Secure communication method of BAM memristor neural network based on adaptive quantization control

Abstract

The present invention belongs to the field of new generation information technology, and specifically relates to a confidential communication method of a BAM memristor neural network based on adaptive quantization control, including: step S1: constructing a fuzzy inertial BAM memristor neural network drive system and a response system; step S2: setting the synchronization error of the drive system and the response system, and establishing a synchronization error system; step S3: designing an adaptive quantization controller, applying the adaptive quantization controller to the response system, so that the response system is synchronized with the drive system at a fixed time, thereby realizing a confidential communication method. The present invention introduces an inertial term into a fuzzy BAM memristor neural network, and designs an adaptive quantization controller, solves the problem of fixed time synchronization of a fuzzy inertial BAM memristor neural network, and applies it to the field of confidential communication. The proposed confidential communication method significantly improves the complexity and security of confidential communication.

Description

Technical Field

[0001] The present invention relates to the field of new generation information technology, and in particular to a secure communication method of a BAM memristor neural network based on adaptive quantization control. Background Art

[0002] Bidirectional associative memory neural network, referred to as BAM neural network. It consists of two layers of neurons and has the characteristics of hetero-associative and content-addressable memory. The two layers of neurons are completely interconnected, and there is no signal transmission between neurons in the same layer. In real life, BAM neural network has powerful information processing capabilities and has good applications in information associative memory, image processing, artificial intelligence and other fields.

[0003] It is well known that memristors have the characteristics of nanometer size, non-volatility and hysteresis loop, which make them work like neuron synapses. Based on the advantages of memristors, they can replace traditional resistors to simulate biological synapses, and the synaptic density in memristor neural networks is closer to that of biological neural networks. Naturally, researchers combined memristors with BAM neural networks to form a type of neural network, called BAM memristor neural networks.

[0004] Since fuzziness is inevitable in the real world, fuzzy theory is used as a suitable processing method to form a fuzzy neural network. The synchronization of fuzzy neural networks plays an important role in image processing and pattern recognition applications. Fuzzy neural networks and BAM memristor neural networks are combined to form a new type of fuzzy BAM memristor neural network.

[0005] The transmission of signals is usually affected by bandwidth and communication channels. Due to the limited transmission capacity, signal distortion occurs, which affects the control performance of the system. Therefore, in order to solve this problem, signal quantization is an effective method to improve communication efficiency. The use of a quantized controller can reduce control costs and save channel resources. The present invention proposes an adaptive quantized controller that enables a neural network to achieve synchronization within a fixed time. Summary of the invention

[0006] In view of this, the purpose of the present invention is to provide a secure communication method for a BAM memristor neural network based on adaptive quantization control, which can achieve fixed time synchronization of a fuzzy inertial BAM memristor neural network, and propose a secure communication method, thereby significantly improving the complexity and security of secure communication.

[0007] The present invention provides a secure communication method of a BAM memristor neural network based on adaptive quantization control, comprising the following steps:

[0008] Step S1: construct a fuzzy inertial BAM memristor neural network drive system and response system; Step S1 specifically includes the following steps:

[0009] Step S11: construct the fuzzy inertial BAM memristor neural network drive system as follows:

[0010]

[0011] Wherein, time t≥0; i=1,2,…,n; j=1,2,…,m; n≥2 and m≥2 are the number of neurons in the i-th layer and the j-th layer of the driving system respectively; x i (t) and y j (t) are the state variables of the neurons in the i-th and j-th layers of the driving system at time t, and They are state variables x i (t) and y j The second-order derivative of (t) is the inertia term of the drive system; and is a positive constant; τ(t) is a discrete time lag and satisfies 0≤τ(t)≤τ, where τ is a positive constant; and represents a feed-forward template; and represents the fuzzy feedforward minimum template; and represents the fuzzy feedforward maximum template; and Represents the elements of the minimal template for fuzzy feedback; and Represents the elements of the maximum template of fuzzy feedback; and denote the inputs of the j-th neural network and the i-th neural network respectively; ∧ and ∨ denote the fuzzy AND and fuzzy OR operators respectively, and for any real numbers a and b, they satisfy:

[0012]

[0013]

[0014]

[0015]

