Collision calculation batch adjudication method based on three-dimensional probability density

Abstract

The present application relates to the field of war game simulation, and discloses a collision calculation war game batch decision method based on three-dimensional probability density, specifically comprising: loading the action effect three-dimensional probability density body of the war game unit, defining the spatial range and probability distribution of the action effect; detecting the collision intersection of the probability density body and the target in the three-dimensional space, obtaining the probability density value of each micro unit in the collision intersection volume of the three-dimensional probability density body and the target, if the intersection volume contains at least two micro units, using one of the average value, parallel success probability or gravity weighted to calculate the corrected probability value; combining the meteorological, terrain, electromagnetic disturbance secondary correction to obtain the final probability; taking the final probability as the threshold, comparing the uniform random number, and determining the success or failure of the action. The present application uses three-dimensional probability density body collision detection and multi-method probability correction to improve the war game decision efficiency and reliability.

Description

Technical Field

[0001] This invention relates to the field of wargaming simulation, and in particular to a method for batch adjudication of wargaming based on collision calculation using three-dimensional probability density. Background Technology

[0002] Wargaming is an important tool for military simulation and combat simulation. In large-scale wargaming, numerous combat units engage in interactive confrontations and action decisions. Traditional solutions mainly fall into two categories. The first category is a linear queuing-based probabilistic decision-making method based on manual wargaming. This method relies on manual operation and pre-set probability tables for rapid decision-making, offering fast computation speed but resulting in significant loss of action details, making it difficult to meet the demands of modern high-precision wargaming. The second category is a simulation calculation method based on mathematical modeling. This method performs concurrent analytical calculations for each action, preserving more interaction details, but it involves a massive computational load, has high requirements for hardware and software environments, and limits the timeliness of wargaming, hindering its widespread adoption.

[0003] To address the aforementioned issues, current technologies lack a batch decision-making method for wargames that can balance computational efficiency with detail preservation. Some improvements attempt to simplify simulation models or employ approximate algorithms, but these often come at the cost of sacrificing probabilistic accuracy. Therefore, a new decision-making method for wargame scenarios is needed that can reduce computational complexity, decrease reliance on high-performance hardware and software environments, and improve the real-time performance and practicality of large-scale wargame simulations while ensuring the reliability of probabilistic calculations. Summary of the Invention

[0004] One objective of this invention is to propose a batch adjudication method for collision calculation wargames based on three-dimensional probability density. This invention loads a three-dimensional probability density volume of action effects onto each wargame unit, triggers adjudication through collision intersection detection, obtains the probability density values ​​of each micro-unit within the intersection volume, and performs probability correction using one of the following methods: average calculation, parallel system success probability calculation, or centroid distance weighted calculation. Further correction is then performed by incorporating meteorological, terrain, and electromagnetic dynamic environmental disturbance factors. Based on the final corrected probability, a probability function is constructed to execute random adjudication. This method possesses advantages such as high computational efficiency in large-scale wargame simulations, complete preservation of action details, low dependence on hardware and software environments, and high reliability of probability results.

[0005] The collision calculation wargaming batch adjudication method based on three-dimensional probability density according to an embodiment of the present invention includes the following steps: For each wargame unit, a corresponding 3D probability density volume of action effects is loaded. This volume represents the influence range of the action effect in 3D space, and the probability density represents the probability of the effect occurring. The collision and intersection of the action probability density volumes of enemy and friendly units in 3D space during wargame unit simulations are detected. The probability density values ​​of each tiny unit within the collision and intersection volume of the 3D probability density volume and the target are then obtained. When the collision intersection volume contains at least two micro-units, a probability correction method is used to calculate the corrected probability value from at least two probability density values. This method involves a comprehensive calculation of the two probability density values, chosen from one of the following: average value calculation, parallel system success probability calculation, or centroid distance weighted calculation. The corrected probability value is then further corrected by considering meteorological, topographical, and electromagnetic dynamic environmental disturbances to obtain the final probability. A probability function is constructed based on this final probability for adjudication calculations. The probability function takes uniformly distributed random numbers from 0 to 1. ,like If the final probability is less than or equal to the final probability, the ruling action is successful. If the probability is greater than the final probability, the ruling action fails.

[0006] Optionally, loading a corresponding three-dimensional probability density volume of action effects for each wargame unit includes: For each action function of each wargaming unit, the effective spatial range of the action function in three-dimensional space is determined. The effective spatial range extends in all directions from the current position of the wargaming unit to form a continuous three-dimensional geometric region. A three-dimensional effect field model is constructed within the functional effective space. The three-dimensional effect field model is a closed three-dimensional geometric body. The three-dimensional effect field model is discretized into several tiny unit volumes according to a preset spatial resolution, and each tiny unit volume corresponds to a cubic mesh with a fixed side length. Each micro-unit is assigned a success probability value. The success probability value is pre-calculated based on the distance and direction between the micro-unit and the center of the wargaming unit, as well as the equipment performance curve. The assigned micro-unit carries a probability density value. All micro-units together constitute a three-dimensional probability density volume of the action function.

[0007] Optionally, the detection of collisions and intersections between the probability density volumes of opposing and friendly units in three-dimensional space during wargame unit simulations includes: During wargaming simulations, the collision detection process is repeatedly executed at fixed time steps. Within each time step, based on the current position and attitude of the friendly unit, calculate the three-dimensional spatial region occupied by the friendly unit's action probability density volume at the current moment. The three-dimensional spatial region is the set of spatial positions of all tiny unit bodies covered by the probability density volume. Calculate the three-dimensional spatial area occupied by the other party unit based on its location and dimensions; Determine whether there is an intersection between the region occupied by the player's probability density volume and the region occupied by the opponent's unit. The intersection determination is carried out by bounding box intersection test in three-dimensional space or precise mesh overlap detection. If an intersection exists, a collision is determined, triggering the adjudication calculation process corresponding to the action function, and recording the index of the tiny unit cell covered by the intersection volume; if no intersection exists, the adjudication calculation is skipped, and the detection continues for the next time step.

