Expert system for AGV managing in bidirectional networks: KADS methodology based approach

This paper presents an expert system prototype for collision avoidance and routing unidirectional automated guided vehicles (AGV) in bidirectional networks. KADS methodology was used for the prototype development. System modelling, concepts employed in expert system development, and the model implementation are presented. The expert system prototype allows to solve some specific conflicts arising in a bidirectional network between AGVs during task accomplishments. It is mainly based on ‘stepping back’ in simulated time concept and Dijkstra’s shortest path method. It routes all vehicles through their shortest paths and uses a conflict managing strategy to solve conflicts occurring during AGV travels. Two illustrative examples are discussed in this manuscript.