Intelligent active fault-tolerant system for multi-source integrated navigation system based on deep neural network
- PDF / 3,859,608 Bytes
- 18 Pages / 595.276 x 790.866 pts Page_size
- 9 Downloads / 231 Views
(0123456789().,-volV)(0123456789().,-volV)
SMART DATA AGGREGATION INSPIRED PARADIGM & APPROACHES IN IOT APPLNS
Intelligent active fault-tolerant system for multi-source integrated navigation system based on deep neural network Chengjun Guo1,2
•
Feng Li1 • Zhong Tian1 • Wei Guo2 • Shusen Tan1
Received: 9 October 2018 / Accepted: 20 December 2018 Ó Springer-Verlag London Ltd., part of Springer Nature 2019
Abstract This paper proposes an intelligent active fault-tolerant system based on deep neural network. That is, an active faulttolerant integrated navigation system is established by adding neural network to the fault-tolerant integrated navigation system based on one-class support vector machine fault detection algorithm. When there is no fault, the neural network trains each sub-filter; when there is a fault, the neural network which has been in the training state will predict the fault time data and use the neural network prediction data to replace the fault data into the main filter for fusion. It can be seen from the simulation analysis that the system can detect the fault of the navigation sub-filtering system well, and when the fault occurs, the prediction data of the neural network is used for information fusion. Simulation results show that the system can provide stable and reliable navigation under the condition of time-varying system and observation noise and complex environment. Keywords Active fault-tolerant One-class SVM Fault detection Deep neural network State estimation
1 Introduction With the development of technology, the requirements for reliability, accuracy, and robustness of multi-sensor navigation systems are getting higher and higher in the recent years [1], and the navigation systems must be developed toward integration and being fault-tolerant, that is, to develop various integrated navigation systems and faulttolerant navigation systems based on inertial navigation system [2]. For fault-tolerant integrated navigation, there are many subsystems inside. Once a subsystem fails, the information fusion feedback may affect each subsystem, resulting in inaccurate or unusable navigation and
& Chengjun Guo [email protected] 1
Research Institute of Electronic Science and Technology, University of Electronic Science and Technology of China, Chengdu 611731, China
2
National Key Laboratory of Science and Technology on Communications, University of Electronic Science and Technology of China, Chengdu 611731, China
positioning information. Therefore, in order to improve the reliability and accuracy of fault-tolerant navigation systems, fault detection and diagnosis are essential. The common fault detection methods include Chisquare detection methods, autonomous integrity monitored extrapolation (AIME), optimal fault detection (OFD), multiple solution separation (MSS) and so on. The Chisquare test method is simulated and validated in paper [3], but this method is based on model, and the accurate realtime system model, measurement model, and noise model are always difficult to get
Data Loading...
