Weapon mounting simulation platform based on ai adaptive coordination

Abstract

The application discloses an AI adaptive coordination-based weapon mounting simulation platform, relates to the field of weapon mounting simulation, and is used for solving the coordination problem of state switching of multiple types of weapons. The weapon mounting simulation platform is used for realizing linkage logic abstraction among states of multiple weapons by constructing a state interaction matrix, accurately combing interference relations in a layered competition mode, and making scheduling basis no longer rely on fixed rules and have adaptive capacity. The scheduling token guides the parallel operation of the state machine, and the switching rhythm of multiple states is unified by means of a dynamic threshold mechanism, so that the problem of inconsistent response in the synchronous control process of multiple weapons is effectively solved. The switch jitter factor and the temperature difference offset span with scene pertinence are introduced, the synergistic risk of the current state combination is quantitatively evaluated by means of a machine learning model, the adaptive rearrangement is driven in real time, and accurate coordination and stable closed circulation of state switching in a high-density parallel mounting scene are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of weapon mounting simulation, more particularly, the present application relates to a weapon mounting simulation platform based on AI adaptive coordination. BACKGROUND

[0002] In the current weapon external mounting simulation platform, when simulating the scene of simultaneously mounting multiple types of weapons, there is often a problem that the state switching of different weapons cannot be accurately coordinated during simulation (Source: A weapon external mounting management simulation method and system based on state machine, application number: CN119475549A). Specifically, due to the interlacing and superposition of state changes such as cooling requirements, target switching modes, attack methods and warming processes of different weapon types, this complex state interaction makes the simulation system prone to state conflicts or coordination failures when controlling in parallel. The reason for this problem is that the simulation system did not finely divide the state machines of the mutual influence of different weapon states at the beginning of design, the linkage logic definition between states is insufficient or ignores the boundary conditions of mutual interference, thereby causing the system to be unable to effectively handle the state alternation of multiple weapons in complex scenarios. Once such state coordination failure occurs, it will cause part of the weapon states to be unable to update in real time during the simulation process, and in severe cases, it may even cause weapon state misreporting or insufficient weapon preparation, which will directly affect the reliability of the simulation result and cannot truly reflect the complex situation of actual combat, thereby reducing the simulation accuracy of the platform to the actual combat process.

[0003] In order to solve the above problems, a technical solution is provided. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a weapon mounting simulation platform based on AI adaptive coordination, which realizes the linkage logic abstraction between multiple weapon states by constructing a state interaction matrix, and precisely sorts out the interference relationship in a hierarchical competition mode, so that the scheduling is no longer dependent on fixed rules and has adaptive ability; the scheduling token guides the parallel operation of the state machine, and unifies the switching rhythm of multiple states by means of a dynamic threshold mechanism, effectively solving the problem of inconsistent response in the synchronous control process of multiple weapons; two parameters of a scene-specific switch jitter factor and a temperature difference offset span are introduced, and the machine learning model is used to quantitatively evaluate the coordination risk of the current state combination, so that the adjustment of priority no longer depends on artificial limiting or simple threshold judgment, but has the ability of dynamic adaptive rearrangement, to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0006] State storage unit: Collects initial state vectors of all weapons in the start-up phase, derives linkage relationship based on type mapping and generates state interaction matrix, which is written into state storage together with original state data;

[0007] Priority sorting unit: Reads interaction matrix from state storage to perform hierarchical competition analysis, sorts potential interference state pairs detected by interference strength to form priority sequence, and outputs corresponding scheduling token queue;

[0008] Synchronization switching unit: Activates state machines of each weapon in parallel according to scheduling token queue, and injects dynamic threshold value into forward switching threshold pool, so that the latest state data is synchronized and switched at a unified time scale and fed back to state storage in real time;

[0009] Risk rearrangement unit: Continuously tracks state machine trajectory and evaluates synergy risk through machine learning, and triggers priority rearrangement and writes back to state storage to drive state machine rescheduling if risk exceeds limit or conflict count reaches limit;

[0010] Snapshot output unit: When state machine trajectory is stable and conflict count is zero, solidify the current state to generate a mounted snapshot, and push the snapshot to the scheduling end to start the next round of simulation.

[0011] In a preferred embodiment, the state storage unit includes the following:

[0012] For all weapons mounted on the same carrier, collect the cooling state, target switch mode, attack mode and warming state of each weapon, and organize these state data into a state matrix. The number of rows of the state matrix corresponds to the number of state characteristics, and the number of columns corresponds to the number of weapons.

[0013] In a preferred embodiment, the state storage unit further includes the following:

[0014] Based on the state matrix, combined with the type characteristics of each weapon and the predefined linkage rules, a linkage matrix is generated by calculating the absolute value of the Pearson correlation coefficient of each pair of weapon state vectors, to quantify the state linkage strength between weapons; The linkage matrix is directly used as the state interaction matrix, and the state matrix and the state interaction matrix are stored in the state storage.

[0015] In a preferred embodiment, the priority sorting unit includes the following:

[0016] The state interaction matrix is read, hierarchical competition analysis is performed on all non-diagonal elements of the state interaction matrix to identify potential interference state pairs, each potential interference state pair consisting of non-zero interaction strength of two different weapons; the identified potential interference state pairs are sorted in descending order of state interaction strength to form a priority sequence; and a scheduling token queue is generated based on the priority sequence, each scheduling token corresponding to a potential interference state pair, containing identifiers of the two weapons and their state interaction strength, and arranged in order of the priority sequence.

[0017] In a preferred embodiment, the synchronization switching unit further comprises the following:

[0018] Based on the scheduling token queue arranged in descending order of state interaction strength between weapon pairs, the weapon state machines are activated in parallel to preferentially process weapon pairs with high state interaction strength; the dynamic threshold value is injected into the forward switching threshold pool; state synchronization switching is realized through a unified global clock, state switching of a certain weapon is triggered when its state change demand value reaches or exceeds its dynamic threshold value, and state switching of other weapons whose state change demand values are close to their dynamic threshold values within a preset tolerance range is also triggered to maintain synchronization; after state switching is completed, the state bin is updated in real time to record new state data of the weapons, and the state interaction matrix is updated to reflect changes in state interaction relationships between weapon pairs.

