Multi-target integrated safety control system for ICPS under dual DoS attack under dual-end ADETCS

Abstract

The present invention discloses a multi-target integrated safety control system for ICPS under dual DoS attack under dual-end ADETCS, including: a first adaptive discrete event generator is provided in the sensing unit, which is used to screen the acquired ICPS system sampling data, and send the ICPS system sampling data satisfying the trigger mechanism to the control unit; a second adaptive discrete event generator is provided on the connection channel between the control unit and the execution unit, which is used to screen the control unit output data, and send the control unit output data satisfying the trigger mechanism to the execution unit; these two adaptive discrete event generators constitute a dual-end ADETCS. Through this system, in the process of data transmission, on the basis of ensuring the saving of network resources of the sensing unit, the network resources and optimization problem capabilities of the execution unit are further improved, so as to better achieve the optimal compromise balance between control performance and communication resource saving.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment safety control, and more particularly to a multi-target integrated safety control system under a dual-end ADETCS where an ICPS is subjected to a double DoS attack. Background Art

[0002] Industrial cyber physical system (ICPS) is a multi-dimensional complex system that integrates computing, networking and physical environment. It realizes real-time perception, dynamic control and information services of large-scale engineering systems through the organic integration and deep collaboration of computing, communication and control technologies.

[0003] In terms of ICPS communication, network bandwidth is a key resource that determines the speed and efficiency of data transmission in the system. In order to effectively manage and utilize bandwidth resources, introducing a trigger mechanism is a common method. It can trigger data transmission based on specific conditions or events, thereby improving bandwidth utilization and system performance. Depending on the trigger conditions, the event-triggered communication mechanisms that can be used are also different. For example, the discrete event trigger mechanism uses pre-set parameters to filter the data transmitted after sampling, thereby saving network resources, but its trigger threshold is fixed and cannot change dynamically according to the behavior of the system; the adaptive discrete event triggered communication mechanism (ADETCS) can dynamically adjust the trigger parameters, and its trigger threshold can be adaptively adjusted according to the current system operation status to meet the needs, ensuring the dynamic performance of the system. Therefore, this trigger mechanism can achieve the desired dynamic performance and save network resources. However, the single-end ADETCS does not consider the issues of network resource saving and system optimization at the execution end.

[0004] Therefore, how to effectively enhance the ability to resist two-end DoS attacks and further improve the network resources and optimization capabilities of the execution end, so as to better achieve an optimal compromise balance between control performance and communication resource conservation, is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of the above problems, the present invention provides a multi-target integrated safety control system for ICPS under dual-end ADETCS under dual DoS attacks, so as to at least solve some of the technical problems mentioned in the above background technology.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] The present invention provides a multi-target integrated safety control system for ICPS under dual-end ADETCS under dual DoS attack, comprising a sensing unit located at the output end of the controlled object, an execution unit located at the input end of the controlled object, and a control unit located between the sensing unit and the execution unit, wherein the sensing unit, the control unit and the execution unit form a closed-loop control;

[0008] The sensing unit is provided with a first adaptive discrete event generator; the first adaptive discrete event generator is used to filter the acquired ICPS system sampling data, and send the ICPS system sampling data that meets the trigger mechanism of the first adaptive discrete event generator to the control unit;

[0009] A second adaptive discrete event generator is provided on the connection channel between the control unit and the execution unit; the second adaptive discrete event generator is used to filter the control unit output data and send the control unit output data that meets the trigger mechanism of the second adaptive discrete event generator to the execution unit;

[0010] The first adaptive discrete event generator and the second adaptive discrete event generator constitute a double-terminal ADETCS.

[0011] Further, the first adaptive discrete event generator triggering mechanism is expressed as:

[0012]

[0013] Among them, Φ y is the event trigger weight matrix to be designed in the sensor unit; is the event triggering error of the first adaptive discrete event generator, and is the ICPS system sampling data input to the first adaptive discrete event generator at time k; is the ICPS system sampling data output by the first adaptive discrete event generator at time k; is an adaptive trigger parameter of the first adaptive discrete event generator;

[0014] The triggering mechanism of the second adaptive discrete event generator is expressed as:

[0015]

[0016] Among them, Φ u is the event trigger weight matrix to be designed in the execution unit; is the event trigger error of the second adaptive discrete event generator, and is the control unit output data input to the second adaptive discrete event generator at time k; The control unit output data output by the second adaptive discrete event generator at time k; is the adaptive trigger parameter of the second adaptive discrete event generator.

[0017] Further, the adaptive trigger parameter of the first adaptive discrete event generator or the second adaptive discrete event generator is expressed as:

[0018] σ(t k h)=max{σ min ,ρσ(t k-1 h)},σ(t k h)∈[σ min ,σ max ]

[0019]

[0020] Among them, σ(t k h) represents the adaptive trigger parameter of the first adaptive discrete event generator or the second adaptive discrete event generator at time k; σ(t k-1 h) represents the adaptive triggering parameter of the first adaptive discrete event generator or the second adaptive discrete event generator at time k-1; σ min Represents the lower limit of the adaptive trigger parameter; σ max Represents the upper limit of the adaptive trigger parameter; and 0≤σ min <σ max <1; real number β>0, Δ is the normalized value of the data transmission error at different times, set Ensure that the threshold coefficient ρ is within the monotonic interval; y(t k h) is the data output by the first adaptive discrete event generator or the second adaptive discrete event generator at time k; y(t k-1 h) is the data output by the first adaptive discrete event generator or the second adaptive discrete event generator at time k-1.

