A data link intelligent anti-interference decision method driven by intention situation and double drive

Abstract

A kind of data chain intelligent anti-interference decision method of intention situation double drive includes the following steps;Step one: the message requirement of task intention translation is received by each node of data chain, and the multi-dimensional resource situation information related to task node is called from situation database by the center node of subtask execution area;Step two: establish multi-objective optimization decision model, and set weight for multiple sub-objective functions according to intention requirement;Step three: based on intention requirement and multi-dimensional resource situation information, existing anti-interference strategy in strategy library is quickly matched, if it does not exist, then call intelligent algorithm to solve multi-objective optimization decision model as needed, and obtain each parameter decision result;Step four: each parameter decision result is verified, when strategy is feasible, anti-interference decision result is updated to strategy library, and is issued to each task node, and the anti-interference strategy is executed by task node.The present application improves transmission capacity in harsh environment, and intelligently resists complex unknown interference threat.

Description

Technical Field

[0001] This invention belongs to the field of data link network technology, specifically relating to an intention- and situation-driven intelligent anti-interference decision-making method for data links. Background Technology

[0002] Data link systems are critical information systems used to transmit high-value real-time intelligence information, playing a vital role in intelligence reconnaissance, electronic warfare, network intrusion, and attack systems. With the continuous improvement of informatization, the rapid growth of wireless communication services, the rapid development of electronic offensive capabilities, the increasing variety of jamming patterns, and the continuous improvement of intelligence levels, the electromagnetic environment faced by wireless communication is becoming increasingly complex and harsh. This makes existing anti-jamming technologies increasingly inadequate to meet the requirements of reliable and effective communication under jamming conditions. In future high-intensity confrontation environments, there is an urgent need to research and develop new anti-jamming technologies to adapt to the ever-growing and evolving electromagnetic threats, enabling data link systems to possess rapid response capabilities and strong anti-jamming capabilities, while maintaining high stability to ensure the reliability and effectiveness of information transmission.

[0003] Currently, communication anti-interference technology is developing towards integration, intelligence, networking, and software integration. Intelligent selection of communication parameters can greatly improve communication anti-interference capabilities. In recent years, the successful application of artificial intelligence algorithms in fields such as speech and visual recognition has solidified the position of artificial intelligence in future technological development. Research on artificial intelligence algorithms is emerging in various fields, and a research boom has also been launched in the field of communications. Intelligent anti-interference communication systems based on artificial intelligence algorithms have thus become a hot research topic.

[0004] One existing technology proposes to use non-cooperative frequency hopping (UFH) and power control methods in combination, so that the transmitter can make anti-interference decisions in both the frequency domain and the power domain, thereby effectively improving the shortcomings of UFH technology and enhancing the performance of the anti-interference system.

[0005] Existing technology 2 addresses the anti-interference problem in frequency-hopping asynchronous networking under complex electromagnetic environments by proposing a multi-agent fuzzy deep reinforcement learning algorithm based on a centralized training and distributed execution framework. The multi-agent deep reinforcement learning algorithm is used to make intelligent decisions on the parameters of each subnet, thereby achieving the goal of avoiding co-channel interference and countering hostile interference.

[0006] Existing technology three proposes an anti-interference decision-making method based on the binary artificial bee colony algorithm. By adding constraints and setting appropriate weights, the algorithm is made more closely aligned with actual communication needs. Compared with the basic genetic algorithm and particle swarm optimization algorithm, this algorithm has better convergence performance and can better meet the real-time performance requirements of anti-interference communication.

[0007] The prior art proposes an intelligent anti-interference system architecture based on online learning. It designs targeted adaptive anti-interference schemes according to different interference environments and jointly optimizes multiple parameters such as modulation, coding, and transmission power to ultimately achieve real-time, efficient, and reliable anti-interference communication.

[0008] Based on the above analysis, the problems and shortcomings of the existing technology are as follows:

[0009] Existing technologies are ill-suited to adapt to complex and ever-changing mission requirements and environmental conditions, and cannot guarantee the real-time performance and reliability of communication in future high-intensity combat environments.

[0010] Existing technologies often employ a single anti-interference method, with each technology operating independently and failing to be effectively combined to achieve a holistic anti-interference effect.

