Digital twin aerial reconnaissance mission simulation support system

Abstract

The present invention relates to the field of digital twin technology, specifically to a digital twin aerial reconnaissance mission simulation support system, the system comprising: an environmental state coding module collects dynamic data based on environmental data in the aerial reconnaissance mission, monitors the target moving speed and direction in real time, processes the dynamic data through time series coding, inputs a graph convolutional network, updates the node state, and obtains an environmental state vector. In the present invention, by real-time monitoring of the target moving speed and direction, and time series coding processing of dynamic data, the updating speed and accuracy of environmental information are greatly improved, and real-time information is imported into a deep Q network. Through the systematic evaluation of risk and success probability, accurate optimization of the flight path is achieved, which not only enhances the adaptability to the complex aerial reconnaissance mission environment, but also optimizes the safety and efficiency in the execution of the mission. By adjusting the flight resources and mission parameters in real time, the effective utilization of resources and the high adaptability of the mission are guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of digital twin technology, and in particular to a digital twin aerial reconnaissance mission simulation support system. Background Art

[0002] Digital twin technology is an advanced computer simulation technology that simulates and analyzes the behavior of a system or product by creating a virtual model of a physical entity. This technology involves collecting large amounts of real-time data and using this data to update the virtual model to ensure that the model accurately reflects the current state of its physical counterpart. Digital twin technology is widely used in industries such as manufacturing, aerospace, automotive industry and building management to optimize product performance, predict maintenance needs and improve operational efficiency. This technology can not only reduce costs and time during the product development and testing phase, but also continue to provide support after the product is put into use, guiding maintenance and operational decisions through simulation analysis.

[0003] Among them, the aerial reconnaissance mission simulation support system is a system that uses digital twin technology to simulate and support aerial reconnaissance missions. This topic involves creating and utilizing digital twin models of aircraft to simulate actual aerial reconnaissance activities such as surveillance, reconnaissance, and other air monitoring missions. The uses of the system include training pilots and mission planners, testing and optimizing flight and reconnaissance strategies, and evaluating potential mission risks and outcomes. Through this simulation, the effectiveness of the mission can be predicted without actual flight, thereby improving the success rate and safety of the mission.

[0004] Although existing digital twin technologies provide simulation and analysis for system or product behavior, they often show problems of slow response and untimely information updates when dealing with rapidly changing and highly dynamic aerial reconnaissance environments. Such technologies rely on static data models and are difficult to effectively capture immediate changes in target actions, such as sudden movements or strategy adjustments, making it difficult for simulated virtual models to accurately reflect real-time physical reality. In addition, due to the lack of flexibility in resource allocation and task scheduling in existing technologies, fixed strategies and preset parameter settings may lead to improper resource allocation and poor task execution in the face of ever-changing battlefield environments. For example, if the target suddenly deploys a new defense system and existing technologies fail to adapt and adjust quickly, the reconnaissance mission may be at risk of failure, increasing operational uncertainty. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art and propose a digital twin aerial reconnaissance mission simulation support system.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solution: The digital twin aerial reconnaissance mission simulation support system includes:

[0007] The environmental state encoding module collects dynamic data based on the environmental data in the aerial reconnaissance mission, monitors the target's moving speed and direction in real time, processes the dynamic data through temporal coding, inputs the graph convolutional network, updates the node status, and obtains the environmental state vector;

[0008] The path optimization module inputs the deep Q network based on the environmental state vector, performs route selection calculation, compares the risk assessment with the strategy score, selects the flight path based on the risk and success probability, and generates an optimized path decision;

[0009] The dynamic task scheduling module refines the reconnaissance route arrangement, adjusts the timetable and resource allocation, optimizes the allocation of flight resources, analyzes the resource input and expected benefits, and obtains the task scheduling plan based on the optimized path decision;

[0010] The real-time event response module monitors changes in environmental status, including the deployment of new defense systems, based on the task scheduling plan, adjusts task parameters and flight strategies, recalculates the task plan, and generates response adjustment instructions.

[0011] As a further solution of the present invention, the environmental state vector includes target dynamics, meteorological data, and terrain information; the optimized path decision specifically includes the selected path, risk level, and success prediction value; the task scheduling plan includes route arrangement, timetable, and resource allocation; and the response adjustment instruction includes strategy modification, parameter update, and aerial reconnaissance route adjustment.

