Research On Multi-sensor Fusion Positioning System Based On Extended Kalman Filter

With the development of self-driving technology for high precision positioning need to be more and more high,therefore,how to improve the positioning precision of the positioning system using various sensors and stability becomes one of the hot spot of current research.In this paper,the multi-sensor fusion positioning based on extended kalman filter is studied,and an improved kalman filtering algorithm is proposed based on the problem of expanding kalman filter measurement noise.First of all,this article analyses the importance of multisensor fusion positioning,comprehensive utilization of multiple sensor fusion can the advantages of each sensor,improve the positioning accuracy,increase the positioning system robustness.Secondly analyzes the theory of multiple sensor fusion,this paper introduces the main satellite positioning system and its positioning principle,and then analyzes the fusion positioning popular algorithm of kalman filter and extended kalman filter algorithm.Next,the robot operating system platform of ROS,and ROS robot_localization fusion positioning package was tested in the experiment,further validated through the experiment the superiority of multi-sensor fusion positioning,improve the positioning accuracy and stability of the system,analyzes the advantages and disadvantages of the current fusion localization algorithm at the same time.Finally aiming at the problems before the kalman filter algorithm,we proposed a modified kalman filtering algorithm,and optimize the measurement noise,by using two GPS,make its dynamic adjustment measurement noise.The results show that the improved EKF fusion results are significantly better than those before the improvement,and the positioning accuracy is further improved.At the same time,the subsequent improvement of the algorithm is also given in the future pROSpect.