[0021] Furthermore, the judgment mechanism of the dual-end ADETCS includes:

[0022] When ||y(t k h)|| and||y(t k-1 h)|| is greater than or equal to zero, 0≤ρ<1, then σ(t k h)<σ(t k-1 h), the data output by the ICPS system increases, when σ(t k h)=σ min When σ(t kh) no longer decreases;

[0023] When ||y(t k h)|| and||y(t k-1 h)|| is less than 0, then σ(t k h)≥σ(t k-1 h), the data output by the ICPS system decreases, when σ(t k h)=σ max When σ(t k h) No longer increase.

[0024] Further, the control unit includes a first attack detector;

[0025] The first attack detector is used to determine the extent to which the sensor unit is subjected to a DoS attack according to a data sampling interval of the ICPS system;

[0026] If the data sampling interval is less than the maximum delay allowed by the ICPS system, it is determined that the sensor unit is subjected to a small-energy DoS attack or is not subjected to a DoS attack; and when the sensor unit is subjected to a small-energy DoS attack, a model mechanism strategy is used to robustly respond to the DoS attack elastic control;

[0027] If the data sampling interval is greater than or equal to the maximum delay allowed by the ICPS system, it is determined that the sensor unit is subjected to a large-energy DoS attack, and principal component analysis-long short-term memory neural network is used to compensate for the data lost due to the DoS attack, that is, data-driven.

[0028] Further, the execution unit includes a second attack detector;

[0029] The second attack detector is used to determine whether the execution unit is subjected to a DoS attack according to the data receiving interval of the execution unit and the data sampling interval of the ICPS system;

[0030] If the data receiving interval of the execution unit is greater than or equal to the data sampling interval, it is determined that the execution unit is subjected to a DoS attack; and the principal component analysis-long short-term memory neural network is used to compensate for the data lost due to the DoS attack;

[0031] If the data receiving interval of the execution unit is smaller than the data sampling interval, it is determined that the execution unit is not subjected to a DoS attack.

[0032] Furthermore, by introducing the safety stability margin and the disturbance suppression capability coefficient, a Lyapunov-krascovskii functional is constructed; and based on the Lyapunov-krascovskii functional, an observer and a controller in the control unit under the double-terminal ADETCS are set.

[0033] The Lyapunov-krascovskii functional is expressed as:

[0034] V(t)=V1(t)+V2(t)+V3(t)+V4(t)+V5(t)

[0035] in:

[0036]

[0037] Wherein, V(t) represents the Lyapunov-krascovskii functional; V1(t) represents the functional at the current moment; V2(t) represents the time-delay functional at the current moment; V3(t) represents the single-integral functional at the current moment; V4(t) represents the single-integral functional of the derivative at the current moment; V5(t) represents the double-integral functional of the derivative at the current moment; Indicates that α Stability is the current state of the system; τ(t) represents the system delay under multi-objective constraints; P, S, Q, R and Z are all positive definite symmetric matrices.

[0038] Furthermore, the setting objectives of the observer include:

[0039] (1) When the ICPS system is free of disturbance, that is, At this time, the augmented error system of the observer has a safety stability margin; wherein, represents the disturbance of the system under multi-objective constraints; represents the sensor measurement noise under multi-objective constraints; τ(t) represents the system delay under multi-objective constraints;

[0040] (2) When the ICPS system is disturbed, that is, At this time, the augmented error system of the observer meets the following performance indicators:

[0041]

[0042] Among them, γ1 is the disturbance suppression rate, ||·||2 is the norm of L2[0,∞); It represents the current state of the system with α stability; α1 represents the stability margin in the observer acquisition process.

[0043] Furthermore, the setting targets of the controller include:

[0044] When the ICPS system is free of disturbance, that is, At this time, the closed-loop fault ICPS model of the controller has a safety stability margin; wherein, Consider α System disturbances during stability; Consider α State error at stability; Consider α Fault error during stability;

[0045] When the ICPS system has disturbances, that is, At this time, the closed-loop fault ICPS model of the controller meets the following performance indicators:

[0046]

[0047] Among them, ζ(t) is the system state quantity after the stability margin α2 is introduced in the controller acquisition process; is the disturbance coefficient under the system state quantity ζ(t); γ′1 is the disturbance suppression rate, ||·||2 is the norm of L2[0,∞); α2 represents the stability margin in the controller acquisition process.

[0048] It can be seen from the above technical solution that, compared with the prior art, the present invention discloses a multi-target integrated safety control system for ICPS under dual-end ADETCS under dual DoS attack, which has the following beneficial effects:

[0049] The present invention uses a two-terminal adaptive discrete event triggered communication mechanism (ADETCS) to ensure that the network resources of the sensing unit are saved during data transmission, while also further improving the network resources and optimization problem capabilities of the execution unit, thereby better achieving an optimal compromise balance between control performance and communication resource saving.