[0012] As a further solution of the present invention, the step of real-time monitoring of the target moving speed and direction is specifically:

[0013] Collect the real-time position coordinates of the target from the surveillance system and radar sensor data of the aerial reconnaissance mission, and add a timestamp to each record to generate a continuous trajectory record;

[0014] The target moving speed and direction are calculated using two consecutive position coordinates in the trajectory record, and the position difference of the real-time position coordinates and the time interval between two adjacent timestamps are substituted into the formula:

[0015] ,

[0016] ,

[0017] Where v represents the moving speed of the target, Indicates the target's moving direction. , as well as , They represent the position coordinates of the target in two consecutive monitorings, and are used to calculate the straight-line distance and direction of the target moving from one point to another. is the time interval between two monitorings, Represents environmental factors, represents the error correction coefficient, Represents the geographic area correction parameter, represents the geographical environment complexity parameter, represents the time interval correction parameter, represents the target speed variance parameter, A correction parameter representing the error correction factor.

[0018] As a further solution of the present invention, the step of acquiring the environment state vector is specifically:

[0019] Dynamic data is processed through time series coding, which combines the timestamp information collected by radar sensors to construct a coded sequence;

[0020] The encoded sequence is input into the graph convolutional network, which uses the connection relationship between nodes and node characteristics to perform feature fusion and update the node status through the formula:

[0021] ,

[0022] in, is the learned weight of the node, is the node weight matrix, is the adjustment parameter for the continuity and variability of the environment, is the influence factor of the degree of change, is a nonlinear activation function, is the old state of the node, and Respectively represent the continuity and change of the environment, is the current node status;

[0023] After feature fusion and nonlinear transformation at the network layer, the output node state vector set is classified as an environment state vector, reflecting the dynamic state of the current environment.

[0024] As a further solution of the present invention, the method for obtaining the optimized path decision is specifically as follows:

[0025] Input the environment state vector to the deep Q network, obtain the value of each estimated flight path by processing the environment state vector, and use the environment state vector to evaluate the current state of each optional route node;

[0026] Calculate the risk assessment value and strategy score for each path using the following formula:

[0027] ,

[0028] in, is the calculated path value, represents the predicted value before taking action a in state s, r is the immediate reward, and are the learning rate and discount factor respectively, Is in the new state The best action to take The maximum Q value, is the adjustment parameter of the action selection probability, is the risk adjustment parameter, is the risk of the current path, is the risk of the ideal path, sigmoid represents the Sigmoid function, is the new state after taking action a, Is in the new state The best action to take under

[0029] By comparing the risk assessment values ​​and strategy scores of multiple paths and combining the values ​​output by the deep Q network, the optimal flight path is selected based on safety and efficiency.

[0030] As a further solution of the present invention, the step of obtaining the task scheduling solution is specifically as follows:

[0031] Based on the optimized path decision, refine the reconnaissance route arrangement, analyze the potential benefits of multiple routes, and select the route with the highest expected benefit as the preliminary route;

[0032] Adjust the schedule and resource allocation, combine the preliminary path with the current schedule, analyze the resource requirements of multiple reconnaissance nodes and the estimated resource conflicts, and substitute them into the formula for resource allocation:

[0033] ,

[0034] Make resource allocations;

[0035] in, is the resource allocation of node i, is the total amount of flight resources available for allocation, n is the number of reconnaissance nodes, is the importance weight of the node, is the schedule factor of the scout node, is the spatial position factor, is the schedule adjustment factor, is the adjustment factor for spatial position;

[0036] Conduct a benefit analysis on resource input and expected benefits, evaluate the resource input and actual benefits of multiple nodes, calculate the resource input-output ratio, and establish a task scheduling plan.

