Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 02, 2018 - June 05, 2018
For service robots operating in complex indoor environments, intelligent sensor based navigation is key for autonomous operation. In this paper, we present semantic navigation for service robots operating in indoor environments using range and vision sensors. We propose a novel mapping and navigation algorithm using topological and metric mapping. For semantic navigation a dictionary of visual objects is stored that allows the system to distinguish between different sections of the map easily. This enables to send easy human like command based navigation goals for the service robot. The proposed method is tested in simulation environment and the results are discussed. One advantage of the present system is that it does not require a large amount of memory for storing the map and based on the application the system can easily be switched between either the topological or metric map.