[0019] In a preferred embodiment, the synchronization switching unit further comprises the following:

[0020] The dynamic threshold value is calculated based on the average state interaction strength of each weapon with other weapons.

[0021] In a preferred embodiment, the risk rearrangement unit comprises the following:

[0022] The state machine trajectory is continuously monitored to record changes in weapon states over time, the switch jitter factor is calculated to assess the stability of state switching and the consistency of command response, the temperature deviation span is calculated to quantify the accuracy of temperature control, and the switch jitter factor and temperature deviation span are analyzed using a random forest model to generate a synergistic risk coefficient for assessing the risk of multi-weapon state combinations. When the synergistic risk coefficient exceeds a preset risk threshold or the conflict count reaches a preset upper limit, a priority reordering mechanism is triggered to move high-risk or high-conflict weapon pairs to the front of the scheduling token queue, and the updated scheduling token queue is written to the state bin and used to drive the state machine to reschedule according to the new queue order.

[0023] In a preferred embodiment, the risk rearrangement unit further comprises the following:

[0024] The number of state switching in a specific time period is counted, compared with an ideal target switching interval, and the absolute value of the difference between the actual switching number and the target switching interval is normalized to obtain a switch jitter factor;

[0025] The actual temperature curve and the target temperature curve are obtained within the same time window, the difference between the two at each time point is calculated and taken as an absolute value, then accumulated, and finally normalized by dividing by the temperature control target amplitude to obtain a temperature deviation span.

[0026] In a preferred embodiment, the snapshot output unit includes the following:

[0027] The stability of the state is judged by calculating the smoothness index of the state machine trajectory, and the conflict count is monitored to ensure that the number of weapon pairs that do not meet the dynamic threshold requirement during state synchronization is zero. When the smoothness index of the state machine trajectory is less than the preset threshold and the conflict count is zero, the current state data is captured and solidified as a carrying snapshot, and then the carrying snapshot is transmitted to the scheduling end. The scheduling end analyzes the data in the carrying snapshot and loads it into the state warehouse as the initial state of the next round of simulation task to start a new round of simulation task.

[0028] In a preferred embodiment, the snapshot output unit further includes the following:

[0029] The smoothness index quantifies the fluctuation of the state machine trajectory by evaluating the average change amplitude of the state value within a preset time window.

[0030] The technical effects and advantages of the AI adaptive coordinated weapon mounting simulation platform of the present application are:

[0031] The present application realizes the linkage logic abstraction between multiple weapon states by constructing a state interaction matrix, accurately sorts the interference relationship in a hierarchical competition mode, so that the scheduling no longer depends on fixed rules and has adaptive ability; the scheduling token guides the parallel operation of the state machine, and unifies the switching rhythm of multiple types of states with the help of a dynamic threshold mechanism, effectively solving the inconsistent response problem in the synchronous control process of multiple weapons; the switch jitter factor and temperature difference offset span with scene specificity are introduced, and the synergistic risk of the current state combination is quantitatively evaluated by means of a machine learning model, so that the adjustment of priority no longer depends on artificial limits or simple threshold judgment, but has the ability of dynamic adaptive rearrangement; the above processing logic runs through the whole simulation cycle, and the triggering, monitoring, judgment and correction of state switching are integrated into an integrated flow process, which not only improves the coordination accuracy of multi-weapon parallel state simulation, but also enhances the discrimination ability of the system in dealing with complex dynamic state combinations, so that the whole platform maintains a coherent and orderly response mechanism in the process of handling high-density state interaction scenarios, thereby significantly improving the stability and realism of the simulation. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 Structure diagram of the weapon mounting simulation platform based on AI adaptive coordination of the application.

[0033] Figure 2 Flow diagram of the risk rearrangement unit of the weapon mounting simulation platform based on AI adaptive coordination of the application. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the application.

[0035] Embodiment 1: Figure 1 The weapon mounting simulation platform based on AI adaptive coordination of the application is given, which comprises:

[0036] The state warehousing unit: the initial state vector of all weapons is summarized in the starting stage, the linkage relationship is derived according to the type mapping, the state interaction matrix is generated, and the original state data is written into the state warehouse together;

[0037] The priority sorting unit: the interaction matrix is read from the state warehouse to execute hierarchical competition analysis, the potential interference state pairs detected are sorted according to the interference intensity to form a priority sequence, and the corresponding scheduling token queue is output;

[0038] The synchronization switching unit: each weapon state machine is activated in parallel according to the scheduling token queue, and the dynamic threshold value is injected into the forward switching threshold pool, so that the latest state data is switched synchronously under the unified time mark and is fed back to the state warehouse in real time;

[0039] The risk rearrangement unit: the state machine trajectory is continuously tracked, the coordination risk is evaluated through machine learning, if the risk exceeds the limit or the conflict count reaches the limit, the priority rearrangement is triggered, and the state warehouse is rewritten to drive the state machine to be rescheduled;

[0040] The snapshot output unit: when the state machine trajectory is stable and the conflict count is zero, the current state is solidified to generate a mounting snapshot, and the snapshot is pushed to the scheduling end to start the next round of simulation.

[0041] In the development of an AI-adaptive coordination-based weapon mounting simulation platform, the parallel mounting of multiple types of weapons presents significant technical challenges. Due to the complex interactions between the state characteristics of different weapons (such as cooling requirements, target switching modes, attack methods, and warming processes), the interweaving and superimposition of these state changes can lead to state conflicts or coordination failures during simulation. For example, the cooling process of a weapon may interfere with the warming state of another weapon, or the switching of target switching modes may conflict with the attack preparation of other weapons. If this mutual influence between states is not adequately managed and coordinated, it can cause the simulation results to deviate from the actual combat scenario, reducing the reliability and simulation accuracy of the platform. To address this issue, the present invention proposes a weapon external mounting management simulation method based on state machines. The state storage unit serves as the starting point of the entire scheme, aiming to lay a solid foundation for subsequent state management and scheduling through systematic state data initialization and interaction modeling. This method not only captures the independent states of each weapon but also quantifies the linkage effects between them, ensuring that the simulation platform accurately reflects the dynamic characteristics of multi-weapon cooperative combat.

