Enhancing Cognitive Virtual Maps for Domestic Service Robots by Adapting the User's Perception on Uncertain Information Based on the Robot Experience: A Neural Network Approach

摘要

Assistive robots possess the competency to provide companionship to the human beings. Navigation is an essential factor of the robot. The robot should have the capability to virtually imagine the description of an unknown environment to prove that it possesses a better navigation skill. Subsequently they should improve the capability in relevance to the imagining with its experience. Therefore this paper proposes a method to improve the virtual imagining capability of a service robot while understanding the uncertain information about an object size using the artificial neural network. The proposed method possesses the capability to create a virtual map of an environment and enhances it using the actual sensory data. The Virtual Map Modeler (VOM) has been introduced in order to create the improved virtual maps. The capabilities of the robots have been demonstrated and evaluated for the performance.