In this research a complete development and implementation of a navigation system for an in-door autonomous mobile robot is presented. The constitution of the developed navigation system consists of: (1) An efficient hierarchical map representation method to model a large in-door environment. This method describes the significant features of an actual in-door world consisting of either a single building or a network of buildings connected through common bridges (a corridor like pass way), in which the robot's environment is divided hierarchically using a tree structure into three levels, Building, Corridor and Room. (2) A prior preparation algorithm to prepare the information necessary for path planning using the robot's environment model for the purpose to have fast and efficient path planning algorithms. (3) Fast and computationally efficient automated path planning strategy that generates collision free paths made of straight lines. This strategy includes two algorithms, (a) hierarchical planning for Building and Corridor levels, and (b) free space planning for a Room level. Consequently, geometrical and scene information along a generated path is abstracted to support real time actions while navigating the path. (4) A real time navigation algorithm to ensure real time control of robot's behavior, and to monitor the actual execution of the planned path in a real environment. For the purpose of having a reliable navigation system, a new real time language based on the Action Mode concept as a unit of robot's motion representation is developed. Using this language the sensor based mobile robot's behavior can be described. The developed navigation system is implemented and demonstrated on an actual mobile robot called Yamabico. M in a real in-door environment. The demonstration results show the effectiveness of the implemented navigation system with its proposed constitution.
ASJC Scopus subject areas