A reactive architecture for autonomous agent navigation using fuzzy logic

This paper presents an implementation of a fuzzy reactive architecture. The application considered is where a virtual agent must navigate to pass obstacles on heading towards a desired target with no prior knowledge of the environment. The agent might not reach the goal and may get trapped in local minima. A behaviour-based fuzzy controller with a local path planning algorithm has been developed. The fuzzy a-level technique, as a behaviour selection method, has been used to decide which behaviour task needs to be executed. Evaluation has shown that the virtual agent can handle such situations without problem. The paths produced are reasonably smooth even though there are some sharp turns and are not too far from the potential shortest paths. This indicates the benefit of our method, where reliable and accurate decisions have been made during navigation tasks.