ABSTRACT
As we all known, networks technologies have been widely applied in NCSs [48, 145, 93]. NCSs are not a class of isolated systems anymore by transmitting control and measurement signals over public networks. Not only are mobility and interoperability increased, but also cost in maintenance and installation is reduced in NCSs [75, 181, 149]. However, NCSs are much vulnerable to various malicious attacks [106, 105, 184], such as data injection attacks [98], replay attacks [188], DoS attacks [21], and so on. It is worth emphasizing that NCSs have a very high demand for real time, however, actuators or controllers do not receive signals on time due to DoS attacks [128, 171, 175]. Therefore, DoS attacks which influence availability of signals cause serious harm on NCSs. Thus a filtering algorithm is designed to monitor malicious attacks firstly. Then combined with the designed filtering algorithm, a control algorithm is designed to deal with malicious attacks. In security research of NCSs, there are two main methods of combining filtering algorithms with control algorithms. One is based on the method of Fault Detection and Isolation (FDI). The FDI which determines whether alarm is mainly based on comparison of a given threshold value and the difference between a measured value and a estimated value [9]. The other is based on the method of IDS. The IDS is placed in transmission networks to filter malicious attacks before NCSs are attacked; thus integrity and availability of signals are ensured in NCSs [170, 193]. Note that both algebraic graph methods and game theory are used to solve security problems on NCSs [193, 143]. Moreover, the game theory which are a class of useful methods on modeling security issues in NCSs is well-established in [8]. Nevertheless, not only do DoS attacks cause damage, but also actuator saturation has an effect on NCSs.