[0042] A state machine is a model used to describe the transitions between different states of a system, and is used as a tool for precisely controlling and managing weapon states in a weapon mounting simulation platform. By defining a series of discrete states (such as "standby", "activate", "cool", and "attack") and the transition conditions and rules between states, the behavior and response of weapons in different operation stages are characterized. Each state corresponds to specific functions and constraints, and the state machine monitors the input signals and current state of the system in real time, determines and executes the switching from one state to another, thereby achieving dynamic management of weapon states. In complex scenarios with multiple weapons running in parallel, state machines can effectively handle state interactions and concurrency issues, ensuring the accuracy and consistency of state switching, and thus improving the stability and simulation accuracy of the simulation platform, providing controllability and predictability for system behavior.

[0043] The mounting weapon analysis described in the present invention is designed and implemented for the case of a single carrier. This premise assumes that the operating environment, state management, and coordination mechanisms of all mounted weapons are subject to the constraints of the same carrier.

[0044] 1-1: Summarize the initial state vector.

[0045] At the start of the simulation task, the system first collects the initial state data of each weapon for all mounted weapons. These data include key parameters such as refrigeration state, target switch mode, attack mode, and warming state, which reflect the running characteristics of each weapon. In order to systematize the management of these information, the state of each weapon is arranged into a state vector. For example, for a certain weapon, its state vector contains multiple state characteristic values, each corresponding to a specific state parameter, such as refrigeration temperature or switch mode identification. Assuming there are several weapons, the state vectors of each weapon are arranged in the same feature order, and finally all the weapon state vectors are aggregated into a state matrix. The number of rows of the state matrix corresponds to the number of state characteristics, and the number of columns corresponds to the number of weapons, forming a complete two-dimensional data structure, which comprehensively records the state baseline of each weapon at the start of the simulation.

[0046] 1-2: Derive linkage relationship according to type mapping.

[0047] Based on the completed state matrix, the system combines the type characteristics of each weapon and the predefined linkage rules to further analyze the mutual influence relationship between different weapon states. For example, the refrigeration state of a weapon may affect the warming process of another weapon through heat transfer, or when the target switch mode of a weapon changes, it may trigger the state adjustment of other weapons. In order to quantify these influences, the system constructs a linkage matrix to describe the state correlation between each pair of weapons. The number of rows and columns of the linkage matrix is equal to the number of weapons, forming a square matrix, where each element represents the linkage strength between a pair of weapons.

[0048] The linkage strength takes the absolute value form of Pearson correlation coefficient, and the calculation process is as follows: for the state vectors of any two weapons, first calculate the average value of each state vector, that is, add all the characteristic values and divide by the number of characteristics. Then, for each pair of characteristic values, calculate the difference between each characteristic value and the corresponding average value, sum the products of these differences, and divide by a standardization factor. This standardization factor consists of two parts: one is the square root of the sum of the squares of the differences between each characteristic value of the first weapon and its average value, and the other is the square root of the sum of the squares of the differences between each characteristic value of the second weapon and its average value, multiplied by each other. Finally, take the absolute value of the calculation result as the linkage strength between the two weapons. This strength value is between zero and one, reflecting the strength of the state correlation between the two weapons, whether the correlation is positive or negative.

[0049] Through this calculation method, the system can scientifically evaluate the state interaction strength between weapons, avoiding the neglect of potential influence relationship due to experience-based judgment. The quantized linkage matrix provides accurate basis for subsequent state management and scheduling, enabling the system to fully consider the mutual interaction between states when handling multiple weapon coordination, improving the realism and coordination of the simulation.

[0050] 1-3: Generating the state interaction matrix.

[0051] With the calculated linkage matrix, a state interaction matrix is directly generated. To simplify subsequent processing, the structure and values of the state interaction matrix are completely inherited from the linkage matrix, i.e., the values of each element remain unchanged. The state interaction matrix is also a square matrix, with the number of rows and columns equal to the number of weapons, and has symmetry. For the diagonal elements of the matrix, since they represent the self-interaction of the same weapon, the value is fixed at zero; while the non-diagonal elements represent the state interaction strength between different weapons, consistent with the corresponding values in the linkage matrix. This matrix provides quantitative interaction data support for subsequent hierarchical competition and cooperation analysis.

[0052] This direct inheritance approach aims to ensure that the calculation results of state interaction relationships can be seamlessly transferred to the next step of analysis, avoiding information loss due to data conversion or redefinition. The state interaction matrix stores the interaction strength of each pair of weapons in a structured form, facilitating the system to quickly read and process in subsequent steps while maintaining data consistency. This matrix form of expression also enhances the system's overall grasp of the state relationships of multiple weapons, providing a reliable reference for dynamic adjustment during simulation.

[0053] 1-4: Writing the state bin.

[0054] After completing the generation of the state matrix and the state interaction matrix, the system stores these two parts of data as the final output of the state warehousing unit to a centralized data structure, called the state bin. The state bin is responsible for saving the state information of all weapons and their interaction relationships during the simulation process. Specifically, the state matrix records the initial state baseline of each weapon, while the state interaction matrix records the quantitative data of the state linkage between weapons. After storage, the data in the state bin can be directly accessed and used in subsequent steps, such as performing hierarchical competition and cooperation analysis or preparing for scheduling.

[0055] The processing process of the state warehousing unit starts with collecting the initial state data of all weapons, and through organizing into a state matrix, achieves structured management of data. Then, based on the state matrix, the linkage relationships between weapons are analyzed, and a linkage matrix is generated and further transformed into a state interaction matrix, which is finally stored in the state bin for subsequent use. From data collection to analysis and storage, a complete process is formed, ensuring systematic processing of multi-weapon state information.

[0056] The state storage unit provides a static state interaction basis for the simulation system by aggregating the initial state vector, deducing the linkage relationship, and generating the state interaction matrix. However, static data is insufficient to cope with the dynamic changes and real-time coordination requirements of multiple weapon states during the simulation process, especially when state conflicts or coordination failures may occur due to the interweaving and superimposing of states such as refrigeration requirements, target switching modes, attack methods, and warming processes. To this end, the prioritization unit focuses on extracting dynamic information from the state interaction matrix, identifying potential interference state pairs through hierarchical competition analysis, and generating scheduling criteria to ensure that subsequent steps can effectively handle state switching in complex scenarios.