[0008] Optionally, the step of obtaining the probability density values ​​of each tiny unit cell within the volume where the three-dimensional probability density volume intersects with the target collision is performed. include: After determining that a collision and intersection have occurred, obtain an index list of all tiny units covered by the intersection volume. The index list is generated during the collision detection phase. Iterate through each tiny unit in the index list and read the pre-assigned success probability value of the unit from memory or storage structure. The success probability value is the probability density value of the tiny unit. ; During reading, the index order remains unchanged, and the read probability density values ​​are stored sequentially into a temporary array, while the number of tiny units n is recorded. If n equals 1, the unique probability density value is directly used as the base probability value; if n is greater than or equal to 2, the probability correction step is initiated.

[0009] Optionally, the comprehensive calculation method is selected from the average value calculation method, including: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. ; Combine all probability density values ​​into a single total value; The total value is divided equally according to the number of tiny units to obtain the average level value of each probability density value; the average level value is then output as the corrected probability value.

[0010] Optionally, the comprehensive calculation method is selected from the parallel system success probability calculation, including: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. ; For each probability density value, determine its corresponding failure probability value, which is the difference between the completely certain event and the probability density value; accumulate and combine all failure probability values ​​to obtain the overall probability that all tiny units fail simultaneously; then determine the complementary probability of the overall probability, use the complementary probability as the success probability of the parallel system, and output the success probability of the parallel system as the corrected probability value.

[0011] Optionally, the comprehensive calculation method is selected from the centroid distance weighted calculation, including: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, first determine the position of the geometric centroid of the collision intersection volume; for each tiny unit, calculate the spatial interval length from the unit's center point to the geometric centroid; determine the initial weight value of each unit based on the spatial interval length, with units having smaller spatial interval lengths receiving larger initial weight values; normalize the initial weight values ​​of all units so that the sum of the normalized weight values ​​equals 1; weight the probability density value of each unit with the corresponding normalized weight value to obtain the centroid-weighted probability value; output the centroid-weighted probability value as the correction probability value.

[0012] Optionally, the step of considering meteorological, topographical, and electromagnetic dynamic environmental disturbance factors to perform a second correction on the corrected probability value to obtain the final probability includes: Read the meteorological data, terrain data, and electromagnetic environment data at the current simulation time. The meteorological data should include at least the visibility value, the terrain data should include at least the level of ground cover, and the electromagnetic environment data should include at least the level of electromagnetic interference intensity. According to the preset environmental correction rules, the meteorological influence coefficient is determined based on the visibility value, the terrain influence coefficient is determined based on the level of ground cover, and the electromagnetic influence coefficient is determined based on the level of electromagnetic interference intensity. The correction probability value is then adjusted sequentially through the meteorological influence coefficient, the terrain influence coefficient, and the electromagnetic influence coefficient to obtain the adjusted probability value. If the adjusted probability value is greater than the threshold for a completely certain event, it is restricted to the threshold. If the adjusted probability value is less than the threshold for an impossible event, it is restricted to the threshold. The probability value obtained after restriction is taken as the final probability.

[0013] Optionally, the preset environmental correction rules include: The meteorological disturbance coefficient mapping table uses visibility range as an index. Each visibility range corresponds to a fixed value of meteorological disturbance coefficient. The lower the visibility value, the smaller the meteorological disturbance coefficient value, and the higher the visibility value, the larger the meteorological disturbance coefficient value. The terrain disturbance coefficient mapping table uses the level of terrain occlusion as an index. The level of terrain occlusion includes four levels: no occlusion, low occlusion, medium occlusion, and high occlusion. Each level corresponds to a fixed value of terrain disturbance coefficient. The higher the occlusion level, the smaller the value of the terrain disturbance coefficient. The electromagnetic disturbance coefficient mapping table uses the electromagnetic interference intensity level as an index. The electromagnetic interference intensity level includes four levels: no interference, weak interference, medium interference, and strong interference. Each level corresponds to a fixed value of electromagnetic disturbance coefficient. The higher the interference intensity, the smaller the value of the electromagnetic disturbance coefficient. All three mapping tables are pre-stored in the configuration file of the simulation system, and are retrieved by looking up the tables based on the environmental data collected in real time during the simulation.

[0014] Optionally, the decision calculation based on the constructed probability function of the final probability includes: The system calls the built-in uniform random number generation function of the wargaming simulation system. The random number generation function is implemented using either a linear congruential algorithm or a Mason swirl algorithm, and generates an independent and identically distributed random number each time it is called. , The value range of is [0,1], inclusive; the final probability is... With random numbers Perform numerical comparisons; if ≤ If the decision is successful, the wargaming system will execute the subsequent logic corresponding to the successful action, such as recording the hit, deducting the opponent's unit's health, or marking the target as discovered; if > If the decision fails, the output result will be "failure". The wargaming system will then execute the subsequent logic corresponding to the failed action, such as recording the miss, maintaining the opponent's unit status, or keeping the target undiscovered.

[0015] The beneficial effects of this invention are: (1) The computational efficiency is significantly improved. The three-dimensional probability density volume and collision intersection detection mechanism are adopted, and the adjudication calculation is triggered only when spatial overlap occurs. This avoids the high overhead of continuous analytical calculation of all action pairs in traditional simulation methods, and the amount of computation is significantly reduced in large-scale wargame simulations.

