Research On Distributed Navigation System Based On Sensors Network

In order to meet the requirement of the low cost and high performance on navigation system and solve high cost and large volume of the traditional single master inertial navigation system(INS) that ignores the impact of the carrier`s local stochastic perturbation on navigation performance, the low cost MEMS inertial sensor was used in skew redundant configuration in this thesis. The distributed navigation system based on sensors network was designed by installing multi skew redundant inertial measurement unit(SRIMU) nodes in different parts of the carrier. For the purpose of improving the performance of navigation system, key technologies on the distributed navigation system based on sensors network were researched. The main content and innovation points are as follows:Firstly, the error calibration of navigation sensors was researched. A calibration method using the three axis turntable and a static multi-position calibration method were presented for the inertial sensors. In addition, A PSO(particle swarm optimization) algorithm-based error calibration method was presented for tri-axial magnetometer, completing the calibration of bias error and scale factor error. With the error analysis of tri-axial magnetometers, the multi-attitude based direct algorithm was proposed for calibrating the main error—bias error. The simulation results showed that the tri-axial magnetometer measurement error was reduced to 150 n T from 400 n T~1500n T by the PSO algorithm-based error calibration method and the multi-attitude based direct algorithm improved measurement accuracy of the heading angle from 2° to 0.3°.Secondly, it was studied in this thesis that the structure of the skew redundant configuration and the multi-sensor data fusion technology for the sensors network node. With the optimal skew configuration rule, eight-sensor redundant configuration was proposed. For the skew redundant configuration, a new optimal data fusion algorithm based on virtual sensor was proposed, which simplified the dynamic model of measurement system, reduced the computational costs and improved the measurement accuracy. In simulation experiment, compared with the traditional fusion method-the least squares, the accelerometer measurement accuracy was improved better by the new fusion method.Thirdly, the distributed navigation system based on sensors network was designed, which consisted of multi skew redundant inertial measurement unit(SRIMU) nodes installed in different parts of the carrier. With analysis on distributed architecture, the measurement fusion of the distribution navigation system is presented using the equivalent model based on virtual sensor. The simulation results verified that the distributed navigation system provided the high precision navigation information and suppressed the impact of carrier`s local stochastic perturbation on navigation accuracy.Fourthly, in order to verify the effectiveness and performance of the distributed navigation system based on sensors network, the digital simulation platform and hardware platform were built, which verified the validity of the algorithm proposed on this thesis. Finally, as a viable method for the relative navigation, a relative navigation device based on distributed ultra-tight GNSS/INS integration was designed. Through the exchange and sharing of information in multiple carriers installed the device, the relative navigation accuracy and capability of fault tolerance were improved.