Research On Indoor Positioning Of Mobile Robot Based On Vision And Inertia

Indoor positioning technology for mobile robots is one of the key research topics in the field of robotics.Due to the severe interference of indoor GPS signals,this has led to the fact that GPS-based positioning methods cannot be widely used in mobile robot indoor positioning.Therefore,indoor mobile robots usually use sensors such as laser sensors,vision sensors,inertial measurement units(IMU),and wheel encoders to assist in positioning.Among them,the vision sensor has problems such as tracking loss when the robot moves too fast,or invalid data during occlusion.IMU can estimate fast movement,but there are problems such as long-term cumulative drift.Therefore,merging the two data to estimate the strengths and complement the weaknesses can effectively improve the indoor positioning accuracy of the robot.The main research content of this paper is to fuse inertia and visual information to solve the indoor positioning problem of robots.The specific research content is as follows:1.The mathematical model of the visual inertial positioning system is studied,including the definition of the coordinate system and the pose representation,the camera’s pinhole imaging model,the binocular camera imaging model,the camera distortion model,and the nonlinear optimization method theory.The research has laid the foundation.2.Research the visual positioning technology.The first is the composition of the visual positioning system and its main processes.The second is the basic principle of Zhang’s camera calibration method.The binocular camera used in this paper is calibrated and the calibration results are given.An improved Harris feature extraction and feature matching algorithm is proposed.The experimental results of feature extraction and matching and the results of visual odometer are given through experiments.3.Researching the positioning technology that combines vision and inertia.Including the overall framework of the visual inertial system,an optimized binocular visual inertia tight coupling algorithm based on VINS-Mono is proposed.First,the expression of IMU pre-integration is derived.Secondly,the estimation principle of the initialization parameters of the visual inertial system is derived.Finally,the derivation of the objective function,the residual term,and the derivation of the objective function with respect to the incremental derivative of each variable to be optimized.4.Build the software and hardware platform required for visual inertial system and conduct experimental analysis.First,the experimental hardware platform consists of sensors,processors,and so on.Secondly,the composition of the software test platform includes the establishment of the ROS platform and the use of common functions.Finally,experimental verification and analysis are performed.The experiment consists of two parts: Eu Roc data set test and actual scene hardware experiment.The reliability and accuracy of the fusion algorithm are verified by simulation and actual platform.