A method for generating attack and defense strategies for underwater acoustic sensor networks based on dynamic psychological game theory

Through dynamic psychological game theory, the offensive and defense model of the water acoustic sensor network is constructed, and the transmission delay and safety are optimized, and the problem of lack of dynamic game theory models in the underwater environment in the existing technology is solved, and more effective confrontational strategy generation and Resource optimization.

CN118827243BActive Publication Date: 2024-11-19XIAMEN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411304780.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-19
Publication Date
2024-11-19
Estimated Expiration
2044-09-19

AI Technical Summary

Technical Problem

The existing technology lacks dynamic game theory-related offensive and defensive models in underwater environments, making it difficult to effectively evaluate and optimize the security of marine acoustic sensing networks, especially confrontation strategies when facing enemy attacks.

Method used

The water acoustic sensor network offensive and defense confrontation strategy generation method based on dynamic psychological game theory is adopted to construct a game model, consider the dynamic psychological effects of both the enemy and us, and optimize the transmission delay and security through psychological expectation utility functions and virtual regret minimization algorithm. Generate the optimal confrontation strategy.

Benefits of technology

It improves the transmission security of the water acoustic sensing network in the face of enemy interference, optimizes resource allocation, enhances the practicality and adaptability of the confrontation strategy, and provides a more comprehensive and practical offensive and defense model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118827243B_ABST
    Figure CN118827243B_ABST
Patent Text Reader

Abstract

A method for generating attack and defense strategies for underwater acoustic sensor networks based on dynamic psychological game theory involves underwater communication attack and defense strategies. First, determine the attack and defense scenarios and game objects of UASNs, the importance of the information to be transmitted, and the predetermined transmission route; secondly, construct a model based on dynamic psychological game theory #imgabs0#; UASNs need to minimize the total transmission delay and the risk of eavesdropping, and the enemy needs to extend the transmission delay of UASNs as much as possible under energy constraints. The two parties in the game define the psychological expected utility function according to the main goal, and use the virtual regret minimization algorithm to train the average strategy, and finally obtain the action strategy of the two parties in the predetermined route. By establishing a network attack and defense model based on dynamic psychological game theory, the optimal confrontation strategy in the attack and defense scenario of the underwater acoustic sensor network is obtained, which can improve the transmission security of the underwater acoustic sensor network.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to underwater safety communication, and in particular to a method for generating an attack and defense confrontation strategy for an underwater acoustic sensor network based on dynamic psychological game theory. Background Art

[0002] With the development of society, people pay more and more attention to the development and utilization of the ocean. Underwater acoustic communication technology is currently the most robust and reliable way to transmit underwater long-distance data. It is also an indispensable key technology for the construction of marine informatization such as submarine military bases and marine observation networks. Countries are paying more and more attention to the marine environment. In special marine environments such as marine observation networks, marine data, like traditional data, faces multiple security issues. If basic information such as marine geology, ocean currents, longitude and latitude are used by the enemy, it may pose a potential security threat to the country. Therefore, establishing an attack and defense strategy model in Underwater Acoustic Sensor Networks (UASNs) is of great significance to ensure the security of marine data.

[0003] Game theory is an effective method to build an attack and defense strategy model. Relevant research on network security based on game theory has been conducted at home and abroad. Zheng Hefeng et al. (Zheng Hefeng, Chen Sijun. Analysis of network attack and defense strategies based on incomplete information games [J]. Firepower and Command and Control, 2023, 48(08): 67-73+80.) built a network attack and defense model based on incomplete information and analyzed the selection of the optimal strategy for both sides. Existing literature proposes an underwater attack and defense model based on evolutionary game theory and designs an algorithm for selecting the optimal strategy. In UASNs, there are few literatures related to game theory, and most of the research is about static game problems. In real confrontation scenarios, dynamic game models are more meaningful for practical research.

[0004] Game theory and human behavior research show that human psychology and emotions have a great influence on behavioral decisions. Existing literature proposes a dynamic psychological game theory model, which quantifies the dynamic psychological effects of both parties in the game and adds them to the game theory model to develop a psychological game theory solution. A dynamic game theory attack and defense model established on the terrestrial IoT battlefield is proposed to analyze and enhance the security of the IoT system in the adversarial battlefield. For the development of marine informatization construction, the security of the underwater IoT also urgently needs to establish relevant attack and defense models for simulation research. Compared with information transmission on land, underwater equipment is basically located in harsh environments such as the sea surface, water, and seabed, which are severely affected by external factors. Therefore, the security confrontation scenarios of UASNs need to be targeted for research and development.

