An autonomous vehicle for surface filling

Abstract

An autonomous mobile robot capable of exploring and navigating through an unknown, partially clustered closed indoor environment is presented. The robot's intelligence and autonomy are demonstrated by its capability of exploring and navigating the entire accessible area. A high level of performance has been reached at low hardware requirements. We have developed the concept of "complementary regions". It is the base of our recursive surface filling navigation strategy. Task planning, motion control and sensory activities are carried out in parallel. Absolute location of the robot using sonar information is also achieved during robot motion. Four environment representations are built and consulted as the robot evolves. A parallel multiprocessor architecture is used and a hierarchical layered distributed software structure has been implemented accordingly. A behaviour inspired design methodology with different abstraction levels is applied. Real time software and hardware parallelism complement each other. The system has been successfully tested.