[0011] Existing anti-interference technologies either focus on adjusting only one or two parameters or use only one algorithm for intelligent decision-making, without taking interference into account, and are not systematic. Summary of the Invention

[0012] To overcome the shortcomings of existing technologies, the present invention aims to provide an intention- and situation-driven intelligent anti-interference decision-making method for data links, addressing the problems of lack of autonomy, low optimization level, slow response speed, and poor anti-interference effect in existing data link networks. Through a joint intention and situation-driven mechanism, the data link system can adaptively generate anti-interference strategies based on resource situation and task requirements, achieving autonomous, intelligent, and agile scheduling of multi-dimensional resources such as time, frequency, and space, improving transmission capabilities in harsh environments, and intelligently resisting complex and unknown interference threats.

[0013] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0014] A data link intelligent anti-interference decision-making method driven by both intent and situation includes the following steps;

[0015] Step 1: Each node in the data link receives the message request for translating the task intent, and the central node in the subtask execution area retrieves the multi-dimensional resource situation information related to the task node from the situation database;

[0016] Step 2: Establish a multi-objective optimization decision model and set weights for multiple sub-objective functions according to the intended requirements;

[0017] Step 3: Based on the stated intent and multi-dimensional resource situation information, use an inference algorithm to quickly match existing anti-interference strategies in the strategy library. If no strategy exists, call the intelligent algorithm as needed to solve the multi-objective optimization decision model and obtain the decision results for each parameter.

[0018] Step 4: Verify the decision results of each parameter. When the strategy is feasible, update the anti-interference decision results to the strategy library and distribute them to each task node so that the task node can execute the anti-interference strategy.

[0019] Furthermore, in step one, the message requirements for translating the task intent and the multi-dimensional resource situation information specifically include:

[0020] The intent layer extracts key entities such as task type, task domain, and task object from the task intent through named entity recognition, and then translates them into message requirements for nodes, bandwidth, latency, reliability, and transmission power.

[0021] The multidimensional resource situation information in the situational layer includes multiple state parameters of data link resources, including spectrum resources, time slot resources, spatial resources, link resources, power domain resources, and so on.

[0022] The decision center of the decision layer connects to the intent layer above and the situation layer below, adaptively generating the optimal anti-interference strategy, which includes coupled optimization decisions on multiple communication parameters.

[0023] Furthermore, step two includes:

[0024] The decision variables of the multi-objective optimization decision model are X={F,P,R,L,T}, which represent the frequency hopping parameters <frequency hopping point, frequency hopping rate, frequency hopping sequence>, transmit power <power level>, transmission rate <modulation method, coding method>, allocated time slot <time slot number>, and threshold <threshold level>, respectively.

[0025] The linear combination of the multiple sub-objective functions forms a multi-objective optimization decision model; the sub-objective function f min-ber f min-power f max-throughout f min-jamming The corresponding communication decision criteria are: minimizing bit error rate, minimizing power consumption, maximizing data throughput, and minimizing interference.

[0026] Y=ω1f min-ber +ω2f min-power +ω3f max-throughout +ω4f min-jamming

[0027] By setting ω i (1≤i≤m) assign certain weights to the sub-objective function. It reflects the degree of preference of the task requirements for the objective function.

[0028] Specifically, the weighting coefficients are determined by the message requirements for intended translation, and the corresponding relationships are as follows: when the reliability requirement is high: ω1 = ω4 > ω2 > ω3; when the latency requirement is high: ω3 > ω1 > ω4 > ω2; when the bandwidth requirement is high: ω3 > ω2 > ω1 = ω4; when the transmission power requirement is high: ω1 > ω4 > ω3 > ω2.

[0029] The constraints in the multi-objective optimization decision model stem from the limited availability and finiteness of data link resources:

[0030] (1) For frequency hopping parameters: Taking into account the spectral situation of each task node's location, a set of frequency points with relatively low interference is selected, i.e. in This indicates the interference level of the i-th frequency point;

[0031] (2) Regarding transmit power: First, the power level limit of the data link terminal must be met, i.e., P i ∈{P1,P2,P3,…,P 11 Secondly, the choice of transmission power will affect the communication coverage, so the distance between the source node and the destination node should be fully considered, that is:

[0032]

[0033] Where, state ij This represents the connection requirement between node i and node j, d ij D is the distance between two nodes. i This indicates the coverage area under the set power, and is also affected by parameters such as receiver sensitivity, radio wave frequency, dielectric loss, and channel bandwidth.

