Method and system for resisting dynamic SSDF attack based on credit mechanism

Abstract

The invention discloses a method and a system for resisting dynamic SSDF attack based on a credit mechanism. The method comprises the following steps of: distributing an initial review value for each user in a network; repeatedly executing the following steps until a preset termination condition is met; each user executes local spectrum sensing, and judges whether a certain frequency band is occupied or not by observing the surrounding environment; calculating a comprehensive review value by using the current review value, the historical review value, the gain factor and the suppression factor; distributing a fusion weight for each node according to the calculated comprehensive sound value; performing weighted fusion on the local spectrum sensing result of each node by using the distributed fusion weight to obtain a decision result; and comparing the global decision value with the local spectrum sensing value, and updating each review value. According to the historical and current sensing behaviors of the secondary users in the users, a Beta voucher system is used for distributing voucher values to the users, and the voucher values are combined into a comprehensive voucher value to measure the reliability of the secondary users in centralized cooperative spectrum sensing.

Description

Technical Field

[0001] The present invention belongs to the technical field of wireless communications, and in particular relates to a method and system for resisting dynamic SSDF attacks based on a credit mechanism. Background Art

[0002] In cognitive radio, although cooperative spectrum sensing (CSS) has advantages in sensing performance, it is vulnerable to attacks due to its low security performance, among which spectrum sensing data falsification (SSDF) attacks are the most common and common attacks. These malicious users (MUs) tamper with the received initial values ​​and send false messages to the fusion center, thereby degrading the fusion decision performance and making it easy for the system to make wrong judgments. To solve this problem, more reliable CSS methods are needed to minimize the impact of attackers on the system. In order to encourage correct sensing data sharing among secondary users (SUs), trust schemes are used to identify MUs in CSS. Most of the research based on trust schemes is based on the fact that MUs always report false sensing data. To avoid the detection of trust schemes, MUs can exhibit reasonable behavior by providing correct sensing data and sometimes partially hiding them by launching SSDF attacks in an intermittent manner. That is, they can maintain high credibility in the alternating process of reporting true or false sensing data. Many research works focus on attack detection based on detection probability, but rarely consider the penalty of attacks, ignoring how to implement effective penalty schemes for attackers. In addition, in dynamic SSDF attacks, MUs submit real and false perception data alternately, and general reputation mechanisms cannot effectively identify such attacks. MUs may always be in a credible state, and a well-established reputation update mechanism should be sensitive to changes in vehicle behavior and be able to punish false perception reports submitted by them.

[0003] Based on the above analysis and research, dynamic attacks in spectrum sensing is an urgent problem that needs to be solved. Summary of the invention

[0004] In order to solve the problems existing in the prior art, the present invention provides a dynamic SSDF attack security method based on a credit mechanism to improve this problem. In this method, the fusion center uses the Beta reputation system to assign corresponding reputation values ​​to secondary users based on their current and historical spectrum perception performance, and then evaluates the credibility of secondary users in centralized collaborative spectrum perception of the Internet of Vehicles through the comprehensive reputation value. At the same time, a penalty scheme for correcting reputation is proposed, in which gain factors and inhibition factors are introduced to encourage secondary users to conduct active and honest perception activities. The entire strategy can effectively alleviate the impact of malicious vehicle behavior on decision-making, ensure the correctness of global decision-making, and greatly improve the fairness and robustness of the system. Considering the attack mode of dynamic attacks, the behavior pattern of intermittent dynamic attacks is introduced to further verify the effectiveness of our proposed defense strategy.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is: a method for resisting dynamic SSDF attacks based on a credit mechanism, comprising the following steps: Assign an initial reputation value to each user in the network; Repeat the following steps until the preset termination condition is met; Each user performs local spectrum sensing and determines whether a frequency band is occupied by observing the surrounding environment; The comprehensive reputation value is calculated using the current reputation value, historical reputation value, gain factor and inhibition factor; According to the calculated comprehensive reputation value, a fusion weight is assigned to each node; Using the assigned fusion weights, the local spectrum sensing results of each node are weighted and fused to obtain the decision result D( t ); Each reputation value is updated based on the comparison between the global decision value and the local spectrum sensing value.

[0006] Furthermore, when initializing the reputation value, the initial reputation value is set according to the basic attributes of the user.

