Simulation Attack Method of Energy Harvesting Cognitive Internet of Things Against Primary Users in Evolutionary Game

Abstract

The present invention proposes a method for simulating attacks by primary users in an energy harvesting cognitive Internet of Things against the evolution game. For the scenario of a cognitive Internet of Things applying energy harvesting technology, where selfish secondary user nodes in this scenario can supplement energy through energy harvesting technology, the method first establishes an incentive and punishment mechanism including credit parameters and punishment parameters, then uses evolutionary game to analyze the dynamics and attack and defense mechanisms of selfish secondary users and normal secondary users, and then reduces the attack probability of selfish secondary users by adjusting the punishment parameters to maximize the utilization rate of idle spectrum and the throughput of normal secondary users; the present invention can improve the throughput of normal secondary users in the cognitive Internet of Things adopting the energy harvesting scheme.

Description

Technical Field

[0001] The present invention relates to the technical fields of cognitive radio and Internet of Things, and in particular to an evolutionary game-based energy harvesting cognitive Internet of Things countermeasure against primary user simulation attacks Background Art

[0002] With the rapid development of wireless communication technology, the wireless Internet of Things is considered to be one of the most promising technologies in many fields. However, due to the explosive growth of Internet of Things devices, wireless communication is very challenging, and many devices that require bandwidth compete for the very limited wireless bandwidth. Therefore, the shortage of spectrum resources has become an important challenge for large-scale deployment of Internet of Things devices. In this regard, cognitive radio is a spectrum sharing technology with spectrum sensing capabilities that can automatically sense the surrounding electromagnetic environment, dynamically detect and effectively utilize idle spectrum, which will greatly reduce the constraints of spectrum and bandwidth limitations on the development of wireless communication technology. Applying cognitive radio to the Internet of Things field is an inevitable trend to alleviate the shortage of spectrum resources in the Internet of Things. Cognitive Internet of Things communication networks have become a hot research direction for the new generation of Internet of Things wireless communication technology.

[0003] In a cognitive Internet of Things network, if the primary user is not on the channel, the secondary user can enter the channel to work after sensing an idle channel and reuse the idle spectrum to improve the utilization rate of spectrum resources. However, selfish secondary user nodes will launch attacks during the spectrum sensing stage, occupy the idle spectrum by imitating the information characteristics of the primary user, creating an illusion of a busy channel, causing normal secondary users to evacuate the spectrum. The primary user simulation attack will reduce the spectrum usage rate, seriously affect the information transmission work of normal secondary users, and cause the performance of the cognitive Internet of Things communication network to decline sharply.

[0004] In addition, there is also an energy supply problem in cognitive Internet of Things communication networks. Energy harvesting technology can harvest energy from radio frequency signals and non-radio frequency signals (such as solar energy, wind energy, vibration, etc.) and use it for data transmission or store it in an energy storage device. Among them, harvesting the energy of radio frequency signals is relatively simple and can be achieved only by multiplexing the antenna of the communication device, so it is more conducive to the application of cognitive Internet of Things. With the rise of energy harvesting technology and its application in various communication fields, energy harvesting technology will hopefully become one of the effective and promising solutions to the energy limitation in cognitive Internet of Things communication networks. However, the application of energy harvesting technology in cognitive Internet of Things will inevitably exacerbate the threat of the primary user simulation attack, a communication security problem.

[0005] Cognitive Internet of Things allows secondary users to use the spectrum of primary users when the primary users are absent. However, when the primary users are absent, selfish secondary users can mislead normal secondary users to leave the spectrum band by launching primary user emulation attacks, so as to selfishly occupy the idle spectrum. In addition, selfish secondary users can continuously collect and store energy through energy harvesting technology, which can offset the energy cost of launching emulation attacks to a certain extent. Therefore, selfish secondary users are more motivated to launch primary user emulation attacks after being equipped with energy harvesting technology.

[0006] In summary, the primary user emulation attack in the context of energy harvesting is one of the most challenging problems in cognitive Internet of Things, and it is also the problem to be solved by this invention. From the existing literature, game theory has been widely applied to the research of security confrontation problems in cognitive communication networks. Game theory mainly studies the decision-making when the behaviors of decision-making subjects directly interact with each other and the equilibrium of such decision-making. It studies competitive phenomena through mathematical theories and methods, and can analyze the optimal strategy selection in the case of mutual influence between decision-making behaviors. Summary of the Invention

[0007] The present invention proposes an evolutionary game-based method for energy harvesting cognitive Internet of Things to counter primary user emulation attacks, which can improve the throughput of normal secondary users in cognitive Internet of Things adopting energy harvesting schemes.

[0008] The present invention adopts the following technical solutions.

