Secret communication method based on multi-modal inertial neural network under spoofing attack
By designing an anti-attack controller in a multimodal inertial neural network, the problem of system synchronization performance degradation under spoofed attacks is solved, exponential synchronization and confidential communication security under attack conditions are achieved, and control costs are reduced.
Patent Information
- Application Number
- CN202410986328.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-23
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-07-23
AI Technical Summary
Under spoofing attacks, multimodal inertial neural network systems face problems of random perturbation and mixed time lag, resulting in a degradation of synchronization performance and affecting the security of confidential communications.
A confidential communication method based on multimodal inertial neural network was designed. By establishing a dynamic equation with random perturbation and mixed time delay, variable substitution reduction processing and synchronization error system were used to design an anti-attack controller to make The system is exponentially synchronized with the target system under attack conditions, thereby realizing the encryption and decryption of plaintext signals.
Effectively resist spoofing attacks, ensure that the multimodal inertial neural network system can be exponentially synchronized under attack conditions, realize the security and effectiveness of confidential communications, reduce control costs, and alleviate the problem of communication resource limitations.
Smart Images

Figure CN119011210A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of new generation information technology, and in particular to a secure communication method based on a multimodal inertial neural network under deception attacks. Background Art
[0002] With the rapid development of communication technology, more and more information interactions rely on network transmission, such as confidential documents, digital images, etc. Therefore, ensuring secure communication has become a very critical topic. Studies have found that the synchronization performance of inertial neural networks significantly improves the security of data communication. The synchronization of inertial neural networks refers to the fact that multiple inertial neural networks reach a common trajectory, such as a common equilibrium, limit cycle or chaotic trajectory. On the other hand, the structures and parameters of many practical systems often face unpredictable changes. In this context, Markov jump systems have demonstrated excellent ability to describe these systems. Markov jump systems are multimodal hybrid systems, and the jumps between different modes follow the principle of Markov chains. This multimodal system can effectively simulate various emergency situations that may be encountered in real-life systems, such as sudden changes in system parameters, failure or maintenance of components. Therefore, the synchronization performance of multimodal inertial neural network systems following Markov chains has broad application prospects in the field of communication security, especially in confidential communication.
[0003] In confidential communications, communication networks are often faced with various threats from adversaries. In particular, due to the technical limitations or constraints of physical components, data transmission between sensors, controllers, and other devices is vulnerable to malicious attacks. For example, spoofing attacks, as a common network attack method, maliciously tamper with the transmitted data, thereby destroying the integrity of the control signal. Therefore, it is of great significance to study the synchronization performance of multimodal inertial neural networks under spoofing attacks in confidential communications. Summary of the invention
[0004] The purpose of the present invention is to provide a secure communication method based on a multimodal inertial neural network under a deception attack, so as to solve the synchronization problem of a multimodal inertial neural network with random perturbations and mixed time delays under a deception attack, and to apply it to the field of information security, thereby achieving the purpose of secure communication.
[0005] To achieve the above object, the present invention adopts the following technical solution: a secure communication method based on a multimodal inertial neural network under a deception attack, comprising the following steps:
[0006] Step S1: Establish a multimodal inertial neural network system with random perturbations and mixed time delays, and its dynamic equation is:
[0007]
[0008] in, N represents the number of nodes of the neural network system; represents the state variable of the pth node at time t; f(x p (t))=[f 1 (x p1 (t)), ..., f n (x pn (t))] T 、f(x p (t-δ(t))=[f 1 (x p1 (t-δ(t))),...,f n (x pn (t-δ(t)))] T and f(x p (s))=[f 1 (x p1 (s)), ..., f n (x pn (s))] T represents the activation function of the neuron and satisfies |f p (y)-f p (x)|≤l p |yx|, where l p is a positive constant, and y≠x is a known parameter; represents the inertia term; δ(t) and τ(t) represent discrete time lag and distributed time lag respectively, and satisfy 0<δ(t)≤δ and 0<τ(t)≤τ, Represents the external input of the system; and represents a positive definite diagonal matrix; represents n-dimensional Euclidean space, represents an n×n real matrix; and represents the connection weight matrix; φp(s) and ψp(s) are continuous bounded functions; r(t) is a right-continuous Markov process with a value of And there are the following modal transition probabilities:
[0009]
