Channel-Based Trust Model for Security in Underwater Acoustic Networks
Alberto Signori,Filippo Campagnaro,Ivor Nissen,Michele Zorzi
DOI: https://doi.org/10.1109/jiot.2022.3176374
IF: 10.6
2022-10-08
IEEE Internet of Things Journal
Abstract:Underwater acoustic networks are often used in mission-critical scenarios, such as military underwater networks and assets deployed for tsunami prevention; hence, an attack performed against these types of networks can easily lead to disastrous consequences. Nevertheless, countermeasures to possible network attacks have not been widely investigated so far. A reputation system, where a node gains trust each time it exhibits a good behavior, and loses trust each time it behaves suspiciously, is an effective way to identify possible attackers in the network. The main challenge when applying a reputation system in an underwater network is to understand whether the network performance degrades because a node is acting maliciously intentionally, or because of the changed channel conditions, causing a large packet drop. For instance, when a ship travels close to an underwater network deployment, it causes an increased packet loss, and so does the change of environmental conditions, such as a drop of temperature, the presence of rain or the increase of the wind speed. This behavior of the acoustic channel can be characterized with a hidden Markov model, whose parameters are obtained observing the time evolution of the acoustic channel in a sea experiment. This article presents a trust model based on the knowledge of the channel state, inferred from the perceived noise and received signal strength, in which misbehavior and correct behavior are considered differently according to the actual channel state. We evaluate the model both analytically and through simulations, implementing the trust mechanism in the DESERT Underwater Network framework.
computer science, information systems,telecommunications,engineering, electrical & electronic