[0009] An evolutionary game-based method for energy harvesting cognitive Internet of Things to counter primary user emulation attacks, aiming at the cognitive Internet of Things scenario applying energy harvesting technology, in which selfish secondary user nodes can supplement energy through energy harvesting technology. The method first establishes an incentive and punishment mechanism including credit parameters and punishment parameters, then uses evolutionary game to analyze the dynamics and attack and defense mechanisms of selfish secondary users and normal secondary users, and then reduces the attack probability of selfish secondary users by adjusting the punishment parameters to maximize the utilization rate of idle spectrum and the throughput of normal secondary users.

[0010] The evolutionary game summarizes the attack and defense mechanisms of secondary users in the cognitive Internet of Things network with a game tree, specifically:

[0011] Half of the branches of the game tree represent: when the channel is busy, selfish secondary user nodes cannot perceive the idle spectrum and will not launch attacks. At this time, normal secondary user nodes do not know whether it is the primary user or the selfish secondary user node launching the emulation attack using the spectrum on the channel. Therefore, normal secondary user nodes will choose to execute or not execute emulation attack detection. If they choose to execute detection, the probability of detection error of the emulation attack detection, that is, the false alarm probability p eer ,

[0012] When a normal secondary user node misdetects a primary user in the channel as a selfish secondary user node, the normal secondary user node will enter the channel's used spectrum and interfere with the primary user's communication, causing the normal secondary user node to be punished u eer , thus reducing the enthusiasm of the normal secondary user node for detection;

[0013] When the channel is idle and the selfish secondary user node launches a simulation attack, the normal secondary user node is afraid of false alarms and dare not perform detection. In order to consider the impact of false alarm penalties in the scenario of the game between the normal secondary user node and the selfish secondary user node, a false alarm impact factor δ is set, and the false alarm penalty when the channel is busy is converted into the loss of performing detection in the face of the simulation attack of the selfish secondary user node, λ = p p ·y·p eer ·u eer ·δ, to simulate the situation where the enthusiasm of the normal secondary user node for performing detection decreases in the face of the simulation attack of the selfish secondary user node; the other half of the branches of the game tree indicates that when the channel is idle, the selfish secondary user node can choose to launch or not launch a simulation attack, and the normal secondary user node can choose to perform or not perform simulation attack detection. When the selfish secondary user node does not launch a simulation attack, the normal secondary user node does not need to perform detection. However, considering that user nodes in the real situation are not completely rational and may make wrong decisions, after the normal secondary user node encounters multiple simulation attacks from the selfish secondary user node, it habitually performs detection first before using the spectrum.

[0014] The payoff matrices of both sides of the game are obtained based on the game tree and system parameters, and are described in a table as follows:

[0015]

[0016] The relevant parameters of the payoff matrix are as follows

[0017] r s ——The payoff of monopolizing the idle channel: If the selfish secondary user node launches a simulation attack and successfully scares away the normal secondary user node, the selfish secondary user node monopolizes the idle channel and obtains the payoff r s ; when the selfish secondary user node does not attack or launches an attack but is detected by the normal secondary user node, the secondary user nodes share the idle spectrum together, and the selfish secondary user node and the normal secondary user node each obtain the payoff c a ——Attack cost: The energy loss of the selfish secondary user node launching a simulation attack;

[0018] c d ——Detection cost: The energy loss of performing the simulation attack detection of the selfish secondary user node;

[0019] u p—— Penalty parameter: The penalty imposed on a selfish secondary user node when its spoofing attack is detected, and the corresponding fine is obtained by a normal secondary user node;

[0020] E - Energy harvesting parameter: The energy gain obtained through energy harvesting;

[0021] H - Credit parameter: The reward obtained by a selfish secondary user node when it does not initiate a spoofing attack and does not cause a normal secondary user node to perform detection;

[0022] p p —— Miss detection probability: The probability that a selfish secondary user node initiates a spoofing attack, and a normal secondary user node performs detection but fails to detect the spoofing attack of the selfish secondary user node;

[0023] p ma —— Miss detection probability: The probability that a selfish secondary user node initiates a spoofing attack, and a normal secondary user node performs detection but fails to detect the spoofing attack of the selfish secondary user node;

[0024] p eer —— False alarm probability: The probability of misdetecting the primary user in the channel as a spoofing attack of a selfish secondary user node when the channel is busy;

[0025] x - Attack probability: The probability that a selfish secondary user node initiates a spoofing attack;

[0026] y - Detection probability: The probability that a normal secondary user node performs spoofing attack detection;

[0027] u eer —— False alarm penalty parameter: The penalty suffered for interfering with the primary user when, when the channel is busy, the primary user in the channel is misdetected as a spoofing attack of a selfish secondary user and then enters the channel

[0028] δ - False alarm influence factor: The influence factor of the penalty for interfering with the primary user on the detection probability of a normal secondary user node; λ - Detection loss parameter: The loss of converting the false alarm penalty into the loss of performing spoofing attack detection.

