Cooperative Perceptual Systems for Partner Robots Based on Sensor Network

Abstract

Summary This paper proposes a method for cooperative perception of partner robots on a sensor network. While the perception of a robot is restricted locally, the observation from an environmental system is global. Therefore, the robot can easily perceive the environment by receiving and integrating the global information from the environment. Furthermore, the size of robot can be small because the robot does not equip with many sensors if the environmental information is easily available through a wireless network. In this paper, we focus on the localization of human and robots by the integration of camera image and infrared line sensor, and we apply steady-state genetic algorithms for extracting humans and target objects from camera images. Furthermore, we propose a sensor network based on the cooperation between partner robots and environmental system. The proposed method is applied for a navigation task of partner robots interacting with humans and environmental system. Finally, we show several experimental results based on our proposed method.