Automatic 3D space initialization system using depth camera and catalogue database

In this paper, we propose a method to initialize a real living space for the purpose of robotics services support in human daily activities. To introduce service robot into unknown or first-see living environment, a robotics system needs environmental information such as furniture 3D appearance and articulation. This functional information can be realized using environment map that can filtering candidates for furniture automatically. The environment map which contains furniture 3D information presented in this paper is developed by using depth camera and online database like furniture catalogue. The realization of this proposed method is done through a plug-in created in layout software, Microsoft Office Visio. This initialization for robotics service generation in human daily-life may provide a better robot's task performance. The result shows the 3D geometric model or environment map of standard Japanese-style living space.

I. Introduction

Robot Technology (RT) has been involved in human activities for many years. In the early years of RT, robots have become human co-worker at factory or working place which also known as industrial robot. Nowadays, instead of just being human ‘co-worker’, robot has become human ‘partner’ which ‘live’ in a more complex environment and also known as domestic robot. One of the function of domestic robot is to ‘do work’ at home where robot needs information resource of the operation environment to interact with human and delivering appropriate service according to human intention [1]. In an environment which is specifically designed for robot such as factory, it is not important for the robot to understand or recognize the environment. The robot just focuses to finish the scheduled robotic routines at the assembly. Meanwhile, human living space is not designed to accommodate robots living together with human. It will be costly for the house developer to build every homes with RT functionalities as well as it is not appropriate to design one robot to each house. Therefore, in order to provide robotic services at home, it is significant to make the robot to be able to integrate with every individual actual home environment.