The Analysis Of Accuracy And Computational Time Of Ins/gps System Based On Kalman Filter

In recent years, with the development of electronic computers, especially the rapid development of computer technology and the research of modern control theory, the application of integrated navigation technology has been more and more widely, to aerospace, spacecraft to navigation or even daily vehicleetc..As the name implies, navigation is will be able to carry out a certain combination of various navigation devices work independently through the computer before, so as to achieve a variety of navigation equipment advantage complementary purposes, so as to get higher precision, better reliability of the results.At present, both at home and abroad for this technology has done a lot of research, has launched a variety of series of integrated navigation system, it has become the most important and basic navigation systemnow.Thisthesis studies the key technology of global positioning system in combination of INS/GPS and influence factors on the amount of calculation and the accuracy of different data fusion algorithm.In order to achieve better navigation parameters of the object, INS/GPS integrated navigation system which can achieve advantage complementarity is obtainedbasing on inertial navigation system andglobal positioning system in the paper. The measurement results of the inertial navigation system and the global positioning system will be filtered fusion using the navigation computer. The navigation parameters this thesis concerning with mainly include: the pose of the object angle, the position of an object in spaceand the object's three-dimensional velocity.Thisthesis focus on thedata fusion process of the INS/GPS integrated navigation system. Four data fusion algorithms including basic Kalmanfilter,the extended Kalman filter,unscented Kalman filter and particle filter are researched.Firstly,this paper does the following research basing on the theoretical of the four filtering algorithms:analyzing the factors influencingthe precision or computation and summarize the whole law basing the same kind offiltering system; analyzing the amount of calculation and accuracy of different filtering algorithm. The study also summarizes the application conditions and the advantages and disadvantages of different algorithmwith the combination of all of the above conclusions.Finally,the matlab simulation results are given.The results show that, the filter step, the state vector dimension and initial value is several key factors of affecting the accuracy of the simulation process in the basic KF filter. This paper verified the result by simulation and the optimalexperience value of the input are given.Running EKF and UKF filtering algorithm for integrated navigation respectivelyin the same kind of 18 dimensional system including the same filtering step and the optimal initial value,results show that average service time of UKF is 1.3 times of EKF time, EKF of the error term is 2 times of UKF. In the classical particle filter, the number of particles mainlyaffect the accuracy and the amount of calculation, the simulations further validate this conclusion.Results about four filtering algorithm show that the more complexity the system isthe more nonlinear it is, so it needs more complex filtering fusion algorithm. So the time it consums is longer, but the accuracy will be increased.