[0005] Dynamic psychological game theory takes into account the dynamic psychological effects of both sides, and can largely restore the action differences between the two sides due to emotional influence in real attack and defense confrontation scenarios. It can provide new ideas for the establishment of attack and defense models under UASNs. By constructing a dynamic psychological game model under UASNs scenarios, the security of UASNs can be further evaluated, and the optimal confrontation strategy can be provided for UASNs under attack. Summary of the invention

[0006] The purpose of the present invention is to provide a method for generating an attack and defense strategy for an underwater acoustic sensor network based on dynamic psychological game theory when UASNs are attacked; by establishing an attack and defense model based on dynamic psychological game theory, the security of UASNs is evaluated and the optimal strategy for confrontation is given. The objects in our UASNs designed by the present invention are transmission nodes, underwater robots (AUVs) equipped with intelligent metasurfaces (RIS), and surface base stations, and the enemy objects are eavesdropping nodes and enemy AUVs; in the game, UASNs need to transmit information based on the importance of the information to be transmitted and the security of the network to minimize the total transmission delay and the risk of being eavesdropped; the enemy needs to eavesdrop and transmit interference information in UASNs to interfere with the opponent to the greatest extent under the condition of energy limitation.

[0007] The present invention comprises the following steps:

[0008] 1) Determine a UASNs scenario, where our UASNs object contains: N transmission nodes, J , 1 surface base station, the enemy targets include: M eavesdropping nodes, I ; for Square underwater robot, is the enemy underwater robot; data is transmitted according to the predetermined data transmission route R, and it takes X hops to complete the data transmission. The number of hops in each communication transmission stage is , ; The information to be transmitted is transmitted by the source node. It is known that the information importance of the information to be transmitted is ; It is divided into five levels, namely ,The higher the level, the higher the importance of the information;

[0009] 2) Construct a model based on dynamic psychological game theory [P,H,Z,I,U,U'] , where P is the set of game users, which are UASNs and enemies in this game, H is the action sequence record taken by each game user before reaching a certain time stage, Z represents the action sequence record taken by the game user at each stage after reaching the end point, and I represents the information set at each decision point. and Represent the psychological expected utility functions of UASNs and the enemy respectively; Indicates that UASNs are The action set, It means that UASNs use node multi-hop to relay data. It means that the UASNs party uses AUV to assist in data relay transmission, and reflects the sound waves to the unaffected nodes through the RIS board; Indicates the enemy at each stage The action set, Indicates that the enemy will send out jamming signals. It means that the enemy does not send interference signals to the surroundings;

[0010] The psychological expected utility function of both parties is defined by the main goal that both parties need to achieve. The UASNs party needs to transmit information according to the importance of the information to be transmitted and the network security, with the optimization goal of minimizing the total transmission delay and the risk of being eavesdropped. The enemy needs to eavesdrop and transmit interference information in UASNs to interfere with the opponent to the greatest extent under the condition of energy limitation. In addition, the two parties will predict the next action of the opponent during the game, and set the predicted behavior of the opponent as the first-order belief. , Denote the first-order beliefs of the UASNs and the enemy, respectively; when the results of the opponent's actions differ from their expected actions, both parties will be frustrated / encouraged psychologically, and the expectation gap is defined as the second-order belief, , Represent the second-order beliefs of the UASNs and the enemy respectively; therefore, the psychological expected utility functions of the two parties in the game can be defined as:

[0011]

[0012]

[0013] in, and Represent the psychological expected utility functions of UASNs and the enemy, respectively. is the material reward of UASNs based on first-order belief, Material rewards for the enemy based on first-order beliefs; That is, UASNs adopt a strategy , predicting the enemy's action to be The probability of The enemy adopts a strategy , predicting UASNs actions as The probability of It represents the action sequence of each stage when reaching the end point; Z represents the action sequence record taken by the game user at each stage after reaching the end point; They represent the first-order belief of UASNs and the second-order belief of UASNs respectively; Respectively represent the enemy’s first-order belief and the enemy’s second-order belief; and Respectively represent the psychological frustration / encouragement effect of the UASNs and the enemy based on the second-order belief action on the other side, which is defined as the gap between the expected material reward and the actual material reward, and They represent the willingness of the UASNs party and the enemy party to frustrate the other party’s belief; is the material reward of UASNs based on first-order belief, is the material reward of the enemy based on the first-order belief, namely:

[0014]

[0015]

[0016] is the maximum delay threshold for UASNs data transmission; Indicate the sequence of actions at each stage to reach the end point; represents the first-order belief of UASNs; represents the enemy's first-order belief; It represents the total transmission delay of UASNs; That is, UASNs adopt a strategy , predicting the enemy's action to be The probability of The enemy adopts a strategy , predicting UASNs actions as The probability of Represents the weight of increasing the data transmission delay on the UASNs side, represents the weight of the enemy's energy consumption on its own node, and , as the enemy's energy level decreases, the enemy's weight on its own node energy consumption Increased accordingly; is the energy limit of the enemy node; The power of the jamming signal transmitted by the enemy node; Respectively represent the UASNs and the enemy in the stage Actions taken when That is, both parties take actions The data transmission delay of UASNs; is the number of hops per communication transmission. ; is the indicator function, when The function value is 1 when , otherwise it is 0; It is a comprehensive evaluation parameter of the node's information importance and network security. Network security is based on the probability of attack Make a judgment, among which, is the number of data transmission failures, L is the total number of data transmissions, and the comprehensive evaluation parameter of the node for information importance and network security The calculation formula is:

[0017]

