Study On System Calibration For Dual IMU And Vision Sensors Fusion Pose Measurement On Moving Base

Pose measurement on moving base is the critical technology of navigation tracking and stability control for airborne,vehicle-mounted,carrier-borne mechanical equipment.The pose measurement speed of IMU is fast,but the measurement results drift with time;visual pose measurement has excellent stability,but its speed is slow,and the measurement will fail due to image blur or occlusion.With dual IMU and vision combined,advantages of two kinds of sensors can be complemented,and the target pose on moving base can be measured quickly,accurately and stably.The basis of information fusion between dual IMU and visual sensor and the key of accurate and stable pose measurement on moving base is to analyze the relationship of each sub coordinate of the system and unify visual and inertial data to a same coordinate system.In this paper,the system calibration technology was studied for dual IMU and vision sensors fusion pose measurement on moving base.The main contents and key technologies are as follows:1.The principle of dual IMU and vision sensors fusion pose measurement on moving base was analyzed;the overall scheme of system calibration was designed,including vision module calibration and vision inertial calibration.Schemes for every part calibration were designed according to their characteristics.2.Calibration method for vision pose measurement module was discussed.This paper analyzed the non-orthogonal and linkage relationship of axes,established the mathematical model of the target coordinate system and deduced the dynamic update model of the target coordinate system.The vision module calibration system,visual algorithm and calibration flow were designed;the accuracy and reliability of vision module calibration were verified by experiments.3.Method for visual-inertial calibration was discussed.A nonlinear optimal calibration framework was designed with fixed iterative steps,which was used for resolving the transformation matrix of camera and slave IMU coordinate systems,target and main IMU coordinate systems.The accuracy and reliability of vision module calibration were verified by experiments.