[0007] Furthermore, the weight of the user in the collaborative spectrum sensing process is dynamically adjusted according to the user's reputation value; for users whose reputation value is lower than a preset threshold, penalty measures such as reducing their weight or removing them from the network are taken.

[0008] Furthermore, each SU receives spectrum sensing data through the antenna and can independently sense and make binary decisions; each SU transmits data to the FC through the reporting channel to achieve data aggregation; the FC uses multi-data fusion technology to comprehensively analyze the collected data and make global decisions.

[0009] in, The signal transmitted by PU, Indicates from PU to The channel gain of each SU, For the The noise at each SU, Indicates that PU does not occupy the authorized spectrum. Indicates that PU is using the licensed spectrum. For the The signal received by each SU is Indicates The energy of the signal received by each SU, Indicates the energy threshold.

[0010] Furthermore, the tth sensing time slot The fusion weight value of each SU can be calculated as:

[0011] in, , is the trust threshold value.

[0012] Furthermore, the final decision is made according to the majority fusion rule as follows:

[0013] in, Indicates that the class is always occupied. Indicates the always idle class, Represents a always-error class.

[0014] Based on the concept of the method, the present invention provides a credit mechanism-based anti-dynamic SSDF attack system, including an initialization module, an iteration module, a perception module, a calculation module, a fusion module, and a reputation update module; The initialization module is used to assign an initial reputation value to each user in the network; The iteration module is used to repeatedly execute the following steps until the preset termination condition is met; The sensing module is used to obtain local spectrum sensing for each user and determine whether a frequency band is occupied by observing the surrounding environment; The calculation module calculates the comprehensive reputation value using the current reputation value, historical reputation value, gain factor and inhibition factor; The fusion module assigns a fusion weight to each node based on the calculated comprehensive reputation value. The assigned fusion weight is used to perform weighted fusion on the local spectrum sensing results of each node to obtain a decision result. The reputation update module is used to update each reputation value based on the comparison between the global decision value and the local spectrum sensing value.

[0015] The present invention can also provide a computer device, including a processor and a memory, the memory is used to store a computer executable program, the processor reads the computer executable program from the memory and executes it, and when the processor executes the computer executable program, the credit mechanism-based anti-dynamic SSDF attack method described in the present invention can be implemented.

[0016] At the same time, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program is executed by a processor, the credit mechanism-based anti-dynamic SSDF attack method described in the present invention can be implemented.

[0017] Compared with the prior art, the present invention has at least the following beneficial effects: In view of the problem of a large number of dynamic SSDF attacks in the system, the application scenario of this method is centralized collaborative spectrum sensing with hard fusion of local energy measurements. In this scheme, according to the historical and current perception behaviors of secondary users among the users, FC uses the Beta reputation system to assign reputation values ​​to them, and combines them into a comprehensive reputation value to measure the reliability of secondary users in centralized collaborative spectrum sensing. At the same time, an incentive mechanism strategy for correcting reputation is proposed, in which gain factors and inhibition factors are introduced, and positive feedback is given to honest secondary users to enhance the motivation and sustainability of submitting real decisions, and sanctions are imposed on malicious users to encourage them to adjust their malicious behaviors and avoid making mistakes again. The entire scheme can effectively alleviate the impact of malicious behaviors of secondary users on decisions, ensure the correctness of global decisions, and greatly improve the security and robustness of cognitive radio networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart for realizing the present invention; Figure 2 is the relationship between SNR and the minimum sample size required; Figure 3 is the relationship between the number of secondary users and the probability of correct perception; Figure 4 This is a graph showing the relationship between the proportion of malicious users and the probability of false perception; Figure 5 This is the relationship between the number of secondary users and the correct perception probability of the proposed method; Figure 6 This is the relationship between the number of secondary users and the probability of false perception of the proposed method; Figure 7 The present invention provides a method for combating malicious users in different scenarios based on the perceived performance simulation test results. DETAILED DESCRIPTION