[0037] As a further solution of the present invention, the step of obtaining the response adjustment instruction is specifically as follows:

[0038] Based on the task scheduling scheme, monitor the changes in environmental status, including the deployment of new defense systems, combine real-time reconnaissance information with environmental dynamic feature data, determine whether it exceeds the preset risk threshold, and substitute it into the formula for monitoring and evaluation:

[0039] ,

[0040] Where R represents the risk assessment value, represents the weighting coefficient of risk assessment, It is the product of the importance weights of multiple nodes, potential threats and urgency levels. represents the importance weight of the node, Represents the potential threat value of the node, Indicates the exponential impact of the node's urgency on the evaluation, is the weighting coefficient of urgency, is the urgency score of the node, is the established risk threshold, is the environmental complexity factor, is the dynamic characteristic data of the current environment, is the complexity adjustment factor;

[0041] If the risk assessment value R exceeds the preset threshold, the mission parameters and flight strategy are adjusted, including replanning the reconnaissance route, changing the flight altitude or speed, or increasing resource allocation;

[0042] Recalculate the mission plan, re-input the adjusted parameters into the dynamic mission scheduling module, generate the current flight path and schedule, and generate corresponding response adjustment instructions.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are:

[0044] In the present invention, the updating speed and accuracy of environmental information are greatly improved by real-time monitoring of the target's moving speed and direction, and time-series coding processing of dynamic data. By importing real-time information into the deep Q network, accurate optimization of the flight path is achieved through systematic evaluation of risk and success probability. This method not only enhances the adaptability to complex aerial reconnaissance mission environments, but also optimizes the safety and efficiency of mission execution. By adjusting flight resources and mission parameters in real time, the innovative solution ensures the effective use of resources and the high adaptability of the mission, significantly improving the execution flexibility and overall success rate of the mission. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a system flow chart of the present invention;

[0046] Figure 2 A flowchart of the steps of real-time monitoring of target moving speed and direction of the present invention;

[0047] Figure 3 A flow chart of the steps for obtaining the environment state vector of the present invention;

[0048] Figure 4 A flow chart of a method for obtaining an optimized path decision of the present invention;

[0049] Figure 5 A flowchart of the steps for obtaining the task scheduling solution of the present invention;

[0050] Figure 6 The present invention is a flow chart of the steps for obtaining the response adjustment instruction. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0053] Example:

[0054] See also Figure 1 , the digital twin aerial reconnaissance mission simulation support system includes:

[0055] The environmental state encoding module collects dynamic data based on the environmental data in the aerial reconnaissance mission, monitors the target's moving speed and direction in real time, processes the dynamic data through temporal coding, inputs the graph convolutional network, updates the node status, and obtains the environmental state vector;

[0056] The path optimization module is based on the environment state vector, inputs the deep Q network, performs route selection calculations, compares the risk assessment with the strategy score, selects the flight path based on the risk and success probability, and generates the optimized path decision;

[0057] The dynamic task scheduling module is based on optimizing path decisions, refining reconnaissance route arrangements, adjusting timetables and resource allocations, optimizing the allocation of flight resources, and conducting a benefit analysis of resource input and expected returns to obtain a task scheduling plan;

[0058] The real-time event response module monitors changes in environmental status based on the task scheduling plan, including the deployment of new defense systems, adjusts mission parameters and flight strategies, recalculates the mission plan, and generates response adjustment instructions.

[0059] The environmental state vector includes target dynamics, meteorological data, and terrain information. The optimized path decision includes the selected path, risk level, and success prediction value. The task scheduling plan includes route arrangement, timetable, and resource allocation. The response adjustment instructions include strategy modification, parameter update, and aerial reconnaissance route adjustment.

[0060] See also Figure 2 ,The steps for real-time monitoring of target moving speed and direction are as follows:

[0061] Collect the real-time position coordinates of the target from the surveillance system and radar sensor data of the aerial reconnaissance mission, and add a timestamp to each record to generate a continuous trajectory record;

[0062] Using two consecutive position coordinates in the trajectory record, calculate the target moving speed and direction, and substitute the position difference of the real-time position coordinates and the time interval between two adjacent timestamps into the formula:

[0063] ,

[0064] ,

[0065] Where v represents the moving speed of the target, Indicates the target's moving direction. , as well as , They represent the position coordinates of the target in two consecutive monitorings, and are used to calculate the straight-line distance and direction of the target moving from one point to another. is the time interval between two monitorings, Represents environmental factors, represents the error correction coefficient, Represents the geographic area correction parameter, represents the geographical environment complexity parameter, represents the time interval correction parameter, represents the target speed variance parameter, A correction parameter representing the error correction factor.