[0016] f j (·) and g i (·) represents an activation function, and each activation function satisfies the Lipschitz condition, that is, for any real number a and b, and |g i (a)-g i (b)|≤L i |ab|, where L i and is a positive number; at the same time, each activation function is bounded, satisfying and |g i (·)|≤F i , where F i and is a positive constant; a ji (x i (t))、b ji (x i (t))、c ij (y j (t)) and d ij (y j (t)) represents the memristor connection weight, and they satisfy:

[0017]

[0018]

[0019] The switching interval of the memristor is T i >0, and are all constants,

[0020] Introducing a replacement function α i and β j As a constant, replace the variables of the drive system and rewrite it into the following form:

[0021]

[0022] Step S12: constructing a response system corresponding to the driving system:

[0023]

[0024] in and Respectively represent the state variables of the neurons in the i-th layer and the j-th layer of the response system at time t; u i (t), v i (t) and An adaptive quantized controller to be designed; and represents the memristor connection weight, and they satisfy:

[0025]

[0026]

[0027] The switching interval of the memristor is Ti >0, and are all constants, The other performance indicators in the response system are the same as the performance indicators in the drive system;

[0028] Step S2: setting the synchronization error between the driving system and the response system, and establishing a synchronization error system;

[0029] Step S3: designing an adaptive quantization controller, and applying the adaptive quantization controller to the response system, so that the response system is synchronized with the driving system at a fixed time, thereby realizing a confidential communication method.

[0030] Furthermore, step S2 specifically includes the following steps:

[0031] Step S21: According to the fuzzy inertial BAM memristor neural network drive system and response system constructed in step S1, the synchronization error of the drive system and the response system is set to:

[0032]

[0033]

[0034] Step S22: According to the driving system and the response system, and the synchronization error set in step S21, a synchronization error system is established as follows:

[0035]

[0036]

[0037] in

[0038]

[0039]

[0040] in

[0041] Furthermore, step S3 specifically includes the following steps:

[0042] Step S31: According to the synchronization error constructed in step S2, an adaptive quantization controller is designed as follows:

[0043]

[0044] in,

[0045]

[0046]

[0047] where i = 1, 2, ..., n, j = 1, 2, ..., m; sign(·) is the sign function; 0 <a<1;b> 1;q(z i (t)) is the error z i The quantized value of (t), is the quantization error, θ is the sector boundary, 0<ρ<1;q(e i (t)) is the error e i The quantized value of (t), Error The quantitative value of Error The quantitative value of and is the adaptive quantized synchronous controller gain; and are all arbitrary normal numbers; ω i , and are all constants and satisfy:

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058] Step S32: applying the adaptive quantization controller to the response system, so that the response system is synchronized with the drive system at a fixed time;

[0059] Step S33: After the drive system is synchronized with the response system, the sending end obtains the chaotic signal generated by the drive system as the encrypted signal x i (t) and y j (t); The receiving end obtains the chaotic signal generated by the response system as a decryption signal and

[0060] Step S34: The transmitting end encrypts the signal x i (t) and y j (t), and the plaintext signal r i (t) and Perform encryption operations respectively to obtain the ciphertext signal h i (t) and h i (t) = x i (t)+r i (t),

[0061] Step S35: The transmitting end sends the ciphertext signal h through the channel i (t) and The receiving end receives the ciphertext signal h through the channel i (t) and

[0062] Step S36: The receiving end receives the ciphertext signal h i (t) and and decryption signal and Perform decryption operation to obtain the decrypted plaintext signal ss i (t) and

[0063] Furthermore, the response system is synchronized with the drive system at a fixed time, and the upper bound T of the fixed time is:

[0064]

[0065] Where 0 <a<1;g> 1;

[0066]

[0067]

[0068]

[0069]

[0070]

[0071] The present invention provides a secure communication method of a BAM memristor neural network based on adaptive quantization control. Compared with the prior art, the present invention has the following beneficial effects:

[0072] 1. The present invention uses memristors to simulate synapses in a neural network and constructs a memristor neural network.

[0073] 2. In the present invention, the fuzzy neural network is combined with the BAM memristor neural network, and the inertia term is introduced to construct a fuzzy inertial BAM memristor neural network.