[0018] and Represent the influencing factors of information importance and network security respectively; Indicates the importance of information; is the probability of attack; when the enemy attacks, if the UASNs do not choose to use the AUV-assisted action, the data transmission will fail and the probability of attack will increase, that is, If the enemy does not conduct interference attacks or the UASNs adopt AUV-assisted transmission actions, the probability of attack will be reduced. rise;

[0019] UASNs adopt strategies , the enemy adopts strategy The probability of , is as follows:

[0020]

[0021] in, Indicates that UASNs adopt a strategy , predict enemy actions (UASNs first-order belief is ), Indicates that the enemy adopts a strategy , predict UASNs actions (The enemy's first-order belief is ) probability; Indicates that UASNs are in the action sequence Next Strategy Take Action The probability of Indicates that UASNs are in the action sequence Next Strategy Take Action The probability of Indicates that the enemy is in action sequence Next Strategy Take Action The probability of Indicates that the enemy is in action sequence Next Strategy Take Action The probability of is the indicator function, when The function value is 1 when , otherwise it is 0; indicator function Indicates when The function value is 1 when , otherwise it is 0; indicator function Indicates when The function value is 1 when , otherwise it is 0; indicator function Indicates when The function value is 1 when , otherwise it is 0;

[0022] make and Represents the psychological frustration / encouragement effect of the UASNs and the enemy based on the second-order belief action on the other side, which is defined as the gap between the expected material return and the actual material return:

[0023] F s d 1 , d 2 , h ̂ X = [ π ¯ ' d 1 , d 2 -( i 1 ∑ x=1 X t a h x ,b h x T m + i 2 E m - ∑ x=1 X P int l b h x =b1 E m )] +

[0024] F a r 1 , r 2 , h ̂ X = [ π ¯ r 1 , r 2 -( T m -t( h ̂ X ) T m )] +

[0025] in, and are the material rewards of UASNs and enemies based on second-order beliefs, respectively; is the maximum delay threshold for UASNs data transmission; Indicate the sequence of actions at each stage to reach the end point; It represents the total transmission delay of UASNs; Represents the weight of increasing the data transmission delay on the UASNs side, Represents the weight of the enemy's energy consumption on its own node; is the number of hops per communication transmission. ; is the energy limit of the enemy node; The power of the jamming signal transmitted by the enemy node; Respectively represent the UASNs and the enemy in the stage Actions taken when That is, both parties take actions The data transmission delay of UASNs; is the indicator function, when The function value is 1 when , otherwise it is 0; It indicates that the enemy will send out jamming signals;

[0026] In the above game, the psychological expected utility function of both parties is , middle, and They represent the willingness of the UASNs and the enemy to frustrate the other party's belief. The enemy's frustration of the UASNs is related to the importance of the information transmitted by the UASNs. The higher the importance of the information, the higher the degree of attack on the UASNs when it is interfered and intercepted.

[0027] 3) Before transmitting information for the first time, UASNs conduct a comprehensive assessment of network security and information importance. The evaluation parameters are: , set the transmission mode selection threshold Take the first action The choice of node is the comprehensive evaluation parameter of information importance and network security. , the node selects node broadcast for transmission, ;otherwise , the node chooses to use the AUV carrying RIS to assist in transmission, ; It means that UASNs use node multi-hop to relay data. It means that UASNs use AUV to assist in data relay transmission;

[0028] At the same time, the enemy eavesdropping nodes all maintain a receiving state. Once they eavesdrop on the information transmitted by the UASNs node, they send the direction information of the relevant nodes to the enemy AUV. , choose to send malicious data to the eavesdropped node to interfere with transmission; set is the enemy interference energy threshold. , continuously transmit a signal of length T to interfere, , otherwise no interference signal is emitted, ; Indicates that the enemy will send out jamming signals. It means that the enemy does not send interference signals to the surroundings;

[0029] 4) Both UASNs and the enemy use the virtual regret minimization algorithm to solve the optimal action strategy of both parties under the predetermined route according to the model built in step 2); set the maximum number of iterations K;

[0030] 4.1) Both parties randomly select their first-order beliefs and second-order beliefs, and decide the actions to be taken in the current iteration based on their first-order beliefs;

[0031] 4.2) Based on the first-order beliefs, second-order beliefs and the actions taken by both parties, the psychological expected utility function of the UASNs and the enemy at the current number of iterations is obtained. , , and then get the virtual value of both parties in the game , ;

[0032] in, , They represent the virtual values ​​of UASNs and the enemy, respectively; represents the sequence of actions at each stage when reaching the end point, Indicates that UASNs adopt a strategy ; Indicates that the enemy adopts a strategy ; The first-order belief of UASNs is, represents the enemy's first-order belief; It represents the action sequence of each stage when reaching the end point; Z represents the action sequence record taken by the game user at each stage after reaching the end point; , They represent the psychological expected utility function values ​​of UASNs and the enemy at the current iteration k respectively;

