Research On Vehicle Positioning Orientation System Of MEMS+ Vision

Facing the harsh environment in the modern battlefield and the large demand for vehicle positioning orientation,under the premise of ensuring the positioning accuracy,the vehicle positioning and orientation system with low cost,small volume,strong autonomy and anti-jamming has become the inevitable development trend of vehicle positioning orientation.This paper mainly studies the combinatorial positioning directional positioning technology based on monocular vision and MEMS inertial measurement components.In this paper,the positioning orientation algorithm of MEMS inertial measurement component is studied firstly.The coordinate system used in the positioning orientation algorithm of MEMS inertial measurement component is defined,and the relationship between coordinate system is analyzed,then the attitude,speed and position update algorithm of Strapdown inertial navigation are analyzed.Secondly,the core content of Monocular visual positioning is studied.This paper expounds the principle of camera imaging and introduces the working principle of Kinect camera,experiments and analyzes the feature point detection and matching technology,the ORB SLAM algorithm and the RGBD SLAM algorithm,and uses the visual location algorithm of ORB SLAM to calculate the visual location information of this paper.On the basis of the above,the research on the fusion algorithm of MEMS inertial measurement components and visual positioning is studied,firstly,the Strapdown inertial navigation is calculated by using Kalman filter to filter the location information of MEMS inertial measurement components,then the ORB SLAM is used to solve the visual information.The visual position information of the solution is filtered by Kalman filter.Finally,the visual reliability judgment is carried out.when the visual information is unreliable,the position information calculated by the filtered MEMS inertial measurement component is used as the position of the system,and when the visual information is reliable,Using the track weighted fusion algorithm,the filtered visual positioning information and MEMS inertial measurement component positioning information are fused,the position information after fusion is used as the position of the system.Finally,according to the actual situation of vehicle positioning orientation,the software and hardware design platform of vehicle positioning and orientation system is constructed.The hardware design part of the system is mainly based on data acquisition module,information processing module and communication interface module.The software design part focuses on the analysis of the complete vehicle positioning and orientation system workflow,and the function module that need to be implement.Through C and VHDL language programming to design and develop the system.Experiments show that the track weighted fusion algorithm effectively implements the combination positioning of the vehicle positioning and orientation system,and the fused system can not only satisfy the requirements of positioning accuracy,but also has a good update frequency.