[0034] (3) Regarding transmission rate: First, the data link terminal's rate limit must be met, i.e., R i ∈{R1,R2,R3}, and secondly, to ensure normal communication, the selected transmission rate should be less than the current channel capacity, that is:

[0035] R i <C i =B i log(1+η)

[0036] Among them, channel capacity is closely related to the available bandwidth B and signal-to-interference-plus-noise ratio η in the link resources;

[0037] (4) For time slots: The time slot resource requirements of each task node are L x (x = 1, 2, ..., n), the time slot allocation needs to satisfy the following constraints:

[0038]

[0039] Where C represents the business volume within period T, I represents the business message capacity, and T represents the periodic report generated by the business.

[0040] (5) For thresholds: The thresholds are adjusted according to the current interference situation of the channel. At the same time, the current threshold level is set by statistically analyzing the proportion of different service levels of the current channel and the channel occupancy, so as to give priority to high-priority intentions.

[0041] Furthermore, step three includes:

[0042] The strategy library is used to store and update the optimal configuration schemes corresponding to different intentions and situations. The initial construction of the strategy library is based on existing interference patterns, including partial frequency band interference and broadband blocking interference. Anti-interference strategies under different interference environments are obtained by constructing simulation scenarios. When making decisions, the inference algorithm is first used to quickly search the strategy library for anti-interference strategies that can cope with the current network situation and meet the intention requirements. If no similar situation exists, the intelligent algorithm is further called to solve the decision model online.

[0043] Furthermore, the intelligent algorithm involves genetic algorithm, artificial bee colony algorithm, binary particle swarm algorithm and reinforcement learning algorithm. The system can flexibly switch decision-making algorithms according to different scenarios, thereby further improving decision-making performance.

[0044] When the message requirement is low latency, the binary particle swarm optimization algorithm with the fastest convergence speed is preferred to solve the multi-objective optimization decision model; when the reliability requirement is high, the reinforcement learning algorithm with strong environmental adaptability is preferred to solve the multi-objective optimization decision model.

[0045] The main parameters involved in the reinforcement learning algorithm are as follows:

[0046] (1) State: denoted as s t = < I t ,R t >, s∈S, where I represents the message requirement to be translated, R represents the set of multi-dimensional resource states in the data link system, and S is the state set of the algorithm input, including multi-domain resources in the frequency domain, time domain, spatial domain, power domain, and information domain;

[0047] (2) Actions: Composed of five types of variable parameters: frequency hopping parameters, transmit power, transmission rate, time slots, and thresholds. Frequency hopping parameters include frequency hopping point, frequency hopping rate, and frequency hopping sequence. Transmission rate depends on modulation and coding methods. Modulation methods include MSK, QPSK, 16QAM, and 64OAM. Coding methods include 1 / 2 convolutional code, 2 / 3 convolutional code, (7,3)RS code, and (15,11)RS code. Whether the action set includes time slots and thresholds depends on the data link node type. For Link-16 nodes, time slots need to be allocated. For TTNT nodes, thresholds need to be set. For gateway nodes, both time slots and thresholds need to be considered.

[0048] (3) Reward: The reward function comprehensively considers the bit error rate, information transmission rate, transmission power, and frequency interference. When the system's bit error rate (r) and transmission power do not meet the intended requirements, the reward is 0; when the information transmission rate (R) is greater than the selected channel capacity (C), a penalty is imposed; when the frequency hopping point is subject to a certain degree of interference, the penalty is based on the frequency interference level (f). state As a penalty; otherwise, the objective function value Y is used as a reward, and the reward function expression is as follows:

[0049]

[0050] Define the evaluation function value For node i in state s i Select action a below i The optimal action's discount reward accumulation value is selected in the next time slot; the Q value is updated as follows:

[0051]

[0052] Here, α is a quantity that weighs the results of the previous and current learning, and γ is a factor that considers the impact of future rewards on the present. During the learning process, the Q value is continuously updated through recursion to obtain the maximum long-term cumulative reward, thereby obtaining the optimal configuration strategy that can meet the intention requirements.

