Research On 3D Motion Trajectory Detection Based On Strapdown Inertial Navigation And Attitude Solution

For a long time,the research of motion trajectory detection relies on external observations.Although the external observation method has high-precision characteristics,some problems,such as the problems of large size,high cost and environmental constraints,limit its scope of application.The method based on the internal sensor has the advantages of low cost,small size,anti-interference,etc.,which meets the current mainstream demand.The research of moving object attitude based on MEMS sensor and strapdown inertial navigation has become a hot topic in positioning navigation.At present,the research mainly focuses on the solution of attitude angle and the detection of simple motion trajectory.However,in the study of motion trajectory detection,there is a widespread problem of large error.In this paper,the diversity of errors in motion trajectory detection is analyzed,including component error,random error and cumulative error.Different errors are processed in this paper.For the component error,this paper adopts the static sensor,and at the same time,the method of averaging multiple sets of data is used to solve the zero offset.Then the zero offset is removed from the acquired data in order to obtain the results that are closer to the original data.For the random error,the improved two-state Kalman algorithm is used to remove the random noise affecting the acceleration and angular velocity values.Then the angular velocity correction algorithm based on Mahony complementary filtering is used to correct the angular velocity by using the normalized acceleration data in order to make angular velocity data more accurate.For the cumulative error,this paper proposes an adaptive acceleration compensation algorithm that is compensated first and then corrected.This algorithm can make the result of the operation better and more accurate,and reduce the influence of cumulative error.In this paper,error processing,including the four methods talked above,is designed in the motion trajectory detection model.The data collected by the smartphone and the MPU9250 sensor are tested respectively,and the obtained trajectory is more accurate than the existing method.