Combining Finite State Machine and Decision-Making Tools for Adaptable Robot Behavior

摘要

Modeling robot behavior is a common task in robot software development. However, its difficulty grows exponentially along with system complexity. To facilitate the development of a modular, rather than monolithic, behavior system, proper software tools need to be introduced. This paper proposes combination of a well-known finite state machine and a custom decision-making tool for implementing adaptive robot behaviors. The notion of automatic behavior adaptation reflects the capability of the robot to adapt during runtime based on the individual end-user, as well as the particular context of use, therefore delivering the most appropriate interaction experience. While each tool on its own can be used towards that aim, a unified approach that combines them simplifies the task at hand and distinguishes the roles of designers and programmers. To demonstrate the methods' applicability, a concrete example of their combined use is presented.