[0066] Detailed explanation of formula parameters:

[0067] , , , : Position coordinates of two consecutive records; : The time interval between two records; : Geographic area correction parameter, used to adjust the error caused by the complexity of the geographical environment; : Geographic environment complexity parameter, indicating the complexity of the geographical environment, such as mountainous areas, cities, etc.; : Environmental factors, used to adjust the impact of time intervals; : Time interval correction parameter, used to adjust the time correction caused by the variance of the target moving speed; : Target speed variance parameter, indicating the degree of change of target speed during the observation period; : Error correction coefficient, used to adjust the accuracy of the calculated direction; : Error correction coefficient correction parameter, used to further adjust the error correction coefficient according to the complexity of the environment.

[0068] Specific calculation example:

[0069] Suppose in two observations, the data obtained are as follows:

[0070] =100 m, =100 m, =300 m, =400 m, = 10 seconds, =0.3 (lower environmental complexity), =0.2 (speed does not change much), =1.5 (medium correction factor), = 0.5 seconds (small time adjustment), = 0.1 m (small error correction), =0.2 (the environment has little influence on error correction), =0.1 (small time interval adjustment);

[0071] Calculation process:

[0072] Compute the sum of squared distances:

[0073] ,

[0074] The sum of squares of distances corrected for geographic area is: ,

[0075] Calculation speed: ,

[0076] Calculation direction: .

[0077] See also Figure 3 , the specific steps for obtaining the environment state vector are:

[0078] Dynamic data is processed through time series coding, which combines the timestamp information collected by radar sensors to construct a coded sequence;

[0079] The encoded sequence is input into the graph convolutional network, which uses the connection relationship between nodes and node characteristics to perform feature fusion and update the node status through the formula:

[0080] ,

[0081] in, is the learned weight of the node, is the node weight matrix, is the adjustment parameter for the continuity and variability of the environment, is the influence factor of the degree of change, is a nonlinear activation function, is the old state of the node, and Respectively represent the continuity and change of the environment, is the current node status;

[0082] After feature fusion and nonlinear transformation at the network layer, the output node state vector set is classified as an environment state vector, reflecting the dynamic state of the current environment.

[0083] Parameter Description:

[0084] : The old state of the node, indicating the state at the previous time point; : The weight matrix of the node, used for state update; : The learned weight of the node; : Adjustment parameters for environmental continuity and variability; : Change degree impact factor, used to adjust the impact of change degree on node status update; : Environmental continuity, which measures the smoothness of environmental changes; : Environmental variability, measuring the rapidity of environmental change; Non-linear activation functions, commonly used ones include ReLU, sigmoid, etc.

[0085] Specific numerical examples:

[0086] Assume a simplified scenario with only two nodes (n=2) and concretize some parameter values ​​for calculation.

[0087] Assume the following parameter values:

[0088] ,

[0089] ,

[0090] ,

[0091] =0.8,

[0092] ,

[0093] ,

[0094] (Using ReLU function);

[0095] Calculation process:

[0096] Calculate the weight influence of each node: ,

[0097] Here, we need to and After matrix multiplication, multiply by , the specific calculation is:

[0098] ,

[0099] Calculate the impact of environmental factors:

[0100] ,

[0101] Calculate the effect of the adjustment factor on the degree of change:

[0102] ,

[0103] Combine all the calculation results and pass them through a nonlinear activation function:

[0104] ,

[0105] See also Figure 4 , the specific method for obtaining the optimized path decision is:

[0106] Input the environment state vector to the deep Q network, obtain the value of each estimated flight path by processing the environment state vector, and use the environment state vector to evaluate the current state of each optional route node;

[0107] Calculate the risk assessment value and strategy score for each path using the following formula:

[0108] ,

[0109] in, is the calculated path value, represents the predicted value before taking action a in state s, r is the immediate reward, and are the learning rate and discount factor respectively, Is in the new state The best action to take The maximum Q value, is the adjustment parameter of the action selection probability, is the risk adjustment parameter, is the risk of the current path, is the risk of the ideal path, sigmoid represents the Sigmoid function, is the new state after taking action a, Is in the new state The best action to take under

[0110] By comparing the risk assessment values ​​and strategy scores of multiple paths and combining the values ​​output by the deep Q network, the optimal flight path is selected based on safety and efficiency.

