Multi-source information fusion method based on factor graph

Abstract

The invention relates to a multi-source information fusion method based on a factor graph. The multi-source information fusion method aims to realize full-source positioning and navigation without relying on satellite navigation in a complex environment, takes an inertial navigation system as the core, utilizes all available navigation information sources, and performs rapid fusion, optimal configuration and self-adaptive switching on asynchronous heterogeneous sensor information. A factor graph model is constructed by means of recursive Bayesian estimation, the factor graph is broadened by means of a variable node and a factor node of the system after measurement information of different sensors are acquired, state recursion and updating are completed based on a set cost function, and thefactor graph optimization problem is solved through sparse QR decomposition by adopting an increment smoothing method. The multi-source information fusion method effectively solves the time-varying state space problem generated between carrier motion and measurement availability, can calculate a solution of precise navigation according to dynamic changes of a carrying platform, realizes plug-and-play of multiple sensors, and meets the requirements of carriers changing in complex environment and different tasks.

Description

technical field [0001] The invention relates to a fusion method, in particular to a multi-source information fusion method based on a factor graph, which is suitable for all-source navigation under complex environments, different requirements and tasks. Background technique [0002] Satellite navigation has vulnerabilities such as weak signal, poor penetrating ability, susceptibility to occlusion, interference and spoofing, which will lead to limited navigation, positioning and timing services of carriers in complex environments. In order to solve the constraints of the inherent vulnerability of satellite navigation systems, the Advanced Military Technology Research Institute led by the US Defense Advanced Research Projects Agency (DARPA) launched the All Source Positioning and Navigation (ASPN) project in 2010. The all-source navigation system takes the timing and inertial measurement unit integrated with inertial navigation and precise clock as the core, and can be flexibl...

AI Technical Summary

Problems solved by technology

[0004] With the increase in the number and types of sensors in the fusion of multi-source information, the complex and changeable application environment and the different requirements and tasks, different sensors have different update frequencies, and it is difficult for all sensors to work at the same time. There is asynchrony or even delay in time; at the same time, with the complex and changeable application environment and the requirements of special tasks, the application of sensors is limited, and there will be situations where one or more sensor measurement information is periodically abnormal or unavailable. Some multi-source information fusion mostly uses filtering methods

However, every time new measurement information is available, the entire extended state variable needs to be updated, increasing the complexity of the algorithm

At the same time, there are also problems such as different update frequencies of multi-source measurement information, time asynchrony, and signal delay when multi-sensor information fusion is performed. For example, the update frequency of most navigation sensors (such as vision, GPS, lidar, etc.) is much lower than that of IMU and The time is not synchronized with each other; some navigation sensors (GPS, etc.) have signals that are not available for a short period of time, and some navigation sensors (indoor navigation, ground-based PNT, underwater acoustic positioning, etc.) can only provide navigation in local areas, etc.

Images