[0057] 2-1: Read the state interaction matrix from the state warehouse.

[0058] The state interaction matrix is read from the state warehouse. The state interaction matrix is a square matrix with the same number of rows and columns as the total number of mounted weapons. Each element in the matrix represents the state interaction strength between a pair of weapons, which is obtained by calculating the correlation of weapon state vectors. The diagonal elements of the matrix are always zero, because there is no interaction interference between each weapon and its own state. When reading the state interaction matrix, the system will extract all elements in the matrix, including the number of rows, the number of columns, and the value of each non-diagonal element, ensuring that subsequent analysis is based on comprehensive state interaction information.

[0059] 2-2: Perform hierarchical competition analysis.

[0060] After obtaining the state interaction matrix, the system enters the hierarchical competition analysis phase. The purpose of this phase is to identify potential interference relationships between each pair of weapons. The system traverses all non-diagonal elements in the state interaction matrix, as these elements represent the state interaction strength between different weapons. For each row and column in the matrix, the system checks each non-diagonal element one by one, extracting the weapon pair associated with each element and the state interaction strength value of the pair. After extraction, the system records these weapon pairs as potential interference state pairs, indicating that their state switching may affect each other. Specifically, if the value of an element in the matrix is not zero, the weapons represented by the row number and column number of the element are identified as a potential interference state pair, and the state interaction strength is the value of the element. After traversal, the system generates a set containing all potential interference state pairs, each of which is bound to its corresponding state interaction strength. Hierarchical competition analysis simplifies complex interaction relationships by converting matrix data into weapon pairs, making it easier to further process and manage state conflicts between weapons.

[0061] 2-3: Sort by state interaction strength to form a priority sequence.

[0062] After identifying all potential interference state pairs, the system ranks these weapon pairs to determine the order of processing. The ranking is based on the state interaction intensity of each potential interference state pair, and the system arranges the intensity values from high to low. The specific process is as follows: the system first compares the state interaction intensity of all potential interference state pairs, finds the pair with the highest intensity, and places it at the beginning of the sequence; then, in the remaining potential interference state pairs, it repeats this process to find the pair with the second highest intensity, and places it at the second position in the sequence; and so on, until all potential interference state pairs are arranged in order of intensity from high to low, forming a priority sequence. Weapon pairs with high state interaction intensity are placed at the front of the sequence, indicating that they have a higher risk of state conflict and need to be prioritized. During the ranking process, if there are potential interference state pairs with the same state interaction intensity, the system can determine their relative positions according to pre-set rules, such as the order of weapon numbers, to ensure the uniqueness of the sequence. The generation of the priority sequence enables the system to process the interaction between weapons in an orderly manner according to the risk level, thereby improving the coordination efficiency of state switching.

[0063] 2-4: Output the scheduling token queue.

[0064] Based on the priority sequence, the system generates a scheduling token queue to guide subsequent scheduling operations. Each token in the scheduling token queue corresponds to a potential interference state pair and contains the identification of the weapon pair and their state interaction intensity. The generation process is as follows: the system starts from the first position of the priority sequence and creates a token for each potential interference state pair in turn; each token records the unique identification of the two weapons in the weapon pair and the corresponding state interaction intensity value in the matrix; the tokens are arranged in the order of the priority sequence to ensure that the tokens generated by weapon pairs with high intensity are located at the front of the queue. After the queue is generated, the system outputs this queue completely for subsequent scheduling processes. The design of the scheduling token queue aims to convert the priority sequence into a form that can be directly executed, and the order in the queue reflects the processing priority of the weapon pairs, enabling high-risk potential interference state pairs to be prioritized for coordination, thereby optimizing the state management effect in the entire simulation process.

[0065] The processing process of the priority ranking unit starts with reading the state interaction matrix, identifies potential interference state pairs through hierarchical competition analysis, then ranks them according to state interaction intensity to form a priority sequence, and finally generates a scheduling token queue. The state interaction matrix provides the initial data, the hierarchical competition analysis converts it into an operable weapon pair set, the ranking process clarifies the priority, and the scheduling token queue converts the analysis results into scheduling instructions. This ensures a seamless transition from data extraction to scheduling preparation, enabling the system to dynamically coordinate the state switching of multiple weapons in simulation tasks and improving the overall coordination and reliability of the simulation platform.

[0066] The prioritization unit performs hierarchical competition analysis based on the state interaction matrix, identifies potential interference state pairs, and generates a scheduling token queue, providing priority basis for subsequent dynamic scheduling. However, in the face of real-time switching needs of multiple weapon states in the simulation process, relying solely on static analysis and scheduling preparation is insufficient to cope with the complex interaction of state such as refrigeration demand, target switching mode, attack mode, and warming process, which may lead to state conflicts or coordination failure. Therefore, the synchronous switching unit focuses on activating each weapon state machine in parallel according to the scheduling token queue, and realizes the synchronous switching and real-time feedback of the state through the dynamic threshold mechanism, to ensure the coordination consistency of multiple weapon states in the simulation process.

[0067] 3-1: Activate weapon state machines in parallel according to the scheduling token queue.

[0068] In the simulation scenario of multiple types of weapons mounted in parallel, to ensure that high-interaction weapons have priority to enter the state management process and reduce potential state conflicts, the system designs a scheduling token queue as the starting point for processing. The scheduling token queue is arranged in order from high to low according to the state interaction intensity, and each token contains the identification of the weapon pair and its corresponding interaction intensity value. Interaction intensity reflects the degree of mutual influence between weapons when the state changes, and is usually derived from the analysis of weapon coordination needs in the simulation task. The system starts from the head of the queue, reads each token in turn, and activates the corresponding weapon state machine according to the weapon pair identification in the token. The activation of the weapon state machine means starting the monitoring and control of the states such as refrigeration, target selection, attack mode, and warming, and this process is realized in parallel. The realization of parallel activation relies on the multi-thread processing capability of the simulation platform, ensuring that each weapon state machine can run independently, while avoiding time delay caused by sequential processing.