[0111] Specific calculation process and derivation of each parameter in the formula:

[0112] parameter (Learning rate):

[0113] Assumptions is 0.9, which means that the newly learned information will account for 90% of the current valuation, while 10% of the original valuation will be retained. This setting allows the algorithm to adapt to recent environmental changes more quickly.

[0114] parameter (predicted value before taking action a in state s):

[0115] Assume that the previous predicted value of taking action a in a certain state s is 100. This value is found in the Q table, which is updated through the previous training iterations.

[0116] Parameter r (immediate return):

[0117] Assume that after the aircraft selects a certain path, because the path is short or the risk is low, the immediate reward r obtained is 50. This value is calculated based on the risk score of the flight path and environmental factors.

[0118] parameter (Discount factor):

[0119] Assumptions It is 0.95, which indicates that future rewards have a high importance in the current value assessment. A high value means that the aircraft will prioritize long-term overall benefits rather than short-term benefits.

[0120] parameter (New Status The predicted value of the best action to take under

[0121] Assuming that in the new state Next, the best action The predicted value is 120, which is also obtained through Q table query.

[0122] parameter (Risk Adjustment Parameters):

[0123] Assumptions is 0.1, which is an empirical setting used to adjust the degree to which risk assessment influences action selection.

[0124] parameter and (The risk of the current path and the risk of the ideal path):

[0125] Assume the risk of the current path 0.8 (e.g. based on target threat level or obstacle density), risk of ideal path is 0.3, and these values ​​are usually obtained through environmental analysis and comparison with historical data.

[0126] parameter :

[0127] The sigmoid function is a commonly used normalization function that maps real numbers to the interval (0,1), which helps control the impact of risk adjustment on action selection. The specific calculation method is:

[0128] ,

[0129] if =0.8-0.3=0.5, then:

[0130] ,

[0131] Specific calculation process:

[0132] By setting the above parameters, substitute the values ​​into the formula for calculation:

[0133] ,

[0134] This gives a new path selection value The value is 106.41, which will be used to decide whether to select this path as the flight path.

[0135] See also Figure 5 , the specific steps for obtaining the task scheduling plan are:

[0136] Based on the optimized path decision, the reconnaissance route arrangement is refined, the potential benefits of multiple routes are analyzed, and the route with the highest expected benefit is selected as the preliminary route;

[0137] Adjust the schedule and resource allocation, combine the preliminary path with the current schedule, analyze the resource requirements of multiple reconnaissance nodes and the estimated resource conflicts, and substitute them into the formula for resource allocation:

[0138] ,

[0139] Make resource allocations;

[0140] in, is the resource allocation of node i, is the total amount of flight resources available for allocation, n is the number of reconnaissance nodes, is the importance weight of the node, is the schedule factor of the scout node, is the spatial position factor, is the schedule adjustment factor, is the adjustment factor for spatial position;

[0141] Conduct a benefit analysis on resource input and expected benefits, evaluate the resource input and actual benefits of multiple nodes, calculate the resource input-output ratio, and establish a task scheduling plan.

[0142] Parameter details:

[0143] : The total amount of flight resources available for allocation, assuming 1000 units; n: The total number of nodes in the reconnaissance mission, assuming there are 10 nodes; : The importance weight of the node, which is based on the strategic position of the node or the analysis of previous task data. Assume that the importance weight of node i is 0.8; : Schedule factor, indicating the time urgency of the task, assumed to be 0.7; : Spatial location factor, indicating the importance of the node's geographical location, assumed to be 0.5; : Schedule adjustment factor, adjusted according to the urgency of the task. Assuming the urgency is high, is 1.2; : The adjustment factor for spatial location, weighted according to the strategic importance of the geographical location, is assumed to be 1.1.