[0029] In the payoff matrix, energy harvesting E is introduced to simulate a scenario where selfish secondary user nodes are more motivated to attack. Specifically, when communication nodes are equipped with energy harvesting capabilities, selfish secondary user nodes obtain energy through energy harvesting. In the spectrum sensing stage, they first harvest energy and then initiate an attack. The harvested energy can partially offset the energy consumption of initiating a spoofing attack, thereby increasing the motivation of selfish secondary user nodes to initiate spoofing attacks. To simulate the scenario of energy harvesting - primary user spoofing attack with a higher attack frequency, let selfish secondary user nodes obtain additional energy E when initiating a spoofing attack;

[0030] Introduce the credit parameter H into the payoff matrix. The specific method is as follows: when the selfish secondary user node does not cause a negative effect on the cognitive Internet of Things network, that is, when it does not initiate a simulation attack and does not cause the normal secondary user node to perform detection, it obtains a reward to encourage the selfish secondary user node to evolve towards a non-attack strategy selection;

[0031] The credit parameter H ensures that when the normal secondary user node does not detect, under the incentive of the credit parameter, the payoff of the selfish secondary user node not attacking is greater than the payoff of attacking, that is

[0032]

[0033] At the same time, the credit parameter ensures that when the selfish secondary user node initiates a simulation attack, the payoff of the normal secondary user node performing detection is greater than the payoff of not performing detection, to ensure the enthusiasm of the normal secondary user node to detect the received signal, that is

[0034]

[0035] The expected payoff of the selfish secondary user node initiating a simulation attack is expressed by the formula:

[0036]

[0037] The expected payoff of the selfish secondary user node not initiating a simulation attack is:

[0038]

[0039] The average expected payoff of the entire population of selfish secondary user nodes is:

[0040]

[0041] The replicator dynamics equation of the selfish secondary user node is:

[0042]

[0043] The expected payoff of the normal secondary user node performing simulation attack detection is:

[0044]

[0045] The expected payoff of the normal secondary user node not performing simulation attack detection is:

[0046]

[0047] The average expected payoff of the entire population of normal secondary user nodes is:

[0048]

[0049] The replicator dynamics equation of the normal secondary user node is:

[0050]

[0051] By using formula 3 and formula 4 of the simultaneous equations, we construct a dynamic system equation group. Let formula 3 = 0 and formula 4 = 0, and we get five local equilibrium points: (0,0), (1,0), (0,1), (1,1) and (x * ,y * ),x * ∈[0,1],y * ∈[0,1], where The stability is judged by the Jacobian matrix. The Jacobian matrix of the selfish secondary user node simulation attack evolution game model is expressed as:

[0052]

[0053] When the local equilibrium point satisfies the conditions tr J < 0, det J > 0, the equilibrium point is an evolutionary stable point; the calculation results of the four elements of the Jacobian matrix corresponding to the five local equilibrium points are expressed in the form of

[0054]

[0055] From formula 1 and formula 2, we can know that:

[0056] Heng established,

[0057] For X1(0,0), we have Therefore, X1(0,0) is an evolutionary stable point;

[0058] For X2(1,0), we have Therefore, point X2(1,0) is unstable;

[0059] For X5(x * ,y * ), with trJ=0, the eigenvalue of J Therefore X5(x * ,y * ) is a saddle point;

[0060] For X3(0,1) and X4(1,1), there are two scenarios;

[0061] Scenario 1: Right now At this time, for X3(0,1), we have Therefore, X3(0,1) is an unstable point. For X4(1,1), Therefore, X4(1,1) is an evolutionary stable point;

[0062] Scenario 2 That is

[0063]

[0064] At this time, for X3(0, 1), there is Therefore, X3(0, 1) is an unstable point; for X4(1, 1), there is

[0065] Therefore, X4(1, 1) is an unstable point;

[0066] In summary, when the inequality holds, both X1(0, 0) and X4(1, 1) are evolutionary stable points. When it does not hold, only X1(0, 0) is an evolutionary stable point;

[0067] The throughput of normal secondary users is:

[0068]

[0069] Since 0 < p ma <1, that is, -1 < y(1 - p ma ) - 1 < 0, then when the attack probability of selfish secondary users is minimized (x = 0), the throughput of normal secondary users reaches the maximum. By setting the penalty parameter, X1(0, 0) becomes the only evolutionary stable point, thereby reducing the attack probability of selfish secondary users and maximizing the throughput of normal secondary users.

