An adaptive federated filter algorithm based on improved GA and its application

Wei Quan,Jiancheng Fang
DOI: https://doi.org/10.1117/12.717597
2006-01-01
Abstract:Pointing to some complex systems, general federated filters can not be suit for rather large changes of system parameters, and various inertial devices all exists the defect of error accumulating as time, which cause inaccuracy of model and bad performance of filters. So, in order to meet the requirements of accurate model building, it is very necessary to adaptively adjust for the noise model parameters of integrated navigation system. Based on SINS/CNS/GPS integrated navigation system model for long flight-time unmanned plane, pointed to low precision of model & filtering and stability & practicability of filtering algorithms, an adaptive federated filter algorithm based on improved GA was established in this paper. This algorithm avoids the premature convergence problem of general GA by improving the fitness function, takes advantage of decimal-coded to improve both the speed and the accuracy of calculating, builds the adaptive federated filter model based on improved GA through analyzing the model parameters of reference-system and local-filters. In the end, the semi-physical simulation is done by using this method. The experimental results show that as compared with adaptive federated filter algorithm, this filter not only increase the navigation system's accuracy and reliability greatly, but also owns quick rapidity of convergence. It has high merits of project application.
What problem does this paper attempt to address?