Underwater combat situation simulation deduction method and platform under complex marine hydrographic environment

Abstract

The application relates to the technical field of underwater combat simulation deduction, and particularly discloses a method and a platform for underwater combat situation simulation deduction under a complex marine hydrological environment, which comprises the following steps: acquiring a sound speed profile and a density profile; extracting a sound line propagation path from the sound speed profile to construct an acoustic connectivity graph; identifying mutually disconnected components in the acoustic connectivity graph caused by sound shadow zones, and defining each connected component as a dominant transmission domain; extracting a thermocline and a density jump layer boundary from the density profile, and constructing a maneuvering constraint graph inside each dominant transmission domain; inputting the acoustic connectivity graph and the maneuvering constraint graph into a combat effectiveness prediction model to obtain an intrinsic support degree; solving a Poisson equation inside each dominant transmission domain to obtain a block-continuous dominant potential field; generating an action strategy in the gradient direction of the dominant potential field; and triggering a detection or attack decision when crossing the boundary of the dominant transmission domain. The application solves the technical problem that the existing potential field method ignores underwater dominant discontinuity.

Description

Technical Field

[0001] This invention relates to the field of underwater combat simulation and deduction technology, and in particular to a method and platform for underwater combat situation simulation and deduction in complex marine hydrological environments. Background Technology

[0002] In underwater combat environments, sound wave propagation is affected by the sound speed profile, forming convergence zones and sound shadow zones. The ocean thermocline and density strata significantly constrain submarine maneuvering and sonar detection. Existing underwater combat situation simulation methods mostly employ artificial potential field methods or improved potential energy models to perform path planning and situation assessment for combat entities. They characterize the distribution of battlefield advantages by constructing globally continuous gravitational and repulsive fields or potential energy fields, and guide the movement of combat entities based on the potential energy gradient direction.

[0003] However, the aforementioned existing technologies all overlook a fundamental technical problem: underwater combat advantages are inherently disconnected due to the presence of acoustic shadow zones. Specifically, acoustic shadow zones block sound ray propagation, disrupting acoustic connectivity between different spatial regions, preventing combat advantages from being continuously transferred across the entire space. Existing potential field methods assume a globally continuous advantage field, which contradicts actual underwater physics. Therefore, they cannot correctly identify the advantage transfer domain isolated by acoustic shadow zones, nor can they automatically generate detection or attack commands at the boundary locations of abrupt advantage changes. This results in significant deviations between the situational analysis results output by the simulation system and the actual underwater physical environment.

[0004] Therefore, this invention proposes a method and platform for simulating underwater combat situations in complex marine hydrological environments. Summary of the Invention

[0005] This invention provides a method and platform for simulating underwater combat situations in complex marine hydrological environments. By identifying the discontinuity of underwater advantage transmission and transforming physical boundaries into command trigger points, it solves the technical defect of existing potential energy field methods that cannot accurately characterize the distribution of segmented advantages due to ignoring the obstruction of the acoustic shadow zone, and realizes simulation and deduction that conforms to the laws of underwater physics.

[0006] This invention provides a method for simulating and extrapolating underwater combat situations in complex marine hydrological environments, including: Obtain sound velocity and density profiles for underwater combat zones; The sound ray propagation path is extracted from the sound speed profile, and an acoustic connectivity graph is constructed using spatial location pairs that have direct sound rays or single-reflection sound rays as edges. Identify the connected components in the acoustic connectivity graph, where the presence of the sound shadow region causes different connected components to be disconnected from each other. The acoustic connectivity graph is decomposed into connected components. Each connected component is defined as a dominant transfer domain. The boundary between dominant transfer domains corresponds to the location of the sound shadow zone. The dominant potential energy field is discontinuous between different dominant transfer domains. Extract the thermocline boundary and density clinch boundary from the density profile, and construct a kinematic constraint map within each dominant transfer domain. By inputting the acoustic connectivity graph and maneuver constraint graph into the combat effectiveness prediction model, the intrinsic support of each spatial location for underwater combat operations can be obtained. Within each dominant transfer domain, the Poisson equation is solved using the intrinsic support as the potential energy value to obtain a block-continuous dominant potential energy field. Using the gradient direction of the advantageous potential energy field as the direction of movement, an action strategy is generated for each combat entity. When the movement path of a combat entity crosses the boundary of the superiority transfer domain, a detection command or an attack command is triggered at the crossing point. When the preset simulation termination condition is met, the simulation result containing all triggered instructions is output. When the preset simulation termination condition is not met and the preset environment update time is reached, update the sound velocity profile and density profile, and return to the step of extracting the sound ray propagation path from the sound velocity profile.