[0033] 4.3) Calculate the action sequence at each stage when reaching the end point In each stage, the two parties in the game take UASNs Action Set , the enemy at each stage Action Set The virtual regret value obtained by the action in UASNs is , the enemy is ;in, , Represents the action sequence of each stage when reaching the end point The two players in the next game are in the action set , The virtual regret value of , Represent the virtual values ​​of UASNs and the enemy, respectively. represents the sequence of actions at each stage when reaching the end point, Indicates that UASNs are The action set, Indicates the enemy at each stage Action set; and represents the action; in the Kth iteration, the regret values ​​of the UASNs and the enemy taking actions are: , enemy ;in, , Indicates the action set taken by both parties in the Kth round of iteration , The regret value, represents the number of initialization trainings, k represents the number of iterations, I represents the information set at each decision point, and Z represents the action sequence record taken by the game user at each stage after reaching the end point;

[0034] When the regret value is negative, it is recorded as zero, that is, , ; Indicates the action set taken by UASNs at the k+1th iteration The regret value, Indicates the enemy's action set at the k+1th iteration The regret value of ; I represents the information set at each decision point;

[0035] 4.4) Update the strategies of both parties according to the regret value. The probability distribution of the game actions of the game users is proportional to the positive regret value. UASNs chooses the action at the k+1 iteration. Probability Defined as:

[0036]

[0037] in, Indicates that UASNs chooses an action at k+1 iterations The regret value, Indicates the action set taken by UASNs at the k+1th iteration The regret value, represents the number of initialization trainings, k represents the number of iterations, and I represents the information set at each decision point. Represents an action set The number of actions in

[0038] Similarly, the enemy chooses action at the k+1th iteration Probability Defined as:

[0039]

[0040] in, Indicates that the enemy chooses an action at the k+1th iteration The regret value, Indicates the enemy's action set at the k+1th iteration The regret value, represents the number of initialization trainings, k represents the number of iterations, and I represents the information set at each decision point. Represents an action set The number of actions in To express action; Indicates that UASNs are The action set, Indicates the enemy at each stage Action set;

[0041] 4.5) Repeat steps 4.1) to 4.4) until the maximum number of iterations K is reached, and the final attack and defense strategies of the two players are obtained;

[0042] 5) After each transmission, the evaluation parameters are updated according to the actions selected by both parties. , increase the weight of data transmission delay for UASNs , the enemy's weight on its own node energy consumption , update the historical action record and repeat steps 2) to 4) until the number of hops X is reached and the transmission step is completed.

[0043] The present invention considers the key issue of establishing an attack and defense confrontation model under the ocean attack and defense confrontation in the underwater acoustic sensor network, uses the basic model of psychological game theory to measure the psychological changes of the two parties in the game, so that the model is more in line with reality, and considers the importance of information and network security to obtain an attack and defense confrontation strategy model applicable to the underwater acoustic sensor network.

[0044] The present invention has the following outstanding advantages:

[0045] 1) Based on the communication characteristics of the underwater acoustic sensor network, the present invention constructs the action set of the two parties in the game, innovatively uses AUV and RIS to complete the anti-eavesdropping action of UASNs, and the enemy completes the signal interception action by broadcasting interference signals.

[0046] 2) Dynamic psychological game theory is used to establish a model in the attack and defense confrontation scenario of underwater acoustic sensor networks. The dynamic psychological effects of both sides are considered, and the psychological expected utility function is introduced to more comprehensively describe the behavior and decision-making process of the two sides in the confrontation. The action differences caused by the emotional influence of the two sides in the real attack and defense confrontation scenario are restored to a large extent, providing new ideas for the establishment of attack and defense models under UASNs.

[0047] 3) In the game analysis process, in addition to considering the main game goals of the two players, further consideration is given to the security performance evaluation of the underwater acoustic sensor network, the importance of transmitted information, and the impact of the importance of transmitted information on the psychological morale of the two players; giving priority to protecting key data will help optimize resource allocation and improve overall security.

[0048] 4) By using the virtual regret minimization algorithm for strategy training, the average optimal strategy of both parties in the game is effectively found, and it performs well in dealing with dynamic and uncertain environments; by establishing a network attack and defense model based on dynamic psychological game theory, the optimal confrontation strategy is generated, thereby improving the transmission security of the underwater acoustic sensor network when facing enemy interference.

[0049] 5) The method of the present invention can be combined with other security technologies and strategies to provide a powerful tool for improving the security of underwater communications; it can be applied to military communications, ocean monitoring, offshore energy protection, underwater communications and sensor networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 This is a scenario diagram of a model of a method for generating attack and defense strategies for an underwater acoustic sensor network based on dynamic psychological game theory.

[0051] Figure 2 The present invention is a flow chart of a model of a method for generating an attack and defense confrontation strategy for an underwater acoustic sensor network based on dynamic psychological game theory.

[0052] Figure 3 It is a schematic diagram of the average strategy iteration of the enemy in an embodiment of the method for generating attack and defense confrontation strategies for underwater acoustic sensor networks based on dynamic psychological game theory of the present invention.