[0010] Among them, χ>0 and satisfies π ij ≥0(i≠j) represents the transition probability from the i-th mode to the j-th mode, c represents the coupling strength, represents the internal coupling positive definite matrix, G = (G pq(r(t))) N×N represents the coupling weight configuration matrix; if the qth node is connected to the pth node (p≠q), then G pq (r(t))>0, otherwise G pq (r(t)) = 0, and the diagonal elements of G are defined as represents the random interference intensity of the system and satisfies |ρ(t, x p (t), r(t))| 2 ≤Ξ p (i)|x p (t)| 2 , where p (i) is a positive constant; W(t) is defined in the complete probability space Brownian motion on , where Ω is the sample space, is a subset of the sample space, is the probability; u p (t) is an anti-attack controller;
[0011] The multimodal inertial neural network with random disturbance and mixed time delay is subjected to variable substitution and order reduction, and the substitution variables are set as Then rewrite the system into the following form:
[0012]
[0013] Where, C(r(t))=C(r(t))+I n -D(r(t)), I n represents the n-order identity matrix; φ p (s), ψ p (s) and Δ p (s) is a continuous bounded function;
[0014] Step S2: According to the multimodal inertial neural network system, a target system is established as:
[0015]
[0016] in, represents the state variable of the target system; φ*(ω) and ψ*(ω) are continuous bounded functions; other performance indicators of the target system are the same as the performance indicators of the multimodal inertial neural network with random disturbance and mixed time delay described in step S1;
[0017] The target system is subjected to variable substitution and order reduction processing, and the substitution variables are set to Then rewrite the system into the following form:
[0018]
[0019] Where, C(r(t))=C(r(t))+I n -D(r(t)), φ * (ω), ψ * (ω) and Δ * (ω) is a continuous bounded function;
[0020] Step S3: According to the multimodal inertial neural network system and the target system, a synchronization error is set to construct a synchronization error system, which specifically includes the following steps:
[0021] Step S3-1: Setting the synchronization error between the multimodal inertial neural network system and the target system to: p (t) = x p (t)-S(t);
[0022] Step S3-2: According to the synchronization error, a synchronization error system is constructed as follows:
[0023]
[0024] The error system is subjected to variable substitution and order reduction, and the substitution variables are set to Then rewrite the system into the following form:
[0025]
[0026] Among them, g(e p (t)) = f(x p (t))-f(S(t)), g(e p (t-δ(t)))=f(x p (t-δ(t)))-f(S(t-δ(t))), g(e p (s)) = f(x p (s))-f(S(s));
[0027] Step S4: Taking full account of the deception attacks in the communication network, based on the synchronization error system constructed in step S3, design an anti-attack controller u under deception attacks p (t), so that the multimodal inertial neural network system is in the anti-attack controller u p Under the action of (t), the index is synchronized with the target system, thereby successfully completing the encryption and decryption of the plaintext signal and further realizing confidential communication.
[0028] Step S5: Build a multimodal inertial neural network model, and verify its exponential synchronization effect under deception attacks and its effectiveness in confidential communication through numerical simulation.
[0029] Furthermore, step S4 specifically includes the following steps:
[0030] Step S4-1: Design a controller u that is resistant to deception attacks p (t) is:
[0031]
[0032] Among them, θ 1 ,θ 2 and θ 3 is a normal number; [t 2k , t 2k+1 ) and [t 2k+1 , t 2k+2 ) represent the control time interval and rest time interval of the anti-attack controller respectively; α(t) is the deception attack signal and satisfies the Bernoulli distribution, which is defined as: The probability is is the mathematical expectation of the random variable a; h(·) is a nonlinear function and satisfies |h(·)|≤h 1 , where h 1 is a positive constant; R p (t) is the anti-attack item, specifically:
[0033] R p (t) = -η p (t)q(z p (t))-msign(q(z p (t)))
[0034] in, is the quantizer, m is a positive integer, represents the set of real numbers, is the initial quantization parameter, To quantify the density, the specific quantization function is defined as follows:
[0035]
[0036] in, q(m * ) is the quantized value, is the exact parameter of the quantizer; η p (t) is the adaptive exponential update rate for attack and satisfies the equation:
[0037]
[0038] Among them, β p , α and ε p is a positive constant and satisfies the inequality m(1+αθ 1 )≥αθ 2 h 1 ; The relevant parameters in the anti-attack controller satisfy the following inequalities:
[0039]
[0040] Among them, η p and is a normal number, γ p (i)>0,
[0041] Step S4-2: applying the anti-attack controller to the multimodal inertial neural network system, so that the multimodal inertial neural network system can also be exponentially synchronized with the target system under a deception attack;
[0042] Step S4-3: After the multimodal inertial neural network system is exponentially synchronized with the target system, the transmitting end obtains the chaotic signal y generated by the multimodal inertial neural network system p (t) as an encryption signal, the receiving end obtains the chaotic signal v(t) generated by the target system as a key signal;
[0043] Step S4-4: The sending end encrypts the signal y p (t) and the plaintext signal Z p (t) Perform encryption operation to obtain the ciphertext signal H p (t) = y p (t)+Z p (t);
[0044] Step S4-5: The sending end sends the ciphertext signal H p (t) sending the ciphertext signal H to a receiving end through a transmission channel, and the receiving end receives the ciphertext signal H through the transmission channel p (t);
[0045] Step S4-6: The receiving end receives the ciphertext signal H p (t) and the key signal v p (t) Perform decryption operation to obtain the decrypted plaintext signal Z' p (t) = H p (t)-v p (t).
