An Efficient Polynomial-based Filtering Against False Data Injection Attack in CPNS

Abstract

Cyber Physical Network System (CPNS) is gaining lot of attention in many applications like, transportation networks, vehicular networks, life-critical applications and many more. Hence, the system needs to be protected from various kinds of attacks that degrade the system’s performance. There are many different types of attacks that are possible on cyber physical systems, among them false data injection attack is a serious threat to the system’s security. In this type of attack, the adversary compromises sensor nodes, inject false data and send them to the controller through compromised nodes. This makes the controller to estimate wrong system states which leads to various serious issues. Therefore, the false data must be filtered out before it reaches the sink. If all the false data flow towards the controller then it will be bottle neck to filter all the false data and this could paralyze the network. To resolve this issue many filtering schemes have been developed in the past, all use Message Authentication Codes (MACs) for report endorsement and en-route filtering. But they are not suitable for CPNS because of static routes and lack resilience to the number of compromised nodes. Hence, an enhanced scheme has been proposed which uses polynomials instead of MAC for report endorsement and also uses bloom filtering along with en-route filtering. Hence, this achieves high resilience to the number of compromised nodes and achieves high filtering efficiency.