Research On GNSS/Vision Integrated Navigation And Positioning Technology

The Global Navigation Satellite System(GNSS)has the characteristics of being all-weather,global,and its error does not accumulate over time.In an open outdoor environment,GNSS can achieve continuous navigation,positioning and timing(Positioning Navigation Timing,PNT)functions.It is widely used in the national economy and defense construction.However,the GNSS signal that reaches the ground is very weak,and it is vulnerable to occlusion,interference,and deception,which makes the estimated position error significantly increase,and the results may even become unusable.With the improvement of computer computing power and the continuous development of image processing technology in recent years,simultaneous localization and mapping(SLAM)technology has developed rapidly,and can achieve continuous navigation functions in complex environments.Aiming at the problem of continuous navigation and positioning in a weak GNSS environment,this paper studies stereo visual SLAM(V-SLAM)and GNSS/vision integrated navigation and positioning technology,which is of great significance for achieving continuous navigation in a weak GNSS environment.Based on the demand for continuous navigation in a weak GNSS environment,this paper focuses on GNSS/vision integrated navigation and positioning technology,including GNSS positioning principles and algorithms,basic principles of stereo vision SLAM based on features,GNSS /vision loose combination algorithm,GNSS/vision tight combination algorithm,and test and analyze the continuous navigation and positioning performance of the integrated navigation system in a weak GNSS environment.The main content and results of the thesis are as follows:1.Studied GNSS positioning and visual SLAM navigation algorithms,and tested their respective navigation performance and complementary features.The coordinate system used in navigation is introduced in detail,the GNSS pseudo-range single point positioning(Standard Point Position,SPP)algorithm is studied,the error equation form of the SPP is derived in detail,and the least square solution is given.Studied the principle of GNSS RTK(Real-time Kinematic).Using the measured data in a weak GNSS environment for calculation and analysis.The basic framework of visual SLAM is studied,the principle of pinhole camera imaging is deduced in detail,the principle of stereo visual pose estimation and the stereo visual navigation tracking,mapping and closed loop detection module based on features are studied,and the effects of visual navigation results are discussed.It also analyzes in detail the effect of closed loop on the improvement of visual navigation.The results show that the static SPP results can reach the decimeter level,and the dynamic weak GNSS environment will degrade to the meter level;the RTK positioning result also degrade to the decimeter level in harsh environments;the multi-system has a little enhancement effect in SPP positioning accuracy;when the satellite signal is blocked in a harsh environment,the GNSS positioning result will be interrupted.Visual navigation is affected by many factors,and the carrier will deviate from the true value by about1?3m every 100m;closed loop greatly improves the accuracy of visual navigation and will make the positioning error converge to near zero;visual navigation will not be interrupted in the normal movement with the ability to continue navigation.These results show that visual navigation and GNSS have complementary advantages and disadvantages,and integrated navigation can be used to achieve continuous navigation.2.Studied the GNSS and visual SLAM loose combination model,and tested the navigation and positioning performance of the loose combination system and the loose combination after simulating interruption.The Kalman filtering and data post-processing RTS(Rauch-Tung-Striebel)smoothing algorithms commonly used in integrated navigation systems are deduced and studied in detail.Studied GNSS/vision loose combined method and its mathematical model is given.Tested and analyzed the positioning performance of the GNSS/vision loose combined system in the GNSS interruption environment in detail.The test results based on the public data set and the measured data show that the GNSS/vision loose combination system can well improve the accuracy of the visual navigation single system,and can better deal with the problem of signal occlusion or interruption in a weak GNSS environment,realize the function of continuous navigation,and verify the feasibility of the GNSS/vision loose integrated navigation system.3.Proposed a tight combination model of GNSS and visual SLAM,and verified the effectiveness of the continuous navigation of the tight combination system.A GNSS/visual tight combination method is proposed and its mathematical model is given.The navigation and positioning performance of a single navigation system,GNSS/visual loose combination,GNSS/visual tight combination is tested and analyzed.The navigation and positioning performance of the GNSS/visual loose combination and the GNSS/vision tight combination when the number of observation satellites is less than 4 is analyzed and compared in detail,the test results based on the measured data show that the GNSS/vision tight combination has the same accuracy as the GNSS/vision loose combination,but the tight combination system can achieve continuous navigation when the number of observation satellites is less than 4.Verified the effectiveness of visual navigation for assisting GNSS continuous navigation and positioning.