Research On Positioning Technology Based On Inertial/Vision/UWB Combination

Mobile robot's exploration of the environment and motion accurately in various environments require different sensors such as camera and inertial measurement unit to recognize the current environment and obtain the exact position of the robot itself.This process includes the steps of the mobile robot acquiring the environmental image to solve its own pose and establishing a map.The accuracy of the mobile robot's estimation of its own pose is one of the most important factors affecting the accuracy of the map.Therefore,based on the ROS system,this paper proposes a multi-sensor fusion algorithm for autonomous positioning and synchronous mapping of IMU,vision and UWB combination,and validates the effectiveness of the fusion algorithm on real mobile robots.The specific research work of this article is as follows:Firstly,the basic principles of inertial navigation,visual navigation,and UWB positioning are studied.In this paper,the method of inertial positioning is realized.The pose estimation based on feature point matching between adjacent frames of images is realized.And the traditional positioning method based on UWB ranging technology is also realized.Secondly,the theory of graph optimization is studied.In this paper,based on the actual situation of mobile robot moving in three-dimensional environment,the expressions of ranging residual and motion-constrained residual are proposed,and the numerical solution of the minimum value of the overall loss function can be used to realize the realtime positioning.The experiment results show that the positioning accuracy of the method is higher than the traditional equation solving method.Thirdly,the multi-sensor fusion positioning theory of IMU,vision and UWB positioning technology is studied.In this paper,the framework of federated Kalman filtering is adopted,and the inertial measurement unit is taken as the reference system.The IMU and visual positioning,IMU and UWB positioning are taken as two subsystems respectively,and the federated Kalman filter equation is established to realize the fusion positioning of multi-sensor.Fourthly,the platform of ROS-based mobile robots is studied.On the hardware platform,a complete prototype experimental environment is set up,including the installation of various types of sensors,the placement and positioning of base stations.On the software platform,the serial port parsing program of UWB ranging module is written,and the various algorithms adopted and proposed in this paper are transplanted into the mobile robot,and the multi-sensor fusion positioning algorithm proposed in this paper are verified in real time.The research of SLAM based on IMU,vision,UWB is of great significance for indoor positioning,navigation and map construction required for subsequent obstacle avoidance.It can improve the positioning accuracy of mobile robot systems indoors,provide reliable maps.It has sufficient application prospects in logistics and transportation,military robots,etc.