Abstract:

According to a network model based on the sequential probability ratio test（SPRT），one detection method of selfish routes with respect to the routing functionality in wireless sensor network was developed.Through the analysis of the observed sample values of the nodes，the calculated routes evaluation values were compared with the known threshold values so as to determine whether it was selfish route.Simulation results show that the sequential sampling scheme based on the sequential probability ratio test had high detection accuracy，and the number of the required observation was greatly reduced，so it could operate faster.A merit of SPRT is that the number of observations required to test statistical hypotheses need not be determined in advance when compared to other methods based on a fixed number of observations.This allows for making prompt decision on the behavior of damage to the network，thus it can limit the scope of the damaged networks.

Key words: wireless sensor network, sequential probability ratio test, network model, misbehavior detection, evaluation, NS-2 simulation