[0053] Furthermore, step four is specifically described as follows:

[0054] Verify whether the obtained anti-interference strategy meets the current intent requirements and is feasible under the current network situation. If it does not meet the requirements or is not feasible, correct the decision result by switching the intelligent algorithm until the verification is successful; otherwise, update the anti-interference decision result to the strategy library and distribute it to each task node so that the task node can execute the anti-interference strategy.

[0055] The aforementioned intelligent anti-interference decision-making method for data links driven by both intent and situation is used to achieve dynamic and flexible scheduling of multi-dimensional resources of data links in a highly adversarial environment, so as to resist unknown interference and improve the transmission performance of communication systems in complex environments.

[0056] The beneficial effects of this invention are:

[0057] This invention proposes an intelligent anti-interference decision-making method for data links driven by both intent and situation. It can adaptively generate anti-interference strategies based on the message requirements translated from task intent and the resource situation obtained in real time, thereby achieving a high degree of adaptation between the underlying data link resources and the upper-layer task requirements.

[0058] The method of this invention solves the decision model quickly by flexibly calling various intelligent algorithms, and optimizes multiple communication parameters of the data link network by coupling. It further realizes autonomous, intelligent and agile scheduling of multi-dimensional resources such as time, space and frequency of the data link, thereby enhancing the communication guarantee capability and anti-interference capability of the system in high-intensity confrontation environment, and providing the possibility for the intelligent defense of complex and unknown interference threats of future data link communication systems. Attached Figure Description

[0059] Figure 1 This is a flowchart of the data link intelligent anti-interference decision-making method driven by intent and situation provided in the embodiments of the present invention.

[0060] Figure 2 This is a diagram of the intent-situation dual-drive anti-interference decision architecture provided in the embodiments of the present invention.

[0061] Figure 3 This is a schematic diagram of a multi-objective optimization decision-making model for data link networks provided in an embodiment of the present invention.

[0062] Figure 4 This is a flowchart of the data link adaptive generation anti-interference strategy provided in an embodiment of the present invention.

[0063] Figure 5 This is a flowchart of the reinforcement learning algorithm provided in an embodiment of the present invention. Detailed Implementation

[0064] The present invention will now be described in further detail with reference to the accompanying drawings.

[0065] like Figure 1 As shown in the embodiment of the present invention, a data link intelligent anti-interference decision-making method driven by both intent and situation is proposed. The method includes the following steps:

[0066] S11, each node in the data link receives the message request for the translation of the task intent, and the central node of the subtask execution area retrieves the multi-dimensional resource situation information related to the task node from the situation database.

[0067] S12, Establish a multi-objective optimization decision model and set weights for multiple sub-objective functions according to the intended requirements.

[0068] S13. Based on the current intent requirements and resource status, the inference algorithm is used to quickly match existing anti-interference strategies in the strategy library. If no strategy exists, the intelligent algorithm is called as needed to solve the multi-objective optimization decision model and obtain the decision results for each parameter.

[0069] S14. Verify the decision result. When the strategy is feasible, update the anti-interference decision result to the strategy library and distribute it to each task node so that the task node can execute the anti-interference strategy.

[0070] The specific implementation of the present invention will be described in more detail below with reference to a specific embodiment.

[0071] Example 1:

[0072] This invention proposes a dual-drive intelligent anti-interference decision-making method for data links, which leverages task intent translation results and multi-dimensional resource situational awareness results. This method can adaptively generate the optimal anti-interference strategy based on task intent and situational information, thereby enhancing communication assurance capabilities and system anti-interference capabilities in high-intensity confrontation environments, and further improving the reliability, flexibility, autonomy, and intelligence of future data link communication systems.

[0073] like Figure 2 As shown in the figure, the intent-situation dual-drive anti-interference decision architecture provided in this embodiment of the invention is as follows:

[0074] The intent layer extracts key entities such as task type, task domain, and task object from the task intent through named entity recognition technology, and then translates them into message requirements for nodes, bandwidth, latency, reliability, transmission power, etc.