[0053] Figure 4 It is a schematic diagram of average strategy iteration of UASNs in an embodiment of the method for generating attack and defense confrontation strategies of underwater acoustic sensor networks based on dynamic psychological game theory of the present invention. DETAILED DESCRIPTION

[0054] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in the following embodiments in conjunction with the accompanying drawings; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] like Figure 1 As shown in the scenario, this embodiment sets a The UASNs contains N=48 UASNs sensor nodes, 1 source node, 1 base station node, J=2 , where 48 sensor nodes are uniformly and randomly distributed, the source node is located at the coordinate (500, 0), the base station node is located at the coordinate (500, 1000), and the 2 The initial position is the coordinates (500, 300) and (500, 600) for standby. The information to be transmitted needs to be transmitted from the source node to the base station node. The importance of the information It is divided into five levels, namely , the higher the level, the more important the information; the enemy contains M=5 eavesdropping nodes, I=1 , eavesdropping nodes and are all randomly distributed; UASNs obtain the pre-set data transmission route R according to the DBR protocol. The routing results are as follows Figure 1 As shown, the simulation hop count .

[0056] like Figure 2 As shown, the embodiment of the method for generating an attack and defense confrontation strategy for an underwater acoustic sensor network based on dynamic psychological game theory of the present invention specifically includes the following steps:

[0057] 1) Model construction: Building a dynamic psychological game theory model [P,H,Z,I,U, U ' ] ; is the game user set, including all objects of UASNs and enemies, represents the action sequence record taken by each game user before reaching a certain time stage, Z represents the action sequence record taken by the game user at each stage after reaching the end point, and I represents the information set at each decision point. and Represent the psychological expected utility functions of UASNs and the enemy respectively; Indicates that UASNs are The action set, It means that UASNs use node multi-hop to relay data. It means that UASNs use AUV to assist in data relay transmission; Indicates the enemy at each stage The action set, Indicates that the enemy will send out jamming signals. It means that the enemy does not send interference signals to the surroundings; , are the first-order beliefs of UASNs and the enemy, respectively. , are the second-order beliefs of the UASNs and the enemy, respectively. After determining the objects of the two parties in the game, the psychological expected utility function of the two parties in the game is constructed according to the game goal:

[0058]

[0059]

[0060] in, and Represent the psychological expected utility functions of UASNs and the enemy, respectively. is the material reward of UASNs based on first-order belief, Material rewards for the enemy based on first-order beliefs; That is, UASNs adopt a strategy , predicting the enemy's action to be The probability of The enemy adopts a strategy , predicting UASNs actions as The probability of Indicate the sequence of actions at each stage to reach the end point; They represent the first-order belief of UASNs and the second-order belief of UASNs respectively; Respectively represent the enemy’s first-order belief and the enemy’s second-order belief; and Respectively represent the psychological frustration / encouragement effect of the UASNs and the enemy based on the second-order belief action on the other side, which is defined as the gap between the expected material reward and the actual material reward, and They represent the willingness of the UASNs party and the enemy party to frustrate the other party’s belief; is the material reward of UASNs based on first-order belief, The material rewards for the enemy based on first-order beliefs, specifically:

[0061]

[0062]

[0063] in, is the maximum delay threshold for UASNs data transmission; Indicate the sequence of actions at each stage to reach the end point; represents the total transmission delay of UASNs; and, , , are the influencing factors of information importance and network security, is the probability of attack occurrence, set when the first transmission is performed in the simulation ; Indicates the weight of increasing the data transmission delay for UASNs. represents the weight of the enemy's energy consumption on its own node, is the energy limit of the enemy node; The power of the jamming signal transmitted by the enemy node; Respectively represent the UASNs and the enemy in the stage actions to be taken when is the indicator function, when When , the function value is 1, otherwise it is 0; That is, the two parties in the game take actions separately The data transmission delay of UASNs; ,and The duration of transmitting jamming signals to the enemy; , is the transmission distance of the i-th hop, is the speed of sound underwater; ,and is the distance traveled by the AUV to the corresponding transmission location, is the AUV driving speed, is the transmission distance of the sound wave reflected by RIS during the i-th hop.

[0064] exist Indicates that UASNs adopt a strategy , the enemy adopts strategy The probability of That is, UASNs adopt a strategy , predict enemy actions (UASNs first-order belief is ), The enemy adopts a strategy , predict UASNs actions (The enemy's first-order belief is ) probability; and,

[0065]

[0066] in, Indicates that UASNs are in the action sequence Next Strategy Take Action The probability of Indicates that UASNs are in the action sequence Next Strategy Take Action The probability of Indicates that the enemy is in action sequence Next Strategy Take Action The probability of Indicates that the enemy is in action sequence Next Strategy Take Action The probability of is the indicator function, when The function value is 1 when , otherwise it is 0; is the indicator function, when The function value is 1 when , otherwise it is 0; is the indicator function, when The function value is 1 when , otherwise it is 0; is the indicator function, when The function value is 1 when , otherwise it is 0;

[0067] make and Representing the psychological frustration / encouragement effect of the UASNs and the enemy based on the second-order belief action on the other party, it can be defined as the gap between the expected material return and the actual material return, that is:

