| In order to achieve high guidance-precision and anti-jamming, informationfusion problems in GPS/INS integrated navigation system are studied. Firstly,the principles of GPS and INS is introduced. The paper discusses the necessaryfor GPS and INS to integrate and their combined modes, especially studies theirinformation algorithms in detail. There has important theoretical and practicalsignificance to improve navigation precision of integrated navigation system. The mathematical model of integrated navigation system is establishedfirstly. An adaptive kalman filter algorithm based on fuzzy theory is used in thesystem. This method changes measure noise covariance matrix R on-line basedon the actual measure data and it can improve the filter precision. A layeredfilter algorithm is also introduced in order to reduce the main filter calculation.Secondly, a fuzzy weight reasoning system is established for many minor filtersto fuse their information. This method avoids calculating reverse matrix of P andlessen the main filter calculation, its framework is simple and can achieve easily.Because of navigation system complexity and the uncertain surroundings, therecan't build2 veracious system model and get system noise covariance matrix Qand measure noise covariance matrix R. To solve this problem, H_∞ robustfilter is attempted to apply in navigation system. With unknown Q,R andimprecise system model we can also gain good estimation precision by using H_∞ robust filter, this just supply a gap to kalman filter. The performance of theH_∞ filter is improved due to alterable γf. There are fault data during the dataacquisition due to uncertain disturbance; these data will degrade the navigationaccuracy of whole system. A continuous degressive function is presented in thispaper. This function can eliminate the influence of disturbance and improvefilter precision. Finally, Some experiments have been done to the GPS/INSsystem and the experiment results show that the above algorithms are usable andeffective.
|
PDF Full Text Request