[0050] Based on the dual-end ADETCS, the present invention combines data-driven and model mechanism methods for DoS attacks. In the sensing unit, small-energy DoS attacks are processed by elastic control based on the model mechanism, while large-energy DoS attacks are processed by PCA-LSTM based on data-driven prediction models to reconstruct and compensate for data packets lost due to DoS attacks. In the execution unit, the attack energy is no longer distinguished. When a DoS attack is detected, the prediction model established based on PCA-LSTM is used to predict and compensate for it. This method effectively improves the stability of data transmission.

[0051] This paper combines the time-delay system theory with less conservative technology, innovatively designs the Lyapunov-krascovskii functional, and imposes multi-objective constraints on the observer and controller to improve the performance of the observer and controller.

[0052] Other features and advantages of the present invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the present invention.

[0053] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0055] Figure 1 A schematic diagram of a multi-target integrated safety control system framework in which an ICPS is subjected to a double DoS attack under a dual-end ADETCS provided in an embodiment of the present invention.

[0056] Figure 2 A schematic diagram of a data transmission update sequence provided by an embodiment of the present invention.

[0057] Figure 3 A schematic diagram of observer state estimation error provided by an embodiment of the present invention.

[0058] Figure 4 A schematic diagram of a system output response curve provided by an embodiment of the present invention.

[0059] Figure 5 A schematic diagram of output error of adaptive trigger parameters provided by an embodiment of the present invention.

[0060] Figure 6 A schematic diagram of an adaptive trigger parameter change curve provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] The embodiment of the present invention discloses a multi-target integrated safety control system in which an ICPS is subjected to a double DoS attack under a double-end ADETCS, comprising a sensing unit located at the output end of a controlled object, an execution unit located at the input end of the controlled object, and a control unit located between the sensing unit and the execution unit, wherein the sensing unit, the control unit and the execution unit form a closed-loop control; wherein a first adaptive discrete event generator is provided in the sensing unit; the first adaptive discrete event generator is used to filter the acquired ICPS system sampling data, and send the ICPS system sampling data satisfying the trigger mechanism of the first adaptive discrete event generator to the control unit; a second adaptive discrete event generator is provided on the connection channel between the control unit and the execution unit; the second adaptive discrete event generator is used to filter the control unit output data, and send the control unit output data satisfying the trigger mechanism of the second adaptive discrete event generator to the execution unit; the above-mentioned first adaptive discrete event generator and the second adaptive discrete event generator constitute a double-end ADETCS.

[0063] The present invention uses a two-terminal adaptive discrete event triggered communication mechanism (ADETCS) to ensure that the network resources of the sensing unit are saved during the data transmission process, and the network resources and optimization problem capabilities of the execution unit are further improved, thereby better achieving an optimal compromise balance between control performance and communication resource saving. The above content is described in detail below.

[0064] 1. Double-ended ADETCS

[0065] (1) Design of two-terminal ADETCS

[0066] ADETCS can automatically adjust the trigger threshold according to the operating status of the ICPS system, so that the system performance and network resource conservation can be balanced. In view of the problem that the single-end ADETCS does not consider the insufficient conservation and optimization of network resources on the execution side, the embodiment of the present invention introduces a dual-end ADETCS, i.e., a first adaptive discrete event generator and a second adaptive discrete event generator, and designs the following adaptive trigger mechanism to determine whether the sampled data is transmitted:

[0067] 1) The first adaptive discrete event generator triggering mechanism is expressed as:

[0068]

[0069] Among them, Φ y is the event trigger weight matrix to be designed in the sensor unit; is the event triggering error of the first adaptive discrete event generator, and is the ICPS system sampling data input to the first adaptive discrete event generator at time k; is the ICPS system sampling data output by the first adaptive discrete event generator at time k, and h is the sampling period; l y The interval between the current sampled data on the sensor side and the latest transmitted data divided by the sampling period; is the ICPS system sampling data obtained at time k; The ICPS system sampling data that satisfies the triggering mechanism of the first adaptive discrete event generator at time k; is an adaptive trigger parameter of the first adaptive discrete event generator.

[0070] 2) The triggering mechanism of the second adaptive discrete event generator is expressed as:

[0071]

[0072] Among them, Φ u is the event trigger weight matrix to be designed in the execution unit; is the event trigger error of the second adaptive discrete event generator, and is the control unit output data input to the second adaptive discrete event generator at time k; The control unit output data output by the second adaptive discrete event generator at time k; The control unit output data obtained at time k; Output data for the control unit that satisfies the triggering mechanism of the second adaptive discrete event generator at time k; and h is the sampling period; l u The interval between the current sampled data on the execution side and the latest transmitted data divided by the sampling period; is the adaptive trigger parameter of the second adaptive discrete event generator.

