Design And FPGA Realization Of Kalman Filter In Vehicle Integrated Navigation System

In rencent years, vehicle is more and more important in our everyday lives. With the development of economy and the progress of electronic technology, computer technology, communication technology, the vehicle navigation system also has been developed. In that any navigation system has dynamic positioning error, there are two ways to reduce error: One is differential GPS technology, and the other is Kalman filter. Because that Kalman filter needs no GPS base stations and data communication devices, neither is it limited by the action scope of signal, we often utilize Kalman filter to reduce the dynamic positioning error.Kalman filter is an efficient recursive filter. With optimization methods, it can be used to estimate the optimal state of a dynamic system from a series of measurements which do not contain noise completely. Kalman filter is the most optimal, and it has the highest efficiency in suppressing positioning error. So the study of Kalman filter of vehicle dynamic navigation and positioning system has great significance. In this paper, the Kalman filter of vehicle GPS/DR integrated navigation is studied in-depth, and following are the results achieved:Firstly, the Kalman filter of vehicle GPS/DR integrated navigation is designed. At the beginning, the specific design of Kalman filter is analyzed. Then the EKF algorithm is applied to vehicle GPS/DR integrated navigation. The system model of GPS/DR navigation is given, while the state equation of system, the observation equation of system and the filtering equation of system are modeled. As the filtering equation of system is established, the optimization algorithm of average acceleration adaptive algorithm is applied to reduce the model error and to improve the accuracy of the filtering estimate The Kalman filter is simulated via Matlab and the simulation results are analyzed. The simulation results show that the Kalman filter is high in positioning accuracy and small in estimation error.Secondly, the Kalman filter is achieved via FPGA. First of all, based on the Matlab simulation results, the system design for hardware implementation of Kalman filter is made. Then the Kalman filtering algorithm is divided into one top-level module and five bottom modules to design. The bottom modules are mykalman, muladd, multi2, div, ramy1, and ramy2. Five bottom modules are designed by using the Megafunction/LPM of Altera, then the top-level module is realized with the five modules. Every model of Kalman filtering algorithm are simulated via Quartua II. The simulation results are given which show that the design meets the requirements, and the Kalman fitering algorithm achieved by FPGA is feasible. Compared to the traditional method of DSP, FPGA loaded with Kalman filtering algorithm has better real-time, parallelism, expansibility, as well as lower power, higher accuracy and computing speed.