Nowadays a major research issue of mobile robots is to develop a robust 3D environment sensing for navigation and task execution. To achieve this, a variety of techniques have been developed for the determination of the 3D scene geometric information such as stereo vision, laser structured light, laser range finder and so on. But these methods have many limitations. To overcome these limitations we introduce a new sensing algorithm, which is based on the moiré technique and stereo vision. To verify the performance of this sensor system we conducted a series of simulation for various simple environments. The result shows the feasibility of successful perception with several environments.Keywords : 3-dimensional sensing, moiré technique, stereo vision, robot sensor system
INTRODUCTIONA major research issue for mobile robots is to develop an environment sensing and recognition system for mobile robot navigation and task execution, which is robust and efficient. A variety of techniques have been developed in the past for the determination of 3D scene geometric information. Especially, although the binocular vision sensors among them have been widely used as representative ones of passive visual sensors for mobile robots, they still suffer from image intensity variation due to the illumination noise, insufficient feature information on environment composed of plain surfaces, and correspondence problem between multiple images. These reasons have led most mobile robot researches on 3D environment reconstruction using visual sensors to deal with just straight line edge and corner as interesting features [1]-[3], but these features are saliently observed in well arranged and structured environment with polygonal objects or polygon-textured surfaces. In addition, this information is not sufficient to describe the whole structure of 3D space. Therefore, robots frequently use active sensors for more reliable range sensing, and have become recognized as a promising alternative proposal, which includes the infrared sensor, the ultrasonic sensor, and the laser sensor [4] In our previous research we developed the novel sensing algorithm based on laser structured light sensing method.[9]. In many approaches to indoor robot applications, laser sensor has been used for detail sensing and modeling objects, which is commonly categorized to the laser sensor and the laser range finder measuring the time-of-flight. Though the laser range finder has more advantages in views of measuring, it need still high cost with high power consumption, heavy weight, and low accuracy in near range. In addition, the latter needs more scanning procedure than the former. This scanning procedure is a time consuming task to limit the sensing time, and needs a precisely controlled scanning mechanism. In order to keep up the advantages of the sensor system using the laser-structured light and to decrease the sensing time without degradation of the sensor resolution, it was necessary to develop a new visual sensor system different from the sensors mentioned a...