[0073] 3) Adaptive trigger parameters of the first adaptive discrete event generator or the adaptive trigger parameter of the second adaptive discrete event generator Can be expressed as:

[0074] σ(t k h)=max{σ min ,ρσ(t k-1 h)},σ(t k h)∈[σ min ,σ max ] (3)

[0075]

[0076] Among them, σ(tk h) represents the adaptive trigger parameter of the first adaptive discrete event generator or the second adaptive discrete event generator at time k; σ(t k-1 h) represents the adaptive triggering parameter of the first adaptive discrete event generator or the second adaptive discrete event generator at time k-1; σ min Represents the lower limit of the adaptive trigger parameter; σ max Represents the upper limit of the adaptive trigger parameter; and 0≤σ min <σ max <1; real number β>0, Δ is the normalized value of the data transmission error at different times, set Ensure that the threshold coefficient ρ is within the monotonic interval; y(t k h) is the data output by the first adaptive discrete event generator or the second adaptive discrete event generator at time k; y(t k-1 h) is the data output by the first adaptive discrete event generator or the second adaptive discrete event generator at time k-1.

[0077] 4) The judgment mechanism of the above two-terminal ADETCS includes:

[0078] When ||y(t k h)|| and||y(t k-1 h)|| is greater than or equal to zero, 0≤ρ<1, then σ(t k h)<σ(t k-1 h), the data output by the ICPS system increases, when σ(t k h)=σ min When σ(t k h) no longer decreases; when ||y(t k h)|| and||y(t k-1 h)|| is less than 0, then σ(t k h)≥σ(t k-1 h), the data output by the ICPS system decreases, when σ(t k h)=σ max When σ(t k h) No longer increase.

[0079] (2) Analysis of ICPS-related delay and data transmission timing

[0080] The transmission of data packets in the ICPS system is as follows Figure 2 As shown, under double-terminal ADETCS, combined Figure 1 The impact of double DoS attacks on system data transmission in the ICPS system is shown in Figure 1. Considering the data transmission law of the system, the sampling period is h, and the timing analysis of data transmission in the ICPS system is as follows:

[0081] 1) In the sensor unit, the sampling time sequence {i1,i2,…,i k ,…}; the sending time sequence that passes through the first adaptive discrete event generator and satisfies the trigger mechanism and is successfully transmitted is The time sequence of successful transmission from the sensor unit wireless network to the control unit is: The data time sequence after control unit compensation is: Then the input and output of the first adaptive discrete event generator are y(i k )∈R m and y(t k )∈R m ; The data after compensation at the sensor end under DoS attack is y b (j k )∈R m ; If the sensor unit has no DoS attack or only has a small energy DoS attack, the data sequence transmitted to the sensor end and the execution end is y(j k )∈R m ; m represents the dimension.

[0082] 2) In the execution unit, the output data of the controller is obtained through the second adaptive discrete event generator and the sending time sequence of the successful transmission of the trigger mechanism is: The time sequence of successful transmission to the execution unit through the execution end wireless network is: The data time sequence after execution unit compensation is: The input and output of the second adaptive discrete event generator are u(i k )∈R m and u(t k )∈R m , the data after compensation at the execution end under DoS attack is u b (j k )∈R m ; If there is no DoS attack on the execution end, the data sequence transmitted to the execution end is u(j k )∈R m .

[0083] 3) Maximum delay allowed by the ICPS system

[0084] Figure 1The controlled object of ICPS is continuous, while the calculations of the observer and controller in the control unit are all performed in the form of digital quantities. The introduction of two-terminal ADETCS enables the data that meets the requirements of the ICPS system to be transmitted in a non-uniform manner. Based on the theoretical analysis method of time-delay systems, the non-uniform transmission properties are converted into time delays, and then the design methods of the observer and controller are studied in a continuous manner. Under the two-terminal ADETCS, considering the impact of dual limited energy DoS attacks, the maximum delay allowed by the ICPS system is expressed as:

[0085] τ(t)=τ y (t)+τ u (t)+τ gc (t) (4)

[0086] in:

[0087] Among them, τ y (t) is the time delay of the sensor unit; τ u (t) is the delay of the execution unit; τ gc (t) is the calculation delay of the control unit; τ(t) represents the system delay under multi-objective constraints, and satisfies h≤τ(t)<h m ,h m is the maximum delay allowed by the ICPS system; is the time when the sensor side starts transmitting at the kth moment; is the time when the execution side starts transmission at the kth moment; is the time series of the successful transmission from the sensor unit to the control unit at the kth moment; is the time sequence that the control unit successfully transmits to the execution unit at the kth moment;

[0088] (3) Combining data-driven and model-based approaches to counter DoS attacks

[0089] 1) The control unit includes a first attack detector, which is used to judge the degree of DoS attack suffered by the sensor unit according to the data sampling interval of the ICPS system; if the data sampling interval is less than the maximum delay allowed by the ICPS system, it is judged that the sensor unit is subjected to a small-energy DoS attack or is not subjected to a DoS attack; and when the sensor unit is subjected to a small-energy DoS attack, a model mechanism strategy is used to compensate for the data lost due to the DoS attack; if the data sampling interval is greater than or equal to the maximum delay allowed by the ICPS system, it is judged that the sensor unit is subjected to a large-energy DoS attack, and a principal component analysis-long short-term memory neural network is used to compensate for the data lost due to the DoS attack, that is, data-driven.