[0070] The present invention can improve the throughput of normal secondary users in the cognitive Internet of Things adopting the energy harvesting scheme. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The present invention will be further described in detail below in conjunction with the drawings and specific embodiments:

[0072] Attached Figure 1 is a schematic diagram of the cognitive Internet of Things network model according to an embodiment of the present invention;

[0073] Attached Figure 2 is a schematic diagram of the game tree describing the main user simulation attack problem in an embodiment of the present invention;

[0074] Attached Figure 3 is a schematic diagram for verifying the influence of energy harvesting on the evolutionary strategy of selfish secondary user nodes in an embodiment of the present invention;

[0075] Attached Figure 4 is a schematic diagram for verifying the influence of the penalty parameter on the evolutionary strategy of selfish secondary user nodes in an embodiment of the present invention;

[0076] Attached Figure 5Schematic diagram of scenario 1 of the two parties' evolutionary stabilization strategy in an embodiment of the present invention;

[0077] Attached Figure 6 Schematic diagram of scenario 2 of the two parties' evolving stable strategies in an embodiment of the present invention;

[0078] Attached Figure 7 This is a schematic diagram of verifying the effect of the method proposed in the present invention on improving the normal secondary user throughput in an embodiment of the present invention. DETAILED DESCRIPTION

[0079] As shown in the figure, the energy harvesting cognitive Internet of Things based on evolutionary game is used to counter the primary user simulation attack method. For the cognitive Internet of Things scenario that applies energy harvesting technology, the selfish secondary user nodes in the scenario can replenish energy through energy harvesting technology. The method first establishes a reward and punishment mechanism including credit parameters and penalty parameters, and then uses evolutionary game to analyze the dynamics and attack and defense mechanisms of selfish secondary users and normal secondary users. Then, the penalty parameters are adjusted to reduce the attack probability of selfish secondary users to maximize the utilization of idle spectrum and the throughput of normal secondary users.

[0080] The evolutionary game summarizes the attack and defense mechanism of secondary users in the cognitive IoT network using a game tree, specifically:

[0081] Half of the branches of the game tree represent: when the channel is busy, the selfish secondary user node cannot sense the idle spectrum and will not launch an attack. At this time, the normal secondary user node does not know whether the spectrum in the channel is used by the primary user or the selfish secondary user node launching a simulated attack. Therefore, the normal secondary user node will choose to perform or not perform simulated attack detection. If it chooses to perform detection, the probability of false detection of simulated attack detection is p eer ,

[0082] When a normal secondary user node misdetects a primary user in the channel as a selfish secondary user node, the normal secondary user node will enter the channel to use the spectrum and interfere with the communication of the primary user, causing the normal secondary user node to be punished u eer , thus reducing the enthusiasm of normal secondary user nodes to perform detection;

[0083] When the channel is idle and the selfish secondary user node launches a simulation attack, the normal secondary user node dares not perform detection for fear of false alarms. In order to consider the impact of false alarm penalties in the game scenario between normal secondary user nodes and selfish secondary user nodes, a false alarm impact factor δ is set to convert the false alarm penalty when the channel is busy into the loss of performing detection when facing the selfish secondary user node simulation attack, λ = p p ·y·p eer ·u eer· δ is used to simulate the situation where the enthusiasm of normal secondary user nodes to perform detection decreases when facing the simulation attack of selfish secondary user nodes; the other half of the branches of the game tree indicates that when the channel is idle, the selfish secondary user node can choose to initiate or not initiate a simulation attack, and the normal secondary user node can choose to perform or not perform simulation attack detection. When the selfish secondary user node does not initiate a simulation attack, the normal secondary user node has no need to perform detection. However, considering that user nodes in real situations are not completely rational and may make wrong decisions, after experiencing multiple simulation attacks from selfish secondary user nodes, the normal secondary user node habitually performs detection before using the spectrum.

[0084] Based on the game tree and system parameters, the payoff matrices of both sides of the game are obtained and described in the following table:

[0085]

[0086] The relevant parameters of the payoff matrix are as follows

[0087] r s —— The payoff for monopolizing the idle channel: If the selfish secondary user node initiates a simulation attack and successfully frightens the normal secondary user node, the selfish secondary user node monopolizes the idle channel and obtains the payoff r s ; when the selfish secondary user node does not attack or the attack is detected by the normal secondary user node, the secondary user nodes share the idle spectrum together, and the selfish secondary user node and the normal secondary user node each obtain the payoff

[0088] c a —— Attack cost: The energy loss of the selfish secondary user node when initiating a simulation attack;

[0089] c d —— Detection cost: The energy loss of performing the simulation attack detection of the selfish secondary user node;

[0090] u p —— Penalty parameter: The penalty suffered by the selfish secondary user node when its simulation attack is detected, and the normal secondary user node obtains the corresponding fine;

[0091] E - Energy harvesting parameter: The energy gain obtained through energy harvesting;

[0092] H - Credit parameter: The reward obtained by the selfish secondary user node when it does not initiate a simulation attack and does not trigger the normal secondary user node to perform detection;

[0093] p p —— False alarm probability: The probability that the selfish secondary user node initiates a simulation attack, and the normal secondary user node performs detection but fails to detect the attack of the selfish secondary user node;