[0016] (2) Action details are preserved. By discretizing the action effects into tiny units in three-dimensional space and assigning them probability density values, the non-uniform distribution characteristics of the effects in space are preserved. Compared with the linear queuing-style decision-making of manual wargames, it can more accurately reflect the spatial differences in detection and hit probabilities.

[0017] (iii) Low dependence on hardware and software environment. This invention does not rely on high-performance computing clusters or dedicated simulation platforms. It can complete probability correction and adjudication by using table lookup correction and simple arithmetic operations. The calculation process is lightweight and easy to deploy and promote on ordinary computing devices.

[0018] (4) The probability results are highly reliable. By using three correction methods—average value, success probability of parallel system, or centroid distance weighting—and combining secondary corrections based on meteorological, topographical, and electromagnetic environmental disturbance coefficients, the final probability is closer to the actual battlefield conditions. The random decision-making method ensures the statistical rationality of the simulation results. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the collision calculation method for batch adjudication of wargames based on three-dimensional probability density proposed in this invention. Figure 2 This is a schematic diagram of the three-dimensional probability density volume collision detection and adjudication method for the batch adjudication of wargames based on three-dimensional probability density proposed in this invention. Figure 3 This is a schematic diagram of the probability correction method for the collision calculation wargaming batch adjudication method based on three-dimensional probability density proposed in this invention. Detailed Implementation

[0020] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0021] refer to Figures 1-3 A batch adjudication method for collision calculation in wargames based on three-dimensional probability density includes the following steps: For each wargame unit, a corresponding 3D probability density volume of action effects is loaded. This volume represents the influence range of the action effect in 3D space, and the probability density represents the probability of the effect occurring. Collisions between the probability density volumes of enemy and friendly units in 3D space are detected during simulated actions. The probability density values ​​Pij of each micro-unit within the collision intersection volume of the 3D probability density volume and the target are obtained. When the collision intersection volume contains at least two micro-units, a corrected probability value is calculated from these two probability density values ​​using a probability correction method. This method involves a comprehensive calculation of the two probability density values, with the comprehensive calculation method selected from one of the following: average value calculation, parallel system success probability calculation, or centroid distance weighted calculation. The corrected probability value is then further corrected by considering meteorological, terrain, and electromagnetic dynamic environmental disturbances to obtain the final probability. A probability function is constructed based on the final probability for adjudication calculations. The probability function uses uniformly distributed random numbers from 0 to 1. ,like If the final probability is less than or equal to the final probability, the ruling action is successful. The final probability determines that the action has failed.

[0022] In this embodiment, loading a corresponding three-dimensional probability density volume of action effects for each wargaming unit includes: For each action function of each wargaming unit, the effective spatial range of the action function in three-dimensional space is determined. The effective spatial range extends in all directions from the current position of the wargaming unit to form a continuous three-dimensional geometric region. A three-dimensional effect field model is constructed within the functional effective space. The three-dimensional effect field model is a closed three-dimensional geometric body. The three-dimensional effect field model is discretized into several tiny unit volumes according to a preset spatial resolution, and each tiny unit volume corresponds to a cubic mesh with a fixed side length. Each micro-unit is assigned a success probability value. The success probability value is pre-calculated based on the distance and direction between the micro-unit and the center of the wargaming unit, as well as the equipment performance curve. The assigned micro-unit carries a probability density value. All micro-units together constitute a three-dimensional probability density volume of the action function.

[0023] Specifically, determining the effective spatial range of the action function in three-dimensional space involves: The largest continuous three-dimensional spatial region in which a certain function of a wargame unit (such as radar detection, artillery hit, jamming release) can produce a non-zero effect, either physically or logically.

[0024] Specifically, the three-dimensional effect field model is discretized into several micro-units according to a preset spatial resolution, and each micro-unit corresponds to a cubic mesh with a fixed side length. The size of the micro-units is not set to a fixed absolute value, but is determined by the developers based on the accuracy requirements of the wargaming simulation and the computational resources to pre-set the spatial resolution. The resolution defines each cubic grid with a fixed side length. The specific side length depends on the physical characteristics of the action function (e.g., 1 to 10 meters for optical detection, 50 to 200 meters for artillery coverage) and the scale of the simulation (smaller values ​​for tactical level, larger values ​​for campaign level). At the same time, the computational load needs to be balanced to ensure real-time performance. After the grid resolution parameters are preset in the system configuration file, the bounding box length, width, and height of the 3D effect field model are divided by the resolution to obtain the number of grids in each direction. Each grid is a micro-unit. If the unit at the boundary is not completely filled, the probability density value can be reduced according to the falling proportion. When the target volume is smaller than the unit, the probability density value of the unit can still be read directly.

[0025] In this embodiment, detecting the collision and intersection of the probability density volumes of opposing units and friendly units in three-dimensional space during wargame unit simulation includes: During wargaming simulations, the collision detection process is repeatedly executed at fixed time steps. Within each time step, based on the current position and attitude of the friendly unit, calculate the three-dimensional spatial region occupied by the friendly unit's action probability density volume at the current moment. The three-dimensional spatial region is the set of spatial positions of all tiny unit bodies covered by the probability density volume. Calculate the three-dimensional spatial area occupied by the other party unit based on its location and dimensions; Determine whether there is an intersection between the region occupied by the player's probability density volume and the region occupied by the opponent's unit. The intersection determination is carried out by bounding box intersection test in three-dimensional space or precise mesh overlap detection. If an intersection exists, a collision is determined, triggering the adjudication calculation process corresponding to the action function, and the index of the tiny unit cell covered by the intersection volume is recorded; if no intersection exists, the adjudication calculation is skipped, and the detection continues for the next time step.