[0019] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] refer to Figure 1 The present invention provides a method for resisting dynamic SSDF attacks based on a credit mechanism, comprising the following steps: S1. Before starting, assign an initial reputation value to each node in the network. This initial value can be a fixed value or set based on some basic attributes of the user to provide an initial value for the spectrum sensing and reputation value update process. The following methods can be used to initialize the reputation value: fixed value initialization, assigning the same fixed reputation value to all nodes, which is simple and easy, but may not be flexible enough in actual applications and does not take into account the differences between nodes; initialization based on node attributes, setting the initial reputation value based on the hardware configuration and geographical location of the node, etc., can better reflect the actual capabilities and value of the node; initialization based on historical behavior, if some nodes in the network have historical behavior records, the initial reputation value can be calculated based on these historical behavior records, which requires historical data support, but is more reasonable in actual applications; random initialization, randomly assigning an initial reputation value to each node, which is simple and has certain flexibility, but may require additional adjustments and optimizations in actual applications. Adaptive initialization, in the process of collaborative spectrum sensing, dynamically adjust the initial reputation value by observing the initial performance of the node. This method is more flexible, but requires a monitoring and feedback mechanism deployed in the network. In order to improve the robustness and security of the system, an initialization method based on node attributes or historical behaviors is adopted.

[0021] S2. Each user performs local spectrum sensing, that is, judging whether a certain frequency band is occupied by observing the surrounding environment. This can be done by setting the detection duration and sampling rate using the energy detection method; S3. Calculate the comprehensive reputation value using the current reputation value, historical reputation value, gain factor and inhibition factor: The user's comprehensive reputation is calculated based on the node's current reputation and historical reputation, combined with a gain factor (used to reward good behavior) and a suppression factor (used to punish bad behavior). A simple weighted average method or a machine learning algorithm can be used. S4. Calculate the fusion center to assign fusion weights: According to the calculated comprehensive reputation value, a fusion weight is assigned to each node, which is used to reflect the importance and credibility of the node in the collaborative spectrum sensing process.

[0022] S5. The fusion center makes a decision and obtains D(t): Using the assigned fusion weights, the local spectrum sensing results of each node are weighted and fused to obtain the final decision result D(t). This process can be completed in the fusion center (FC) or through a distributed algorithm.

[0023] S6. Update each reputation value based on the comparison between the global decision value and the local spectrum sensing value: The global decision value D(t) is compared with each user's local spectrum perception value. If a user's local perception result is consistent with the global decision, it is considered that the user has shown good behavior and its reputation value can be appropriately increased; if it is inconsistent, it is considered that the user may have bad behavior and its reputation value needs to be reduced. In this way, the reputation value of the node can be continuously updated and adjusted, thereby improving the reliability and security of the entire system.

[0024] A dynamic SSDF attack security method based on a credit mechanism provided in this embodiment includes the following steps: Step 1, initialization, that is, initializing the reputation value, assuming that each user is honest; Step 2: Get local spectrum sensing. Local spectrum sensing uses multi-user-based collaborative spectrum sensing to take advantage of spatial diversity. Each SU receives spectrum sensing data through the antenna and can independently sense and make binary decisions; each SU transmits data to the FC through the reporting channel to achieve data aggregation; the FC uses multi-data fusion technology to comprehensively analyze the collected data to make global decisions.

[0025]

[0026] in, The signal transmitted by PU, Indicates from PU to The channel gain of each SU, For the The noise at each SU, Indicates that PU does not occupy the authorized spectrum. Indicates that PU is using the licensed spectrum. For the The signal received by each SU is Indicates The energy of the signal received by each SU, Indicates the energy threshold.

[0027] Local spectrum sensing allows unlicensed users (secondary users or cognitive users) to opportunistically access idle frequency bands that have been allocated to licensed users (or primary users) but are rarely used or even unused. Local spectrum sensing mainly involves the physical layer and the link layer. The physical layer focuses on various specific local detection algorithms, while the link layer focuses on collaboration between users and optimization of local perception, collaborative perception, and perception mechanisms. Energy detection algorithms, matched filter detection methods, or cyclostationary feature detection methods can be used; Step 3: The reputation system can help identify and isolate malicious users, thereby reducing the impact of malicious users on the system. Integrate the reputation system into the collaborative spectrum sensing process. in, is the probability of the perceived behavior, , >0, > 0. The current reputation value, historical reputation value, gain factor and inhibition factor are further used to calculate the comprehensive reputation value.