[0090] 2) The execution unit includes a second attack detector, which is used to determine whether the execution unit is subjected to a DoS attack based on the interval at which the execution unit receives data and the data sampling interval of the ICPS system; if the interval at which the execution unit receives data is greater than or equal to the data sampling interval, it is determined that the execution unit is subjected to a DoS attack; and a principal component analysis-long short-term memory neural network (PCA-LSTM) is used to compensate for the data lost due to the DoS attack; if the interval at which the execution unit receives data is less than the data sampling interval, it is determined that the execution unit is not subjected to a DoS attack.

[0091] Relatively speaking, there are three reasons why energy classification is not performed in the execution unit: first, the data passing through the sensor unit has been compensated; second, the controller's elastic control robust response has improved the system's performance and enhanced its adaptability to external interference and uncertainty; third, the data received by the execution end has been processed by the integrated security controller, effectively reducing the risk of attack. Therefore, in order to improve efficiency and simplify defense strategies, differentiated responses are implemented for DoS attacks on the two sensor units and the execution unit.

[0092] 2. Co-design of ICPS security control and communication under multi-objective constraints under DoS attacks

[0093] In addition to stability, ICPS is also expected to have good dynamic performance. Therefore, this section will construct the Lyapunov-krascovskii functional by introducing the safety stability margin and the disturbance suppression capability coefficient; and based on the Lyapunov-krascovskii functional, the observer and controller in the control unit of the two-terminal ADETCS are set.

[0094] (1) Constructing an augmented error system for state and fault estimation with α-stability

[0095] The ICPS integrated safety control framework under dual-terminal ADETCS describes the sampling data system and introduces a linear ICPS model with actuator failure under limited dual DoS attacks:

[0096]

[0097] Where: x(t)∈R n is the ICPS system state vector; is the derivative of x(t); u(t)∈R n is the control input vector; y(i k )∈R m To obtain the ICPS system sampling data, is the ICPS system disturbance; Measuring noise for sensors; For a continuous time-varying actuator fault, it is assumed that the derivative norm of the fault is bounded, that is, there exists a constant f1 and A, B, E f , B w , C, E v are all known matrices of suitable dimension.

[0098] Next, the dynamic performance of the observer is improved by introducing the α-stability constraint (α1 is used in this paper to represent the stability margin in the observer acquisition process). Specifically, the ICPS system state vector Then the linear ICPS model after introducing α-stability is: It means that the stability margin α1 is introduced in the form of an exponential function when designing the observer.

[0099]

[0100] in, represents the derivative of the system state estimate ψ(t) under multi-objective constraints; represents the fault estimation value under multi-objective constraints The derivative of represents the disturbance of the system under multi-objective constraints; represents the sensor measurement noise under multi-objective constraints;

[0101]

[0102] Where I represents the identity matrix; A represents a known appropriate-dimensional matrix;

[0103] Finally, we can obtain the state and fault estimation augmented error system with α-stability:

[0104]

[0105] in: A quantity that represents the current state; Indicates that α Stability is the current state of the system; for The derivative of

[0106]

[0107] in, represents the disturbance of the system under multi-objective constraints; L represents the observer gain matrix; F represents the fault estimation observer gain matrix; Consider α Fault error during stability;

[0108] (2) Setting the observer in the control unit of the double-ended ADETCS

[0109] 1) Design goal of an α-stable observer

[0110] Under the dual-end ADETCS, for multi-objective constrained ICPS subjected to dual DoS attacks and actuator failures, if the system is to have the ability to resist the dual potential threats of attack intrusion and actuator failures and have α-stability, the design goal of the α-stable observer is:

[0111] When the ICPS system is free of disturbance, that is, At this time, the augmented error system of the observer has a safety stability margin; wherein, represents the disturbance of the system under multi-objective constraints; represents the sensor measurement noise under multi-objective constraints; τ(t) represents the system delay under multi-objective constraints;

[0112] When the ICPS system has disturbances, that is, At this time, the augmented error system of the observer meets the following performance indicators:

[0113]

[0114] Among them, γ1 is the disturbance suppression rate, ||·||2 is the norm of L2[0,∞); L2[0,∞) represents the Euclidean norm; It represents the current state of the system with α stability; α1 represents the stability margin in the observer acquisition process.

[0115] 2) Design method of state and fault observer with α-stability

[0116] Theorem 1: Under double-ended ADETCS, given scalars α1,h1,γ1,n i (i=1,2,3,4), if there exist suitable matrices M, Y and positive symmetric matrix P, which satisfy the ratio of inequality formula (9) and formula (10), then there exist state estimation observer gain matrix L and fault estimation observer gain matrix F, which can ensure that the augmented error system (i.e., formula (7)) is asymptotically stable when the system is under double DoS attack, and has α-stable performance, and satisfies the performance index (i.e., formula (8)) when there is disturbance. L and F can be obtained by Find; where P -1 is the inverse matrix of a positive definite matrix; Y is a matrix of suitable dimension; is the augmented matrix of state estimation gain and fault estimation gain. It is obtained in the functional solution and the operation of MATLAB linear matrix inequality.

