基于惯性网络的分布式信息融合算法

A Distributed Information Fusion Algorithm Based on Inertial Network

摘要:

分布式导航系统是飞机多传感器导航系统设计的新概念,可以大幅提高系统导航性能和容错水平,并能动态配置传感器功能,但是目前并无完善的信息融合算法与之对应;文章在构建惯性传感器网络的基础上,将多个低成本惯性传感器系统配置在飞机的多个位置以作为网络节点,设计了分阶段处理的分布式信息融合算法,综合利用各节点所测量的惯性信息,最后得到本节点的局部状态估计;通过仿真实验表明,采用此方法,有效降低了导航滤波估计误差,因此,系统导航性能及容错能力得到大幅提高。

Abstract:

Distributed navigation system is a new concept designed for the aircraft multi-sensor navigation system. And it can significantly improve the performance and fault-tolerance level of the system. What’s more, distributed navigation system can configure the sensors dynamically. However, there is no corresponding information fusion algorithm to deal with the new method. To solve the problem, a two-step distributed information fusion algorithm is presented in this paper. The distributed information fusion algorithm is based on the structure of inertial network, which consists of IMU (Inertial measurement unit) nodes installed in different parts of the carrier to provide redundant navigation information. With this algorithm, each node will take full use of the other nodes’ inertial information. The simulation results show that using this method, the navigation filtering estimation is effectively reduced, therefore, the performance of navigation system and fault tolerant capability is improved greatly. 