Embedded Strapdown Inertial Measurement Unit Based On MEMS Sensors And Its Application

Railway track geometry is one of the key factors affecting the quality of the train running. Therefore there are many countries around the world attaching great importance to develop railway track geometrical parameters testing technology and equipment. One of the key technologies in railway track geometrical parameters detection is to obtain the datum of measurements. This paper researches in the theme of founding datum plane of the railway track geometrical parameter mea-surements. And the solution which bases on MEMS inertial measurement unit(IMU) is proposed after contrasting and analyzing a variety of methods of founding datum plane.Based on the depth study of basic principles of strapdown inertial measure-ment, the thesis models the datum plane by the theory of strapdown inertial mea-surement. To facilitate the verification of the algorithm, the theoretical trajectory, attitude and output values of sensors associated with the carrier coordinate system are deduced under certain assumptions through the reverse process of strapdown inertial measurement solving.In order to choose the right attitude solving algorithm, in the interference of random noise at different intensity, the effects under coning motion of fourth-order Runge-Kutta method, fourth-order Adam Williams method and fourth-order Ham-ming method are compared. The results show that the algorithm accuracy of fourth-order Adam Williams method and fourth-order Hamming method are more better than the fourth-order Runge-Kutta method when the interference is small, while the fourth-order Adam Williams method and fourth-order Hamming method are worse and the Hamming algorithm are worst in the case of large disturbances.The solver of strapdown inertial measurement based on fourth-order Runge-Kutta method is tested by numerical simulation with the theoretical output values of sensors obtained in the carrier motion simulation. The results show that the curve obtained by numerical simulation coincides with the curve obtained by theoretical calculations in a very high degree. It is proves that the algorithm pro-posed is feasible in principle.Because of the low speed and constraints for ground tracks of the train, and the low precision of the MEMS sensors, the algorithm of strapdown inertial mea-surement is simplified moderately by approximating the surface of earth as inertial space. The numerical simulation shows that the accuracy of simplified algorithm proposed is still high in a short time, but to be worse and worse after a long time.The embedded systems based on high-performance ARM9core microcontrol-ler LPC3250and C program of algorithm which bases on the fourth-order Runge-Kutta method are designed. And a PC application is also designed to facili-tate System initialization and sensor bias errors settings as well as historical data records. After hardware and software debugging, the system obtains a highest solver frequency of512Hz, and an upload frequency of128Hz.At the end of the thesis, the system is calibrated and tested in laboratory. The tests show that the simplified algorithm mentioned in this thesis has a low error drift speed than the full algorithm. And it proves that the embedded inertial mea-surement unit designed in this thesis is able to track the local level datum dynami-cally with a high accuracy in a short time, but the accuracy would fall sharply in a long time.