[0117]

[0118]

[0119] in:

[0120]

[0121] Η 13 =3M-8M T , Η 14 =3M+12M T , Η 15 =3M

[0122]

[0123] Η 23 =-M-8M T

[0124] Η 24 =-M+12M T , Η 25 =-M, Η 26 =-h1n2YE v ,

[0125] Η 33 =-8M-8M T , Η 34 =-8M+12M T , Η 35 =-8M

[0126] Η 44 =12M+12M T , Η 45 =12M, Η 55 =-n3P,

[0127] Η 88 =-n1P

[0128]

[0129] Η′ 22 =(n3-n2)PMM T

[0130]

[0131] Where h1 is the maximum allowable delay obtained by the computational observer under given performance conditions; M is an appropriate-dimensional matrix in the observer derivation process; n i(i=1,2,3,4) refers to the parameters set in the derivation process. Since nonlinear terms will appear in the derivation process, the parameter setting method is used to convert them into linear terms for solution. In the Matlab linear matrix inequality, the system is kept stable by adjusting the parameters.

[0132] (3) Setting up the controller based on multi-objective constraints in the control unit of the double-ended ADETCS

[0133] Based on the observer designed above, for ICPS suffering from double DoS attacks and actuator failures, this section will give a method for attack active-passive hybrid intrusion tolerance, fault active fault tolerance and communication collaborative design under a two-terminal ADETCS.

[0134] First, construct a comprehensive safety controller:

[0135]

[0136] Among them, t∈(j k ,j k+1 ), the controller gain matrix Fault Mitigation Matrix Satisfy (I-BB + )E f = 0. Based on system measurement output data Estimated state feedback term Under the dual-end ADETCS, the active and passive hybrid intrusion tolerance control of DoS attacks with different energies can be realized based on the original measurement data or reconstructed compensation data under the dual-side DoS attack; Unknown continuous faults can be actively adjusted.

[0137] In this section, we also introduce α-stability (α2 is used in this paper to represent the stability margin in the process of obtaining the integrated safety controller) and H ∞ The performance index constitutes a multi-objective constraint, combining formula (6) and formula (7) and state transformation The following closed-loop ICPS model can be obtained:

[0138]

[0139] Where:

[0140]

[0141] is the internal sampling characteristic term of the actuator fault, which is processed by the cross-term inequality technique.

[0142] 1) Design goal of controller with α-stability under two-terminal ADETCS

[0143] Under the dual-end ADETCS, for ICPS coupled with double-sided DoS attacks and actuator failures, if the system is to have the ability to resist both potential threats of attack intrusion and actuator failures. For DoS attacks on the execution side and large-energy DoS attacks on the sensor side, the PCA-LSTM prediction compensation model is used to effectively compensate for the data packet loss caused by them. Based on the maximum safety adaptive event trigger interval, the gain matrix K and event trigger weight matrix Φ1, Φ2 of the comprehensive safety controller with passive tolerance of small-energy DoS attacks and active adjustment of actuator failures on the sensor side are collaboratively obtained, so that the closed-loop fault ICPS meets the following conditions:

[0144] When the ICPS system is free of disturbance, that is, The closed-loop fault ICPS model (i.e., formula (12)) is asymptotically stable, i.e., has α-stability. Consider α System disturbances during stability; Consider α State error at stability; Consider α Fault error during stability;

[0145] When the ICPS system has disturbances, that is, The closed-loop fault ICPS meets the following performance indicators:

[0146]

[0147] Among them, γ′1 is the disturbance suppression rate, ||·||2 is the norm of L2[0,∞)

[0148] 2) Design method of integrated safety controller with α-stability

[0149] Theorem 2: Given a scalar If there is a suitable matrix K1, and positive definite symmetric matrix At the same time, if the ratio of formula (14) to formula (16) is satisfied, there exists a controller gain matrix K and adaptive event trigger weight matrices Φ1 and Φ2, which can ensure the asymptotic stability of the closed-loop fault system (i.e., formula (12)) for different energy levels of the two-sided DoS attack and satisfy H ∞ Performance index (i.e., Formula 13), It can be obtained in conjunction with Φ1 and Φ2.

[0150]

[0151] in:

[0152]

[0153]

[0154]

[0155] in, To solve the maximum allowable delay of the controller.

[0156] 3. Simulation experiment on DoS attack state and fault estimation and design of integrated safety controller based on multi-objective constraints under dual-end ADETCS.

[0157] Figure 3 The observer state estimation error curve under multi-objective constraints is given when α1≠0.

[0158] In different time intervals, the specific situations of DoS attacks, disturbances and actuator failures are different. The following will analyze the observer's estimation effect on faults and states in detail through Table 1.

[0159] Table 1 Analysis of simulation results of observer state and fault estimation error

[0160]

[0161]

[0162] In summary, the α-stability, H ∞ The multi-objective constrained observer of the performance indicators provides a more accurate estimate of the system state and actuator failures.

[0163] Figure 4 It is the system output response curve when α2≠0 under multi-objective constraints.

[0164] In different time intervals, the specific situations of DoS attacks, disturbances and actuator failures are different. The following will analyze in detail the control effect of the comprehensive safety controller on the system in different time periods through Table 2.

[0165] Table 2 Description of simulation results of controller design

[0166]

[0167] In summary, this paper introduces the α-stability, H ∞ The multi-objective constraints of performance indicators and the integrated safety controller designed based on the PCA-LSTM data-driven method improve the disturbance suppression capability, stability performance and dynamic performance level of ICPS under the threat of double DoS attacks and actuator failures.