[0075] The multi-dimensional resource situation information in the situational awareness layer includes multiple state parameters of data link resources such as spectrum resources, time slot resources, spatial resources, link resources, and power domain resources. Among them, spectrum resources include frequency availability at different times and spaces; time slot resources include time slot occupancy status and time slot remaining; spatial resources include node location information and antenna configuration; link resources include signal-to-interference-plus-noise ratio and link connectivity; and power domain resources include selectable power levels and upper and lower power limits.

[0076] The decision-making center of the decision-making layer connects to the intent layer above and the situational awareness layer below, adaptively generating the optimal anti-interference strategy, which includes coupled optimization decisions for multiple communication parameters. This dual-driven anti-interference decision-making based on intent and situational awareness enhances the data link network's adaptability to complex and ever-changing task requirements and environmental situations, better ensuring the real-time performance and reliability of communication in future high-intensity confrontation environments.

[0077] like Figure 3 As shown, this is a schematic diagram of a multi-objective optimization decision model for data link networks provided in an embodiment of the present invention:

[0078] Furthermore, a multi-objective optimization decision model is established, wherein the decision variables of the multi-objective optimization decision model are X={F,P,R,L,T}, which represent frequency hopping parameters, transmit power, transmission rate, time slot allocation, and threshold, respectively.

[0079] Furthermore, the objective function of the multi-objective optimization decision model is a linear combination of multiple sub-objective functions, where the sub-objective function f min-ber f min-power f max-throughout f min-jamming The corresponding communication decision criteria are the minimum bit error rate criterion, the minimum power consumption criterion, the maximum data throughput criterion, and the minimum interference criterion, which respectively ensure the accuracy, security, real-time performance, and reliability of message transmission in the data link network, thereby improving transmission performance.

[0080] Y=ω1f min-ber +ω2f min-power +ω3f max-throughout +ω4f min-jamming

[0081] By setting ω i (1≤i≤m) assign certain weights to the sub-objective function. It reflects the degree of preference of the task requirements for the objective function.

[0082] Specifically, the weighting coefficients are determined by the message requirements for intended translation, with the following corresponding relationships: When reliability requirements are high, minimizing the bit error rate and minimizing interference criteria take precedence, followed by minimizing power consumption criteria, i.e.: ω1 = ω4 > ω2 > ω3; When latency requirements are high, a faster transmission rate is required, i.e., maximizing data throughput criteria takes precedence, followed by minimizing the bit error rate and minimizing interference criteria, while ensuring transmission reliability, i.e.: ω3 > ω1 > ω4 > ω2; When bandwidth requirements are high, maximizing data throughput criteria takes precedence, i.e.: ω3 > ω2 > ω1 = ω4; When transmission power requirements are high, minimizing power consumption criteria takes precedence, i.e.: ω2 > ω1 > ω4 > ω3. Furthermore, the constraints in the multi-objective optimization decision model stem from the finiteness and availability of data link resources:

[0083] (1) For frequency hopping parameters: Taking into account the spectral situation of each task node's location, a set of frequency points with relatively low interference is selected, i.e. in This indicates the interference level of the i-th frequency point;

[0084] (2) Regarding transmit power: First, the power level limit of the data link terminal must be met, i.e., P i ∈{P1,P2,P3,…,P 11Secondly, the choice of transmission power will affect the communication coverage, so the distance between the source node and the destination node should be fully considered, that is:

[0085]

[0086] (3) Regarding transmission rate: First, the data link terminal's rate limit must be met, i.e., R i ∈{R1,R2,R3}, and secondly, to ensure normal communication, the selected transmission rate should be less than the current channel capacity, that is:

[0087] R i <C i =B i log(1+η)

[0088] (4) For time slots: The time slot resource requirements of each task node are L x (x = 1, 2, ..., n), the time slot allocation needs to satisfy the following constraints:

[0089]

[0090] (5) For thresholds: The thresholds are adjusted according to the current interference situation of the channel. At the same time, the current threshold level is set by statistically analyzing the proportion of different service levels of the current channel and the channel occupancy, so as to give priority to high-priority intentions.