[0046] Compared with the prior art, the present invention has the following beneficial effects:
[0047] 1. The present invention takes into account factors such as communication delay and random interference faced by the system, and specifically introduces random disturbance and mixed time delay to form a more general multimodal inertial neural network, so that the present invention has a wider application prospect.
[0048] 2. In order to resist the deception attack of modifying the transmission data with a certain probability, the present invention designs an adaptive anti-attack controller for the multimodal inertial neural network, which has the ability to automatically adjust its own parameters according to external attacks.
[0049] 3. The present invention takes into account that signal transmission of real-time systems is often restricted by communication capacity and bandwidth limitations, and performs quantization processing on the anti-attack controller. This processing method helps to reduce control costs and effectively alleviate the problems caused by communication resource limitations. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.
[0051] In the attached picture:
[0052] Figure 1 It is a flowchart of a first specific embodiment of the secure communication method based on a multimodal inertial neural network under a deception attack of the present invention;
[0053] Figure 2 is the system state y without the anti-attack controller in the numerical simulation of the present invention p The state trajectory of v(t) and v(t), where (a) is the system state y without the anti-attack controller p1 (t) and v 1 (t) is the state trajectory of the system; (b) is the system state y without the anti-attack controller p2 (t) and v 2 (t) state trajectory;
[0054] Figure 3 is the system state y under the action of the anti-attack controller in the numerical simulation of the present invention p (t) and v(t), where (a) is the system state y under the action of the anti-attack controller p1 (t) and v 1 (t) is the state trajectory of the system; (b) is the state y of the system under the action of the anti-attack controller p2 (t) and v 2 (t) state trajectory;
[0055] Figure 4 is the state trajectory of the synchronization error system without the action of the controller in the numerical simulation of the present invention, where (a) is the state trajectory of the synchronization error system without the action of the anti-attack controller p1 (t) is the state trajectory of the synchronous error system z without the anti-attack controller; (b) is the state trajectory of the synchronous error system z without the anti-attack controller p2 (t) state trajectory;
[0056] Figure 5 is the state trajectory of the synchronization error system under the action of the anti-attack controller in the numerical simulation of the present invention, where (a) is the state trajectory of the synchronization error system z under the action of the anti-attack controller p1 (t) is the state trajectory of the synchronous error system z under the action of the anti-attack controller; (b) is the state trajectory of the synchronous error system z under the action of the anti-attack controller p2 (t) state trajectory;
[0057] Figure 6 is a curve diagram of the Markov process r(t) in the numerical simulation of the present invention;
[0058] Figure 7 is a curve diagram of the deceptive attack signal α(t) in the numerical simulation of the present invention;
[0059] Figure 8 A schematic diagram of signal transmission during a confidential communication process provided by an embodiment of the present invention;
[0060] Fig. 9 The numerical simulation result diagram of the secure communication method provided by the embodiment of the present invention, wherein (a) is the encrypted signal y p (t) is the state trajectory; (b) is the plaintext signal Z p (t) is the state trajectory; (c) is the ciphertext signal H p (t) is the state trajectory of the key signal v(t); (d) is the state trajectory of the key signal v(t); (e) is the decrypted signal Z′ p (t) is the state trajectory; (f) is the plaintext signal Z p (t) and the decrypted signal Z′ p (t) State trajectory of the error. DETAILED DESCRIPTION
[0061] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.
