Light-Weight Hidden Markov Trust Evaluation Model for IoT network

Abstract

The open-ended nature of the Internet of Things (IoT) had whipped them vulnerable to a variety of attacks, therefore the need of securing and stabilizing the network while keeping the integrity intact has become the most prominent requirement. Traditionally cryptographic methods were employed to secure networks but the demand of undesirable code size and processing time had given rise to trust-based schemes for addressing the misbehavior of attacks in the IoT networks. With reference to it, several trust-based schemes have been proposed by researchers. However, the prevailing schemes require high computational power and memory s pace; which weakens the network integrity and control. In this context, the paper presents a Light-weight Hidden Markov Model (L/W- HMT) for trust evaluation to alleviate the effect of compromised nodes and restricts the storage of unnecessary data to reduce overhead, memory, and energy consumption. This research work has presented a 2state HMM with Trusted state and compromised state together with essential and unessential output as observation state. Amount of packets forwarded, dropped, modified, and received are the parameters for state transition and emission matrices while the forward likelihood function evaluates the trust value of the node. Simulation performed on MATLAB indicates that the intended L/W-HMT scheme outperforms in connection with detection rate, packet delivery rate and energy consumption, on an average by 6% , 8% and 70% respectively when compared to the similar OADM trus t model.