[0094] pma —— False alarm probability: The probability that a selfish secondary user node launches a simulation attack and a normal secondary user node performs detection but fails to detect the attack of the selfish secondary user node;

[0095] p eer —— Missed alarm probability: The probability of misdetecting the primary user in the channel as an attack by a selfish secondary user node when the channel is busy;

[0096] x - Attack probability: The probability that a selfish secondary user node launches a simulation attack;

[0097] y - Detection probability: The probability that a normal secondary user node performs simulation attack detection;

[0098] u eer —— False alarm penalty parameter: The penalty suffered when interfering with the primary user after misdetecting the primary user in the channel as an attack by a selfish secondary user when the channel is busy;

[0099] δ - False alarm influence factor: The influence factor of the penalty for interfering with the primary user on the detection probability of a normal secondary user node; λ - Detection loss parameter: The loss of performing simulation attack detection converted from the false alarm penalty.

[0100] In the payoff matrix, the energy harvesting E is introduced to simulate a scenario where selfish secondary user nodes are more motivated to attack. Specifically, when communication nodes are equipped with energy harvesting capabilities, selfish secondary user nodes obtain energy through energy harvesting. In the spectrum sensing stage, they first harvest energy and then launch an attack. The harvested energy can partially offset the energy consumption of launching a simulation attack, thus increasing the motivation of selfish secondary user nodes to launch simulation attacks. To simulate the scenario of energy harvesting - primary user simulation attack with a higher attack frequency, let selfish secondary user nodes obtain additional energy E when launching a simulation attack;

[0101] In the payoff matrix, the credit parameter H is introduced. The specific method is as follows: When a selfish secondary user node does not cause negative effects on the cognitive Internet of Things network, that is, when it does not launch a simulation attack and does not cause normal secondary user nodes to perform detection, it obtains a reward to encourage selfish secondary user nodes to evolve towards the strategy of not attacking;

[0102] The credit parameter H ensures that when normal secondary user nodes do not detect, under the incentive of the credit parameter, the payoff of a selfish secondary user node not attacking is greater than the payoff of attacking, that is

[0103]

[0104] At the same time, the credit parameter ensures that when a selfish secondary user node launches a simulation attack, the payoff of a normal secondary user node performing detection is greater than the payoff of not performing detection, to ensure the enthusiasm of normal secondary user nodes to detect the received signal, that is

[0105]

[0106] The expected payoff of a selfish secondary user node launching a simulation attack is expressed by the formula:

[0107]

[0108] The expected payoff of a selfish secondary user node not launching a simulation attack is:

[0109]

[0110] The average expected payoff of the entire population of selfish secondary user nodes is:

[0111]

[0112] The replicator dynamics equation of a selfish secondary user node is:

[0113]

[0114] The expected payoff of a normal secondary user node performing simulation attack detection is:

[0115]

[0116] The expected payoff of a normal secondary user node not performing simulation attack detection is:

[0117]

[0118] The average expected payoff of the entire population of normal secondary user nodes is:

[0119]

[0120] The replicator dynamics equation of a normal secondary user node is:

[0121]

[0122] By simultaneously solving Equation 3 and Equation 4 of the formula, a system of dynamic equations is constructed. Let Equation 3 = 0 and Equation 4 = 0, and five local equilibrium points are obtained: (0, 0), (1, 0), (0, 1), (1, 1) and (x * , y * ), x * ∈[0, 1], y * ∈[0, 1], where Its stability is judged through the Jacobian matrix. The Jacobian matrix of the evolutionary game model of selfish secondary user node simulation attack is expressed as:

[0123]

[0124] When the local equilibrium point satisfies the conditions tr J < 0 and det J > 0, this equilibrium point is an evolutionarily stable point; the calculation results of the four elements of the Jacobian matrix corresponding to the five local equilibrium points are presented in a table as

[0125]

[0126] It can be seen from Formula 1 and Formula 2 that:

[0127] always holds,

[0128] For X1(0, 0), there is Therefore, X1(0, 0) is an evolutionarily stable point;

[0129] For X2(1, 0), there is Therefore, the point X2(1, 0) is unstable;

[0130] For X5(x * , y * ), there is trJ = 0, and the eigenvalues of J are Therefore, X5(x * , y * ) is a saddle point;

[0131] For X3(0, 1) and X4(1, 1), there are two scenarios;

[0132] Scenario 1, That is, At this time, for X3(0, 1), there is Therefore, X3(0, 1) is an unstable point. For X4(1, 1), there is Therefore, X4(1, 1) is an evolutionarily stable point;

[0133] Scenario 2, That is,

[0134]

[0135] At this time, for X3(0, 1), there is Therefore, X3(0, 1) is an unstable point; for X4(1, 1), there is

[0136] Therefore, X4(1, 1) is an unstable point;

[0137] In summary, when the inequality holds, both X1(0, 0) and X4(1, 1) are evolutionarily stable points. When it does not hold, only X1(0, 0) is an evolutionarily stable point;

[0138] The throughput of normal secondary users is as follows:

[0139]

[0140] Since 0 < p ma <1, that is, -1 < y(1 - p ma ) - 1 < 0, then when the attack probability of selfish secondary users is minimized (x = 0), the throughput of normal secondary users reaches the maximum. By setting the penalty parameter, X1(0, 0) becomes the only evolutionary stable point, thereby reducing the attack probability of selfish secondary users and maximizing the throughput of normal secondary users.

