Research On Security Detection And Control Of Networked Visual Control Systems

The vulnerability of the network makes visual information(e.g.,visual images)and non-visual information(e.g.,control signals)in Networked Visual Control Systems(NVCSs)be highly susceptible to cyber-attacks,especially attacks on images(e.g.,crop attacks,noise attacks and data stealing,etc.).This will lead to information leakage,forgery,or even missing,which will damage the accuracy and integrity of system state and control information and even cause system instability.The current research of non-visual information mainly focuses on cyber-attack detection and security control of networked control systems.The security of visual image information is mainly studied in terms of information confidentiality and forensic technology,etc.However,image processing and transmission usually take a lot of time due to the large amount of image data,which is hard to meet real-time requirements of image encryption,security detection and control.Therefore,this dissertation firstly analyzes the mechanism of image encryption algorithm in conjunction with high real-time requirements of NVCSs.The fast encryption method is then investigated for NVCSs images to guarantee the security of NVCSs image transmission.Furthermore,a security detection method that integrates image encryption and digital watermarking is developed to achieve NVCSs image information protection and integrity detection synchronously.Finally,the security control strategy with collaborative attack detection and data compensation is studied to improve the safe and stable operation of NVCSs.The main work of this dissertation is summarized as follows:(1)To address the information leakage problem of visual information in network transmission,a fast image encryption method based on ranks intersection is proposed.Firstly,the real-time and security requirements of NVCSs for image encryption algorithms is analyzed.Then,a fast image encryption algorithm based on rank-crossing is designed.In the rank-crossing cyclic encryption,the encryption key is cyclically shifted to the right to guarantee the image confidentiality while reducing the image processing delay to meet the real-time requirements.Furthermore,the performance of the proposed algorithm is analyzed to verify the security,high real-time performance and robustness against image tampering attacks such as noise attacks and crop attacks.Finally,a real-time control experiment based on networked inverted pendulum visual control platform are employed to verify that the proposed algorithm can meet real-time requirements in NVCSs.(2)To address the problems of information leakage and tampering faced by image information transmission in NVCSs,a security detection method with integrating image encryption and digital watermarking is proposed to improve the image attack detection capability.Firstly,based on digital watermarking technique and combined with fast image encryption algorithm proposed in(1),the security detection method is presented.The intersections of image encryption ranks is used as the watermark embedding location to achieve image information protection and integrity detection synchronously.Finally,the security,functionality and rapidity of the proposed security detection method are verified by both simulation and practical experiments to demonstrate that the proposed method can achieve effective detection and real-time warning of cyber-attacks.(3)The incomplete data caused by cyber-attacks will affect the control performance and even lead to system instability.To address this problem,security control of NVCSs based on the cooperative mechanism of attack detection and data compensation is studied to improve the safe and stable operation performance.Firstly,the cooperative mechanism between attack detection and data compensation is designed by combining security detection method proposed in(2).Then,an attack detector for image real-time and integrity discrimination is proposed.According to the detection results,an online compensation scheme based on cubic spline interpolation is proposed to reduce the effect of cumulative errors and improve the robustness.Finally,the online compensation scheme is optimized by considering the characteristics of networked inverted pendulum visual control system.Experimental results confirm the feasibility and effectiveness of the proposed compensation mechanism.