Abstract
Recently the studies of autonomous mobile robot are increasing. As a task of these robots, navigation of humans is considered. Such a robot needs to find people who need a navigation service. In this paper, we propose two methods to distinguish humans who need navigation services from other people. The first one is using an environment map that presents history of human activity to find humans who stand in not-often used areas. The second method is based on evaluations of human walking paths. An experiment is conducted to evaluate the usefulness of the first method at a public space. To evaluate the usefulness of the first method experiments are conducted using the proposed environment maps that are built on three different times of a day. In addition, the accuracy of the second method is also evaluated.
