The Research Of SLAM Algorithm Based On The Sky Polarized Light

Highly intelligent robots are becoming more and more in people’s life and work since the20th century, and make people’s life and work more convenient. They can move autonomously between different obstacles in the real environment and perceive the surrounding environment through their sensors.Navigation and positioning technology is the key and foundation of the autonomous moving for the mobile robots. It mainly includes three aspects: positioning methods, creating map methods and path planning methods. And navigation and positioning is the most important among them.The application environments of mobile robots are very complex. And the robots usually need to solve the problems in the area where human rarely reach, such as:deserts, oceans, spaces and some other special environments. In addition, the functional requirements of mobile robots are increasing, too. So the navigation and positioning system of the mobile robot also has developed some new features: composition of the multiple navigation methods, distributed intelligent architecture, information networking, multiple sensors fusion technology and the cross of different technologies and so on. There are many biological in nature. And they usually have some special navigation and positioning capabilities. For example, some ants and bees can navigation and positioning according to the scattering of the skylight. In addition, some other insects have more exotic navigation and positioning capabilities. These capabilities have greatly inspired the researchers and provided a new way for the research on the navigation and positioning.The research work in this paper is supported by the Intelligent Vehicle Institute, Institutes of Advanced Manufacturing Technology, Hefei Institutes of Physical Science, Chinese Academy of Sciences. The paper starts from the bionic mechanism of animal, and carries out a series of researches on the navigation and positioning methods based on the polarized skylight, which include the following aspects:(1) Analyzing the scattering principle when the sunlight pass through the atmosphere, and surveying the works of the foreigners and domestic researchers, we understand three mathematical description methods of the polarized light, two polarization distribution models of the skylight and two polarized light measurement methods (include their advantages and disadvantages). At the same time, we understand the impact factors of the polarization measurement and the impacts of the polarization distribution model of the skylight in different atmosphere conditions. At the end, we summarize the methods and laws to obtain the ideal measurement data of the sky polarized light in the actual work.(2) We analyze and summarize the methods, laws and current research lever in the field of navigation and positioning. We analyze the differences between the differential model and the Ackerman model of vehicle motion. We analyze the dead reckoning algorithm and the SLAM algorithm, especially EKF-SLAM, particle filter SLAM, UKF-SLAM and Fast-SLAM. We do detail analysis on the differences among them.(3) Through the analysis of the current measurement devices made by foreigners and domestic researchers, we design and make three generations polarized light measurement devices. Especially, we develope and design a set of calibration methods for the Ⅱ generation polarized light measurement device, and now we already could obtain the high-precision measurement data from the device.(4) We study the navigation and positioning methods and design a method which could combine the navigation method with the measurement data of the polarization sensor. How to achieve high-precision navigation and positioning which based on the polarization measurement data is a major problem that we care. We analyze the dead reckoning and the SLAM algorithm, and propose a method that we can navigate based on the polarization information of skylight and create the polarization distribution database simultaneously. Finally, we do many experiments and analyze the experimental results, which show that the method proposed in this paper is effective. It can not only improve the accuracy of navigation and positioning based on the polarization information of the skylight, but also can improve the measurement accuracy of the polarization information of the skylight.We use the "intelligent Pioneer I" unmanned vehicle as the experimental platform, and do all the experiment in a real road environment. Finally, we summarize the weaknesses of our research work, and propose the prospect of the further research.