IEEE Project Abstract

With the increasing connectivity enabled by theInternet of Things (IoT), security becomes a critical concern,and the users should invest to secure their IoT applications.Due to the massive devices in the IoT network, users cannotbe aware of the security policies taken by all its connectedneighbors. Instead, a user makes security decisions based onthe cyber risks he perceives by observing a selected numberof nodes. To this end, we propose a model which incorporatesthe limited attention or bounded rationality nature of playersin the IoT. Specifically, each individual builds a sparse cognitivenetwork of nodes to respond to. Based on this simplified cognitivenetwork representation, each user then determines his securitymanagement policy by minimizing his own real-world securitycost. The bounded rational decision-makings of players and theircognitive network formations are interdependent and thus shouldbe addressed in a holistic manner. We establish a games-ingamesframework and propose a Gestalt Nash equilibrium (GNE)solution concept to characterize the decisions of agents, andquantify their risk of bounded perception due to the limitedattention. In addition, we design a proximal-based iterativealgorithm to compute the GNE. With case studies of smartcommunities, the designed algorithm can successfully identifythe critical users whose decisions need to be taken into accountby the other users during the security management.