[0074] 3. In the present invention, in order to synchronize the response system with the driving system at a fixed time, an adaptive quantization controller is considered, which effectively reduces the control cost and saves channel resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 This is a flow chart of the secure communication method of the BAM memristor neural network based on adaptive quantization control of the present invention;

[0076] Figure 2 : is a synchronous error variation trajectory diagram without the synchronous controller in Example 2 of the present invention, where (a) is the synchronous error z without the synchronous controller i (t) and e i (t), i = 1, 2 change trajectory diagram, (b) is the synchronization error without the synchronization controller and j=1,2 change trajectory diagram;

[0077] Figure 3 is a state change trajectory diagram of the driving system and the response system under the action of the adaptive quantization controller in Example 2 of the present invention, wherein (a) is x1(t) and s1(t) and State change trajectory diagram, (b) is x2(t) and s2(t) and Change trajectory diagram, (c) is the adaptive quantization controller under the action of y1(t) and r1(t) and Change trajectory diagram, (d) is y2(t) and r2(t) and Change trajectory diagram;

[0078] Figure 4: is a diagram showing the change trajectory of the synchronization error under the action of the adaptive quantization controller in Example 2 of the present invention, wherein (a) is the synchronization error z under the action of the adaptive quantization controller i (t) and e i (t), i = 1, 2 change trajectory diagram, (b) is the synchronization error under the action of adaptive quantization controller and j=1,2 change trajectory diagram;

[0079] Figure 5 : is a diagram showing the change trajectory of the quantized synchronization error under the action of the adaptive quantized controller in Embodiment 2 of the present invention, wherein (a) is the quantized synchronization error q(z i (t)) and q(e i (t)), i = 1, 2 change trajectory diagram, (b) is the quantized synchronization error under the action of the adaptive quantization controller and j=1,2 change trajectory diagram;

[0080] Figure 6 This is a control gain change trajectory diagram under the action of an adaptive quantization controller in Example 2 of the present invention, where (a) is the control gain under the action of an adaptive quantization controller and i=1,2 change trajectory diagram, (b) is the control gain under the action of adaptive quantization controller and j=1,2 change trajectory diagram. DETAILED DESCRIPTION

[0081] 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.

[0082] Embodiment 1:

[0083] like Figure 1 As shown, this embodiment provides a secure communication method of a BAM memristor neural network based on adaptive quantization control. The synchronization control method includes the following steps:

[0084] Step S1: construct a fuzzy inertial BAM memristor neural network drive system and response system; Step S1 specifically includes the following steps:

[0085] Step S11: construct the fuzzy inertial BAM memristor neural network drive system as follows:

[0086]

[0087] Wherein, time t≥0; i=1,2,…,n; j=1,2,…,m; n≥2 and m≥2 are the number of neurons in the i-th layer and the j-th layer of the driving system respectively; x i (t) and y j (t) are the state variables of the neurons in the i-th and j-th layers of the driving system at time t, and They are state variables x i (t) and y j The second-order derivative of (t) is the inertia term of the drive system; and is a positive constant; τ(t) is a discrete time lag and satisfies 0≤τ(t)≤τ, where τ is a positive constant; and represents a feed-forward template; and represents the fuzzy feedforward minimum template; and represents the fuzzy feedforward maximum template; and Represents the elements of the minimal template for fuzzy feedback; and Represents the elements of the maximum template of fuzzy feedback; and denote the inputs of the j-th neural network and the i-th neural network respectively; ∧ and ∨ denote the fuzzy AND and fuzzy OR operators respectively, and for any real numbers a and b, they satisfy:

[0088]

[0089]

[0090]

[0091]

[0092] f j (·) and g i (·) represents an activation function, and each activation function satisfies the Lipschitz condition, that is, for any real number a and b, and |g i (a)-g i (b)|≤L i |ab|, where L i and is a positive number; at the same time, each activation function is bounded, satisfying and |g i (·)|≤F i, where F i and is a positive constant; a ji (x i (t))、b ji (x i (t))、c ij (y j (t)) and d ij (y j (t)) represents the memristor connection weight, and they satisfy:

[0093]

[0094]

[0095] The switching interval of the memristor is T i >0, and are all constants,

[0096] Introducing a replacement function α i and β j As a constant, replace the variables of the drive system and rewrite it into the following form:

[0097]

[0098] Step S12: constructing a response system corresponding to the driving system:

[0099]

[0100] in and Respectively represent the state variables of the neurons in the i-th layer and the j-th layer of the response system at time t; u i (t), v i (t) and An adaptive quantized controller to be designed; and represents the memristor connection weight, and they satisfy:

[0101]

[0102]

[0103] The switching interval of the memristor is T i >0, and are all constants, The other performance indicators in the response system are the same as the performance indicators in the drive system;

[0104] Step S2: setting the synchronization error between the driving system and the response system, and establishing a synchronization error system;

[0105] Step S3: designing an adaptive quantization controller, and applying the adaptive quantization controller to the response system, so that the response system is synchronized with the driving system at a fixed time, thereby realizing a confidential communication method.

