The Research Of Rapid Algorithm Of Attitude Determination Based On Gradient Descent And Complementary Filter

The technology of attitude determination is very useful.Especially,attitude determination of moving carrier has played a very important role in the Military and Civilian application such as aeronautical engineer,medical equipment and robot technology.The technology of strap down inertial navigation system is the most widely used method of attitude determination.At present,high-precision inertial measurement units are all too expensive to satisfy demand for low cost.The accuracy and property of MEMS inertial sensor(mems gyro and accelerometer)has been improved step by step with the development of MEMS techniques.Since MEMS component features small size and low cost,strap-down attitude measurement system can be much less expensive.In recent years,the research enthusiasm of low-cost attitude measurement system with medium and low accuracy for consumer and industrial application has been improved.This article designed a low-cost twin-biaxial strap-down attitude measurement embedded system,base on MEMS componert,MPU6050 which is produced by InvenSense company.The article also designed one kind of data fusion algorithm based on complementary filter algorithm to fuse gyroscope data and accelerometer data.In order to improve accuracy,it uses gradient descent to process the accelerometer data before the fusion.A strap-down attitude measurement system has at least three gyroscope and three accelerometer to make up its sensing elements.At first the article set forth the research status of MEMS gyroscope and MEMS accelerometer and the basic concepts of strap down inertial system.Then it leads the concepts of attitude rotation matrix and attitude angle of pitch,roll and yaw on the basis of kinds of usual coordinates.Then it introduced kinds of usual attitude updating algorithm.The next part designs a fast data fusion algorithm on the basis of gradient descent and complementary filter.This part also compare this algorithm and the useful data fusion algorithm base on kalman filtering.In the end it designs hardware circuit of every module,using the K10 single chip microcomputer produced by Freescale Semiconductor as the data processor.The parameter and communication protocol used in the program have been given out.Design some static tests and dynamic test to verify the attitude measuring system designed by the article,using the laboratory existing conditions,and use matlab to analyze testing data.Experimental result indicate that The experimental results indicate that the low-cost system can work in real time and output stable attitude.It has practical value of engineering.The data fusion algorithm on the basis of gradient descent and complementary filter expend low computational so that it is fit for low-cost embedded system and it has good result.