[0028] In the tth sensing time slot The current reliability values ​​of SUs are as follows:

[0029] in, and Respectively represent The local perception results of each SU and the global decision result of FC.

[0030] In the tth sensing time slot Historical reputation of SU is evaluated as the following rules:

[0031] The gain factor is calculated as follows:

[0032] in, is the first round of perception The reputation of each SU, To perceive the timing of events honestly and continuously.

[0033] The inhibitory factor for suppressing malicious attack behavior is expressed as:

[0034] in, is the time when consecutive misperception events occur.

[0035] In summary, The comprehensive reputation value of a SU at time t can be expressed as: .

[0036] Step 4, the data fusion part implements the assignment of different fusion weights to the comprehensive reputation values ​​corresponding to the SUs. The SUs with higher reputations have a greater influence on the final decision, thus improving the perception accuracy of the CSS.

[0037] The tth sensing time slot The fusion weight value of each SU can be calculated as:

[0038] in, , is the trust threshold value.

[0039] Step 5: Global decision-making stage.

[0040] Assumptions SUs are eligible to participate in the collaboration, and the final decision is made mathematically according to the majority fusion rule. as follows:

[0041] in, Indicates always occupied class (AY), Indicates the always idle class (AN), Indicates always error class (AF).

[0042] Step 6, based on the comparison between the global decision value and the local spectrum sensing value, update each reputation value, and then return to step 2 for the next stage of sensing.

[0043] The effect of the present invention can be further illustrated by the following simulation example.

[0044] The present invention is simulated based on the above scheme and embodiment, and the simulation is performed in a cycle-based manner. All simulations are performed in the MATLAB R2019a environment, and the numerical results are obtained by running Monte Carlo simulations for more than 10,000 times; Figure 2The graph depicts the relationship between the minimum required sample size and SNR, with the horizontal axis representing SNR in decibels (dB) ranging from -10dB to 0dB. The vertical axis represents the minimum required sample size, which is the minimum number of observations required for spectrum sensing at a given false alarm probability. As can be seen from the figure, as the SNR increases, the minimum required sample size gradually decreases. This is because a higher SNR means that the signal strength is stronger relative to the noise level, so it is easier to distinguish between the signal and the noise, thereby reducing the number of samples required for correct perception. When the SNR is -10dB, the minimum required sample size is 724, and when the SNR is 0dB, the minimum required sample size is significantly reduced to 15. This shows that under better SNR conditions, fewer observations can be used to achieve the same perception performance. Figure 2 It provides an important benchmark for the subsequent simulation experiment design. Figure 2 The information in the simulation is used to determine the sample size setting under different SNR conditions to ensure the validity and accuracy of the simulation results.

[0045] In order to verify the perception performance of the proposed method, a comparative analysis is performed between it and the reputation method and the BSPRT method. Figure 3 The relationship between the correct perception probability of the three methods and the number of secondary users is given when the malicious users are set to 50%, and the number of secondary users increases from 10 to 60. As the number of secondary users increases, the perception performance of the three methods is improved. The correct perception probability of the reputation method is low when the initial number of secondary users is small. Under the same number of secondary users, the correct perception probability of the BSPRT method is higher than that of the reputation method, and the effect is better than that of the reputation method. The performance of the proposed method has been very stable throughout the process of the number of secondary users increasing from small to large. When the number of secondary users is greater than 35, the performance of the proposed method gradually approaches 1.

[0046] Figure 4 The relationship between the false perception probability of the three methods and the proportion of malicious users is given when the number of secondary users is set to 60. When the proportion of malicious users is small, the impact on the system is small, so the false perception probability of the three methods is close to 0. The false perception probability increases with the increase of the proportion of malicious users. When the proportion of malicious users exceeds 40%, the false perception probability of the three methods increases to varying degrees. Among them, the reputation method increases the fastest and has poor performance, followed by BSPRT. The proposed method shows good perception performance. When malicious users account for half of all secondary users, the false perception probability does not exceed 0.05.

