error detection in wireless sensor networks using hidden markov model and fuzzy logic
error detection in wireless sensor networks using hidden markov model and fuzzy logic
MohammadReza Khaleghi1 Atefeh Rezamanesh2 Fatemeh Hasani3
1) Department of Electrical & Computer, Shahrood Science & Research Branch, Islamic Azad University, Shahrood, Iran-
2) Department of Electrical & Computer, Shahrood Science & Research Branch, Islamic Azad University, Shahrood, Iran-
3) Department of Electrical & Computer, Shahrood Science & Research Branch, Islamic Azad University, Shahrood, Iran-
IOI Code
:
XYZG-ZAAFC
Publication :
2nd International Conference on Science and Engineering(2icesconf.com)
Abstract :
In this project, a method for fault diagnosis in telecommunication networks using hidden Markov model is proposed. The rating on the proposed independence from hardware and system structure is discussed Regardless of the type of technology used and the type of network coding can be relatively large range of defects detected on the network.
Hidden Markov models is a method to predict the time series due to its high flexibility and capacity become more widespread. This model learned networking conventional pattern in different scenarios then based on the current situation predict the output of the system. Working out of the normal range can be easily detect by checking multi-interval sequence of a sensor or node easily and faulty communications equipment removed from the system.
The method proposed in this thesis, including network troubleshooting is a centralized approach. This means that the network information send to a supervisor independent of the network and finally the processing is carried out by this supervisor identified faulty nodes and removed from the network. The combination of hidden Markov models with fuzzy logic is also considered one of the strengths of the proposed approach. HMM should be trained based on conditions in the environment and the structure of the network.
Finally, proposed method in a very practical structure were analyzed in the presence of noise and results show its capability to detect the faults and we gain up to 97.11% accuracy.
Keywords : telecommunication networks error detection hidden markov model fuzzy clusting