[0062] Embodiment 1
[0063] like Figure 1 As shown, this embodiment provides a secure communication method based on a multimodal inertial neural network under a deception attack. The secure communication method includes the following steps:
[0064] Step S1: Establish a multimodal inertial neural network system with random perturbations and mixed time delays, and its dynamic equation is:
[0065]
[0066] in, N represents the number of nodes of the neural network system; represents the state variable of the pth node at time t; f(x p (t))=[f 1 (x p1 (t)), ..., f n (x pn (t))] T 、f(x p (t-δ(t))=[f 1 (x p1 (t-δ(t))),...,f n (x pn (t-δ(t)))] T and f(x p (s))=[f 1 (x p1 (s)), ..., f n (x pn (s))] T represents the activation function of the neuron and satisfies |f p (y)-f p (x)|≤l p |yx|, where l p is a positive constant, and y≠x is a known parameter; represents the inertia term; δ(t) and τ(t) represent discrete time lag and distributed time lag respectively, and satisfy 0<δ(t)≤δ and 0<τ(t)≤τ, Represents the external input of the system; and represents a positive definite diagonal matrix; represents n-dimensional Euclidean space, represents an n×n real matrix; and represents the connection weight matrix; φ p (s) and ψ p (s) is a continuous bounded function; r(t) is a right-continuous Markov process with value And there are the following modal transition probabilities:
[0067]
[0068] Among them, χ>0 and satisfies π ij ≥0(i≠j) represents the transition probability from the i-th mode to the j-th mode, c represents the coupling strength, represents the internal coupling positive definite matrix, G = (G pq (r(t))) N×N represents the coupling weight configuration matrix; if the qth node is connected to the pth node (p≠q), then G pq (r(t))>0, otherwise G pq (r(t)) = 0, and the diagonal elements of G are defined as represents the random interference intensity of the system and satisfies |ρ(t, x p (t), r(t))| 2 ≤Ξ p (i)|x p (t)| 2 , where p (i) is a positive constant; W(t) is defined in the complete probability space Brownian motion on , where Ω is the sample space, is a subset of the sample space, is the probability; u p (t) is an anti-attack controller;
[0069] The multimodal inertial neural network with random disturbance and mixed time delay is subjected to variable substitution and order reduction, and the substitution variables are set as Then rewrite the system into the following form:
[0070]
[0071] Where, C(r(t))=C(r(t))+I n -D(r(t)), I n represents the n-order identity matrix; φ p (s), ψ p (s) and Δ p (s) is a continuous bounded function;
[0072] Step S2: According to the multimodal inertial neural network system, a target system is established as:
[0073]
[0074] in, represents the state variable of the target system; φ*(ω) and ψ*(ω) are continuous bounded functions; other performance indicators of the target system are the same as the performance indicators of the multimodal inertial neural network with random disturbance and mixed time delay described in step S1;
[0075] The target system is subjected to variable substitution and order reduction processing, and the substitution variables are set to Then rewrite the system into the following form:
[0076]
[0077] Where, C(r(t))=C(r(t))+I n -D(r(t)), φ * (ω), ψ * (ω) and Δ * (ω) is a continuous bounded function;
[0078] Step S3: According to the multimodal inertial neural network system and the target system, a synchronization error is set to construct a synchronization error system, which specifically includes the following steps:
[0079] Step S3-1: Setting the synchronization error between the multimodal inertial neural network system and the target system to: p (t) = x p (t)-S(t);
[0080] Step S3-2: According to the synchronization error, a synchronization error system is constructed as follows:
[0081]
[0082] The error system is subjected to variable substitution and order reduction, and the substitution variables are set to Then rewrite the system into the following form:
[0083]
[0084] Among them, g(e p (t)) = f(x p (t))-f(S(t)), g(e p (t-δ(t)))=f(x p (t-δ(t)))-f(S(t-δ(t))), g(e p (s)) = f(x p (s))-f(S(s));
[0085] Step S4: Taking full account of the deception attacks in the communication network, based on the synchronization error system constructed in step S3, design an anti-attack controller u under deception attacks p (t), so that the multimodal inertial neural network system is in the anti-attack controller u p Under the action of (t), the index is synchronized with the target system, thereby successfully completing the encryption and decryption of the plaintext signal and further realizing confidential communication.
[0086] Step S5: Build a multimodal inertial neural network model, and verify its exponential synchronization effect under deception attacks and its effectiveness in confidential communication through numerical simulation.
