Building of Human Frequency Map and Flexible Path Planning Based on Human Observation by Mobile Robot

Abstract

This paper describes a method for autonomous navigation of mobile robots considering environmental rules based on human tracking. In most cases of traditional mobile robot navigation, static environmental maps are used for global path planning. However, dynamic environmental characteristics such as human frequency maps acquired from tracking many humans should be also utilized for robot navigation coexisting with humans. For example, robots should not pass through areas under construction where any humans don't walk even if the areas are passable from static environmental maps based on just building structure. This paper introduces a robot system that observes human walking for generating human frequency maps without any sensors placed in environments. Also, a path planning method based on environmental rules by human frequency maps is described.