[0047] Figure 5The relationship between the correct perception probability and the number of secondary users of the proposed method under different malicious user ratios is given. The malicious user ratio is set to 40%, 50%, and 60%, and the number of secondary users is set to 10 to 60. It can be seen that the smaller the malicious user ratio, the better the effect. When the malicious user ratio is 40%, the correct perception probability is close to 1, and the perception performance is stable. When the malicious user ratio is 50%, the overall performance deteriorates not very fast; when the malicious user ratio is 60%, the system can maintain good perception performance by increasing the number of secondary users. This is because in spectrum sensing, an increase in the number of secondary users usually means an increase in the number of entities that can participate in spectrum sensing and data analysis in the network, which helps to improve the overall performance of the network. When there are malicious users in the network, more secondary users mean more data points for analysis and comparison. This makes it easier for the network to identify and exclude malicious data, improving the robustness and anti-interference ability of the system.

[0048] Figure 6 is the relationship between the false perception probability of the proposed method and the number of secondary users, where the proportion of malicious users is set to 50%, 60%, and 70%, respectively, and the number of secondary users is set to 10 to 60. As the number of secondary users increases, the false perception probability shows a downward trend and tends to zero regardless of the proportion of malicious users. This shows that as the number of secondary users participating in the perception in the network increases, the perception performance of the system is significantly improved. This trend is due to the fact that more secondary users mean that more spectrum sensing data can be collected and fused, thereby improving the overall judgment accuracy of the PU state. Among them, the perception performance of malicious users = 50% is the best, malicious users = 60% is second, and malicious users = 70% is the worst. This reflects that as the proportion of malicious users increases, the challenges faced by the system are also increasing. However, even when the proportion of malicious users is as high as 70%, the false perception probability of the system can still be reduced as the number of secondary users increases, which proves the strong robustness and effectiveness of the proposed method. Large-scale malicious attacks have a very large impact on the system, but by increasing the number of secondary users, the method can obtain good perception effects.

[0049] In order to better evaluate the perceived performance of the proposed method in combating malicious users, different scenarios are set up to test the performance of the method. Figure 7As shown in the figure, in scenario 1, malicious users are set as static attackers, and the number of AN attackers and AF attackers is half each. In each perception, they report PU non-existence and false perception decisions respectively; in scenario 2, malicious users are set as static AY attackers and AN attackers are half each. In each perception, they report PU existence and non-existence decisions respectively; in scenario 3, malicious users are set as dynamic AF attackers, and they report true and false perception decisions alternately in each perception. As can be seen from the figure, the perception performance of scenario 3 is the best, which shows that this method is more applicable to scenarios against dynamic attacks. Because the historical perception results are taken into account, and the gain and suppression factors are introduced, once the malicious user performs intermittent false perception, its trust value will drop sharply under the action of the comprehensive trust value. When the proportion of malicious users is greater than 35%, scenario 2 performs better than scenario 1 because the damage intensity of AF attackers to the system is greater than that of AY and AN.

[0050] Based on the concept of the method, the present invention can also provide a credit mechanism-based anti-dynamic SSDF attack system, including an initialization module, an iteration module, a perception module, a calculation module, a fusion module, and a reputation update module; The initialization module is used to assign an initial reputation value to each user in the network; The iteration module is used to repeatedly execute the following steps until the preset termination condition is met; The sensing module is used to obtain local spectrum sensing for each user and determine whether a frequency band is occupied by observing the surrounding environment; The calculation module calculates the comprehensive reputation value using the current reputation value, historical reputation value, gain factor and inhibition factor; The fusion module assigns a fusion weight to each node based on the calculated comprehensive reputation value. The assigned fusion weight is used to perform weighted fusion on the local spectrum sensing results of each node to obtain a decision result. The reputation update module is used to update each reputation value based on the comparison between the global decision value and the local spectrum sensing value.

[0051] On the other hand, the present invention provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the credit mechanism-based anti-dynamic SSDF attack method described in the present invention can be implemented.

[0052] The present invention can also provide a computer device, including a processor and a memory, the memory is used to store a computer executable program, the processor reads the computer executable program from the memory and executes it, and when the processor executes the computer executable program, it can implement the credit mechanism-based anti-dynamic SSDF attack method described in the present invention.

[0053] The computer device may be a laptop computer, a desktop computer or a workstation.

[0054] The processor can be a central processing unit (CPU), a digital signal processor (DSP), an application specific integrated circuit (ASIC), or an off-the-shelf field programmable gate array (FPGA).