[0087] In this embodiment, step S4 specifically includes the following steps:
[0088] Step S4-1: Design a controller u that is resistant to deception attacks p (t) is:
[0089]
[0090] Among them, θ 1 ,θ 2 and θ 3 is a positive constant; [t 2k , t 2k+1 ) and [t 2k+1 , t 2k+2 ) represent the control time interval and rest time interval of the anti-attack controller respectively; α(t) is the deception attack signal and satisfies the Bernoulli distribution, which is defined as: The probability is is the mathematical expectation of the random variable a; h(·) is a nonlinear function that satisfies | h (·)|≤ h 1, where h 1 is a positive constant; Rp(t) is the anti-attack term, specifically:
[0091] R p (t) = -η p (t)q(z p (t))-msign(q(z p (t)))
[0092] in, is the quantizer, m is a positive integer, represents the set of real numbers, is the initial quantization parameter, To quantify the density, the specific quantization function is defined as follows:
[0093]
[0094] in, q(m * ) is the quantized value, is the exact parameter of the quantizer; η p (t) is the adaptive exponential update rate for attack and satisfies the equation:
[0095]
[0096] Among them, β p , α and ε p is a positive constant and satisfies the inequality m(1+αθ 1 )≥αθ 2 h 1 ; The relevant parameters in the anti-attack controller satisfy the following inequalities:
[0097] -Θ p1 -ε p >0,η p (1-θ)(1+αθ 1 )-Θ p -ε p >0,
[0098] Among them, η p and is a normal number,
[0099] Step S4-2: applying the anti-attack controller to the multimodal inertial neural network system, so that the multimodal inertial neural network system can also be exponentially synchronized with the target system under a deception attack;
[0100] Step S4-3: After the multimodal inertial neural network system is exponentially synchronized with the target system, the transmitting end obtains the chaotic signal y generated by the multimodal inertial neural network system p (t) as an encryption signal, the receiving end obtains the chaotic signal v(t) generated by the target system as a key signal;
[0101] Step S4-4: The sending end encrypts the signal y p (t) and the plaintext signal Z p (t) Perform encryption operation to obtain the ciphertext signal H p (t) = y p (t)+Z p (t);
[0102] Step S4-5: The sending end sends the ciphertext signal H p(t) sending the ciphertext signal H to a receiving end through a transmission channel, and the receiving end receives the ciphertext signal H through the transmission channel p (t);
[0103] Step S4-6: The receiving end receives the ciphertext signal H p (t) and the key signal v p (t) Perform decryption operation to obtain the decrypted plaintext signal Z′ p (t) = H p (t)-v p (t).
[0104] It is worth noting that the present invention takes into account factors such as communication delays and random interference faced by the system, and specifically introduces random disturbances and mixed time delays to form a more general multimodal inertial neural network, thereby making the present invention have a wider application prospect; in order to resist deception attacks that modify transmission data with a certain probability, an adaptive anti-attack controller is designed for the multimodal inertial neural network, and the controller has the ability to automatically adjust its own parameters according to external attacks; considering that the signal transmission of real-time systems is often restricted by communication capacity and bandwidth limitations, the anti-attack controller is quantized, and this processing method helps to reduce control costs and effectively alleviate the problems caused by communication resource limitations.
[0105] Embodiment 2
[0106] This embodiment mainly includes two parts:
[0107] The first is to theoretically prove the effectiveness of the anti-attack controller in Example 1 in achieving synchronization between a multi-modal inertial neural network system with random disturbances and mixed time delays and a target system.
[0108] The second is to verify the exponential synchronization effect of the multimodal inertial neural network system with random perturbations and mixed time delays and the target system in Example 1 and its effectiveness in confidential communication through numerical simulation.
[0109] (Neither the theoretical proof nor the simulation experiment is used to limit the present invention. In other embodiments, the simulation experiment may not be performed, and other experimental schemes may be used to perform experiments to verify the performance of the neural network system.)
[0110] 1. Theoretical Proof
[0111] We first give the definition of synchronization and the lemma that will be used.
[0112] Definition 1: If the system error state satisfies t≥0, where W>0, This indicates that the multimodal inertial neural network system with random disturbances and mixed time delays is exponentially synchronized with the target system under the action of the anti-attack controller.
[0113] Lemma 1: Define a random function V p (t,e p The infinitesimal operator of (t), i) is:
[0114]
[0115] in,
[0116] Lemma 2: Assume that V(t, e(t), r(t)) is A non-negative continuous function on , and satisfies the following conditions:
[0117]
[0118] Among them, ε 1 , ε 2 , ν and μ are positive constants, If ε 1 +ε 2 >0,ε 1 >ν+μδ, Then there is
[0119]
[0120] Where t ≥ t 0 , ρ>0 is the equation The root,
[0121] Next, we will establish some coefficient-based criteria to ensure that the multimodal inertial neural network system with random perturbations and mixed time delays and the target system are exponentially synchronized;
[0122] The Lyapunov function is constructed as:
[0123]
[0124] When t∈[t 2k , t 2k+1 ), according to Lemma 1, we can calculate have to:
[0125]
[0126] Where trace(A) is the trace of matrix A.
