Authors : V. Uma Rani and K. Soma Sundaram
Abstract: In recent days, misbehaving node or malicious node detection in Wireless Sensor Networks (WSN) becomes essential, due to its distributed nature and its increasing demand in various applications. Malicious attacks damages communication between sensor nodes causing the loss of packets, reduced forwarding behaviour of nodes and creating insecure data transmission. Trust model is one of the solutions to provide security in WSN but most of the trust models are susceptible to bad mouthing and ballot attack. In this study, we propose a Dirichlet Distribution based Model (DDTM) to detect malicious attacks, like black hole attack, selective forwarding attack and on/off attack. DDTM uses trinomial Dirichlet distribution for trust evaluation of sensor nodes. DDTM uses Dirichlet fusion rule to combine the opinions gathered from neighbouring nodes and standard deviation rule to overcome bad mouthing and the ballot attack of the trust models. Further, in our proposed DDTM, we include a penalty scheme and a dynamic sliding window scheme to find attacks quickly and provide malicious behaviour feedback to the routing model for secure data transmission. The results of proposed DDTM shows an increased ability compared to present trust models to detect node based attacks and an increase in packet delivery ratio of wireless sensor networks.
V. Uma Rani and K. Soma Sundaram, 2019. Dirichlet Distribution Based Trust Model for Malicious Node Detection in Wireless Sensor Network. Journal of Engineering and Applied Sciences, 14: 4191-4199.