[0091] like Figure 4 As shown, the flowchart of the data link adaptive generation anti-interference strategy provided in this embodiment of the invention is as follows:

[0092] Furthermore, the strategy library is used to store and update the optimal configuration schemes corresponding to different intentions and situations. The initial construction of this library is based on existing interference patterns, including partial frequency band interference and broadband blocking interference. Partial frequency band interference is when a portion of a frequency in a certain frequency band or frequency band is affected by external signals or interference sources, while broadband blocking interference is when a signal in a relatively wide frequency band is interfered with. Anti-interference strategies under different interference environments are obtained by constructing simulation scenarios. When making decisions, the inference algorithm is first used to quickly match existing anti-interference strategies in the strategy library that can cope with the current network situation and meet the intention requirements. If no similar situation exists, the intelligent algorithm is further called to solve the decision model online.

[0093] The intelligent algorithms described involve genetic algorithms, artificial bee colony algorithms, binary particle swarm optimization algorithms, and reinforcement learning algorithms. The system can flexibly switch decision-making algorithms according to different scenarios, thereby further improving decision-making performance. For example, when the message requirement is low latency, the binary particle swarm optimization algorithm, which has the fastest convergence speed, is preferentially used to solve the multi-objective optimization decision model; when the reliability requirement is high, the reinforcement learning algorithm, which has strong environmental adaptability, is preferentially used to solve the multi-objective optimization decision model; when the task requirements are more complex, combining genetic algorithms and particle swarm optimization algorithms can better handle the solution of complex problems. Figure 5 As shown, the flowchart of the reinforcement learning algorithm provided in this embodiment of the invention is as follows:

[0094] The main parameters involved in the reinforcement learning algorithm are as follows:

[0095] (1) State: denoted as s t = < I t ,R t >, s∈S, where I represents the message requirement to be translated, and R represents the set of multi-dimensional resource states in the data link system, including multi-domain resources such as frequency domain, time domain, spatial domain, power domain, and information domain.

[0096] (2) Actions: Composed of five types of variable parameters: frequency hopping parameters, transmit power, transmission rate, time slots, and thresholds. Frequency hopping parameters include frequency hopping point, frequency hopping rate, and frequency hopping sequence. Transmission rate depends on modulation and coding methods. Modulation methods include MSK, QPSK, 16QAM, 64OAM, etc., and coding methods include 1 / 2 convolutional code, 2 / 3 convolutional code, (7,3)RS code, (15,11)RS code, etc. Whether the action set includes time slots and thresholds depends on the data link node type. For the Link-16 data link, Time Division Multiple Access (TDMA) is used to allocate time slots for communication, so time slots need to be allocated for nodes. The TTNT data link uses Single Frequency Single Carrier Multiple Access (SPMA) to allocate different frequency channels to avoid spectrum collisions, so thresholds need to be set for nodes. For gateway nodes, which are both TDMA and TTNT nodes, time slots and thresholds need to be considered.

[0097] (3) Reward: The reward function comprehensively considers the bit error rate, information transmission rate, transmission power, and frequency interference. When the system's bit error rate and transmission power do not meet the intended requirements, the reward is 0; when the information transmission rate is greater than the selected channel capacity, a penalty is imposed; when the frequency hopping point is subject to a certain degree of interference, the interference level is used as the penalty; otherwise, the objective function value Y is used as the reward. The reward function expression is as follows:

[0098]

[0099] Furthermore, define the evaluation function value. For node i in state s i Select action a below i The optimal action's accumulated discount reward value is then selected in the next time slot. The Q value is updated as follows:

[0100]

[0101] Here, α is a measure that weighs the results of the previous and current learning, and γ is a factor that considers the impact of future rewards on the present.

[0102] Furthermore, during the learning process, the Q-value is continuously updated recursively to obtain the maximum long-term cumulative reward, thereby obtaining the optimal configuration strategy that can meet the intent requirements.

[0103] Furthermore, the obtained anti-interference strategy is verified to see if it meets the current intent requirements and is feasible under the current network situation. If it does not meet the requirements or is not feasible, the decision result is corrected by switching the intelligent algorithm until the verification is successful. Otherwise, the anti-interference decision result is updated to the strategy library and distributed to each task node, which then executes the anti-interference strategy.