[0069] High-interaction weapons can enter the management process as soon as possible, providing sufficient time window for subsequent state coordination, thereby improving the response efficiency and stability of the system.

[0070] 3-2: Inject dynamic threshold values into the forward switching threshold pool.

[0071] To realize the coordination and consistency of different weapon state machines in time, the system sets dynamic threshold values for each weapon state machine to accurately control the triggering time of state switching. The determination of dynamic threshold values is based on the state interaction intensity in the scheduling token queue, ensuring that weapons with high interaction intensity have higher priority in switching. The process of calculating dynamic threshold values first needs to determine a baseline threshold value, which represents the default length of time required for state switching without the influence of other weapon interactions. Then, the average interaction intensity of a weapon with all other weapons is calculated, which is done by adding the interaction intensity values of the weapon with each of the other weapons and dividing the sum by the total number of weapons minus one to obtain the average interaction intensity.

[0072] The final value of the dynamic threshold is determined by the relationship between the reference threshold and the average interaction intensity: first subtract the average interaction intensity from the reference threshold, and then multiply the result by the reference threshold. This calculation method makes the dynamic threshold of the weapon with higher interaction intensity smaller, so that the state switching is triggered earlier, which is consistent with the switching time of other weapons. The unit of the dynamic threshold is time, which is consistent with the trigger condition of state switching. By quantifying the interaction relationship, the adaptive adjustment of the switching time is ensured, and the accuracy of multi-weapon cooperative operation is improved.

[0073] 3-3: Synchronize the switching of each state under the unified time scale.

[0074] In the simulation task of multi-weapon parallel management, in order to avoid state conflicts caused by time misalignment, the system establishes a unified time scale based on the global clock to ensure that the state switching of each weapon can be completed at the same time point. The global clock provides a unified timing reference for the simulation task. In each time step, the system continuously monitors the state change demand value of each weapon, such as temperature change demand or mode switching demand. When the state change demand value of a weapon reaches or exceeds its dynamic threshold, the system triggers the state switching of the weapon. At the same time, the system checks the state change demand values of other weapons, and if it finds that the demand values of other weapons are close to their dynamic thresholds, such as reaching a preset proportion of the threshold, the system triggers the state switching of these weapons together.

[0075] The judgment of approaching the threshold is realized by a synchronization factor less than one, which defines the tolerance range between the demand value and the threshold, and is used to determine whether to trigger synchronously. Through this mechanism, the state updates of all related weapons can be completed at the same time point, ensuring the coordination and consistency of state switching. This synchronization processing method fully utilizes the accuracy of the global clock and enhances the reliability of multi-weapon state management.

[0076] 3-4: Real-time feedback to the state warehouse.

[0077] After the state switching is completed, in order to support subsequent cooperative risk assessment and scheduling optimization, the system needs to record the latest state data to the state warehouse. After each state switching is completed, the system writes the latest state data of the corresponding weapon to the state warehouse, including the current values of the cooling, target selection, attack mode, and warming states. If the state change affects the interaction relationship between weapons, such as the mode switching of a weapon changing the cooperative demand with other weapons, the system will update the state interaction matrix and store the updated matrix to the state warehouse.

[0078] The state interaction matrix records the interaction intensity changes between all weapon pairs, reflecting the dynamic evolution process in the simulation process. The state bin maintains the timeliness and integrity of the data through real-time updating, providing accurate data support for subsequent risk assessment and scheduling adjustment.

[0079] The processing process of the synchronous switching unit starts with activating the weapon state machine according to the scheduling token queue, controls the state switching time through injecting dynamic threshold value, realizes the synchronous update of the state based on the global clock, and finally feeds back the latest state data to the state bin in real time. Activating the weapon state machine prepares the running conditions for subsequent switching, the dynamic threshold value determines the accurate switching time, the synchronous switching ensures the coordination consistency of multiple weapons, and the real-time feedback provides data guarantee for the continuous optimization of the system.

[0080] The state storage unit to the synchronous switching unit provides basic support for parallel state management of multiple types of weapons by summarizing the initial state vector, generating the state interaction matrix, performing hierarchical competition analysis, and realizing state synchronous switching. However, the dynamic changes and real-time coordination requirements of multiple weapon states in the simulation process put higher requirements on the adaptability of the system. Relying only on static analysis and preset scheduling of the above steps cannot cope with the complexity in high-density state interaction scenarios, which may lead to state conflicts or coordination failure. Therefore, as shown in Figure 2 the risk rearrangement unit focuses on dynamically tracking the state machine trajectory and introduces machine learning technology to evaluate the coordination risk, ensuring the stability and authenticity of the simulation platform when multiple weapons run in parallel through priority rearrangement and state machine rescheduling.

[0081] 4-1: Continuously monitor the state machine trajectory and calculate the switch jitter factor and temperature deviation span.

[0082] During the simulation process, the system continuously tracks the state machine trajectory of each weapon, which records the changes of weapon states over time, such as the switching of cooling, target selection, attack mode, and heating states. In order to analyze the dynamic behavior of state switching, it is necessary to first extract the switch jitter factor from these trajectories.

[0083] The calculation process of the switch jitter factor is as follows: select a time period, count the number of state switches in that time period, then compare it with the ideal target switching interval. The specific operation is to divide the actual switching number by the target switching interval, then take the absolute value of the result and perform normalization processing. This factor reflects the stability of the weapon pair's response to control instructions and whether there is redundant conflict, and is used to represent the sensitivity of scheduling in parallel state.

[0084] Next, the temperature deviation span is calculated to assess the accuracy of temperature control. The process is as follows: within the same time window, the actual temperature curve and the target temperature curve are obtained, the absolute value of the difference between the two at each time point is taken, and all the differences are accumulated and then normalized by dividing by the temperature control target amplitude to obtain the temperature deviation span. This index quantifies the deviation of temperature control from the expected target, which is particularly critical for weapons that require strict heat control.

[0085] By continuously monitoring and calculating these two indicators, the system can real-time grasp the dynamic characteristics of state switching and temperature management, providing data support for subsequent analysis, while ensuring that the simulation process can timely discover abnormalities or inefficiencies.

[0086] 4-2: Generate synergy risk coefficient using random forest model.