[0141] Example:

[0142] The present invention will be further described below in conjunction with the accompanying drawings and examples.

[0143] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs.

[0144] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0145] As Figure 1 shown, the present invention is directed to a cognitive Internet of Things scenario applying energy harvesting technology, in which selfish secondary user nodes can continuously replenish energy through energy harvesting technology. Compared with ordinary cognitive Internet of Things scenarios, the initial attack probability of selfish secondary user nodes is higher. To solve the negative problems brought about by the application of energy harvesting, the present invention proposes a new reward and punishment mechanism based on evolutionary game theory and effectively reduces the attack probability of selfish secondary users by adjusting the penalty parameter, thereby maximizing the utilization rate of idle spectrum and the throughput of normal secondary users. Figure 2 Regarding the attack and defense mechanism between selfish secondary user nodes and normal secondary user nodes in the described scenario, the following will verify the influence of energy harvesting on the attack probability of selfish secondary user nodes, the role of the penalty parameter in suppressing the attack enthusiasm of selfish secondary user nodes, and the improvement effect of the present invention on the throughput of normal secondary users. The initial settings of the model parameters are r s = 60, H = 30, p ma = 0.2, c a = 5, E = 4, u p = 5, c d= 5, λ = 3.

[0146] A. Verify the influence of the energy harvesting parameter E

[0147] In the cognitive Internet of Things scenario of the present invention, the influence of energy harvesting E is considered. Except for E, the initial settings of the model parameters remain unchanged. The given value of the detection probability y is (y = 0.55), and the initial value of the attack probability x is set to 0.6. The energy harvesting E is set to 3.2, 3.5, 3.8, 4, and 4.5 respectively. As Figure 3 shown, when E is small (E = 3.2), the evolution trend of the simulation attack probability x of the selfish secondary user node tends to 0. When the harvested energy increases, the speed at which x converges to 0 slows down, and when E is greater than a certain value, the evolution trend of x tends to 1, and the greater E is, the faster x converges to 1. The simulation verifies that the energy harvesting parameter E can improve the enthusiasm of the selfish secondary user node to initiate a simulation attack.

[0148] B. Verify the penalty parameter u p 's influence

[0149] A new reward and punishment mechanism is proposed in the energy harvesting cognitive Internet of Things scenario of the present invention. The following verifies the influence of the penalty parameter u p . Except for u p , the initial settings of the model parameters remain unchanged. The given value of y is (y = 0.4), and the initial value of x is set to 0.6. The penalty parameter u is set to 2, u p = 3, u p = 4, u p = 5, u p = 6, u p = 6. As Figure 4 shown, when u p is small (u p = 2), the evolution trend of the simulation attack probability x of the selfish secondary user node tends to 1. When u p increases, the speed at which x converges to 1 slows down, and when u p is greater than a certain value, the evolution trend of x tends to 0, and the greater u p is, the faster x converges to 0. The simulation verifies that the penalty parameter can effectively inhibit the attack enthusiasm of the selfish secondary user node.

[0150] C. Verify the evolutionary stable strategies of both sides

[0151] As described above, for whether X3(0, 1) and X4(1, 1) are equilibrium points, two cases need to be discussed. Therefore, the evolutionary stable strategies of both sides need to be verified for these two cases.

[0152] Case 1: The following inequality needs to be satisfied The parameter settings for Case 1 are as follows: rs = 60, H = 30, p ma = 0.2, c a = 5, E = 4, u p = 5, c d = 5, λ = 3. x and y are set with different initial values within the [0 1] interval, and the simulation results are as Figure 5 shown, verifying that X1(0, 0) and X4(1, 1) are the evolutionary stable points for both sides of the game.

[0153] Case 2: The inequality needs to be satisfied The parameter settings for Case 2 are as follows: r s = 60, H = 30, p ma = 0.2, c a = 5, E = 4, u p = 7, c d = 5, λ = 3. x and y are set with different initial values within the [0 1] interval, and the simulation results are as Figure 6 shown, verifying that X1(0, 0) is the evolutionary stable point for both sides of the game.

