Research Of Filter Algortighm In Civil Aircraft Navigation Sensor

As the development of transportation, aviation, aerospace and navigation, many different kinds of navigation sensor systems emerge to accomplish a more complicated and precise navigation mission. For example, there has been inertial navigation system, wireless transmission navigation system and satellite navigation system. How to make a consummate management scheme for the sensor system to compose different single system to make up an integrated navigation system is a research in great demand in the navigation theory.Global position system and inertial navigation system are both the most advanced navigation system in the world; each has distinctive advantage and can't replace the other. Based on the INS/GPS integrated navigation system, this dissertation research the civil aircraft navigation sensor system, analyze the work principle and compose model of the integrated navigation system and establish a model of the tight coupling of INS/GPS integrated navigation system. Based on the nonlinear model, the unscented kalman filter (UKF) and particle filter are introduced to solve the filtering problem in the tight coupling of INS/GPS integrated navigation system.UKF is different from traditional extended kalman filter (EKF), needn't to compute the Jacobian matrix and achieve second-rank precision or higher, so the estimation precision could be improved greatly. Although UKF overcome the problem of EKF, it still has some problem of the traditional kalman filter. Particle filter is a Mont Carlo algorithm based on recursive propagation. A particle set, which is randomly sampled from probability function and has corresponding weights, is introduced to approach the posterior distribution. Therefore it can handle nonlinear and non-Gaussian model without any limits. Though the particle filter has so many advantages, there are still some existing problems, such as sample impoverishment.Some improved algorithms have been presented in this dissertation since the disadvantage of algorithm above. The adaptive estimation principle is applied for UKF to set up an adaptive UKF. The unscented kalman filter (UKF) is selected as the importance distribution of PF to set up the unscented kalman particle filter (UPF). These algorithms have been analyzed and illuminated separately, and finally are applied to integrated navigation model which is combined into the Flight Management Simulation System to analyze. The simulation result show that the reliability of these algorithms is greatly increased compared to the general algorithm. Furthermore, the real-time of different algorithm has been analyzed by the comparison of the calculated amount.