Research On Mobile Robot Control Based On Multi Sensor Information Fusion

With the development of science and technology, the technology of mobile robot is becoming more and more mature, and the application of robot is becoming more and more extensive. In a certain degree, the level of application of robot technology is reflected a country’s automation level, so it has very important strategic significance for the research of robot. The tracking problem of mobile robot is that depending on the robot’s own sensors to perceive the surrounding environment, according to the path information to achieve autonomous tracking. In the field of robot research, how to perceive and understand the environment is the difficult and hot spot of the robot to realize the autonomous movement. In addition, with the continuous expansion of the application of mobile robot, individual sensors are often difficult to perceive the surrounding environment information, so the application of multi sensor technology is born. Compared with individual sensor, multi sensors can complement and fuse information to improve the robot ability of perceive the surrounding environment, and obtain more comprehensive and accurate information, so as to improve the robustness of robot autonomous motion. Therefore, the technology of multi sensor information fusion of mobile robot has become one of the hot topics in the field of robotics. The main research of this paper is robot tracking movement show based on the ultrasonic sensor, infrared sensor and visual sensor information fusion, including research of data registration and information fusion, path extraction and control algorithms and other related issues. The main work of this paper is as follows:(1) Firstly, the fusion model of multi sensor information is studied. Three kinds of models of multi sensors are analyzed. Secondly, correct the possible errors of acquisition data. And for the problem of differ of the data acquisition frequency, we registered the time of data to synchronize the data. Experimental results show that the modified and registered data is closer to the theoretical value and useful for information fusion, which is helpful for the robot to judge the distance and avoid obstacles.(2) Research on robot path extraction. For tracking path of mobile robot is colorized, and Inspired by the idea of threshold segmentation of gray image, this paper propose an optimal threshold method for path extraction based on HSI color space. Convert the image from RGB to HIS color space, because the independence between the three components of HSI is strong, so we can process the component of H alone. The component of H is regarded as the gray value, according to the proposed improved image segmentation algorithm IOtsu to find the optimal H component. This method can improve the accuracy of path tracking of the robot.(3) The whole process of the system is designed, For infrared and ultrasonic obstacle detection system, a heterogeneous sensor information fusion method is proposed, information fusion used to constrains the robot’s movement. and propose an algorithm based on fuzzy control according to the characteristics of tracking path. This paper presents a new fuzzy rules, which can transform the environment into instructions that robot can recognize, so as to realize the intelligent control of robot.At the end of the paper, some conclusions of the research work are made and some outlook of the next research direction and problems are discussed, including multi robot cooperative control and dynamic environment motion control, etc.