[0168] Furthermore, the data transmission under the dual-end ADETCS

[0169] This paper introduces the dual-terminal ADETCS, Figure 5 represents the normalized value Δ of the transmission data error in the designed adaptive trigger parameter, and its size depends on the norm of the deviation between the system state at the current moment and the system state at the previous moment; specifically, Figure 5 Figures (a) and (b) are the output errors of the trigger parameters of the first adaptive discrete event trigger and the second adaptive discrete event trigger, respectively. Because the trigger parameters are assumed to be sinusoidal functions, Make ρ in the monotonic interval. From the simulation, it can be seen that except for t=250s, the actuator failure occurs, which causes the transmission error to increase instantly, but it quickly recovers to the error range. It can be seen that the adaptive discrete event trigger mechanism designed in this paper meets the performance requirements.

[0170] Figure 6 Figures (a) and (b) respectively show the adaptive trigger parameters of the dual-end ADETCS designed by the present invention. and It is obvious that the trigger parameters in ADETCS change dynamically.

[0171] from Figure 6 It can be seen that when the system is attacked by DoS, the trigger parameters of the dual-end event trigger will gradually decrease, and the threshold of the second adaptive discrete event trigger is smaller than that of the first adaptive discrete event trigger. The reason is that the DoS attack on the execution end does not distinguish between large and small energies, and the data screened by trigger B is the output value of the controller in the intelligent control end element, and the value fluctuates greatly, so a smaller trigger parameter is required to increase the data transmission, but at the same time, unnecessary load and noise should be avoided. When the actuator of the system fails at t=250s, the event trigger parameter automatically decreases instantly with the system fluctuation, which increases the data transmission volume; and when the system runs smoothly under the action of the integrated safety controller designed in this paper, the trigger threshold changes adaptively accordingly.

[0172] In order to reflect the performance advantages of the dual-end ADETCS designed in this paper, the data transmission rates of DETCS in the first reference, ADETCS in the second reference, dual-end ADETCS in the third reference, and dual-end ADETCS in this paper are given in Table 3. Too high a transmission rate will cause network congestion and data packet loss, which will reduce the efficiency of data transmission.

[0173] in:

[0174] First reference: "ICPS integrated security control and communication collaborative design based on DoS attack energy classification";

[0175] Second reference: "ICPS dual security control based on mechanism analysis and data-driven integration under adaptive event-triggered communication mechanism"

[0176] Third reference: "et al. Adaptive event-triggered resilient control of industrial cyber physical systems under asynchronous data injection attack"

[0177] Table 3 Data transmission rate of different trigger schemes

[0178]

[0179] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-target integrated safety control system for ICPS under dual-end ADETCS under dual DoS attack, characterized by: The system comprises a sensing unit located at the output end of the controlled object, an execution unit located at the input end of the controlled object, and a control unit located between the sensing unit and the execution unit, wherein the sensing unit, the control unit and the execution unit form a closed-loop control; The sensing unit is provided with a first adaptive discrete event generator; the first adaptive discrete event generator is used to filter the acquired ICPS system sampling data, and send the ICPS system sampling data that meets the trigger mechanism of the first adaptive discrete event generator to the control unit; A second adaptive discrete event generator is provided on the connection channel between the control unit and the execution unit; the second adaptive discrete event generator is used to filter the control unit output data and send the control unit output data that meets the trigger mechanism of the second adaptive discrete event generator to the execution unit; The first adaptive discrete event generator and the second adaptive discrete event generator constitute a double-terminal ADETCS.

2. The multi-target integrated safety control system for ICPS under dual-end ADETCS according to claim 1 is characterized in that: The first adaptive discrete event generator triggering mechanism is expressed as: Among them, Φ y is the event trigger weight matrix to be designed in the sensor unit; is the event triggering error of the first adaptive discrete event generator, and is the ICPS system sampling data input to the first adaptive discrete event generator at time k; is the ICPS system sampling data output by the first adaptive discrete event generator at time k; is an adaptive trigger parameter of the first adaptive discrete event generator; The second adaptive discrete event generator triggering mechanism is expressed as: Among them, Φ u is the event trigger weight matrix to be designed in the execution unit; is the event trigger error of the second adaptive discrete event generator, and is the control unit output data input to the second adaptive discrete event generator at time k; The control unit output data output by the second adaptive discrete event generator at time k; is the adaptive trigger parameter of the second adaptive discrete event generator.

3. The multi-target integrated safety control system for ICPS under double-end ADETCS according to claim 2 is characterized in that: The adaptive trigger parameter of the first adaptive discrete event generator or the second adaptive discrete event generator is expressed as: σ(t k h)=max{σ min ,ρσ(t k-1 h)},σ(t k h)∈[σ min ,s max ] Among them, σ(t k h) represents the adaptive trigger parameter of the first adaptive discrete event generator or the second adaptive discrete event generator at time k; σ(t k-1 h) represents the adaptive triggering parameter of the first adaptive discrete event generator or the second adaptive discrete event generator at time k-1; σ min Represents the lower limit of the adaptive trigger parameter; σ max Represents the upper limit of the adaptive trigger parameter; and 0≤σ min <σ max <1; real number β>0, Δ is the normalized value of the data transmission error at different times, set Ensure that the threshold coefficient ρ is within the monotonic interval; y(t k h) is the data output by the first adaptive discrete event generator or the second adaptive discrete event generator at time k; y(t k-1 h) is the data output by the first adaptive discrete event generator or the second adaptive discrete event generator at time k-1.