[0154] D. Verifying the improvement effect of the method proposed by the invention on the throughput of normal secondary users

[0155] In a cognitive Internet of Things network adopting energy technology, the attack probability of selfish secondary users is relatively high. In the research of traditional methods, when the initial attack probability is high, the attack probability x of users will evolve to 1, while the present invention can reduce the attack probability of selfish secondary users. The change in the attack probability will affect the throughput of normal secondary users. The model parameter settings are as follows: p p = 0.2, v = 1, t t = 1, x = 0.6, p ma = 0.2. Figure 7 Shows the comparison of the throughput of normal secondary users based on the traditional method and the method proposed by the present invention. After reaching the equilibrium state, the throughput of normal secondary users obtained by the traditional method is 7.68, while the throughput obtained by the present invention is 24. In an energy harvesting environment with a high initial attack probability, this method can increase the throughput of ordinary SUs by 212.5%.

[0156] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0157] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce means for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0158] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0159] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0160] As mentioned above, it is only the preferred embodiments of the present invention, and not a limitation to the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes. However, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution content of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An energy - harvesting cognitive Internet of Things (IoT) counter - primary - user simulation attack method based on evolutionary game, aiming at the cognitive IoT scenario applying energy - harvesting technology. In this scenario, selfish secondary user nodes can supplement energy through energy - harvesting technology, and it is characterized in that: The method first establishes an incentive and penalty mechanism including credit parameters and penalty parameters, then uses evolutionary game theory to analyze the dynamics and attack and defense mechanisms of selfish secondary users and normal secondary users, and then reduces the attack probability of selfish secondary users by adjusting the penalty parameters to maximize the utilization rate of idle spectrum and the throughput of normal secondary users; The evolutionary game summarizes the attack and defense mechanism of secondary users in the cognitive IoT network with a game tree. Specifically, half of the branches of the game tree represent that when the channel is busy, the selfish secondary user node cannot sense the idle spectrum and will not launch an attack. At this time, the normal secondary user node does not know whether the spectrum in the channel is used by the primary user or the selfish secondary user node to launch a simulated attack. Therefore, the normal secondary user node will choose to execute or not execute the simulated attack detection. If it chooses to execute the detection, the probability of false detection of the simulated attack detection is p. eer , When a normal secondary user node misdetects the primary user in the channel as a selfish secondary user node, the normal secondary user node will enter the channel to use the spectrum and interfere with the communication of the primary user, causing the normal secondary user node to be punished, and the corresponding false alarm penalty parameter is u eer , thus reducing the enthusiasm of the normal secondary user node for detection; When the channel is idle and selfish secondary user nodes launch simulation attacks, normal secondary user nodes are afraid to perform detections due to fear of false alarms. To consider the impact of false alarm penalties in the scenario of the game between normal secondary user nodes and selfish secondary user nodes, a false alarm impact factor δ is set, and the false alarm penalty when the channel is busy is converted into the loss of performing detections in the face of simulation attacks by selfish secondary user nodes, λ = p p ·y·p eer ·u eer ·δ, to simulate the situation where the enthusiasm of normal secondary user nodes to perform detections decreases in the face of simulation attacks by selfish secondary user nodes; The other half of the branches of the game tree indicates that when the channel is idle, the selfish secondary user node can choose to initiate or not initiate a simulation attack, and the normal secondary user node can choose to execute or not execute a simulation attack detection. When the selfish secondary user node does not initiate a simulation attack, the normal secondary user node does not need to execute the detection. However, considering that user nodes in real situations are not completely rational and may make wrong decisions, after experiencing multiple simulation attacks by the selfish secondary user node, the normal secondary user node habitually executes the detection before using the spectrum; The payoff matrix of both sides of the game is obtained according to the game tree and system parameters, and is presented in a table as follows: The relevant parameters of the payoff matrix are as follows r s —— Revenue from monopolizing the idle channel: When a selfish secondary user node launches a simulation attack and successfully frightens away the normal secondary user nodes, the selfish secondary user node monopolizes the idle channel and obtains a revenue of r s ; When the selfish secondary user node does not attack or launches an attack but is detected by the normal secondary user nodes, the secondary user nodes share the idle spectrum together, and the selfish secondary user node and the normal secondary user nodes each obtain a revenue c a —— Attack cost: the energy loss of selfish secondary user nodes launching simulation attacks; c d —— Detection cost: the energy loss for performing selfish secondary user node simulation attack detection; u p —— Penalty parameter: the penalty that a selfish secondary user node receives when its simulation attack is detected, and the corresponding fine is obtained by a normal secondary user node; E - Energy harvesting parameter: The energy gain obtained through energy harvesting; H - Credit parameter: The selfish secondary user node obtains a reward when it does not initiate a simulation attack and does not cause the normal secondary user node to execute a detection; p p —— False alarm probability: the probability that a selfish secondary user node launches a simulation attack and a normal secondary user node performs detection but fails to detect the attack of the selfish secondary user node; p ma —— False alarm probability: the probability that a selfish secondary user node launches a simulation attack and a normal secondary user node performs detection but fails to detect the attack of the selfish secondary user node; p eer —— False alarm probability: When the channel is busy, the probability of misdetecting the primary user in the channel as an attack by a selfish secondary user node; x - Attack probability: The probability that the selfish secondary user node initiates a simulation attack; y - Detection probability: The probability that the normal secondary user node executes a simulation attack detection; u eer —— False alarm penalty parameter: When the channel is busy, after misdetecting the primary user in the channel as a selfish secondary user attack and then entering the channel to cause interference to the primary user, the penalty suffered δ - False alarm impact factor: The impact factor of the penalty for interfering with the primary user on the detection probability of the normal secondary user node; λ - Detection loss parameter: The loss of executing a simulation attack detection converted from the false alarm penalty.