[0106] In this embodiment, step S2 specifically includes the following steps:

[0107] Step S21: According to the fuzzy inertial BAM memristor neural network drive system and response system constructed in step S1, the synchronization error of the drive system and the response system is set to:

[0108]

[0109]

[0110] Step S22: According to the driving system and the response system, and the synchronization error set in step S21, a synchronization error system is established as follows:

[0111]

[0112]

[0113] in

[0114]

[0115]

[0116] in

[0117] In this embodiment, step S3 specifically includes the following steps:

[0118] Step S31: According to the synchronization error constructed in step S2, an adaptive quantization controller is designed as follows:

[0119]

[0120] in

[0121]

[0122]

[0123] where i = 1, 2, ..., n, j = 1, 2, ..., m; sign(·) is the sign function; 0 <a<1;b> 1;q(z i (t)) is the error z i The quantized value of (t), is the quantization error, θ is the sector boundary, 0<ρ<1;q(e i (t)) is the error e i The quantized value of (t), Error The quantitative value of Error The quantitative value of and is the adaptive quantization controller gain; and are all arbitrary normal numbers; ω i , and are all constants and satisfy:

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130]

[0131]

[0132]

[0133]

[0134] Step S32: applying the adaptive quantization controller to the response system, so that the response system is synchronized with the drive system at a fixed time;

[0135] Step S33: After the drive system is synchronized with the response system, the sending end obtains the chaotic signal generated by the drive system as the encrypted signal x i (t) and y j (t); The receiving end obtains the chaotic signal generated by the response system as a decryption signal and

[0136] Step S34: The transmitting end encrypts the signal x i (t) and y j (t), and the plaintext signal r i (t) and Perform encryption operations respectively to obtain the ciphertext signal h i (t) and h i (t) = x i (t)+r i (t),

[0137] Step S35: The transmitting end sends the ciphertext signal h through the channel i (t) and The receiving end receives the ciphertext signal h through the channel i (t) and

[0138] Step S36: The receiving end receives the ciphertext signal h i (t) and and decryption signal and Perform decryption operation to obtain the decrypted plaintext signal ss i (t) and

[0139] In this example, the response system is synchronized with the drive system at a fixed time, and the upper bound T of the fixed time is:

[0140]

[0141] Where 0 <a<1;b> 1;

[0142]

[0143]

[0144]

[0145]

[0146]

[0147] It is worth noting that in the present invention, memristors are used to simulate synapses in neural networks. In neural networks, synapses are responsible for information storage and calculation, so they are indispensable. Memristors have memory characteristics similar to synapses and can simulate synapses well. According to the characteristics of memristors, the BAM memristor neural network model is particularly considered, and the fuzzy neural network is combined with the BAM memristor neural network, the inertia term is introduced, and the fuzzy inertial BAM memristor neural network is considered, so that the confidential communication method of the BAM memristor neural network has a wider application background and the complexity of confidential communication is further improved. In the present invention, an adaptive quantization controller is considered, which can effectively reduce the control cost and save channel resources compared to the traditional feedback synchronization controller.

[0148] Embodiment 2:

[0149] This embodiment mainly includes two parts:

[0150] The first is to theoretically prove the effectiveness of the secure communication method of the BAM memristor neural network based on adaptive quantization control proposed in Example 1.

[0151] The second is to verify the synchronization performance of the fuzzy inertial BAM memristor neural network drive system and response system in Example 1 through numerical simulation methods.

[0152] (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.)

[0153] 1. Theoretical Proof

[0154] The following are the definitions and lemmas that will be used in the proof:

[0155] Definition 1: If there exists a constant t * >0 so that the error between the drive system and the response system satisfies At the same time for ||z(t)||≡0 and ‖e(t)‖≡0, then the driving system and the response system are said to be able to achieve finite-time synchronization. Where ||·‖ is defined as the Euclidean norm, t * is called the finite synchronization time,

[0156] Definition 2: If the driving system and the response system can achieve finite time synchronization and there is a fixed constant T max>0, so that for any initial value of the system z(0), e(0), t * (z(0), e(0))≤T max , then the drive system and the response system are said to be able to achieve fixed time synchronization.