[0144] Calculation process and examples:

[0145] Calculate the basic resource allocation for each node:

[0146] ,

[0147] For node i, consider its importance weight:

[0148] Resource allocation of node i = 100 · 0.8 = 80 units,

[0149] Calculate the adjusted resource allocation:

[0150] Adjust using timetable and spatial location factors:

[0151] ,

[0152] The overall resource allocation to node i, combined with time and space adjustment factors:

[0153] ,

[0154] See also Figure 6 , the specific steps for obtaining the response adjustment instruction are:

[0155] Based on the task scheduling plan, monitor the changes in environmental status, including the deployment of new defense systems, combine real-time reconnaissance information with environmental dynamic feature data, determine whether it exceeds the preset risk threshold, and substitute it into the formula for monitoring and evaluation:

[0156] ,

[0157] Where R represents the risk assessment value, represents the weighting coefficient of risk assessment, It is the product of the importance weights of multiple nodes, potential threats and urgency levels. represents the importance weight of the node, Represents the potential threat value of the node, Indicates the exponential impact of the node's urgency on the evaluation, is the weighting coefficient of urgency, is the urgency score of the node, is the established risk threshold, is the environmental complexity factor, is the dynamic characteristic data of the current environment, is the complexity adjustment factor;

[0158] If the risk assessment value R exceeds the preset threshold, the mission parameters and flight strategy are adjusted, including replanning the reconnaissance route, changing the flight altitude or speed, or increasing resource allocation;

[0159] Recalculate the mission plan, re-input the adjusted parameters into the dynamic mission scheduling module, generate the current flight path and schedule, and generate corresponding response adjustment instructions.

[0160] Parameter explanation and acquisition method:

[0161] (Risk assessment weighting coefficient): used to adjust the weight of the overall risk assessment. This coefficient is set based on historical data and expert experience, for example, assuming it is 0.5;

[0162] (Importance weight of the node): Determined based on the strategic location of the node and the influence of past events. For example, assume that the importance of the node is divided into 0.2 and 0.3;

[0163] (Potential threat value of a node): evaluated by analyzing past data and current intelligence, for example, possible values ​​are 10 or 20;

[0164] (Weighting factor of urgency): Similar to , set according to the situation, such as 0.1

[0165] (Emergency score of node i): Based on recent activities and intelligence analysis, the score may be 3 or 4, for example;

[0166] (Established risk threshold): Set a baseline risk value, such as 50;

[0167] (Environmental complexity factor): Determined based on the complexity of the geographical and political environment, for example, 0.2;

[0168] (Dynamic characteristic data of the current environment): real-time data obtained from environmental monitoring equipment and sensors, such as 15;

[0169] (Complexity adjustment coefficient): used to adjust the impact of the complexity factor, for example 1.5.

[0170] Calculation process and examples:

[0171] Assuming there are two nodes in the system (n=2), use the above assumed parameter values ​​to calculate:

[0172] Contribution of Compute Nodes:

[0173] Node 1: ,

[0174] Node 2: ,

[0175] Calculate the risk contribution of each node:

[0176] ,

[0177] ,

[0178] Calculation of the risk assessment denominator:

[0179] ,

[0180] Overall risk assessment:

[0181] ,

[0182] The above steps and calculation examples show how to use specific data and parameters to calculate the risk assessment value R. This result can be used to determine whether it is necessary to adjust the mission parameters and flight strategy. If R exceeds the set threshold (for example, 0.2), adjustments are required.