2. The method for simulating the attack of the energy harvesting cognitive Internet of Things against the primary user in the evolutionary game according to claim 1, wherein: In the payoff matrix, the energy harvesting E is introduced to simulate a scenario where the selfish secondary user node has a higher attack enthusiasm. Specifically, when the communication node is equipped with the energy harvesting ability, the selfish secondary user node obtains energy through energy harvesting, first harvests energy and then initiates an attack in the spectrum sensing stage. The harvested energy can partially offset the energy loss of initiating a simulation attack, thereby increasing the enthusiasm of the selfish secondary user node to initiate a simulation attack. To simulate the scenario of energy harvesting - primary user simulation attack with a higher attack frequency, let the selfish secondary user node obtain an additional energy E when initiating a simulation attack; The credit parameter H is introduced into the payoff matrix. The specific method is: when the selfish secondary user node does not cause negative effects on the cognitive Internet of Things network, that is, when it does not initiate a simulation attack and does not cause the normal secondary user node to execute a detection, it obtains a reward to encourage the selfish secondary user node to evolve towards the strategy of not attacking; The credit parameter H ensures that when the normal secondary user node does not detect, under the incentive of the credit parameter, the benefit of the selfish secondary user node not attacking is greater than the benefit of attacking, that is At the same time, the credit parameter ensures that when the selfish secondary user node initiates a simulation attack, the benefit of the normal secondary user node executing a detection is greater than the benefit of not executing a detection, to ensure the enthusiasm of the normal secondary user node to detect the received signal, that is The expected benefit of the selfish secondary user node initiating a simulation attack is expressed by the formula: The expected benefit of the selfish secondary user node not initiating a simulation attack is: The average expected benefit of the entire population of selfish secondary user nodes is: The replicator dynamics equation of the selfish secondary user node is: The expected benefit of a normal secondary user node performing simulation attack detection is: The expected benefit of a normal secondary user node not performing simulation attack detection is: The average expected benefit of the entire population of normal secondary user nodes is: The replicator dynamics equation of a normal secondary user node is: By using formula 3 and formula 4 of the simultaneous equations, we construct a dynamic system equation group. Let formula 3 = 0 and formula 4 = 0, and we get five local equilibrium points: (0,0), (1,0), (0,1), (1,1) and (x * ,y * ),x * ∈[0,1],y * ∈[0,1], Among them The stability is judged by the Jacobian matrix, and the Jacobian matrix of the simulation attack evolutionary game model of selfish secondary user nodes is expressed as: When the local equilibrium point satisfies the conditions trJ < 0 and detJ > 0, this equilibrium point is an evolutionarily stable point; the calculation results of the four elements of the Jacobian matrix corresponding to the five local equilibrium points are presented in a table as It can be seen from Equation 1 and Equation 2 that: always holds For X1(0, 0), there is Therefore, X1(0, 0) is an evolutionarily stable point; For X2(1,0), there is Therefore, the point X2(1,0) is unstable; For X5(x * , y * ), there is trJ = 0, and the eigenvalues of J Therefore, X5(x * , y * ) is a saddle point; For X3(0, 1) and X4(1, 1), there are two scenarios; Scenario 1 That is At this time, for X3(0, 1), there is Therefore, X3(0, 1) is an unstable point; for X4(1, 1), there is Therefore, X4(1, 1) is an evolutionarily stable point; Scenario 2 That is At this time, for X3(0, 1), there is Therefore, X3(0, 1) is an unstable point; for X4(1, 1), there is Therefore, X4(1, 1) is an unstable point; In summary, when the inequality holds, both X1(0, 0) and X4(1, 1) are evolutionarily stable points; when it does not hold, only X1(0, 0) is an evolutionarily stable point; The throughput of a normal secondary user is: Since 0 < p ma < 1, that is, -1 < y(1 - p ma ) - 1 < 0. Then when the attack probability of the selfish secondary user is minimized (x = 0), the throughput of the normal secondary user reaches the maximum. By setting the penalty parameter, X1(0, 0) becomes the only evolutionarily stable point, thereby reducing the attack probability of the selfish secondary user and maximizing the throughput of the normal secondary user.

Images