4. The multi-target integrated safety control system for ICPS under double DoS attack under double-end ADETCS according to claim 3 is characterized in that: The judgment mechanism of the dual-end ADETCS includes: When ||y(t k h)|| and||y(t k-1 h)|| is greater than or equal to zero, 0≤ρ<1, then σ(t k h)<σ(t k-1 h), the data output by the ICPS system increases, when σ(t k h)=σ min When σ(t k h) no longer decreases; When ||y(t k h)|| and||y(t k-1 h)|| is less than 0, then σ(t k h)≥σ(t k-1 h), the data output by the ICPS system decreases, when σ(t k h)=σ max When σ(t k h) No longer increase.

5. The multi-target integrated safety control system for ICPS under dual-end ADETCS according to claim 1 is characterized in that: The control unit includes a first attack detector; The first attack detector is used to determine the extent to which the sensor unit is subjected to a DoS attack according to a data sampling interval of the ICPS system; If the data sampling interval is less than the maximum delay allowed by the ICPS system, it is determined that the sensor unit is subjected to a small energy DoS attack or is not subjected to a DoS attack; When the sensor unit is attacked by a small-energy DoS attack, a model mechanism strategy is used to compensate for the data lost due to the DoS attack; If the data sampling interval is greater than or equal to the maximum delay allowed by the ICPS system, it is determined that the sensor unit is subjected to a large-energy DoS attack, and principal component analysis-long short-term memory neural network is used to compensate for the data lost due to the DoS attack, that is, data-driven.

6. The multi-target integrated safety control system for ICPS under dual-end ADETCS according to claim 5 is characterized in that: The execution unit includes a second attack detector; The second attack detector is used to determine whether the execution unit is subjected to a DoS attack according to the data receiving interval of the execution unit and the data sampling interval of the ICPS system; If the data receiving interval of the execution unit is greater than or equal to the data sampling interval, it is determined that the execution unit is subjected to a DoS attack; And use principal component analysis-long short-term memory neural network to compensate for the data lost due to DoS attacks; If the data receiving interval of the execution unit is smaller than the data sampling interval, it is determined that the execution unit is not subjected to a DoS attack.

7. The multi-target integrated safety control system for ICPS under dual DoS attack under dual-end ADETCS according to claim 1 is characterized in that: By introducing the safety stability margin and the disturbance suppression capability coefficient, a Lyapunov-krascovskii functional is constructed; and based on the Lyapunov-krascovskii functional, an observer and a controller in the control unit under the double-terminal ADETCS are set.

8. The multi-target integrated safety control system for ICPS under dual-end ADETCS according to claim 7 is characterized in that: The Lyapunov-krascovskii functional is expressed as: V(t)=V1(t)+V2(t)+V3(t)+V4(t)+V5(t) in: Among them, V(t) represents the Lyapunov-krascovskii functional; V1(t) represents the functional at the current moment; V2(t) represents the time-delay functional at the current moment; V3(t) represents the single-integral functional at the current moment; V4(t) represents the single-integral functional of the derivative at the current moment; V5(t) represents the double-integral functional of the derivative at the current moment; represents the current state of the system with α stability, for The derivative of ; τ(t) represents the system delay under multi-objective constraints; P, S, Q, R and Z are all positive definite symmetric matrices.

9. The multi-target integrated safety control system for ICPS under dual DoS attack under dual-end ADETCS according to claim 7 is characterized in that: The setting objectives of the observer include: When the ICPS system is free of disturbance, that is, At this time, the augmented error system of the observer has a safety stability margin; wherein, represents the disturbance of the system under multi-objective constraints; represents the sensor measurement noise under multi-objective constraints; τ(t) represents the system delay under multi-objective constraints; When the ICPS system has disturbances, that is, At this time, the augmented error system of the observer meets the following performance indicators: Among them, γ1 is the disturbance suppression rate, ||·||2 is the norm of L2[0,∞); It represents the current state of the system with α stability; α1 represents the stability margin in the observer acquisition process.

10. The multi-target integrated safety control system for ICPS under dual DoS attack under dual-end ADETCS according to claim 7, characterized in that: The setting objectives of the controller include: When the ICPS system is free of disturbance, that is, At this time, the closed-loop fault ICPS model of the controller has a safety stability margin; wherein, represents the system disturbance when considering α stability; represents the state error when considering α stability; represents the fault error when considering α stability; When the ICPS system has disturbances, that is, At this time, the closed-loop fault ICPS model of the controller meets the following performance indicators: Among them, ζ(t) is the system state quantity after the stability margin α2 is introduced in the controller acquisition process; is the disturbance coefficient under the system state quantity ζ(t); γ′1 is the disturbance suppression rate, ||·||2 is the norm of L2[0,∞); α2 represents the stability margin in the controller acquisition process.

Images