Optimization Method for Wireless Control System Fault Detection

Abstract:

Wireless network control system has high failure rate, and is difficult to be diagnosed. Wireless network transmission signal effectively reflect the failure categories. In order to effectively detect the wireless network control system fault, this paper presents a fault detection method of correlation dimensional nonlinear timing characteristics for wireless network transmission signal, which mainly improves the traditional correlation dimension extraction algorithm. The method processes and analyzes the collected transmission signal of four types wireless network control system in fault condition, and then extract fault feature through an improved correlation dimension algorithm. It improves the calculation accuracy of the correlation dimension with a standard deviation 15% -30% than that of the traditional algorithm, and it significantly enhances the clustering distribution characteristics, reflecting its superiority in fault detection. Fault detection results show that the improved feature extraction method for correlation dimension can effectively detect failure in wireless network control system, whose accuracy is improved by 21.4%, and has great practical value.

• Periodical: Advanced Materials Research (Volumes 846-847)
• Pages: 782-785
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.846-847.782
• Online since: November 2013
• Authors: Shuang Ye*
• Keywords: Correlation Dimension, Fault Detection, Nonlinear Time Series, Wireless Network

References

[1] Zhongcai Wang, Yongbi Li. The intrusion detection system based on data mining research, [J] Bulletin of Science and Technology, 2012, 28(8): 150-152.

[2] Rabin Bhusal, Sangman Moh, Effects of Directional Antennas in Ad Hoc Networks with Contentionbased Channel Access Mechanism, JCIS, Vol. 3, No. 1, p.27 ~ 37, (2013).

DOI: 10.4156/jcis.vol3.issue1.4

[3] Jian Chen, Hewu Li, Alleviating Near-Far Effect on Spectrum Allocation Optimization in Dynamic Spectrum Access, JCIT, Vol. 8, No. 3, p.399 ~ 409, (2013).

DOI: 10.4156/jcit.vol8.issue3.47

[4] Keda Lu, Li Wan, Jieming Wu. Based on data mining technology of network security event prediction research, [J] Bulletin of Science and Technology, 2012, 28(6): 37-40.

[5] Jie Jia, Yanyan Zhou, Yong Yang, Xiaona Luo, Fast Recursive Identification Algorithm for Nonlinear Time Series model based on improved Extreme Learning Machine, IJACT, Vol. 5, No. 4, p.761 ~ 768, (2013).

DOI: 10.4156/ijact.vol5.issue4.91