Research On Securityestimation Of Wireless Sensor Networks Under Malicious Attacks

With the development of science and technology,performance of sensors has been greatly improved.Wireless sensor networks have a wide range of applications in the fields of smart grid,intelligent transportation,aerospace and so on.However,wireless sensor networks are more vulnerable to malicious cyber attacks and more frangible in security due to the intrinsic interconnections among the sensors.Therefore,ensuring the security of wireless sensor networks is more important and urgent than ever.At present,three categories of cyber attacks are commonly investigated: denial-of-service attacks,false data injection attacks and replay attacks.This paper focuses on the security issues of wireless sensor networks and mainly includes three parts:First,an unscented Kalman filtering problem is studied for a nonlinear system with sensor saturation and randomly occurring false data injection attacks.A random variable obeying the Bernoulli distribution is employed to characterize the phenomena of the randomly occurring false data injection attacks.The aim of this paper is to design a modified unscented Kalman filter by minimizing an upper bound of filtering error covariance.A sufficient condition is provided to ensure exponentially bounded of filtering error in the mean square sense.Secondly,the problem of distributed fusion estimation in multi-sensor nonlinear networked system with random access protocol and cyber attacks is considered.For each sensor,the random access protocol is employed to determine which channel has the priority of accessing to the communication networks at each transmission instant.In order to decrease the damage of cyber attacks,a detector based on cross-correlation of causal sequence is designed for local estimators.In addition,a distributed fusion estimation algorithm based on sequential covariance intersection fusion technique is designed using the local estimates calculated by the modified unscented Kalman filter.A sufficient condition is obtained to guarantee the boundedness of the fusion estimation error.Finally,the problem of network security for distributed filtering under malicious cyber attacks over a wireless sensor networks is investigated.A protector based on Dempster-Shafer theory for each sensor is proposed,which according to the online information from its neighboring sensors to decide whether to use the received data.Further,a sufficient condition is presented for the stability of the estimator equipped with the proposed protector under hostile cyber attacks.