Indoor mobile robot navigation system architecture

Abstract

This paper deals with the design and implementation of an indoor mobile robot navigation system architecture. The two main parts of our navigation system are a global path planner and a local path planner. For the global path planner, a new strategy has been proposed which does not consider the robot as a point and does not expand the obstacles. The local path planner is based essentially on an ultrasonic perception system, where the covered region of sight is widened and the distance between any two adjacent sensors can be adjusted. Hence, good resolution can be obtained and laterally positioned obstacles with respect to the robot line of sight are well identified. Furthermore, we propose a technique to improve the odometric method, to reduce the systematic errors and to detect the floor irregularities.