Edge-Node Map Based Localization without External Sensor Data

In this paper, we propose a novel localization method which utilize a edge-node map and information from only inner sensors. To reduce integration errors, we estimate the passed edge from measured moving distance and direction, and integrate the trajectory of the odometry with the estimated edge. We performed two kinds of experiments. First, we operated the robot in four kinds of environments. Second, the robot travel autonomously using this method. From these experiments, we confirmed that our localization system works properly in various kinds of environments and for autonomous driving.