Abstract
This paper is to solve the problem stating that applying Intrusion Detection System (IDS) to guarantee security of Wireless Sensor Networks (WSNs) is computationally costly for sensor nodes due to their limited resources. For this aim, we obtain optimal strategies to save IDS agents’ power, through Quantal Response Equilibrium (QRE) that is more realistic than Nash Equilibrium. A stage Intrusion Detection Game (IDG) is formulated to describe interactions between the Attacker and IDS agents. The preference structures of different strategy profiles are analyzed. Upon these structures, the payoff matrix is obtained. As the Attacker and IDS agents interact continually, the stage IDG is extended to a repeated IDG and its payoffs are correspondingly defined. The optimal strategies based on QRE are then obtained. These optimal strategies considering bounded rationality make IDS agents not always be inDefend. Sensor nodes’ power consumed in performing intrusion analyses can thus be saved. Experiment results show that the probabilities of the actions adopted by the Attacker can be predicted and thus the IDS can respond correspondingly to protect WSNs.
Highlights
Wireless Sensor Networks (WSNs) have attracted considerable concerns owing to their broad applications
To save sensor nodes’ power, we have put forward a method based on Quantal Response Equilibrium (QRE) to make Intrusion Detection System (IDS) agents not always be in Defend
A stage Intrusion Detection Game (IDG) that is able to reflect interactions between the Attacker and IDS agents has been formulated, where we have thoroughly considered players’ preferences and have assigned payoffs of players according to Binmore’s method
Summary
Wireless Sensor Networks (WSNs) have attracted considerable concerns owing to their broad applications. Different methods including anomaly-, misuse-, and specification-based detection are computationally expensive, which are costly for small sensor nodes This situation motivates us to seek optimal strategies of intrusion detection to possibly save sensor nodes’ resources. QRE is adopted to seek optimal strategies of saving IDS agents’ power in WSNs. Considering the characteristics of sensor nodes, we construct a stage Intrusion Detection Game to describe interactions between the Attacker (a player) and IDS agents (the other player). (1) we formulate a stage Intrusion Detection Game according to Binmore’s method to study strategies of malicious sensor nodes and IDS agents, which is able to reflect interactions between the Attacker and IDS agents as well as their preferences;. (4) we realize an implementation of applying the repeated Intrusion Detection Game to WSNs based on the algorithm of calculating QRE-based strategies that can predict the Attacker’s future behavior.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