[0157] Lemma 1: If there exists a continuous, positive definite and radially unbounded function V(t): And the following conditions are met:

[0158] (1)

[0159] (2) For k, 0<a≤1,b> 1. Any solution e(t) of the error system satisfies:

[0160]

[0161] Then the zero solution of the error system is fixed-time stable, and the value of the fixed-time upper bound T is:

[0162]

[0163] Lemma 2: Let x1, x2, …, x n is a positive constant and 0<a≤1,b> 1, then

[0164]

[0165] Next, construct the Lyapunov functional:

[0166] V(t)=V1(t)+V2(t)

[0167]

[0168]

[0169]

[0170] Take the derivative of the Lyapunov functional:

[0171]

[0172]

[0173] and Through mathematical derivation, we can get:

[0174]

[0175]

[0176] According to the lemma, we can finally get:

[0177]

[0178]

[0179] in

[0180]

[0181]

[0182]

[0183]

[0184] make

[0185] This allows us to obtain a fixed time upper bound

[0186] Therefore, according to the definition and the lemma, under the action of the adaptive quantized controller, the response system is synchronized with the drive system in fixed time, and the fixed time upper bound is

[0187] 2. Numerical Simulation

[0188] In this embodiment, the following fuzzy inertial BAM memristor neural network system model is considered, and the driving system is:

[0189]

[0190] The corresponding response system is:

[0191]

[0192] The specific parameter settings are as follows: i = 1, 2; j = 1, 2; t ≥ 0; the activation function is set to f j (·) = g i (·)=tanh(·); discrete time lag a=0.5, b=1.5; α1=1; β1=1.2, α2=1.2, β2=1.3;

[0193] Initial values ​​x1(s)=x2(s)=0.8, y1(s)=y2(s)=1, s1(s)=-1.2, s2(s)=1.2, r1(s)=-0.8, r2(s)=0.8, s∈[-2,0]; switching interval T i =1, According to the memristor connection weight, we can get:

[0194] According to the above parameter settings and inequalities, the parameter value ranges are: ω1≥5.2, ω2≥7.7, Select ω1=5.5,ω2=7.8, The other parameter values ​​are and Taking the value as 1, where i = 1, 2, j = 1, 2, ρ = 0.7, we can get

[0195] The specific simulation experiment results are as follows: Figure 2 : is a synchronous error variation trajectory diagram without the synchronous controller in Example 2 of the present invention, where (a) is the synchronous error z without the synchronous controller i (t) and e i (t), i = 1, 2 change trajectory diagram, (b) is the synchronization error without the synchronization controller and j=1,2 change trajectory diagram; Figure 3 is a state change trajectory diagram of the driving system and the response system under the action of the adaptive quantization controller in Example 2 of the present invention, wherein (a) is x1(t) and s1(t) and State change trajectory diagram, (b) is x2(t) and s2(t) and Change trajectory diagram, (c) is the adaptive quantization controller under the action of y1(t) and r1(t) and Change trajectory diagram, (d) is y2(t) and r2(t) and Change trajectory diagram; Figure 4 : is a diagram showing the change trajectory of the synchronization error under the action of the adaptive quantization controller in Example 2 of the present invention, wherein (a) is the synchronization error z under the action of the adaptive quantization controller i (t) and e i (t), i = 1, 2 change trajectory diagram, (b) is the synchronization error under the action of adaptive quantization controller and j=1,2 change trajectory diagram; Figure 5 : is a diagram showing the change trajectory of the quantized synchronization error under the action of the adaptive quantized controller in Embodiment 2 of the present invention, wherein (a) is the quantized synchronization error q(z i (t)) and q(e i (t)), i = 1, 2 change trajectory diagram, (b) is the quantized synchronization error under the action of the adaptive quantization controller and j=1,2 change trajectory diagram; Figure 6 This is a control gain change trajectory diagram under the action of an adaptive quantization controller in Example 2 of the present invention, where (a) is the control gain under the action of an adaptive quantization controller and i=1,2 change trajectory diagram, (b) is the control gain under the action of adaptive quantization controller and j=1,2 change trajectory diagram. Figure 2 It shows that the drive system and the response system cannot achieve synchronization without the action of the synchronization controller; Figure 3-Figure 5 It shows that under the action of the adaptive quantized controller, the response system is synchronized with the drive system, verifying the synchronization performance.

