This paper addresses a modified localization scheme of a wheeled mobile robot. When it navigates in a square given by the operator, the position error of a robot accumulates and it never ends up with the goal position where the robot intends to go to initially. The objective of localization is to estimate the position of a robot precisely. Many algorithms were developed and still are being researched for localization of a mobile robot at present. Among them, a localization algorithm named continuous localization proposed by Schultz [2] has some merits on real-time navigation and is easy to be implemented compared to other schemes. Continuous Localization (CL) is based on map-matching algorithm with global and local maps using only ultrasonic sensors for making grid maps following a given path by us (e.g. squares). In addition to CL, we here propose systematic error correction and fast powerful map-matching algorithm for localization of a mobile robot by experiments.
展开▼