[0087] Based on the switch jitter factor and temperature deviation span, the system uses a random forest model to evaluate the synergy risk of multiple weapon state combinations. Random forest model is a method of combining multiple decision trees for prediction, which has the advantage of handling complex feature interactions and generating reliable output.

[0088] The training of the model is based on historical simulation data, which contains switch jitter factor, temperature deviation span and corresponding risk labels such as high risk or low risk. The specific process of generating synergy risk coefficient is as follows: the currently calculated switch jitter factor and temperature deviation span are taken as input features and input into the random forest model; each decision tree inside the model makes an independent judgment based on these features and outputs a risk prediction value; then, the prediction values of all decision trees are averaged to obtain the final synergy risk coefficient, which ranges from 0 to 1, and the larger the value, the higher the possibility of coordination failure.

[0089] The parameter settings of random forest are crucial to the performance of the model, involving key parameters such as the number of trees, the maximum depth of the tree and the feature selection method. For example, in a classification task, the number of trees can be adjusted to balance accuracy and computational efficiency, usually set to 100 to 500; the maximum depth of the tree is used to control the complexity of the model, too deep (such as more than 20 layers) may lead to overfitting, too shallow (such as less than 10 layers) may lead to underfitting, the common range is 10 to 20 layers; the feature selection method (such as random feature subset or information gain-based selection) improves the generalization ability by reducing the influence of related features. To optimize these parameters, grid search or random search methods are usually used. For example, on a dataset with 1000 samples and 50 features, the number of trees can be set to 100, 200, 300, the maximum depth can be set to 10, 15, 20, and the best parameter combination can be determined by combining cross-validation accuracy to improve the performance and prediction ability of the model.

[0090] This process generates a dynamic and accurate risk quantification result by considering the potential correlation between switch jitter factors and temperature deviation spans through multi-tree integration, providing the system with comprehensive risk assessment capabilities to adapt to changing simulation scenarios.

[0091] 4-3: Trigger priority reordering based on risk and conflict conditions.

[0092] The system continuously tracks the coordination risk coefficient and the conflict count, which refers to the number of weapon pairs that do not meet the dynamic threshold requirements during state synchronization. To ensure that the simulation can respond to potential coordination issues in a timely manner, the scheduling order needs to be adjusted based on specific conditions. Specifically, compare the coordination risk coefficient with the preset risk threshold, and compare the conflict count with the preset upper limit value. When the coordination risk coefficient exceeds the risk threshold or the conflict count reaches the upper limit value, trigger the priority reordering mechanism.

[0093] The reordering process is as follows: from the scheduling token queue generated by the priority ordering unit, identify weapon pairs with high coordination risk coefficients or involved in conflicts, and adjust the token positions of these weapon pairs to the front of the queue. The specific method is to traverse the queue in order, find the token of the target weapon pair and move it to the front position, while keeping the relative order of other tokens unchanged. This adjustment ensures that high-risk or high-conflict weapon pairs can be prioritized in the next scheduling period, effectively managing potential coordination failures in the simulation and enabling the system to dynamically adapt to complex interactions.

[0094] 4-4: Write updates to the state repository and drive state machine rescheduling.

[0095] After completing the priority reordering, the system writes the updated scheduling token queue to the state repository to drive the state machine's rescheduling. The specific process is as follows: transfer the adjusted scheduling token queue to the state repository in its entirety, overwriting the original queue data; then, the state repository updates the dynamic threshold and synchronization factor in the synchronization switching unit according to the new queue order. The update method is to recalculate the priority parameters for state switching of the weapon pairs at the front of the queue and adjust the parameters of subsequent weapon pairs in order; finally, the state machine guides the execution of the next round of state synchronization switching based on the updated scheduling token queue, specifically by activating the switching logic of the corresponding weapon state machine in order according to the queue order. This forms a closed-loop feedback mechanism that converts monitoring data into scheduling adjustments and applies them to state management, ensuring that the simulation system can quickly respond to state changes, maintain the continuity and orderliness of multi-weapon coordination, and thus adapt to the needs of high-density interactive scenarios.

[0096] The risk rearrangement unit closely links the aforementioned state warehousing unit to the synchronous switching unit, and introduces dynamic adjustment capabilities on the basis of static initialization and preliminary scheduling. Continuous monitoring of the state machine trajectory provides real-time data for risk assessment, which is then converted into actionable risk judgments in collaboration with the generation of risk coefficients. Priority reordering optimizes the scheduling order based on risk judgments, while writing to the state warehouse and rescheduling apply the optimization results to actual state management.

[0097] The state warehousing unit to the risk rearrangement unit lays a solid foundation for parallel state simulation of multiple types of weapons by aggregating initial states, generating state interaction matrices, performing hierarchical competition analysis, implementing state synchronous switching, and dynamically coordinating management. These steps effectively address state conflict and coordination failure, ensuring real-time tracking and optimization of state machine trajectories during simulation. However, the integrity of the simulation task not only depends on the dynamic adjustment of states, but also requires timely capturing of results when the state reaches stability to support subsequent analysis and continuous improvement. To this end, the snapshot output unit focuses on freezing the current state when the state machine trajectory is smooth and the conflict count is zero, generating a mounted snapshot and pushing it to the scheduling end, starting the next round of simulation, thereby achieving data accumulation and task continuity in the context of multiple weapon parallel simulation.

[0098] 5-1: Determine the stability of the state machine trajectory and the conflict count is zero.

[0099] During simulation, it is necessary to accurately determine whether the state machine trajectory of all weapons has reached a stable state, while confirming that the conflict count is zero, to ensure that the system is in a suitable operating condition. The stability of the state machine trajectory is evaluated by calculating the smoothness index, while the conflict count is derived from the state synchronization feedback data in the previous step.

