International Journal of Advanced Robotic Systems (May 2013)
Conceptual Framework for Knowledge-Based Decision Migration in Multi-Component Robots
Abstract
Abstract This paper presents a conceptual framework for dynamically migrating the decisions of multi-component robots within different physical subcomponents. At the higher layer, a decision-making distributed migration system embeds the rules written in a knowledge-based system (KBS). The rules are written and decisions are made at one component, which can be migrated by the underlying distributed system. This eventually results in intelligent migration of the decisions and robot survival, irrespective of a particular physical component failure. Moreover, a robot does not need an exclusive control system for each component. The implementation is a two-step process. Initially, component-specific facts are identified and mapped to suit to respective components and to write rules and explore different gestures and behaviour of the robot. The second step is the embedding of these rules within the distributed or agent-based system and allowing the decisions to float around the system. The Jess (Java Expert System Shell) expert system is used as the knowledge-based tool and different gestures are generated to realize the proposed method.