The 10-DOF Inertial Measurement Unit Based On Extened Kalman Filter

With the rapid development of economy, land resources will soon run out, people gradually focus on ocean resources development. Ocean development technology is more difficult than land because ocean environment is complicated and there is more emergency. The Inertial Navigation is one of the most important technology and the attitude estimation is a part of Inertial Navigation.This paper designs a small, low-power real-time Inertial Measurement Unit which is used for attitude estimation. The hardware platform of the unit is consist of MEMS 3-axis gyroscope, 3-axis accelerometer, 3-axis magnetometer, temperature sensor and a high-speed micro control unit. And the software platform is equiped with high accuracy data fusion algorithm. The unit can be attached to other system with a RS232 interface. Operator can see a 3D model of the system directly through upper computer software and so can know how to control the system.The main work of this paper is choosing of the inertial sensors and the micro control unit, how to design and build the hardware platform, embedded software programming and upper computer software programming, analysis and compare of the data fusion algorithm and test of the system.Chapter 1 shows current situation and development tend of the Inertial Navigation, and the main research contents of this paper.Chapter 2 shows the theory of attitude estimation, one part is expression of attitude, like Euler angle, rotation matrix and unit quaternion, another part is data fusion algorithm, the paper will compare several data fusion algorithm like complementary filter, gradient descent and Kalman filter.Chapter 3 to chapter 5 discuss the design of the system. This paper chooses MEMS inertial sensors and STM32F303VCT6 as its core hardware platform. And the data fusion algorithm is based on Kalman filter which is famous because of its good performance, we optimize it so it can be run in the embedded system. We use Visual Studio 2012 to develop the upper computer software and we draw a 3D model by using the Opern GL graph interface.The last chapter shows test of the system, including filter parameter, static performance, dynamic performance and accuracy of attitude estimation.