| Fleet and missile groups require accurate and reliable navigation information. Forinertial network systems composed of multiple nodes, using master-slave sensor layoutand structure of data fusion, not only can reduce the cost, but realize sensor failuredetection, isolation and reconstruction.This thesis is based on the need of inertialnetwork for pursuing low cost, high accuracy and high reliability, and studies thestructure of inertial network and the data fusion algorithm.The point of this thesis is to improve the accuracy of the navigation information ofthe slave node. The main contents include.Firstly, the general structure of inertial network and two kinds of measurementmodels-static model and dynamic model are introduced.Considered of the backgroundof this thesis, the static model is chosen as measurement model and a master-slaveinertial network model is established.Secondly, the inertial network system is modularized and the the main factorsaffecting the accuracy of inertial network is analysed from different perspectives. Wefocus on the SINS error propagation mechanism,and in accordance with theinsufficiency of strapdown algorithm itself, we use Doppler instrument and barometricaltimeter and some other aid systems to ensure the accuracy of the inertial networksystems. And we establish the error models of these aid systems. We also contrast theerror of master node with slave nodes by simulation and prepare for the data fusionalgorithm.Thirdly, the correction program of position and velocity information of slave nodesis offered. And combined with data fusion theory, we research and summarize the datafusion algorithm of multi-sensor systems. According to the structure of master-slaveinertial network, we take the structure of single master node and three slave nodes asexample to introduce the distributed data fusion algorithm for the system of low costinertial network. The results of simulations show that the algorithm of inertial networkbased on master-slave structure can effectively improve the accuracy of IMU of theslave node, thus ensuring the overall accuracy of the inertial network in the allowablerange.Finally, the inertial network simulation system is structured. According to thesimulation, we study the effect to the entire inertial network of IMU of different levels of precision. This simulation system uses a modular structure, having the advantages ofscalability, it can be the platform of follow-up inertial network simulation system. |