[0068] F s d 1 , d 2 , h ̂ X = [ π ¯ ' d 1 , d 2 -( i 1 ∑ x=1 X t a h x ,b h x T m + i 2 E m - ∑ x=1 X P int l b h x =b1 E m )] +

[0069] F a r 1 , r 2 , h ̂ X = [ π ¯ r 1 , r 2 -( T m -t( h ̂ X ) T m )] +

[0070] in, and are the material rewards of UASNs and enemies based on second-order beliefs, respectively; is the maximum delay threshold for UASNs data transmission; Indicate the sequence of actions at each stage to reach the end point; It represents the total transmission delay of UASNs; Represents the weight of increasing the data transmission delay on the UASNs side, Represents the weight of the enemy's energy consumption on its own node; is the energy limit of the enemy node; The power of the jamming signal transmitted by the enemy node; Respectively represent the UASNs and the enemy in the stage Actions taken when That is, both parties take actions The data transmission delay of UASNs; is the indicator function, when The function value is 1 when , otherwise it is 0.

[0071] The psychological expected utility function of both parties in the game , middle, and Respectively represent the willingness of the UASNs and the enemy to frustrate the other party’s belief, and set .

[0072] 2) According to the above game model, the virtual regret minimization algorithm is used to solve the strategy. The number of iterations for solving each hop strategy is set to K=5000. The specific solution steps are as follows:

[0073] 2.1) UASNs updates the network security assessment parameters, and the enemy updates the remaining energy parameters. Both parties predict each other's actions to form first-order beliefs, and both parties predict each other's first-order beliefs, obtain the expected gap, and form second-order beliefs; both parties in the attack and defense game randomly select their own first-order beliefs and second-order beliefs, and decide the actions taken in the current iteration k based on their first-order beliefs. When UASNs predicts that the enemy takes action , take action Otherwise, take action ; When the enemy predicts that UASNs will take action , take action Otherwise, take action .

[0074] 2.2) Based on the first-order beliefs, second-order beliefs and the actions taken by both parties, the psychological expected utility function of the UASNs and the enemy at the current number of iterations k is obtained. , , and then get the virtual value of both parties in the game , ;in, , They represent the virtual values ​​of UASNs and the enemy, respectively; It represents the action sequence of each stage when reaching the end point, and Z represents the action sequence record taken by the game user at each stage after reaching the end point; Indicates that UASNs adopt a strategy ; Indicates that the enemy adopts a strategy ; , They represent the psychological expected utility function values ​​of UASNs and the enemy at the current iteration k respectively.

[0075] 2.3) Calculate the action sequence at each stage when reaching the end point The virtual regret values ​​obtained by the two parties in the game taking actions in action sets A and B are: , the enemy is ; The regret values ​​of taking actions in the Kth iteration are: , enemy ; When the regret value is negative, it is recorded as zero, that is, , . Indicates the action set taken by UASNs at the k+1th iteration Regret value; Indicates the enemy's action set at the k+1th iteration Regret value;

[0076] 2.4) At each iteration, the strategies of both parties are updated according to the regret value. The probability distribution of the game actions of the game users is proportional to the positive regret value. UASNs chooses the action at the k+1 iteration. Probability Defined as:

[0077]

[0078] in, Indicates that UASNs chooses an action at k+1 iterations The regret value, Indicates the action set taken by UASNs at the k+1th iteration The regret value, represents the number of initialization trainings, k represents the number of iterations, and I represents the information set at each decision point. Indicates action, Represents an action set The number of actions in .

[0079] Similarly, the enemy chooses action at the k+1th iteration The probability of is defined as:

[0080]

[0081] in, Indicates that the enemy chooses an action at the k+1th iteration The regret value, Indicates the enemy's action set at the k+1th iteration The regret value, represents the number of initialization trainings, k represents the number of iterations, and I represents the information set at each decision point. Represents an action set The number of actions in .

[0082] Repeat the above steps 2.1) to 2.4) until the number of iterations reaches K, and the Nash equilibrium solution of the model is obtained. The two parties in the game get the best action strategy under the current number of hops; before each hop transmission, the two parties in the game repeat step 2) until the number of transmission hops reaches , when the transmission is completed, the optimal game strategy for both parties is obtained.

[0083] 3) The two parties in the game select actions based on the optimal game strategy obtained, and draw an average strategy graph. The horizontal axis of the average strategy graph can be the number of iterations, the vertical axis can be the average strategy action probability, and the data point is the average probability of the two parties in the game selecting different actions after each iteration. Based on the average strategy graph, the two parties in the game can select actions with higher average probability under different transmission hops to form the action strategy for this transmission.

[0084] The feasibility of the method of the present invention is verified by computer simulation below.