[0007] Preferably, identifying connected components in an acoustic connectivity graph includes: Using each spatial location as a sound source point, a preset number of sound rays with different initial emission angles are emitted using a ray tracing algorithm; Based on whether each sound ray is reflected by the sea surface or the seabed during its propagation, the sound ray connectivity type between each spatial location pair is recorded. The sound ray connectivity types include direct connectivity, single-reflection connectivity, and no connectivity. When there is no direct connection and no single reflection connection between two spatial locations, these two spatial locations are marked as sound-shadow zone blocking relationships; All spatial pairs with sound shadow zone blocking relationships are considered as sound shadow zone blocking relationships. Based on the sound shadow zone blocking relationships, disconnected connected components in the acoustic connectivity graph are identified.

[0008] Preferably, constructing a maneuver constraint graph within each advantage transfer domain includes: Calculate the temperature gradient of the density profile along the vertical direction, and mark the continuous depth interval where the temperature gradient exceeds the preset temperature gradient threshold as a thermocline. The spatial region located above the thermocline and above the density climax is marked as the upper mixing region; the depth region located below the thermocline and below the density climax is marked as the deep isodense region; and the spatial region located between the thermocline and the density climax is marked as the climax transition region. A maneuver constraint map is generated using the region type to which each spatial location belongs as the maneuver constraint information.

[0009] Preferably, obtaining the intrinsic support for underwater operations at each spatial location includes: The dominant transit domain identifier of each spatial location in the acoustic connectivity graph is used as a constraint condition for the intrinsic support calculation. For sonar detection operations, based on the judgment that there is a direct sound ray or a single reflected sound ray reaching the target area from the current spatial location, and the judgment that the current spatial location is located below the thermocline, a nonlinear weighting function is used to calculate the intrinsic support of each spatial location for the sonar detection operation. For submarine maneuvering operations, based on the judgment results of whether the current spatial position is located in the transition zone between different layers and whether the current spatial position is located in the sound shadow zone, the intrinsic support degree of each spatial position for maneuvering stealth is calculated using the fuzzy membership function. For torpedo attack operations, the intrinsic support of each spatial location for the torpedo attack operation is calculated based on the judgment results of the location of the sound convergence zone from the current spatial location to the target area, and the judgment results of whether the current spatial location has a preset maneuver space range. For underwater acoustic communication operations, based on the judgment results of whether there is a stable acoustic ray channel between the current spatial location and the communication target, and the judgment results of whether there is multipath interference, the signal-to-noise ratio prediction model is used to calculate the intrinsic support of each spatial location for the underwater acoustic communication operation. The calculated intrinsic support is grouped and stored according to the dominant transit domain identifier to ensure that the intrinsic support between different dominant transit domains does not participate in cross-domain coupling calculation.

[0010] Preferred methods for identifying crossing points in the boundary of the advantage transfer domain include: The acoustic connected graph is traversed using either breadth-first search or depth-first search to identify all maximal connected subgraphs. Each maximally connected subgraph is mapped to a dominant transit region; Record the boundary positions between different dominant transmission domains, and the boundary positions correspond to the acoustic shadow region in the acoustic connectivity graph; The boundary location of each dominant transfer domain is overlaid with underwater topographic data, and the mutation rate of intrinsic support on both sides of the boundary is calculated. When the mutation rate exceeds the preset mutation threshold, the corresponding boundary segment is marked as a crossing point.

