Hybrid-driven-based fuzzy secure filtering for nonlinear parabolic partial differential equation systems with cyber attacks

The problem of hybrid-driven fuzzy filtering for nonlinear semi-linear parabolic partial differential equation systems with dual cyber attacks is investigated. First, a Takagi-Sugeno (T-S) fuzzy model is employed to reconstruct the original nonlinear systems. Second, a hybrid-driven mechanism is applied for filter design to balance the systems' performance and the limited network resource consumption under dual cyber attacks composed of deception and denial of service attacks. Based on the Lyapunov direct method, sufficient conditions to guarantee the stability of the augmented system are obtained, and the parameters of the designed filter are earned with explicit form. Finally, a simulation example with robust analysis and comparative analysis is provided to illustrate the effectiveness of the proposed method.     

Two-channel false data injection attacks on multi-sensor remote state estimation

This paper investigates a security problem in cyber-physical systems (CPSs) under multisensory framework. In the considered system, the feedback and feed-forward channels may suffer from false data injection (FDI) attacks. The attacker aims at disturbing the systems by modifying the sensors' measurements and injecting external control inputs while keep ( $ϵ,\delta$)-stealthiness. Two types of multi-sensor fusion Kalman filters, which are equivalent in the absence of attacks, are adopted by the remote state estimator. When the system is maliciously attacked, the necessary and sufficient condition is obtained to guarantee the equivalence of the sequential Kalman filter and data compression Kalman filter under two-channel attacks. Moreover, suboptimal ( $ϵ,\delta$)-stealthiness attack strategies for these Kalman filters are obtained by applying two-stage optimization method, convex analysis and monotonic optimization techniques. Finally, simulations on a numerical example of a quadrotor are carried out to validate the feasibility of theoretical results.     

面向网络取证的网络攻击追踪溯源技术分析

首先定位网络攻击事件的源头,然后进行有效的电子数据证据的收集,是网络取证的任务之一.定位网络攻击事件源头需要使用网络攻击追踪溯源技术.然而,现有的网络攻击追踪溯源技术研究工作主要从防御的角度来展开,以通过定位攻击源及时阻断攻击为主要目标,较少会考虑到网络取证的要求,从而导致会在网络攻击追踪溯源过程中产生的大量有价值的数...     

Hybrid-triggered scheme asynchronous control for hidden Markov jump systems with partially unknown probabilities and cyber attacks

This paper is concerned with the issue of hybrid-triggered scheme $H_{\infty }$ asynchronous control for networked Markov jump systems (MJSs) with probabilistic cyber attacks. First, in view of the subsistent phenomenon that the controller cannot capture the system modes synchronously, the hidden Markov model (HMM) with partly unknown probabilities is introduced in this article to describe such asynchronous phenomenon. In addition, the hybrid-triggered scheme includes time-triggered scheme and event-triggered scheme, in which the switching signal between two schemes obeys random Bernoulli distribution. By utilizing Lyapunov stability theory and matrix inequality technique, some sufficient conditions are developed, which can guarantee the networked MJSs mean-square asymptotically stable (MSAS) and $H_{\infty }$ performance; also, the desired controller gains are obtained by singular value decomposition (SVD) methods. Moreover, as the special cases of the results above, three corollaries are given. Finally, a numerical example and a pulse width-modulation-driven boost converter (PWMDBC) model are provided to demonstrate the usefulness and reliability of our developed approaches.     

Security alert: Generalized deep multi-view representation learning for crime forecasting

Crime is a focal problem in modern society, affecting social stability, public safety, economic development, and life quality of residents. Promptly predicting crime occurrence places in a relatively high accuracy is a very important and meaningful research direction. Via the rapid development of social media (e.g., Twitter), the online information can act as a strong supplement for the offline information (crime records). Additionally, the geographic information and taxi flow between communities can model the spatial relationship between communities, which has already been confirmed effective in previous work. In order to efficiently solve crime prediction problem, we propose a generalized deep multi-view representation learning framework for crime forecasting. Our extensive experiments on a 4-month city-wide dataset that consists of 77 communities and 22 crime types show our model improve the prediction accuracy on most crime types.     

Dynamic event-triggered H∞$$ {H}_{\infty } $$ control for continuous-time singular system with cyber attacks

This essay discusses the $H_{\infty }$ control for continuous-time singular systems with two types of stochastic cyber attacks and dynamic event-triggered scheme. So as to economize network resources and reduce communication burden, the novel dynamic event-triggered mechanism is given. Cyber attacks occur randomly during network information transmission, especially the denial of service (DoS) attacks and the deception attacks. By using Lyapunov–Krasovskii functional, some novel sufficient conditions for the impulse free, regular, and asymptotically stable of the studied system are obtained, and a corresponding controller is devised. In addition, a desired controller gain is received through resolving linear matrix inequalities. Eventually, a few numerical examples are supplied to certify the usefulness of provided design ways and results.     

基于开源信息平台的威胁情报挖掘综述

网络空间新生威胁日趋复杂多变,传统安全防御手段已经捉肘见襟.网络安全威胁情报作为直接或潜在安全威胁的外部信息资源,可帮助安全人员快速甄别恶意威胁攻击并及时作出响应防御.开源威胁情报挖掘技术可从多方开源情报中获取高质量情报,极大弥补了传统威胁情报挖掘信息量单薄等不足.美国及欧洲是最早在政府层面开展开源情报挖掘技术研究的国...     

Resilient control of cyber-physical systems using adaptive super-twisting observer

In this work, the problem of online secure state estimation and attack reconstruction in the face of offensives that corrupt the sensor measurements and modify the actuator commands of cyber–physical systems is investigated for designing a resilient controller for the system. The states of cyber–physical system and its actuator attacks are estimated/reconstructed online using a novel adaptive line-by-line super-twisting observer, whereas sparse stealth attacks on unprotected sensors are reconstructed using a sparse recovery algorithm. The estimated attacks are used for attack compensation by a resilient controller. The efficacy of the proposed technique is illustrated via simulation on a real electric power system under deception actuator attack and stealth sensor attack.