[0183] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. Digital twin aerial reconnaissance mission simulation support system, characterized by: The system comprises: The environmental state encoding module collects dynamic data based on the environmental data in the aerial reconnaissance mission, monitors the target's moving speed and direction in real time, processes the dynamic data through temporal coding, inputs the graph convolutional network, updates the node status, and obtains the environmental state vector; The steps of real-time monitoring of the target moving speed and direction are specifically as follows: Collect the real-time position coordinates of the target from the surveillance system and radar sensor data of the aerial reconnaissance mission, and add a timestamp to each record to generate a continuous trajectory record; The target moving speed and direction are calculated using two consecutive position coordinates in the trajectory record, and the position difference of the real-time position coordinates and the time interval between two adjacent timestamps are substituted into the formula: ， ， Where v represents the moving speed of the target, Indicates the target's moving direction. , as well as , They represent the position coordinates of the target in two consecutive monitorings, and are used to calculate the straight-line distance and direction of the target moving from one point to another. is the time interval between two monitorings, Represents environmental factors, represents the error correction coefficient, Represents the geographic area correction parameter, represents the geographical environment complexity parameter, represents the time interval correction parameter, represents the target speed variance parameter, A correction parameter representing the error correction coefficient; The steps of obtaining the environment state vector are specifically as follows: Dynamic data is processed through time series coding, which combines the timestamp information collected by radar sensors to construct a coded sequence; The coding sequence is input into the graph convolutional network, which uses the connection relationship between nodes and node characteristics to perform feature fusion and update the node status. The reference formula is: ， in, is the learned weight of the node, is the node weight matrix, is the adjustment parameter for the continuity and variability of the environment, is the influence factor of the degree of change, is a nonlinear activation function, is the old state of the node, and Respectively represent the continuity and change of the environment, is the current node status; After feature fusion and nonlinear transformation at the network layer, the output node state vector set is classified as the environment state vector; The path optimization module inputs the deep Q network based on the environmental state vector, performs route selection calculation, compares the risk assessment with the strategy score, selects the flight path based on the risk and success probability, and generates an optimized path decision; The method for obtaining the optimized path decision is specifically as follows: Input the environment state vector to the deep Q network, obtain the value of each estimated flight path by processing the environment state vector, and use the environment state vector to evaluate the current state of each optional route node; Calculate the risk assessment value and strategy score for each path using the following formula: ， in, is the calculated path value, represents the predicted value before taking action a in state s, r is the immediate reward, and are the learning rate and discount factor respectively, Is in the new state The best action to take The maximum Q value, is the adjustment parameter of the action selection probability, is the risk adjustment parameter, is the risk of the current path, is the risk of the ideal path, sigmoid represents the Sigmoid function, is the new state after taking action a, Is in the new state The best action to take under Compare the risk assessment values ​​and strategy scores of multiple paths, combine the values ​​output by the deep Q network, and select the optimal flight path based on safety and efficiency; The dynamic task scheduling module refines the reconnaissance route arrangement, adjusts the timetable and resource allocation, optimizes the allocation of flight resources, analyzes the resource input and expected benefits, and obtains the task scheduling plan based on the optimized path decision; The steps for obtaining the task scheduling scheme are specifically as follows: Based on the optimized path decision, refine the reconnaissance route arrangement, analyze the potential benefits of multiple routes, and select the route with the highest expected benefit as the preliminary route; Adjust the schedule and resource allocation, combine the preliminary path with the current schedule, analyze the resource requirements of multiple reconnaissance nodes and the estimated resource conflicts, and substitute into the formula: ， Make resource allocations; in, is the resource allocation of node i, is the total amount of flight resources available for allocation, n is the number of reconnaissance nodes, is the importance weight of the node, is the schedule factor of the scout node, is the spatial position factor, is the schedule adjustment factor, is the adjustment factor for spatial position; Conduct a benefit analysis on resource input and expected benefits, evaluate resource input and actual benefits of multiple nodes, calculate resource input-output ratio, and establish a task scheduling plan; The real-time event response module monitors changes in environmental status based on the mission scheduling plan, including the deployment of new defense systems, adjusts mission parameters and flight strategies, recalculates mission plans, and generates response adjustment instructions; The steps of obtaining the response adjustment instruction are specifically as follows: Based on the task scheduling scheme, monitor the changes in environmental status, including the deployment of new defense systems, and combine real-time reconnaissance information with environmental dynamic feature data to determine whether the preset risk threshold is exceeded and conduct monitoring and evaluation. The reference formula is: ， Where R represents the risk assessment value, represents the weighting coefficient of risk assessment, It is the product of the importance weights of multiple nodes, potential threats and urgency levels. represents the importance weight of the node, Represents the potential threat value of the node, Indicates the exponential impact of the node's urgency on the evaluation, is the weighting coefficient of urgency, is the urgency score of the node, is the established risk threshold, is the environmental complexity factor, is the dynamic characteristic data of the current environment, is the complexity adjustment factor; If the risk assessment value R exceeds the preset threshold, the mission parameters and flight strategy are adjusted, including replanning the reconnaissance route, changing the flight altitude or speed, or increasing resource allocation; Recalculate the mission plan, re-enter the adjusted parameters into the dynamic mission scheduling module, generate the current flight path and schedule, and generate corresponding response adjustment instructions.

Images