Title :
Research on fault tolerance based on neural networks
Author :
Rongxue, Knag ; Jian, Song ; Youyun, Zhang
Author_Institution :
Dept. of Automotive Eng., Tsinghua Univ., Beijing, China
fDate :
31 Aug.-4 Sept. 2004
Abstract :
An architecture of fault tolerance based on multi-layer feedforward neural networks is presented in this paper. An algorithm for the fault-tolerance analysis in fault detection is presented based on building a stochastic fault model for feedforward neural networks and analyzing the features fault propagation. A controller for fault-tolerance is designed by controller reconfiguration. Satisfactory simulation results show that this architecture of redundancy has a much higher reliability.
Keywords :
fault tolerant computing; feedforward neural nets; fault detection; fault-tolerance analysis; multilayer feedforward neural network; stochastic fault model; Algorithm design and analysis; Condition monitoring; Control systems; Fault tolerance; Fault tolerant systems; Feedforward neural networks; Multi-layer neural network; Neural networks; Redundancy; Remote monitoring;
Conference_Titel :
Signal Processing, 2004. Proceedings. ICSP '04. 2004 7th International Conference on
Print_ISBN :
0-7803-8406-7
DOI :
10.1109/ICOSP.2004.1442260