[0104] Application Examples:

[0105] Imagine a wireless communication system containing multiple communication nodes, such as aircraft, ships, and ground stations. These nodes communicate using data links to transmit critical intelligence and commands. Situational data is collected from sensors and monitoring equipment, and through further analysis and processing, multi-dimensional resource situational information of the data link is obtained. Each communication node receives task intents from the control center node, including the node's task, action plan, communication requirements, and sensitivity to interference. Based on the situational information and intent requirements, this invention, a dual-driven intention-situational intelligent anti-interference decision-making method for data links, is implemented. Task nodes can then obtain corresponding anti-interference communication strategies, enabling the data link network to maintain reliable communication in complex adversarial environments.

[0106] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be included within the scope of protection of the present invention.

Claims

1. A data link intelligent anti-interference decision-making method driven by both intent and situation, characterized in that, Includes the following steps; Step 1: Each node in the data link receives the message request for translating the task intent, and the central node in the subtask execution area retrieves the multi-dimensional resource situation information related to the task node from the situation database; Step 2: Establish a multi-objective optimization decision model and set weights for multiple sub-objective functions according to the intended requirements; Step 3: Based on the stated intent and multi-dimensional resource situation information, use an inference algorithm to quickly match existing anti-interference strategies in the strategy library. If no strategy exists, call the intelligent algorithm as needed to solve the multi-objective optimization decision model and obtain the decision results for each parameter. Step 4: Verify the decision results of each parameter. When the strategy is feasible, update the anti-interference decision results to the strategy library and distribute them to each task node so that the task node can execute the anti-interference strategy. In step one, the message requirements for translating the task intent and the multi-dimensional resource situation information specifically include: The intent layer extracts key entities such as task type, task domain, and task object from the task intent through named entity recognition, and then translates them into message requirements for nodes, bandwidth, latency, reliability, and transmission power. The multidimensional resource situation information in the situational layer includes multiple state parameters of data link resources, including spectrum resources, time slot resources, spatial resources, link resources, power domain resources, and so on. The decision center of the decision layer connects to the intent layer above and the situation layer below, and adaptively generates the optimal anti-interference strategy, which includes coupled optimization decisions on multiple communication parameters. Step two includes: The decision variables of the multi-objective optimization decision model are: These represent the frequency hopping parameters (frequency hopping point, frequency hopping rate, frequency hopping sequence), transmit power (power level), transmission rate (modulation method, coding method), allocated time slot (time slot number), and threshold (threshold level), respectively. The linear combination of the multiple sub-objective functions forms a multi-objective optimization decision model; sub-objective functions , , , The corresponding communication decision criteria are: minimizing bit error rate, minimizing power consumption, maximizing data throughput, and minimizing interference. By setting , Assign certain weights to the sub-objective function. This reflects the degree of preference of the task requirements for the objective function; The weighting coefficients are determined by the message requirements for intent translation, and the corresponding relationships are as follows: When reliability requirements are high: When latency requirements are high: When bandwidth requirements are high: When higher transmission power is required: .

2. The intention-situation dual-driven intelligent anti-interference decision-making method for data links according to claim 1, characterized in that, The constraints in the multi-objective optimization decision model stem from the limited availability and finiteness of data link resources: (1) For frequency hopping parameters: Taking into account the spectral situation of each task node, a set of frequency points with relatively less interference is selected, i.e. ,in Indicates the first The interference level of each frequency point; (2) Regarding transmission power: First, the power level limit of the data link terminal must be met, i.e. Secondly, consider the distance between the source node and the destination node, that is: in, Represents a node With nodes The connection requirements The distance between the two nodes. This indicates the coverage area under the set power, and is also affected by receiver sensitivity, radio wave frequency, dielectric loss, and channel bandwidth parameters. (3) Regarding transmission rate: First, the data link terminal's rate limit must be met, i.e. Secondly, the transmission rate should be less than the current channel capacity, that is: Among them, channel capacity and available bandwidth in link resources and signal-to-interference-plus-noise ratio Closely related; (4) For time slots: The time slot resource requirements of each task node are L X， The time slot allocation needs to meet the following constraints: in, Indicates period Internal business volume Indicates the capacity of business messages. This indicates periodic reports generated by business operations; (5) For the threshold: The threshold is adjusted according to the current interference situation of the channel. At the same time, the current threshold level is set by statistically analyzing the proportion of different service levels of the current channel and the channel occupancy, so as to give priority to high-priority intentions.

