Anomaly detection based on an enhanced projection distance

Recent developments in the field of wireless communications have led to a renewed interest in Mobile Ad hoc Networks (MANETs). The dynamic nature of MANETs makes the routing protocols exposed to various attacks. So, it is becoming difficult to ignore security threats in this area. In this paper, we present a novel anomaly detection method based on an enhanced projection distance of the first principal component by utilizing Mahalanobis distance. Furthermore, we have tended to handle the heterogeneity in training data in model construction using cluster analysis. In the pre-processing step of the method, cluster analysis is adopted to divide data into subsets (groups) of data. Then, the first principal component of each subset is extracted in order to calculate the enhanced projection distance. The results show that the proposed approach is promising in achieving a higher detection rate with low false alarm rate.