[0026] The bounding box intersection test specifically includes: First, determine the axial bounding box boundary of the probability density volume of your own unit at the current moment. This boundary is determined by the minimum and maximum coordinate values ​​of all the tiny units occupied by the probability density volume in three-dimensional space. This gives you the minimum and maximum values ​​of the bounding box in the three axes of horizontal, vertical and horizontal directions. Then, determine the axial bounding box boundary of the area occupied by the opponent's unit in three-dimensional space. Similarly, this gives you the minimum and maximum values ​​of the area in the three axes. The first comparison is performed horizontally: if the maximum value of the user's bounding box in this direction is less than the minimum value of the opponent's bounding box in this direction, or vice versa, then the two bounding boxes have no overlapping area in the horizontal direction, and it is determined that no collision or intersection has occurred, so the test is terminated directly. If there is overlap in the horizontal direction, the comparison continues in the vertical direction: if the maximum value of the user's bounding box in the vertical direction is less than the minimum value of the opponent's bounding box in this direction, or vice versa, then it is determined that no collision or intersection has occurred, so the test is terminated. If there is also overlap in the vertical direction, the comparison continues in the vertical direction: if the maximum value of the user's bounding box in the vertical direction is less than the minimum value of the opponent's bounding box in this direction, or vice versa, then it is determined that no collision or intersection has occurred. Two bounding boxes are considered to intersect only when the projection intervals in the horizontal, vertical, and lateral directions overlap. At this point, a collision intersection is considered possible, which triggers a more refined mesh overlap detection or directly enters the adjudication calculation process. This test only involves numerical comparison operations, with a very small computational load. It is often used in the coarse screening stage to quickly eliminate a large number of non-collision interactions and reduce the computational load of subsequent fine detection.

[0027] Specifically, calculating the three-dimensional spatial region occupied by the probability density volume of friendly units at the current moment includes: Obtain the three-dimensional spatial coordinates of your unit at the current moment. , , ) and attitude angles (heading angle, pitch angle, roll angle); Based on the geometric parameters of the three-dimensional probability density volume corresponding to the action function (such as the radius of a sphere, the length of the semi-axis of an ellipsoid, the angle and direction of a cone, the side length of a cube, etc.), combined with the current position and attitude of the unit, the spatial position set of each tiny unit covered by the probability density volume in the world coordinate system is calculated through coordinate transformation. The specific method is as follows: traverse all grid points in the predefined probability density volume local coordinate system, and transform the local coordinates (relative to the center of the wargame unit) of each tiny unit to the world coordinate system through a rotation matrix and a translation vector to obtain the three-dimensional cube boundary of each tiny unit in the world. , , , , , Finally, the spatial boundaries of all the tiny unit cells are merged to obtain the three-dimensional spatial region occupied by the probability density volume at the current moment. The region can be represented by an axis-aligned bounding box or a set of discrete grid index lists.

[0028] Specifically, calculating the three-dimensional spatial area occupied by the other party unit based on its location and dimensions involves: Obtain the three-dimensional spatial coordinates of the opponent unit at the current moment (usually its geometric center or reference point coordinates), as well as the external dimensions of the opponent unit. The external dimensions can be represented by the length, width, and height of the cuboid bounding box, or by the radius of a sphere, the semi-axis of an ellipsoid, or the set of vertices of a polyhedron. Then, based on the position coordinates, expand along the three axes according to the external dimensions to determine the boundary of the continuous region occupied by the opponent unit in three-dimensional space. For example, if a cuboid bounding box is used, then the calculation... = - Half-length, = +half-length, , , , Similarly, if a sphere is used, the spherical region is obtained with the location as the center and the radius as the size. For complex shapes, a fine mesh model can be pre-built, and each mesh point can be transformed to the world coordinate system through coordinate transformation to form a point cloud set or a mesh index list. The resulting occupied region can be represented by an axis-aligned bounding box, sphere parameters, or discrete voxel set, and is used for collision and intersection detection with the own probability density volume.

[0029] In this embodiment, the probability density values ​​of each tiny unit cell within the volume where the three-dimensional probability density volume intersects with the target collision are obtained. include: After determining that a collision and intersection have occurred, obtain an index list of all tiny units covered by the intersection volume. The index list is generated during the collision detection phase. Iterate through each tiny unit in the index list and read the pre-assigned success probability value of the unit from memory or storage structure. The success probability value is the probability density value of the tiny unit. ; During reading, the index order remains unchanged, and the read probability density values ​​are stored sequentially into a temporary array, while the number of tiny units n is recorded. If n equals 1, the unique probability density value is directly used as the base probability value; if n is greater than or equal to 2, the probability correction step is initiated.

[0030] Specifically, obtaining the index list of all tiny units covered by the intersection volume is as follows: After determining that there is an intersection between the region occupied by the probability density volume of the friendly unit and the region occupied by the opposing unit, the intersection part of the two three-dimensional spatial regions (i.e., the collision intersection volume) is spatially overlapped with the discretized mesh of the probability density volume to determine the position number of all tiny units contained in the intersection volume. Each tiny unit cell has a unique index (i,j,k) in the 3D mesh, where i, j, and k correspond to the mesh numbers in the x, y, and z directions, respectively. By traversing all tiny units within the coverage area of ​​the probability density volume, it is determined whether the cube boundary corresponding to each unit overlaps with the intersection volume. If overlap exists, the index (i,j,k) is added to the list. For cases where the intersection volume partially overlaps with the unit boundary, the unit is still included in the list, and its probability density value is retained as the complete value of the unit itself, without needing to be reduced according to the overlap ratio. The final index list is used for subsequent reading of probability density values. Input.