[0085] The simulation was performed using the MATLAB R2022b platform; the simulation parameters were set as follows: the underwater acoustic sensor network was set to The UASNs contains N = 48 UASNs sensor nodes, 1 source node, 1 base station node, and J = 2 , where 48 UASNs sensor nodes are uniformly and randomly distributed, the source node is located at the coordinate (500, 0), the base station node is located at the coordinate (500, 1000), and the 2 The initial position is the coordinates (500, 300) and (500, 600) for standby. The information to be transmitted needs to be transmitted from the source node to the base station node. The importance of the information It is divided into five levels, namely , the higher the level, the more important the information; the enemy contains M=5 eavesdropping nodes, I=1 , eavesdropping nodes and are all randomly distributed; UASNs obtain the pre-set data transmission routes according to the DBR protocol. The routing results are as follows Figure 1 As shown, the simulated hop count ; The number of iterations of the virtual regret minimization algorithm is set to pseudo K=5000.

[0086] Figure 1 This is the attack and defense scene diagram of this simulation experiment. Figure 1In the figure, two different data transmission modes can be clearly seen, where: the dotted line path is a predetermined route obtained according to the depth-based routing protocol (DBR), which is a schematic diagram of the multi-hop transmission mode, representing the traditional mode of data transmission through multiple intermediate nodes hop by hop. This mode relies on fixed node paths in the network. Although it has strong flexibility and adaptability, it may face greater security risks when facing enemy interference; the solid line path connecting an AUV and two nodes represents the transmission mode assisted by the present invention using AUV. The AUV not only acts as a mobile relay node, but also carries a reconfigurable intelligent surface (RIS) board, which can flexibly reflect and regulate acoustic signals, thereby changing the transmission path of the signal. Due to the mobility of the AUV and the regulation ability of the RIS, the transmission path is in a broken line shape. When facing enemy interference signals, the AUV can actively avoid the interference area to ensure the continuity and reliability of data transmission.

[0087] Figure 3 and Figure 4 Schematic diagrams of the average strategy iterations of the enemy and UASNs during five transmissions. It means that UASNs use node multi-hop to relay data. It means that UASNs use AUV to assist in data relay transmission; Indicates that the enemy will send out jamming signals. Indicates that the enemy does not send interference signals to the surroundings. The strategies chosen by the two parties in the final game are .

[0088] Depend on Figure 3 The probability distribution results of the five enemy strategies show that in the initial stage, due to the high energy level, the enemy is more willing to increase the delay of the UASNs, so it tends to adopt a more aggressive interference strategy, that is, to transmit interference signals in the first three transmissions ( ); As the number of jamming signal transmissions increases, the energy level of the enemy node continues to decrease. This energy consumption directly affects the enemy's subsequent strategy selection; although the enemy adopts jamming strategies in the first three transmissions, as the number of transmissions increases, the enemy's probability of jamming actions decreases due to insufficient energy, until the fourth transmission when the enemy node's own energy is insufficient to support continued jamming, resulting in no jamming actions ( ) The probability ratio increases. This simulation result clearly shows the dynamic adjustment of the enemy's strategy during the game, and the important influence of energy level on strategy selection. At the same time, it also verifies the effectiveness and accuracy of the method described in the present invention in modeling attack and defense scenarios of underwater acoustic sensor networks.

[0089] Depend on Figure 4The probability distribution results of the five UASNs transmission strategies show that the first two times adopted conservative AUV-assisted transmission. Although the same strategy was chosen twice, the probability of AUV-assisted transmission in the second time was higher. This is because in the initial stage, the attack probability was set. Considering safety, UASNs prefer to choose the more reliable AUV-assisted transmission. With the feedback of the results of the enemy and our actions after the first transmission, the probability of subsequent attacks is adjusted according to the results of the first actions selected by the enemy and ours. Since the enemy transmits interference signals in the first transmission, the probability of attack is set in the next four transmissions. , in the fourth transmission, although the UASNs still adopt AUV-assisted transmission, the probability of multi-hop transmission and AUV-assisted transmission is closer than the probability distribution of the first three actions. This is because the UASNs predicts that the enemy's energy level is low based on the first three times that the enemy takes interference actions, so the risk of taking multi-hop transmission is relatively reduced, and its action probability increases accordingly. In the fifth transmission, the UASNs' fifth action strategy is based on the enemy's non-interference action in the fourth transmission. It is believed that the probability of network attack is reduced, and the enemy's energy level is insufficient, so the final multi-hop transmission action probability is higher. The simulation results show that the method of the present invention is effective in modeling the attack and defense scenarios of underwater acoustic sensor networks. The simulation results fully prove the effectiveness of the method of the present invention in modeling the attack and defense scenarios of underwater acoustic sensor networks. The constructed model can dynamically adjust the transmission strategy according to the real-time changes in the battlefield situation, thereby optimizing the use efficiency of network resources while ensuring transmission security.

