notion of analog circuitry, actions and behaviors are intended to be more continuous over periods of time. However, the use of Boolean logic can sometimes inadvertently revert behaviors to be reminiscent of the state transition model since its limited bivalent capability is not consistent with the continuous nature of durative actions and circuit semantics. This paper is a report of work in progress on a Teleo-Reactive extension where Zadehan (fuzzy) logic is substituted for Boolean logic.