[0031] Specifically, storing the read probability density values ​​into a temporary array sequentially is as follows: After obtaining the index list of tiny unit cells covered by the intersection volume, the data storage structure of each tiny unit cell is accessed one by one according to the order of the index list, and the pre-assigned function success probability value (i.e., ); Each probability density value is read and placed into a pre-declared linear data structure (i.e., a temporary array) in the order of reading. The array occupies contiguous storage space in memory, and each array element corresponds to the probability density value of a tiny unit cell. Each time a value is stored, the counter is incremented by one, recording the total number of stored values ​​n. The temporary array is used to temporarily store all the probability density values ​​required for this collision and intersection, for subsequent probability correction steps (such as average value calculation, parallel system success probability calculation, or centroid distance weighted calculation) to read and calculate in batches. After this adjudication calculation is completed, the temporary array can be cleared or overwritten for use in the next collision and intersection event.

[0032] In this embodiment, the comprehensive calculation method selected is the average value calculation, including: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. All probability density values ​​are aggregated into a single total value; The total value is divided equally according to the number of tiny units to obtain the average level value of each probability density value; the average level value is then output as the corrected probability value.

[0033] In this embodiment, the comprehensive calculation method selected from the parallel system success probability calculation includes: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. ; For each probability density value, determine its corresponding failure probability value, which is the difference between the completely certain event and the probability density value. By accumulating and combining all the failure probability values, the overall probability of all tiny units failing simultaneously is obtained; then, the complementary probability of the overall probability is determined, and the complementary probability is used as the success probability of the parallel system, and the success probability of the parallel system is output as the correction probability value.

[0034] Specifically, the comprehensive calculation method selected from the success probability calculation of parallel systems is: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. ; Each probability density value is converted into a failure probability value, which is equal to 1 minus... ; Calculate the product of all failure probabilities, and denote the product as Q; Subtracting Q from the value 1 gives the success probability C of the parallel system. The success probability C of the parallel system is output as the corrected probability value.

[0035] In this embodiment, the comprehensive calculation method selected from the centroid distance weighted calculation includes: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, first determine the position of the geometric centroid of the collision intersection volume; for each tiny unit, calculate the spatial interval length from the center point of the unit to the geometric centroid. The initial weight value of each unit is determined based on the spatial interval length of each unit; the smaller the spatial interval length, the larger the initial weight value is obtained for each unit. The initial weight values ​​of all units are normalized so that the sum of the normalized weight values ​​equals 1; the probability density value of each unit is weighted and combined with the corresponding normalized weight value to obtain the centroid weighted probability value. The centroid-weighted probability value is output as the corrected probability value.

[0036] Specifically, the comprehensive calculation method selected from the centroid distance weighted calculation is: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, first calculate the geometric centroid coordinates G of the collision intersection volume. Each coordinate component of G is equal to the arithmetic mean of the coordinates of the center points of all tiny units. For each tiny unit cell i, calculate the Euclidean distance from its center point to its centroid G. ; Set the initial weights for each tiny unit cell equal to 1 divided by If di equals 0, then the weight of the unit cell is set to a preset upper limit. Calculate the sum of all initial weights. Calculate the normalized weights for each unit cell. equal Divide by ; The centroid-weighted probability value H is calculated to be equal to the probability density value of each unit cell. Multiply by the corresponding normalized weight The sum of all subsequent additions; The centroid-weighted probability value H is output as the corrected probability value.

[0037] In this embodiment, the final probability is obtained by secondary correction of the corrected probability value considering meteorological, topographical, and electromagnetic dynamic environmental disturbance factors, including: Read the meteorological data, terrain data, and electromagnetic environment data at the current simulation time. The meteorological data should include at least the visibility value, the terrain data should include at least the level of ground cover, and the electromagnetic environment data should include at least the level of electromagnetic interference intensity. According to the preset environmental correction rules, the meteorological influence coefficient is determined based on the visibility value, the terrain influence coefficient is determined based on the level of ground cover, and the electromagnetic influence coefficient is determined based on the level of electromagnetic interference intensity. The correction probability value is then adjusted sequentially through the meteorological influence coefficient, the terrain influence coefficient, and the electromagnetic influence coefficient to obtain the adjusted probability value. If the adjusted probability value is greater than the threshold for a completely certain event, the probability value is restricted to the threshold. If the adjusted probability value is less than the threshold for an impossible event, the probability value is restricted to the threshold. The probability value obtained after restriction is taken as the final probability.

[0038] In this embodiment, the preset environmental correction rules include: The meteorological disturbance coefficient mapping table uses visibility range as an index. Each visibility range corresponds to a fixed value of meteorological disturbance coefficient. The lower the visibility value, the smaller the meteorological disturbance coefficient value, and the higher the visibility value, the larger the meteorological disturbance coefficient value. The terrain disturbance coefficient mapping table uses the level of terrain occlusion as an index. The level of terrain occlusion includes four levels: no occlusion, low occlusion, medium occlusion, and high occlusion. Each level corresponds to a fixed value of terrain disturbance coefficient. The higher the occlusion level, the smaller the value of the terrain disturbance coefficient. The electromagnetic disturbance coefficient mapping table uses the electromagnetic interference intensity level as an index. The electromagnetic interference intensity level includes four levels: no interference, weak interference, medium interference, and strong interference. Each level corresponds to a fixed value of electromagnetic disturbance coefficient. The higher the interference intensity, the smaller the value of the electromagnetic disturbance coefficient. All three mapping tables are pre-stored in the configuration file of the simulation system, and are retrieved by looking up the tables based on the environmental data collected in real time during the simulation.

