Research And Implementaion Of Kalman Filter Based On FPGA

With excellent performance, integrated navigation system has been successfully used in aviation,aerospace, marine, seafaring and other fields,and the first choice for their combination algorithm isKalman filter.For the request of the navigation system performance increased,the Kalman filterimplemented in current way can not meet the real-time,high precision,smaller size and lower powerrequirements in some circumstances.Therefore, a scheme was designed and implemented based onFPGA in this paper.After studying the way of Kalman filter implenmented in project and analysing its real-time, thisdissertation accomplished FPGA-based implenmentations of fixed point and floating point math, thendemonstrated the feasibility of implemention of digital signal processing on FPGA, from wherepropose the research programs of Kalman filter based on FPGA.Among the Kalman filter implementation programs, the IP core and AccelDSP techniques werestudied and implemented respectively. In the situation of making the IP cores as the maincalculation compents, the key techniques of multiplication of matrixes, the application of Finite StateMachine and the same source used at different time were studied.Through exporting the results to PCand comparing them to the results from MATLAB and DSP respectively in terms of real-time andaccuracy, demonstrated the Kalman filter implemented on IP core can achive the real-time andaccuracy.Through a series of performance optimization processes, the Kalman filter wasimplenmented on the AccelDSP and better results were got. After that, the two ways’s characteristicsin implementing Kalman filter were analyzed and evaluated.IP core and AccelDSP were integrated to implemented the eight order Kalman filter inSINS/GPS.Through comparing with MATLAB simulation results, get the conclusion that it canachieve high real-time and precision simultaneity.In conclusion, Kalman filter can get high precision and real-time simultaneity based on FPGA, so itis an effective way to improve the integrated navigation system’s performance.