Wireless Sensor Networks with Efficient Clone Detection in Terms of Energy and Memory

Abstract

We propose an energy-efficient location-aware clone detection protocol for densely deployed WSNs to ensure successful clone attack detection and satisfactory network lifetime. Using the sensors' geolocation data, we randomly select witnesses within a ring area to attest to the sensors' legitimacy and report any clone attacks they may have uncovered. The witnesses and the sink along the path can receive data with minimal power consumption thanks to the ring topology. For the first time, we theoretically show that the proposed protocol can achieve a clone detection probability of one hundred percent with trustworthy witnesses. In this expanded work, we examine clone detection performance with untrustful witnesses and find that, even with 10% of witnesses being compromised, the clone detection probability is still very close to 99%. The proposed protocol also requires buffer storage of sensors that depends not on the number of sensors, n, but on the network's radius, h, i.e. Oh, whereas the required buffer storage of sensors is typically dependent on the node density in existing clone detection protocols with random witness selection scheme. Extensive simulations show that our proposed protocol can ensure a long network lifetime by evenly distributing the traffic load across the network.