[0100] The calculation of the smoothness index is based on the changes in state values within a predetermined time window. Specifically, the system first collects multiple consecutive state sampling points within the time window, such as the cooling temperature or switch state of the weapon. Then, for each pair of adjacent sampling points, the absolute difference between the state values of the two points is calculated, i.e., the absolute value of the state value of the current sampling point minus the state value of the previous sampling point. These absolute differences reflect the magnitude of state changes between adjacent time points. Subsequently, the absolute differences between all adjacent sampling points are added to obtain the total variation. Since the total variation is related to the number of sampling points, for normalization, the total variation is divided by the number of adjacent sampling point pairs (i.e., the total number of sampling points minus one), resulting in an average variation. This average variation is the smoothness index. If the average variation is less than a pre-set threshold, it is considered that the state machine trajectory has a small change within the time window and has reached a smooth and stable state. This calculation method can effectively reflect the fluctuation of the state machine trajectory, ensuring the reliability of the judgment result.

[0101] At the same time, the system needs to check if the conflict count is zero. The conflict count represents the number of weapon pairs that do not meet the dynamic threshold requirement in the previous state synchronization process (synchronization switching unit). This data is accumulated from the feedback of the synchronization switching unit. If the conflict count is not zero, it indicates that the state of some weapons has not been fully coordinated, which may affect the accuracy of the simulation. Only when the conflict count is zero can it be confirmed that the state of all weapons has been synchronized and there is no potential conflict.

[0102] By combining the smoothness indicator and the conflict count, the system can determine that the simulation has reached an ideal state when the state machine trajectory is stable and all weapon states are coordinated. This approach not only ensures the comprehensiveness of the judgment, but also avoids subsequent processing deviations caused by state fluctuations or conflict leftovers, providing a solid prerequisite for generating snapshots.

[0103] 5-2: Solidify the current state to generate a snapshot.

[0104] When both the stability of the state machine trajectory and the conflict count are zero, the system immediately extracts the current state data from the state warehouse and solidifies it into a complete snapshot. This process aims to record key information at the stable moment of the simulation to support subsequent analysis and initialization of simulation tasks.

[0105] In specific operations, the system first obtains multiple key data from the state warehouse, including the state matrix, the state interaction matrix, the dynamic threshold, and the coordination risk coefficient. The state matrix contains the current state vector of all mounted weapons, such as cooling level, target selection mode, attack mode, and heating state, fully describing the operation of each weapon.

[0106] Subsequently, the system organizes the above data into a structured format, such as a JSON file. This format stores data in key-value pairs, facilitating parsing and retrieval. For example, the state matrix is a key, and its value is an array containing all weapon state vectors; the state interaction matrix is another key, and its value is a two-dimensional array recording the interaction strength between weapons. The snapshot file is named based on the simulation time point when it is captured, such as "snapshot_stable_time_point", where "stable_time_point" represents the specific time identifier when the state reaches stability. This naming method facilitates tracking and managing snapshots of multiple simulation rounds.

[0107] By solidifying the current state into a snapshot, the system not only preserves the complete information of the simulation at the stable moment, but also provides a directly usable data foundation for subsequent simulation tasks.

[0108] 5-3: Push the snapshot to the scheduling end to start the next round of simulation.

[0109] After generating the snapshot, the system transmits the snapshot file to the scheduling end through the internal network to initialize and start the next round of simulation task. This process ensures that the simulation task can be continuously executed and gradually optimizes the running results based on the stable state of the previous round.

[0110] Specifically, after the snapshot file is transmitted to the scheduling end, the scheduling end first parses the structured data of the file. The data in the parsing: the system extracts the state matrix, state interaction matrix, dynamic threshold and coordination risk coefficient and other data from the snapshot, and loads them into the state warehouse as the initial state of the next round of simulation. The parsing process reads the key-value pairs one by one according to the file format, for example, obtains the state vector of each weapon from the state matrix, and assigns it to the corresponding weapon state storage unit. This way ensures that the starting state of the next round of simulation is exactly the same as the stable state of the previous round.

[0111] Subsequently, the scheduling end configures a new simulation task according to the parsed data, including setting the simulation start time and weapon configuration parameters. The new task is added to the execution queue, which manages all simulation tasks in order to ensure that each round of simulation runs in sequence according to the plan. For example, if the current snapshot is a stable state at a certain time point, the next round of simulation will start from this point to simulate the subsequent state evolution.

[0112] By pushing the snapshot to the scheduling end and starting the next round of simulation, the system realizes seamless connection and data accumulation of the simulation process. Each round of simulation runs on the optimized state of the previous round, thereby improving the performance of the platform in high-density state interaction scenarios.

[0113] In summary, the snapshot output unit generates a snapshot by judging the stability of the state machine trajectory and the zero conflict count, and pushes the snapshot to the scheduling end to start the next round of simulation, realizing the continuity of the simulation process and efficient use of data. This process ensures state stability and zero conflict, records simulation information completely and supports the initialization of subsequent tasks, providing reliable technical support for the weapon-mounted simulation platform.

[0114] The above formulas are dimensionless values calculated, and the formulas are obtained by collecting a large amount of data to simulate the most recent real situation. The preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0115] It should be noted that the system of the present application can be deployed on the device itself to realize embedded application, or can be run on PC or other terminal with user interface, so as to meet various hardware environments and use requirements.

[0116] It is apparent that the foregoing description of certain exemplary embodiments of the application has been presented for the purposes of illustration and description and is not intended to be exhaustive or to limit the application to the precise form disclosed. Many modifications and variations are possible in light of the above teachings. The described embodiments were chosen and described in order to best explain the principles of the application and its best mode of operation. It will be readily understood that the modules, processes, systems, mechanisms, methods, etc. described herein can be embodied remotely from, in-vivo, or in-vitro, or embodied in hybrid combinations of these. The scope of the application is not limited to that of the specific illustrated embodiment. The accompanying drawings and description are

[0117] It is to be understood that the phraseology or terminology herein is for the purpose of description and not of limitation. The use of terms such as first and second, etc., is not intended to imply or constitute a essential or an ordinal determination, but rather to more distinguis one from another. Additionally, the use of terms such as including, comprising, or having of their derivatives is not intended to exclude other items in a claimed process, method, article, or apparatus that are not specifi ed. The use of terms such as including, including but not limited to, comprising, comprising but not limited to, having, having but not limited to are intended to cover non-exclusive inclusions such that a process, method, article, or apparatus that compri a list of elements is not limited to those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0118] The above description is presented for purposes of illustration and description only. It is not intended to be exhaustive or to limit the application to the precise form described. Many modifications and variations are possible in light of the above teaching. It is intended that the scope of the application be limited not with this detailed description, but rather by the claims appended hereto.