[0039] In this embodiment, the decision calculation based on the probability function constructed from the final probability includes: The system calls the built-in uniform random number generation function of the wargaming simulation system. The random number generation function is implemented using either a linear congruential algorithm or a Mason swirl algorithm, and generates an independent and identically distributed random number each time it is called. , The value range of is [0,1], inclusive; the final probability is... With random numbers Perform numerical comparisons; if ≤ If the decision is successful, the wargaming system will execute the subsequent logic corresponding to the successful action, such as recording the hit, deducting the opponent's unit's health, or marking the target as discovered; if > If the decision fails, the output result will be "failure". The wargaming system will then execute the subsequent logic corresponding to the failed action, such as recording the miss, maintaining the opponent's unit status, or keeping the target undiscovered.

[0040] The specific follow-up logic for a failed decision and the wargaming system's failure to execute an action is as follows: For detection operations, failure means that the target was not detected. The system maintains the state that the other party unit is undetected in the field of vision, does not update the situation map, and does not trigger alarms or tracking records. For a hit, a failure means that the target was not hit. The system keeps the enemy unit in good condition, does not deduct health or armor, and records the information of this miss for subsequent statistical analysis. For other action functions (such as interference, identification, etc.), failure indicates that the action did not produce the expected effect. The system does not change the target's state attributes, nor does it change the judgment result of the interaction relationship between the two parties.

[0041] Example 1: This method was practically validated in a company-level war game simulation. The simulation area was a hilly region, 12 kilometers long from east to west and 8 kilometers wide from north to south. The Red Force had 2 main battle tanks, 1 infantry fighting vehicle, and 1 armed reconnaissance vehicle. The Blue Force had 1 anti-tank missile launcher, 1 self-propelled anti-aircraft gun, and 1 radar reconnaissance vehicle. Normally, using manual wargames, a single artillery engagement decision requires consulting an average of 3 tables and takes about 40 seconds, often resulting in incorrect rows being consulted. Running a full-scale simulation on a regular laptop, the game starts to lag when there are more than 8 units, making it unplayable. Our method aims to solve these two problems: it needs to be both fast and accurate.

[0042] In the specific simulation, a three-dimensional probability density volume is first assigned to the "armor-piercing shell firing" of the Red Force tank. This density volume starts from the muzzle and extends forward along the direction of the gun barrel, reaching a maximum range of 2500 meters. It is divided into small squares with dimensions of 4 meters each, and each square stores a hit probability. For example, at a distance of 800 meters, the hit probability of the square directly in the center of the aiming point is 0.88; shifting one square to the left, the probability drops to 0.76; shifting another square to the left, it is only 0.61. The further away, the smaller the number: at 1800 meters, the probability of the center square is 0.52, and the probability of the edge squares is only 0.34. The Red Force reconnaissance vehicle also has a density volume, which is a 90-degree fan-shaped detection area in front. The detection probability is 0.97 at the closest 300 meters, dropping to 0.21 at the furthest 900 meters, and gradually decreasing with distance.

[0043] Three minutes after the simulation began, the Red Team tank detected the Blue Team anti-tank missile vehicle at a distance of 1400 meters. The system detected that the Red Team tank's hit probability density cube collided with the space occupied by the Blue Team vehicle. The collided area contained five density cubes with hit probabilities of 0.63, 0.66, 0.60, 0.65, and 0.62, respectively. Since there were more than one small cube, the system used an average method to add these probability values ​​together and divide by 5, obtaining 0.632 as the base correction probability.

[0044] Next, let's look at the environment. There was a level 3 wind, visibility was approximately 1900 meters, and the meteorological coefficient was 0.97. The terrain was a gentle slope with little obstruction, so the terrain coefficient was 0.99. There was no electronic interference, so the electromagnetic coefficient was 1.00. Multiplying these three coefficients together and then multiplying by the base correction probability of 0.632 gives a final probability of 0.607. Finally, the system drew a random number 0.44, which is less than 0.607, and determined it was a hit. The blue team's anti-tank missile vehicle was hit and exited the simulation.

[0045] The entire adjudication process, from collision detection to result, took 13 milliseconds. In the same scenario, manual wargaming would take 42 seconds (first looking up the base value for distance versus hit rate, then the weather correction table, then the terrain correction table, and then flipping through pages to find the random number table), a difference of over 3000 times. Using full-scale simulation software, a single adjudication takes 220 milliseconds, still more than ten times slower than our method.

[0046] To make the results clearer, we conducted a load test, running simulations with 6, 10, 15, 20, 25, and 30 units respectively, and recording the total decision time for each. The results are shown in the table below.

[0047] Table 1: Comparison of total adjudication time for the three methods at different scales

[0048] Table 1 shows the total time required to complete all decisions for manual wargaming, full-solution simulation, and the method of this invention at six different simulation scales. As the simulation scale gradually increases from 6 units to 30 units, the number of collision decisions increases from 28 to 268. The increase in the number of decisions stems from the superlinear growth in the detection and hit interactions between units as the number of units increases.

[0049] The adjudication mechanism in manual wargames is based on a fixed process of table lookup. The time taken for a single adjudication is not affected by the number of participating units and remains stable at around 0.75 seconds. Therefore, the total time is strictly linearly related to the number of adjudications: 28 adjudications take 21.0 seconds, and 268 adjudications take 201.0 seconds. 201.0 seconds is equivalent to 3.35 minutes. In dynamic combat simulations, delays will cause commanders to frequently wait for adjudication results, forcing the simulation rhythm to be interrupted.