[0127] According to 0<δ(t)≤δ, 0<τ(t)≤τ and |fp(y)-fp(x)|≤lp|yx|, we can get the following inequality:
[0128]
[0129]
[0130] Therefore, we further get:
[0131]
[0132] Among them, Expand to get:
[0133]
[0134] Among them, the quantization function satisfies the following inequality:
[0135]
[0136] Therefore, we can get:
[0137]
[0138] Similarly, when t∈[t 2k+1 , t 2k+2 ), we can get:
[0139]
[0140] in,
[0141] Therefore, when t∈[t 2k , t 2k+1 ), according to the above derivation conclusions, we can get:
[0142]
[0143] in,
[0144] According to Ito's formula and the above formula, we can get:
[0145]
[0146] When t∈[t 2k+1 , t 2k+2 ), we can get:
[0147]
[0148] in,
[0149] Therefore, according to Lemma 2, we can get:
[0150]
[0151] For any There exists a positive constant w p (i) Make the inequality hold: V p (t,e p (t),i)≥w p (i)|e p (t)| 2 ;
[0152] So, we can get:
[0153]
[0154] in,
[0155] Therefore, we can get:
[0156]
[0157] in, Therefore, from Definition 1, we know that the multimodal inertial neural network system with random perturbations and mixed time delays is exponentially synchronized with the target system under the action of the anti-attack controller.
[0158] 2. Numerical Simulation
[0159] In this embodiment, the following multimodal inertial neural network system with 4 nodes is considered:
[0160]
[0161] Where p = 1, 2, 3, 4, r(t) is a Markov chain with a value of x p (t) = (x p1 (t), x p2 (t)) T ,f(x p (t)) = (tanh(x p1 (t)), tanh(x p2 (t))) T ; ρ(t, x p (t), 1) = (0.15sin(x p1 (t)), 0.15sin(x p2 (t))) T ,ρ(t,x p (t),2)==(0.2x p1 (t), 0.2x p2 (t)) T ; I(1) = (1.2, 1.8)T , I(2) = (1.6, 2.2) T , where (1) is the value in mode 1, (2) is the value in mode 2; mixed time-varying lag δ(t) = 0.2|cost|, τ(t) = 0.15sin 2 t; l can be calculated p =1, δ=0.2, τ=0.15.
[0162] Substituting variables in the above system, we get:
[0163]
[0164] The corresponding target system description is:
[0165]
[0166] in, Coupling strength c = 1, G 1,1 (1) = G 1,1 (2) = -1, G 1,2 (1) = G 1,2 (2) = 1, G 1,3 (1) = G 1,3 (2) = 0, G 1,4 (1) = G 1,4 (2) = 0, G 2,1 (1) = G 2,1 (2) = 1, G 2,2 (1) = G 2,2 (2) = -2, G 2,3 (1) = G 2,3 (2) = 1, G 2,4 (1) = G 2,4 (2) = 0, G 3,1 (1) = G 3,1 (2) = 0, G 3,2 (1) = G 3,2 (2) = 1.5, G 3,3 (1) = G 3,3 (2) = -2, G 3,4 (1) = G 3,4 (2) = 0.1, G 4,1 (1) = G 4,1 (2) = 0.5, G 4,2 (1) = G 4,2 (2) = 0, G 4,3 (1) = G 4,3 (2) = 0, G 4,4 (1) = G 4,4 (2) = -1;
[0167] Select parameter γ p (1) = 1, γ p (2) = 1.1; we can calculate v = 0.0968, μ = 0.0231, Θ p1 =-1.118, θ p =1.217; the adaptive index update rate is:
[0168]
[0169] Among them, β p =0.2,ε p =0.6, quantized density θ 1 =0.05,θ 2 =0.5,θ 3 = 0.5, m = 0.5, nonlinear function h(·) = tanh(·), deception attack probability α = 0.5, and the control interval interval is set to Through calculation, the inequality satisfies: -Θ p1 -ε p >0, Therefore, ε 1 =0.6,ε 2 =1.25, satisfying ε 1 >ν+μδ, equation The root ρ=0.49, which satisfies
[0170] The specific simulation results of the multi-modal inertial neural network system with random disturbance and mixed time delay and the target system under the above-mentioned parameters are as follows: Figure 2 is the system state y without anti-attack controller p The state trajectory of v(t) and v(t), where (a) is the system state y without the anti-attack controller p1 (t) and v 1 (t) is the state trajectory of the system; (b) is the system state y without the anti-attack controller p2 (t) and v 2 (t) state trajectory; Figure 3 is the system state y under the action of the anti-attack controller p (t) and v(t), where (a) is the system state y under the action of the anti-attack controller p1 (t) and v 1 (t) is the state trajectory of the system; (b) is the state y of the system under the action of the anti-attack controller p2 (t) and v 2 (t) state trajectory; Figure 4is the state trajectory of the synchronization error system without the action of the controller, where (a) is the state trajectory of the synchronization error system without the action of the anti-attack controller z p1 (t) is the state trajectory of the synchronous error system z without the anti-attack controller; (b) is the state trajectory of the synchronous error system z without the anti-attack controller p2 (t) state trajectory; Figure 5 is the state trajectory of the synchronization error system under the action of the anti-attack controller, where (a) is the state trajectory of the synchronization error system z under the action of the anti-attack controller p1 (t) is the state trajectory of the synchronous error system z under the action of the anti-attack controller; (b) is the state trajectory of the synchronous error system z under the action of the anti-attack controller p2 (t) state trajectory; Figure 4 and Figure 5 It can be seen that without the anti-attack