3. The intention-situation dual-driven intelligent anti-interference decision-making method for data links according to claim 1, characterized in that, Step three includes: The strategy library is used to store and update the optimal configuration schemes corresponding to different intentions and situations. The initial construction of the strategy library is based on existing interference patterns, including partial frequency band interference and broadband blocking interference. Anti-interference strategies under different interference environments are obtained by constructing simulation scenarios. When making decisions, the inference algorithm is first used to quickly search the strategy library for anti-interference strategies that can cope with the current network situation and meet the intention requirements. If no similar situation exists, the intelligent algorithm is further called to solve the decision model online. The intelligent algorithms mentioned include genetic algorithms, artificial bee colony algorithms, binary particle swarm algorithms, and reinforcement learning algorithms; When the message demand is low latency, the binary particle swarm optimization algorithm with the fastest convergence speed is preferred to solve the multi-objective optimization decision model; when the reliability requirement is high, the reinforcement learning algorithm with strong environmental adaptability is used to solve the multi-objective optimization decision model.

4. The intention-situation dual-driven intelligent anti-interference decision-making method for data links according to claim 3, characterized in that, The parameters involved in the reinforcement learning algorithm are as follows: (1) State: denoted as , ,in This indicates the need to translate the intended message. This represents a set of multi-dimensional resource statuses in a data link system. The state set input to the algorithm includes multi-domain resources such as frequency domain, time domain, spatial domain, power domain, and information domain; (2) Actions: consist of five types of variable parameters: frequency hopping parameters, transmit power, transmission rate, time slots, and thresholds. Frequency hopping parameters include frequency hopping points, frequency hopping rates, and frequency hopping sequences. Transmission rates depend on the modulation and coding methods. Modulation methods include MSK, QPSK, 16QAM, and 64OAM. Coding methods include 1 / 2 convolutional codes, 2 / 3 convolutional codes, (7,3) RS codes, and (15,11) RS codes. Whether the action set includes time slots and thresholds depends on the data link node type. For Link-16 nodes, time slots need to be allocated. For TTNT nodes, thresholds need to be set. For gateway nodes, both time slots and thresholds need to be considered. (3) Reward: The reward function comprehensively considers the bit error rate, information transmission rate, transmission power, and frequency interference. When the system's bit error rate is... When the transmission power does not meet the intended requirements, the reward is 0; when the information transmission rate... Greater than the selected channel capacity When frequency hopping is subjected to interference to a certain degree, a penalty will be imposed; when the frequency hopping point is subjected to interference to a certain degree, the penalty will be based on the level of interference. As a penalty; otherwise, the objective function value... As a reward, the reward function expression is as follows: Define the evaluation function value For nodes In state Select action below And in the next time slot, select the accumulated discount reward value for the optimal action; The values ​​are updated as follows: in, It is a quantitative measure of the results of the previous and current learning sessions. It is a factor that considers the impact of future rewards on the present; during the learning process, it is continuously updated through recursion. The value is determined to obtain the maximum long-term cumulative reward, thereby obtaining the optimal configuration strategy that can meet the intention requirements.

5. The intention-situation dual-driven intelligent anti-interference decision-making method for data links according to claim 1, characterized in that, Step four is specifically described as follows: Verify whether the obtained anti-interference strategy meets the current intent requirements and is feasible under the current network situation. If it does not meet the requirements or is not feasible, correct the decision result by switching the intelligent algorithm until the verification is successful; otherwise, update the anti-interference decision result to the strategy library and distribute it to each task node so that the task node can execute the anti-interference strategy.

6. The application of the intention-situation dual-driven data link intelligent anti-interference decision-making method according to any one of claims 1-5, characterized in that, The intention-situation dual-driven intelligent anti-interference decision-making method for data links is used to realize the dynamic and flexible scheduling of multi-dimensional resources of data links in a highly adversarial environment, so as to resist unknown interference and improve the transmission performance of communication systems in complex environments.

Images