Study On Integrated Navigation System Based On Multiscale Estimation Theory

Considering practical request of high precision inertial integrated navigation system, this dissertation put forward the integrated navigation system based on multiscale estimation theory, which was studied systemically and deeply. The main sections were as follows:The dissertation analyzed multiscale estimation theory of single-sensor and single-model of dynamic system. Combining dynamic system analysis method with multicale signal transform, a method of random signal multiscale decompose and estimation algorithm was researched. The algorithm that was used in integrated navigation system could get better estimation effect than at single scale.Basing multiscale estimation theory of multi-sensors and single-model of dynamic system, the multiscale fuse algorithm and multiscale distribute fuse algorithm were studied respectively. The former was the estimation at scale i based on observation at scales N,N —1,---i. The latter was the estimation at the least scale based on observation at all scales. In order to satisfy request of practical integrated navigation system, this dissertation put forward extend single-type to multi-type of the observation system. The algorithm made multiscale estimation theory more extensive in application.Making use of multiscale model for multisensor observation system, multiscale recursive state fusion estimation algorithm was given and the estimation was gotten based on all observation finally. Theoretically the estimation was unbiased and the estimation error variance was optimal. Integrated navigation system simulation proved that the algorithm could improve estimation precision effectively.Combining discrete wavelet transform, dynamic system theory and stochastic process, multiscale stochastic model was set up regarding scale as variable and recursive data fuse algorithm was given. The algorithm could...