[0090] The above embodiments are only preferred embodiments of the present invention and cannot be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for generating attack and defense strategies for underwater acoustic sensor networks based on dynamic psychological game theory, characterized by The following steps are involved: 1) Determine the attack and defense scenarios in the underwater acoustic sensor network UASNs. The two parties in the game include UASNs and the enemy. The UASNs includes N transmission nodes and J Square underwater robot , 1 surface base station; the enemy includes: M eavesdropping nodes, I enemy underwater robot ; Determine the predetermined data transmission route R, including X hops, is the number of hops per communication transmission. ; Information importance of hydroacoustic data It is divided into five levels, namely ,The higher the level, the more important the information; 2) Construct an attack and defense model based on dynamic psychological game theory , where P is the set of game users, including UASNs and enemies, H is the action sequence record taken by each game user before reaching a certain time stage, Z represents the action sequence record taken by the game user at each stage after reaching the end point, and I represents the information set at each decision point. and Represent the psychological expected utility functions of UASNs and the enemy respectively; UASNs at each stage Action Set ,in, It means that UASNs use node multi-hop to relay data. It means that UASNs use AUV to assist in data relay transmission; the enemy uses AUV to assist in data relay transmission at each stage. Action Set ,in, Indicates that the enemy will send out jamming signals. It means that the enemy does not send interference signals to the surroundings; 3) The UASNs minimize the total transmission delay and the risk of being eavesdropped, and the enemy extends the transmission delay of UASNs as much as possible under the condition of energy limitation; the psychological expected utility function of the UASNs and the enemy is defined as follows: ; ; in, and Represent the psychological expected utility functions of UASNs and the enemy, respectively. is the material reward of UASNs based on first-order belief, Material rewards for the enemy based on first-order beliefs; That is, UASNs adopt a strategy , predicting the enemy's action to be The probability of The enemy adopts a strategy , predicting UASNs actions as The probability of is the information importance of hydroacoustic data; It represents the action sequence of each stage when reaching the end point; Z represents the action sequence record taken by the game user at each stage after reaching the end point; They represent the first-order belief of UASNs and the second-order belief of UASNs respectively; Respectively represent the enemy’s first-order belief and the enemy’s second-order belief; and Respectively represent the psychological frustration / encouragement effect of the UASNs and the enemy based on the second-order belief action on the other side, which is defined as the gap between the expected material reward and the actual material reward, and They represent the willingness of the UASNs and the enemy to frustrate the other party's belief. The enemy's frustration of the UASNs is related to the importance of the information transmitted by the UASNs. The higher the importance of the information, the higher the degree of attack on the UASNs when it is interfered and intercepted. 4) Both parties use the virtual regret minimization algorithm to train the average strategy and obtain the action strategies of both parties under the predetermined route. The specific method is as follows: Initialize the number of training times K. At the beginning of training, both parties randomly select their first-order beliefs and second-order beliefs, and decide the actions taken under the current iteration k based on their first-order beliefs, and obtain the virtual values ​​of UASNs and the enemy: ; ; in, , They represent the virtual values ​​of UASNs and the enemy, respectively; It represents the action sequence of each stage when reaching the end point, and Z represents the action sequence record taken by the game user at each stage after reaching the end point; Indicates that UASNs adopt a strategy ; Indicates that the enemy adopts a strategy ; , They represent the psychological expected utility function values ​​of UASNs and the enemy at the current iteration k respectively; Calculate the action sequence for each stage when reaching the end point The two parties in the game take UASNs at each stage. Action Set , the enemy at each stage Action Set The virtual regret values ​​obtained by the actions in , for UASNs and the enemy are: ; ; in, , Represents the action sequence of each stage when reaching the end point The two players in the next game are in the action set , The virtual regret value of , Represent the virtual values ​​of UASNs and the enemy, respectively. Indicate the sequence of actions at each stage to reach the end point; Indicates that UASNs are The action set, Indicates the enemy at each stage Action set; and To express action; In the Kth iteration, the regret values ​​of the UASNs and the enemy taking actions are: ; ; in, , They represent the action sets taken by both parties in the Kth round of iteration. , Regret value; represents the number of initialization trainings, k represents the number of iterations, I represents the information set at each decision point, and Z represents the action sequence record taken by the game user at each stage after reaching the end point; Indicates that UASNs are The action set, Indicates the enemy at each stage Action set; When the regret value is negative, it is recorded as zero, that is, , ; Indicates the action set taken by UASNs at the k+1th iteration The regret value, Indicates the enemy's action set at the k+1th iteration Regret value; At each iteration, the strategies of both parties are updated according to the regret value. The probability distribution of the game actions of the game users is proportional to the positive regret value. UASNs chooses the action at the k+1 iteration. Probability Defined as: ; in, Indicates that UASNs chooses an action at k+1 iterations The regret value, Indicates the action set taken by UASNs at the k+1th iteration The regret value, represents the number of initialization trainings, k represents the number of iterations, and I represents the information set at each decision point. Indicates action, Indicates that UASNs are The action set, Indicates the enemy at each stage Action set; Represents an action set The number of actions in Similarly, the enemy chooses action at the k+1th iteration Probability Defined as: ; in, Indicates that the enemy chooses an action at the k+1th iteration The regret value, Indicates the enemy's action set at the k+1th iteration The regret value, represents the number of initialization trainings, k represents the number of iterations, and I represents the information set at each decision point. Indicates action, Indicates that UASNs are The action set, Indicates the enemy at each stage Action set; Represents an action set The number of actions in Repeatedly select first-order beliefs, second-order beliefs and actions, and update the regret value until the number of iterations reaches K, and obtain the optimized game strategy for both parties.