Claims

1. An AI adaptive coordination based weapon mounting simulation platform, characterized in that, The method comprises the steps of: The state storage unit: the initial state vector of all weapons is summarized in the starting stage, the linkage relationship is derived according to the type mapping, and the state interaction matrix is generated, which is written into the state warehouse together with the original state data; The priority sorting unit: the interaction matrix is read from the state warehouse to perform hierarchical competition analysis, the potential interference state pairs detected are sorted according to the interference strength to form a priority sequence, and the corresponding scheduling token queue is output; The synchronous switching unit: the weapon state machine is activated in parallel according to the scheduling token queue, and the dynamic threshold value is injected into the forward switching threshold pool, so that the latest state data is switched synchronously under the unified time mark and is fed back to the state warehouse in real time; The risk rearrangement unit: the state machine trajectory is continuously tracked, the coordination risk is evaluated through machine learning, and if the risk exceeds the limit or the conflict count reaches the limit, the priority rearrangement is triggered and the state warehouse is rewritten to drive the state machine to be rescheduled; The snapshot output unit: when the state machine trajectory is stable and the conflict count is zero, the current state is solidified to generate a mounted snapshot, and the snapshot is pushed to the scheduling end to start the next round of simulation.

2. The AI adaptive coordination based weapon mounting simulation platform according to claim 1, wherein, The state storage unit comprises the following contents: For all weapons mounted on the same carrier, the cooling state, target switch mode, attack mode and warming state of each weapon are collected, and these state data are arranged into a state matrix, the number of rows of the state matrix corresponds to the number of state characteristics, and the number of columns corresponds to the number of weapons.

3. The AI adaptive coordination based weapon mounting simulation platform according to claim 2, characterized in that, The state storage unit further comprises the following contents: Based on the state matrix, the type characteristics of each weapon and the pre-defined linkage rules are combined, the absolute value of the Pearson correlation coefficient of the state vector of each pair of weapons is calculated, a linkage matrix is generated to quantify the state linkage strength between weapons; The linkage matrix is directly used as the state interaction matrix, and the state matrix and the state interaction matrix are stored in the state warehouse.

4. The AI adaptive coordination based weapon mounting simulation platform according to claim 1, wherein, The priority sorting unit comprises the following contents: Read the state interaction matrix, perform hierarchical competition analysis by traversing all non-diagonal elements of the state interaction matrix to identify potential interference state pairs, each potential interference state pair consists of two different weapons with non-zero interaction strength; the identified potential interference state pairs are sorted according to the state interaction strength from high to low to form a priority sequence; based on the priority sequence, a scheduling token queue is generated, each scheduling token corresponds to a potential interference state pair, contains the identifiers of the two weapons and their state interaction strength, and is arranged in the order of the priority sequence.

5. The AI adaptive coordination based weapon mounting simulation platform according to claim 4, wherein, The synchronous switching unit further comprises the following contents: On the basis of the scheduling token queue arranged in descending order according to the state interaction strength between weapon pairs, the weapon state machine is activated in parallel to preferentially process the weapon pairs with high state interaction strength; the dynamic threshold value is injected into the forward switching threshold pool; the state synchronization switching is realized through a unified global clock, when the state change demand value of a certain weapon reaches or exceeds its dynamic threshold value, the state switching of the corresponding weapon is triggered, and the state change demand value of other weapons within a predetermined tolerance range close to their dynamic threshold value is also triggered to maintain synchronization; after the state switching is completed, the state warehouse is updated in real time to record the new state data of the weapons, and the state interaction matrix is updated to reflect the changes of the state interaction relationship between the weapon pairs.

6. The AI adaptive coordination based weapon mounting simulation platform according to claim 5, characterized in that, The synchronous switching unit further comprises the following contents: Wherein the dynamic threshold is calculated according to the average state interaction intensity of each weapon with other weapons.

7. The AI adaptive coordination based weapon mounting simulation platform according to claim 1, wherein, The risk rearrangement unit includes the following: The state machine trajectory is continuously monitored to record the changes of weapon states over time, the switch jitter factor is calculated to evaluate the stability of state switching and the consistency of instruction response, the temperature deviation span is calculated to quantify the accuracy of temperature control, the switch jitter factor and the temperature deviation span are analyzed by using a random forest model to generate a synergistic risk coefficient for evaluating the risk of multi-weapon state combination, when the synergistic risk coefficient exceeds the preset risk threshold or the conflict count reaches the preset upper limit, the priority rearrangement mechanism is triggered to move the high-risk or high-conflict weapon pair to the front of the scheduling token queue, and the updated scheduling token queue is written to the state warehouse and drives the state machine to reschedule according to the new queue order.

8. The AI adaptive coordination based weapon mounting simulation platform according to claim 7, characterized in that, The risk rearrangement unit also includes the following: The number of state switches in a certain period of time is counted, compared with the ideal target switching interval, and then the absolute value of the difference between the actual switching number and the target switching interval is normalized to obtain the switch jitter factor; The actual temperature curve and the target temperature curve are obtained within the same time window, the difference between the two at each time point is calculated and then the absolute value is taken, then the sum is divided by the temperature control target amplitude to normalize, and the temperature deviation span is obtained.

9. The AI adaptive coordination based weapon mounting simulation platform according to claim 7, wherein, The snapshot output unit includes the following: The stability of the state is determined by calculating the smoothness index of the state machine trajectory, while monitoring the conflict count to ensure that the number of weapon pairs that do not meet the dynamic threshold requirement during state synchronization is zero, when the smoothness index of the state machine trajectory is less than the preset threshold and the conflict count is zero, the current state data is captured and solidified as a launch snapshot, then the launch snapshot is transmitted to the scheduling end, the scheduling end analyzes the data in the launch snapshot and loads it to the state warehouse, as the initial state of the next round of simulation task to start a new round of simulation task.

10. The AI adaptive coordination based weapon mounting simulation platform according to claim 9, wherein, The snapshot output unit also includes the following: Wherein the smoothness index quantifies the fluctuation of the state machine trajectory by evaluating the average change amplitude of the state value within a preset time window.

Images