controller, the state trajectory of the synchronization error system continues to oscillate. After the anti-attack controller is added, the state trajectory of the synchronization error system gradually converges to 0. Figure 6 is the graph of the Markov process r(t); Figure 7 is a graph of the deceptive attack signal α(t); Figure 2-Figure 5 It is shown that the multimodal inertial neural network system with random disturbances and mixed time delays can be exponentially synchronized with the target system under the action of the anti-attack controller, and the exponential synchronization performance is verified; Figure 8 A schematic diagram of signal transmission during the confidential communication process of this embodiment; Fig. 9 The numerical simulation result of the secure communication method of this embodiment is shown in FIG., where (a) is the encrypted signal y p (t) is the state trajectory; (b) is the plaintext signal Z p (t) is the state trajectory; (c) is the ciphertext signal H p (t) is the state trajectory of the key signal v(t); (d) is the state trajectory of the key signal v(t); (e) is the decrypted signal Z′ p (t) is the state trajectory; (f) is the plaintext signal Z p (t) and the decrypted signal Z′ p (t) The state trajectory of the error; Fig. 9 It can be seen that the ciphertext signal effectively ensures the security of the plaintext signal; the error between the decrypted signal and the initial plaintext signal is 0, and the decryption function is successfully realized. In summary, when the multimodal inertial neural network system is synchronized with the target system, the function of confidential communication can be realized.
[0171] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A secure communication method based on a multimodal inertial neural network under deception attacks, characterized in that: The following steps are involved: Step S1: Establish a multimodal inertial neural network system with random perturbations and mixed time delays, and its dynamic equation is: in, N represents the number of nodes of the neural network system; represents the state variable of the pth node at time t; f(x p (t)) = [f1(x p1 (t)),…,f n (x pn (t))] T 、f(x p (t-δ(t))=[f1(x p1 (t-δ(t))),…,f n (x pn (t-δ(t)))] T and f(x p (s)) = [f1(x p1 (s)),…,f n (x pn (s))] T represents the activation function of the neuron and satisfies |f p (y)-f p (x)|≤l p |yx|, where l p is a positive constant, and y≠x is a known parameter; represents the inertia term; δ(t) and τ(t) represent discrete time lag and distributed time lag respectively, and satisfy 0<δ(t)≤δ and 0<τ(t)≤τ, Represents the external input of the system; represents a positive definite diagonal matrix; represents n-dimensional Euclidean space, represents an n×n real matrix; represents the connection weight matrix; φ p (s) and ψ p (s) is a continuous bounded function; r(t) is a right-continuous Markov process with value And there are the following modal transition probabilities: Among them, X>0 and satisfies π ij ≥0(i≠j) represents the transition probability from the i-th mode to the j-th mode, c represents the coupling strength, represents the internal coupling positive definite matrix, G = (G pq (r(t))) N×N represents the coupling weight configuration matrix; if the qth node is connected to the pth node (p≠q), then G pq (r(t))>0, otherwise G pq (r(t)) = 0, and the diagonal elements of G are defined as represents the random interference intensity of the system and satisfies |ρ(t,x p (t),r(t))| 2 ≤Ξ p (i)|x p (t)| 2 , where p (i) is a positive constant; W(t) is defined in the complete probability space Brownian motion on , where Ω is the sample space, is a subset of the sample space, is the probability; u p (t) is an anti-attack controller; The multimodal inertial neural network with random disturbance and mixed time delay is subjected to variable substitution and order reduction, and the substitution variables are set as Then rewrite the system into the following form: Where, C(r(t))=C(r(t))+I n -D(r(t)), I n represents the n-order identity matrix; φ p (s),ψ p (s) and Δ p (s) is a continuous bounded function; Step S2: According to the multimodal inertial neural network system, a target system is established as: in, represents the state variable of the target system; φ * (ω) and ψ * (ω) is a continuous bounded function; the other performance indicators in the target system are the same as the performance indicators in the multimodal inertial neural network with random disturbance and mixed time delay in step S1; The target system is subjected to variable substitution and order reduction processing, and the substitution variables are set to Then rewrite the system into the following form: Where, C(r(t))=C(r(t))+I n -D(r(t)), φ * (ω), ψ * (ω) and Δ * (ω) is a continuous bounded function; Step S3: According to the multimodal inertial neural network system and the target system, a synchronization error is set to construct a synchronization error system, which specifically includes the following steps: Step S3-1: Setting the synchronization error between the multimodal inertial neural network system and the target system to: p (t) = x p (t)-S(t); Step S3-2: According to the synchronization error, a synchronization error system is constructed as follows: The error system is subjected to variable substitution and order reduction, and the substitution variables are set to Then rewrite the system into the following form: where, g(e p (t)) = f(x p (t)) - f(S(t)), g(e p (t - δ(t))) = f(x p (t - δ(t))) - f(S(t - δ(t))), g(e p (s)) = f(x p (s)) - f(S(s)); Step S4: Taking full account of the deception attacks in the communication network, based on the synchronization error system constructed in step S3, design an anti-attack controller u under deception attacks p (t), so that the multimodal inertial neural network system is in the anti-attack controller u p (t) is synchronized with the target system, thereby successfully completing the encryption and decryption of the plaintext signal, further realizing confidential communication; Step S5: Build a multimodal inertial neural network model, and verify its exponential synchronization effect under deception attacks and its effectiveness in confidential communication through numerical simulation.
2. The secure communication method based on multimodal inertial neural network under deception attack according to claim 1, characterized in that: Step S4 specifically includes the following steps: Step S4-1: Design a controller u that is resistant to deception attacks p (t) is: Among them, θ1, θ2 and θ3 are positive constants; [t 2k ,t 2k+1 ) and [t 2k+1 ,t 2k+2 ) represent the control time interval and rest time interval of the anti-attack controller respectively; α(t) is the deception attack signal and satisfies the Bernoulli distribution, which is defined as: The probability is is the mathematical expectation of the random variable a; h(·) is a nonlinear function and satisfies |h(·)|≤h1, where h1 is a positive constant; R p (t) is the anti-attack item, specifically: R p (t)=-η p (t)q(z p (t))-msign(q(z p (t))) in, is the quantizer, m is a positive integer, represents the set of real numbers, is the initial quantization parameter, To quantify the density, the specific quantization function is defined as follows: in, q(m * ) is the quantized value, is the exact parameter of the quantizer; η p (t) is the adaptive exponential update rate for attack and satisfies the equation: Among them, β p , α and ε p is a positive constant and satisfies the inequality m(1+αθ1)≥αθ2h1; the relevant parameters in the anti-attack controller satisfy the following inequality: -I p1 -e p >0, Among them, η p and is a normal number, Step S4-2: applying the anti-attack controller to the multimodal inertial neural network system, so that the multimodal inertial neural network system can also be exponentially synchronized with the target system under a deception attack; Step S4-3: After the multimodal inertial neural network system is exponentially synchronized with the target system, the transmitting end obtains the chaotic signal y generated by the multimodal inertial neural network system p (t) as an encryption signal, the receiving end obtains the chaotic signal v(t) generated by the target system as a key signal; Step S4-4: The sending end encrypts the signal y p (t) and the plaintext signal Z p (t) Perform encryption operation to obtain the ciphertext signal H p (t) = y p (t)+Z p (t); Step S4-5: The sending end sends the ciphertext signal H p (t) sending the ciphertext signal H to a receiving end through a transmission channel, and the receiving end receives the ciphertext signal H through the transmission channel p (t); Step S4-6: The receiving end receives the ciphertext signal H p (t) and the key signal v p (t) Perform decryption operation to obtain the decrypted plaintext signal Z' p (t) = H p (t)-v p (t).
Citation Information
Patent Citations
Cyber-attack detection, localization, and neutralization for unmanned aerial vehicles
CN111727432A
Secret communication method based on inertia complex value memristor neural network
CN115169539A
Hysteretic synchronization control method, system and application of complex valued inertial neural network
CN116909147A
Quantitative synchronous control method and system of multi-modal inertial neural network
CN118092162A
Hypercomplex deep learning methods, architectures, and apparatus for multimodal small, medium, and large-scale data representation, analysis, and applications
US20190087726A1
