Robust execution of robot plans using fuzzy logic

Abstract

Intelligent agents embedded in a dynamic, uncertain environment should incorporate capabilities for both planned and reactive behavior. Many current approaches to this dual need are based on a two-level model: at the upper level, a planner generates a sequences of operations whose performance is expected (in an ideal world!) to satisfy the robot's goals; at the lower level, a controller tries to achieve these goals while dealing with the environmental contingencies. This solution requires that the controller be able to: accept a specific goal as a parameter; cope with the uncertainty and imprecision in the environment; and consider several goals simultaneously—e.g., traverse a hallway while avoiding the obstacles and people and tracking a specific landmark with its camera. A major problem in the design of such a controller is how to resolve conflicts between simultaneous goals. We describe a reactive controller for an autonomous mobile robot that uses fuzzy logic for trading off conflicting goals. This controller has been implemented on Flakey, the mobile robot of SRI International.