[0050] The full-scale simulation employs a continuous mathematical modeling method, solving the differential equations for each pair of interactions in real time. It performs well with a small number of units: 6 units and 28 decisions take only 5.0 seconds, with a single decision taking approximately 0.18 seconds. However, as the number of units increases, the number of interaction pairs rises sharply, leading to queuing and congestion of computational resources. With 10 units, the total time rises to 13.5 seconds, with a single decision taking approximately 0.26 seconds; with 20 units, the total time is 71.6 seconds, with a single decision taking approximately 0.53 seconds; and with 30 units, the total time is 198.3 seconds, with a single decision taking approximately 0.74 seconds. At this point, the total time of the full-scale simulation approaches that of manual wargames, losing its computational efficiency advantage and failing to meet the requirements of real-time simulation.

[0051] The method of this invention maintains millisecond-level single-execution time across all test scales. For 6 units, the total time is 0.33 seconds, with a single execution time of approximately 11.8 milliseconds; for 10 units, the total time is 0.78 seconds, with a single execution time of approximately 15.0 milliseconds; for 15 units, the total time is 1.60 seconds, with a single execution time of approximately 18.0 milliseconds; for 20 units, the total time is 3.04 seconds, with a single execution time of approximately 22.5 milliseconds; for 25 units, the total time is 5.43 seconds, with a single execution time of approximately 28.0 milliseconds; and for 30 units, the total time is 9.38 seconds, with a single execution time of approximately 35.0 milliseconds. Compared to manual wargames, the total time of this invention is only 4.7% of its time at a scale of 30 units; compared to full solution simulation, the total time of this invention is only 4.7% of its time. As the scale increases, the speed-up advantage of this invention becomes more significant: from 6 units to 30 units, the total time for manual wargames increases by 9.6 times, and the time for full solution simulation increases by 39.7 times, while this invention only increases by 28.4 times, but the absolute increase is only 9.05 seconds.

[0052] Besides improving efficiency, this invention also preserves the details of the probability distribution in three-dimensional space. In hit determination at a distance of 1400 meters, when the aiming point is aligned with the target center, the intersection volume contains high-probability microcells, and after weighted correction, the final probability can reach 0.71; when the aiming point is offset by one grid, the intersection volume is mixed with low-probability cells, and the final probability drops to 0.53. Manual wargames cannot distinguish between these two cases, giving a fixed hit rate of 0.60. Comparison with live-fire statistics shows that the method of this invention has a 92% agreement rate in distinguishing different aiming deviations, while manual wargames only achieve 67%. These data demonstrate that this invention significantly improves determination efficiency while ensuring the reliability of probability calculations and the ability to preserve details.

[0053] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A batch adjudication method for collision calculation wargaming based on three-dimensional probability density, characterized in that, Includes the following steps: For each wargame unit, a corresponding 3D probability density volume of action effects is loaded. The 3D probability density volume represents the range of influence of the action effect in 3D space, and the probability density represents the probability of the effect occurring. The collision and intersection of the action probability density volumes of enemy units and friendly units in 3D space are detected during the wargame unit's simulated actions. Obtain the probability density values ​​of each tiny unit cell within the volume where the three-dimensional probability density volume intersects with the target. ; When the collision intersection volume contains at least two micro-units, a probability correction method is used to calculate the corrected probability value from at least two probability density values. The probability correction method is to perform a comprehensive calculation on at least two probability density values. The comprehensive calculation method is selected from one of the following: average value calculation, parallel system success probability calculation, or centroid distance weighted calculation. The corrected probability value is then further corrected by considering meteorological, topographical, and electromagnetic dynamic environmental disturbance factors to obtain the final probability. The adjudication is performed based on a probability function constructed from the final probability, where the probability function takes a uniformly distributed random number between 0 and 1. ,like If the final probability is less than or equal to the final probability, the ruling action is successful. If the probability is greater than the final probability, the ruling action fails.

2. The collision calculation wargaming batch adjudication method based on three-dimensional probability density according to claim 1, characterized in that, The loading of the corresponding three-dimensional probability density volume of action effects for each wargaming unit includes: For each action function of each wargaming unit, the effective spatial range of the action function in three-dimensional space is determined. The effective spatial range extends in all directions from the current position of the wargaming unit to form a continuous three-dimensional geometric region. A three-dimensional effect field model is constructed within the functional effective space. The three-dimensional effect field model is a closed three-dimensional geometric body. The three-dimensional effect field model is discretized into several tiny unit volumes according to a preset spatial resolution, and each tiny unit volume corresponds to a cubic mesh with a fixed side length. Each micro-unit is assigned a success probability value. The success probability value is pre-calculated based on the distance and direction between the micro-unit and the center of the wargaming unit, as well as the equipment performance curve. The assigned micro-unit carries a probability density value. All micro-units together constitute a three-dimensional probability density volume of the action function.

3. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The detection of collisions and intersections between the probability density volumes of opposing and friendly units in three-dimensional space during wargame unit simulations includes: During wargaming simulations, the collision detection process is repeatedly executed at fixed time steps. Within each time step, based on the current position and attitude of the friendly unit, calculate the three-dimensional spatial region occupied by the friendly unit's action probability density volume at the current moment. The three-dimensional spatial region is the set of spatial positions of all tiny unit bodies covered by the probability density volume. Calculate the three-dimensional spatial area occupied by the other party unit based on its location and dimensions; To determine whether there is an intersection between the region occupied by the player's probability density volume and the region occupied by the opponent's unit, the intersection determination is performed using three-dimensional space. Cross-section testing or mesh overlap detection; If an intersection exists, a collision is determined, triggering the adjudication calculation process corresponding to the action function, and recording the index of the tiny unit volume covered by the intersection volume; If there is no intersection, skip the decision calculation and continue with the detection at the next time step.

4. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The probability density values ​​of each micro-unit within the volume where the three-dimensional probability density volume intersects with the target collision are obtained. include: After determining that a collision and intersection have occurred, obtain an index list of all tiny units covered by the intersection volume. The index list is generated during the collision detection phase. Iterate through each tiny unit in the index list and read the pre-assigned success probability value of the unit from memory or storage structure. The success probability value is the probability density value of the tiny unit. ; During reading, the index order remains unchanged. The read probability density values ​​are stored sequentially into a temporary array, and the number of tiny units n is recorded. If n equals 1, the unique probability density value is directly used as the base probability value. If n is greater than or equal to 2, the probability correction step is entered.

5. The collision calculation wargaming batch adjudication method based on three-dimensional probability density according to claim 1, characterized in that, The comprehensive calculation method selected from average value calculation includes: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. All probability density values ​​are aggregated into a total value. The total value is then divided equally according to the number of tiny units to obtain the average level value of each probability density value. The average level value is then output as the corrected probability value.

6. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The comprehensive calculation method selected from the success probability calculation of parallel systems includes: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, all n probability density values ​​are read. For each probability density value, determine its corresponding failure probability value, which is the difference between the completely certain event and the probability density value. By accumulating and combining all the failure probability values, the overall probability of all tiny units failing simultaneously is obtained. Then, the complementary probability of the overall probability is determined, and the complementary probability is used as the success probability of the parallel system. The success probability of the parallel system is output as the correction probability value.

7. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The comprehensive calculation method selected from the centroid distance weighted calculation includes: When the number of tiny units n covered by the collision intersection volume is greater than or equal to 2, first determine the position of the geometric centroid of the collision intersection volume, and for each tiny unit, calculate the spatial interval length from the center point of the unit to the geometric centroid. The initial weight value of each unit is determined based on the spatial interval length of each unit. Units with smaller spatial interval lengths receive larger initial weight values. The initial weight values ​​of all units are normalized so that the sum of the normalized weight values ​​equals the value 1. The probability density value of each unit cell is weighted and combined with the corresponding normalized weight value to obtain the centroid-weighted probability value, which is then output as the corrected probability value.

8. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The process of considering meteorological, topographical, and electromagnetic dynamic environmental disturbances to perform a secondary correction on the probability value to obtain the final probability includes: Read the meteorological data, terrain data, and electromagnetic environment data at the current simulation time. The meteorological data should include at least the visibility value, the terrain data should include at least the level of ground cover, and the electromagnetic environment data should include at least the level of electromagnetic interference intensity. According to the preset environmental correction rules, the meteorological influence coefficient is determined based on the visibility value, the terrain influence coefficient is determined based on the level of ground cover, and the electromagnetic influence coefficient is determined based on the level of electromagnetic interference intensity. The correction probability value is then adjusted sequentially through the meteorological influence coefficient, the terrain influence coefficient, and the electromagnetic influence coefficient to obtain the adjusted probability value. If the adjusted probability value is greater than the threshold for a completely certain event, it is restricted to the threshold. If the adjusted probability value is less than the threshold for an impossible event, it is restricted to the threshold. The probability value obtained after restriction is taken as the final probability.

9. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The preset environmental correction rules include: The meteorological disturbance coefficient mapping table uses visibility range as an index. Each visibility range corresponds to a fixed value of meteorological disturbance coefficient. The lower the visibility value, the smaller the meteorological disturbance coefficient value, and the higher the visibility value, the larger the meteorological disturbance coefficient value. The terrain disturbance coefficient mapping table uses the level of terrain occlusion as an index. The level of terrain occlusion includes four levels: no occlusion, low occlusion, medium occlusion, and high occlusion. Each level corresponds to a fixed value of terrain disturbance coefficient. The higher the occlusion level, the smaller the value of the terrain disturbance coefficient. The electromagnetic disturbance coefficient mapping table uses the electromagnetic interference intensity level as an index. The electromagnetic interference intensity level includes four levels: no interference, weak interference, medium interference, and strong interference. Each level corresponds to a fixed value of electromagnetic disturbance coefficient. The higher the interference intensity, the smaller the value of the electromagnetic disturbance coefficient. All three mapping tables are pre-stored in the configuration file of the simulation system, and are retrieved by looking up the tables based on the environmental data collected in real time during the simulation.

10. The method for batch adjudication of collision calculation wargaming based on three-dimensional probability density according to claim 1, characterized in that, The decision calculation based on the constructed probability function of the final probability includes: The system calls the built-in uniform random number generation function of the wargaming simulation system. The random number generation function is implemented using either a linear congruential algorithm or a Mason swirl algorithm, and generates an independent and identically distributed random number each time it is called. , The value range of is [0,1], inclusive; Final probability With random numbers Perform numerical comparisons; like Less than If the decision is successful, the wargaming system will execute the subsequent logic corresponding to the successful action, such as recording the hit, deducting the opponent's unit's health, or marking the target as discovered. like Greater than If the decision fails, the output result will be "failure". The wargaming system will then execute the subsequent logic corresponding to the failed action, such as recording the miss, maintaining the opponent's unit status, or keeping the target undiscovered.

Images