[0196] 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 of a BAM memristor neural network based on adaptive quantization control, characterized in that: The following steps are involved: Step S1: construct a fuzzy inertial BAM memristor neural network drive system and response system; Step S1 specifically includes the following steps: Step S11: construct the fuzzy inertial BAM memristor neural network drive system as follows: Wherein, time t≥0; i=1,2,…,n; j=1,2,…,m; n≥2 and m≥2 are the number of neurons in the i-th layer and the j-th layer of the driving system respectively; x i (t) and y j (t) are the state variables of the neurons in the i-th and j-th layers of the driving system at time t, and They are state variables x i (t) and y j The second-order derivative of (t) is the inertia term of the drive system; and is a positive constant; τ(t) is a discrete time lag and satisfies 0≤τ(t)≤τ, where τ is a positive constant; and represents a feed-forward template; and represents the fuzzy feedforward minimum template; and represents the fuzzy feedforward maximum template; and Represents the elements of the minimal template for fuzzy feedback; and Represents the elements of the maximum template of fuzzy feedback; and denote the inputs of the j-th neural network and the i-th neural network respectively; ∧ and ∨ denote the fuzzy AND and fuzzy OR operators respectively, and for any real numbers a and b, they satisfy: f j (·) and g i (·) represents an activation function, and each activation function satisfies the Lipschitz condition, that is, for any real number a and b, and |g i (a)-g i (b)|≤L i |ab|, where L i and is a positive number; at the same time, each activation function is bounded, satisfying and |g i (·)|≤F i , where F i and is a positive constant; a ji (x i (t))、b ji (x i (t))、c ij (y j (t)) and d ij (y j (t)) represents the memristor connection weight, and they satisfy: The switching interval of the memristor is T i >0, and are all constants, Introducing a replacement function α i and β j As a constant, replace the variables of the drive system and rewrite it into the following form: Step S12: constructing a response system corresponding to the driving system: in and Respectively represent the state variables of the neurons in the i-th layer and the j-th layer of the response system at time t; u i (t), v i (t) and An adaptive quantized controller to be designed; and represents the memristor connection weight, and they satisfy: The switching interval of the memristor is T i >0, and are all constants, The other performance indicators in the response system are the same as the performance indicators in the drive system; Step S2: setting the synchronization error between the driving system and the response system, and establishing a synchronization error system; Step S3: designing an adaptive quantization controller, and applying the adaptive quantization controller to the response system, so that the response system is synchronized with the drive system at a fixed time, thereby realizing a confidential communication method; Step S2 specifically includes the following steps: Step S21: According to the fuzzy inertial BAM memristor neural network drive system and response system constructed in step S1, the synchronization error of the drive system and the response system is set to: Step S22: According to the driving system and the response system, and the synchronization error set in step S21, a synchronization error system is established as follows: in in Step S3 specifically includes the following steps: Step S31: According to the synchronization error constructed in step S2, an adaptive quantization controller is designed as follows: in where i = 1, 2, ..., n, j = 1, 2, ..., m; sign(·) is the sign function; 0 <a<1;b> 1;q(z i (t)) is the error z i The quantized value of (t), is the quantization error, θ is the sector boundary, q(e i (t)) is the error e i The quantized value of (t), Error The quantitative value of Error The quantitative value of and is the adaptive quantization controller gain; and are all arbitrary normal numbers; ω i , and are all constants and satisfy: Step S32: applying the adaptive quantization controller to the response system, so that the response system is synchronized with the drive system at a fixed time; Step S33: After the drive system is synchronized with the response system, the sending end obtains the chaotic signal generated by the drive system as the encrypted signal x i (t) and y j (t); The receiving end obtains the chaotic signal generated by the response system as a decryption signal and Step S34: The transmitting end encrypts the signal x i (t) and y j (t), and the plaintext signal ri(t) and Perform encryption operations respectively to obtain the ciphertext signal h i (t) and h i (t) = x i (t)+r i (t), Step S35: The transmitting end sends the ciphertext signal h through the channel i (t) and The receiving end receives the ciphertext signal h through the channel i (t) and Step S36: The receiving end receives the ciphertext signal h i (t) and and decryption signal and Perform decryption operation to obtain the decrypted plaintext signal ss i (t) and 2. The secure communication method of the BAM memristor neural network based on adaptive quantization control according to claim 1 is characterized in that: The response system is synchronized with the drive system at a fixed time, and the upper bound T of the fixed time is: Where 0 <a<1;b> 1;

Images