Research Of Indoor Mobile Robot Path Tracking Based On Monocular Vision And RFID

With the rapid development of technology,the indoor mobile robots have broad application prospect,which can be used in home services,entertainment,safe checking and so on.In the mobile robot related technology research,monocular vision has several advantages such as simple structure,easy to calibration and flexible motion,so monocular vision navigation has become an important research content of mobile robots.In indoor environment,RFID technology is widely used because of its non-line-of-sight,contact-free,recognition speed,high accuracy,high security and low cost and so on.This thesis focuses on the research of indoor mobile robot path tracking based on monocular vision and RFID.Firstly,path detection and processing is realized by using the mobile robot to acquire image of path information in real time.In the indoor known path environment,image preprocessing is performed by gray scale,filtering,image binarization and morphological changes methods.Filter out the interference information in the image and get the edge feature.Contrast of the whole image is enhanced by histogram equalization to increase the dynamic range of pixel gray value.Simulation results show that the processed images have good effect for path identification tracking.Then path identification and tracking is realized by Hough transform and Kalman filter.Firstly,road constraints such as the straight road model are adopted.Hough transform is used of extracting line features.The real time and robustness of Hough transform is very good.By the parameter of road line,the deviation of site and angle are known between the path line and mobile robot.Then Kalman filter is used for path tracking.The algorithm proposed by the paper are used to process actual image,simulated results show that the algorithm is right and effective.Finally,tracking algorithm of RFID mobile robot based on improved particle filter is realized.Firstly,this thesis takes RFID as the technical realization of indoor positioning and introduces the RFID technology and RFID positioning technology about algorithm mathematical models in the indoor unknown path environment.Then it is to estimate target state by particle filter algorithm.However the traditional particles are easy to loss the diversity,so a RFID indoor tracking algorithm based on improved particle filter is proposed.The simulation results show that the refinement ability of the particle is improved and the diversity of the particle is increased,and the improved filtering algorithm can obtain better tracking performance.