[0055] The memory of the present invention may be an internal storage unit of a laptop computer, a desktop computer or a workstation, such as a memory or a hard disk; or an external storage unit, such as a mobile hard disk or a flash memory card.

[0056] Computer-readable storage media may include computer storage media and communication media. Computer storage media include volatile and non-volatile, removable and non-removable media implemented by any method or technology for storing information such as computer-readable instructions, data structures, program modules or other data. Computer-readable storage media may include: read-only memory (ROM), random access memory (RAM), solid-state drive (SSD) or optical disk, etc. Among them, random access memory may include resistance random access memory (ReRAM) and dynamic random access memory (DRAM).

[0057] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A method for resisting dynamic SSDF attacks based on a credit mechanism, characterized in that: The following steps are involved: Assign an initial reputation value to each user in the network; Repeat the following steps until the preset termination condition is met; Each user performs local spectrum sensing and determines whether a frequency band is occupied by observing the surrounding environment; The comprehensive reputation value is calculated using the current reputation value, historical reputation value, gain factor and inhibition factor; According to the calculated comprehensive reputation value, a fusion weight is assigned to each node; Using the assigned fusion weights, the local spectrum sensing results of each node are weighted and fused to obtain the decision result D( t ); Each reputation value is updated based on the comparison between the global decision value and the local spectrum sensing value.

2. The anti-dynamic SSDF attack method based on the credit mechanism according to claim 1 is characterized in that: When initializing the reputation value, the initial reputation value is set according to the user's basic attributes.

3. The anti-dynamic SSDF attack method based on the credit mechanism according to claim 1 is characterized in that: The weight of the user in the collaborative spectrum sensing process is dynamically adjusted according to the user's reputation value; for users whose reputation value is lower than the preset threshold, penalty measures such as reducing their weight or removing them from the network are taken.

4. The anti-dynamic SSDF attack method based on the credit mechanism according to claim 1 is characterized in that: Each SU receives spectrum sensing data through the antenna and can independently sense and make binary decisions; each SU transmits data to the FC through the reporting channel to achieve data aggregation; the FC uses multi-data fusion technology to comprehensively analyze the collected data and make global decisions. in, The signal transmitted by PU, Indicates from PU to The channel gain of each SU, For the The noise at each SU, Indicates that PU does not occupy the authorized spectrum. Indicates that PU is using the licensed spectrum. For the The signal received by each SU is Indicates The energy of the signal received by each SU, Indicates the energy threshold.

5. The anti-dynamic SSDF attack method based on the credit mechanism according to claim 1 is characterized in that: The tth sensing time slot The fusion weight value of each SU can be calculated as: in, , is the trust threshold value.

6. The anti-dynamic SSDF attack method based on the credit mechanism according to claim 1 is characterized in that: Make final decision based on majority rule as follows: in, Indicates that the class is always occupied. Indicates the always idle class, Represents a always-error class.

7. A credit mechanism-based anti-dynamic SSDF attack system, characterized in that: It includes initialization module, iteration module, perception module, calculation module, fusion module and reputation update module; The initialization module is used to assign an initial reputation value to each user in the network; The iteration module is used to repeatedly execute the following steps until the preset termination condition is met; The sensing module is used to obtain local spectrum sensing for each user and determine whether a frequency band is occupied by observing the surrounding environment; The calculation module calculates the comprehensive reputation value using the current reputation value, historical reputation value, gain factor and inhibition factor; The fusion module assigns a fusion weight to each node based on the calculated comprehensive reputation value; Using the assigned fusion weights, the local spectrum sensing results of each node are weighted and fused to obtain the decision result; The reputation update module is used to update each reputation value based on the comparison between the global decision value and the local spectrum sensing value.

8. A computer device, characterized in that: It includes a processor and a memory, the memory is used to store computer executable programs, the processor reads part or all of the computer executable programs from the memory and executes them, and when the processor executes part or all of the computer executable programs, it can implement the credit mechanism-based anti-dynamic SSDF attack method described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that: A computer program is stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the credit mechanism-based anti-dynamic SSDF attack method described in any one of claims 1-6.

10. A communication system, characterized in that: Communication is performed using the anti-dynamic SSDF attack method described in any one of claims 1 to 6.

Images