[0011] Preferably, within each dominant transfer domain, the Poisson equation is solved using the intrinsic support as the potential energy value to obtain a block-continuous dominant potential energy field, including: Within each dominance transfer domain, the intrinsic support value of each spatial location is used as the source term of the Poisson equation; The boundary of the dominant transfer domain is used as the first type of boundary condition, and the potential energy value on the boundary is set as the intrinsic support value of the adjacent spatial location inside the boundary. The Poisson equation is discretized into a system of linear equations using the finite difference method. The multigrid method is used to solve the linear equations to obtain a block-continuous dominant potential energy field.

[0012] Preferably, triggering probe or attack commands at the crossing point includes: Record the first superiority transfer domain identifier before the combat entity crosses the superiority transfer domain boundary and the second superiority transfer domain identifier after crossing it. Extract the first eigensupport of the nearest position to the boundary within the first advantage transfer domain and the second eigensupport of the nearest position to the boundary within the second advantage transfer domain from the hydro-operation coupling tensor. Calculate the ratio of the first intrinsic support to the second intrinsic support; When the ratio exceeds the preset detection threshold, it is determined that the area has been crossed, triggering an active sonar detection command. When the ratio is lower than the preset concealment threshold, it is determined that a passive listening command or an evasion maneuver command will be triggered after crossing the area. When the ratio is between the preset concealment threshold and the preset detection threshold, it is determined that the area has been crossed and a regular detection command is triggered.

[0013] Preferably, the action strategy generated for each combat entity, using the gradient direction of the advantageous potential energy field as the direction of movement, includes: Within each advantage transfer domain, multiple candidate movement paths are generated by integrating along the negative gradient direction of the advantage potential field. Calculate the second derivative of the potential energy gradient at each point on each candidate movement path, solve the potential energy curvature based on the second derivative of the potential energy gradient at each point, and identify points whose potential energy curvature exceeds a preset curvature threshold as maneuver adjustment points. When there is a bifurcation point, calculate the matching degree between the potential energy gradient in each bifurcation direction and the maneuver constraint diagram, and select the direction with the highest matching degree as the main movement direction. When the paths of multiple combat entities intersect at the same spatial location, the path allocation order is determined based on the mission priority of each combat entity and the potential energy value of its dominant potential energy field at the corresponding intersecting spatial location. Based on the main direction of movement and the order of path allocation, generate the action strategy for each combat entity.

[0014] Preferably, obtaining the sound speed profile and density profile of the underwater combat zone includes: Collect satellite remote sensing data of sea surface temperature, Argo buoy temperature, salinity and depth profile data, real-time data from underwater sensor network, and historical hydrological observation data; All collected data are interpolated to the same spatial grid. For each spatial grid point, calculate the temperature gradient in the vertical direction between the satellite remote sensing sea surface temperature data and the real-time data from the underwater sensor network; When the temperature gradient exceeds the preset gradient threshold, the inversion parameters of the satellite remote sensing sea surface temperature data are calibrated in reverse using the real-time data of the underwater sensor network. The calibrated satellite remote sensing sea surface temperature data is fused with Argo buoy temperature, salinity, and depth profile data to generate a three-dimensional temperature, salinity, and depth field. Sound velocity profiles and density profiles were extracted from a three-dimensional thermo-salinity field.

[0015] This invention provides an underwater combat situation simulation platform for complex marine hydrological environments, comprising: The acoustic connectivity graph construction module is used to obtain the sound speed profile of the underwater combat area, extract the sound ray propagation path from the sound speed profile, and construct the acoustic connectivity graph with the spatial location pairs of the sound ray that have direct access or first reflection as edges. The connected component decomposition module is used to identify connected components in the acoustic connected graph. Due to the existence of the sound shadow region, different connected components are not connected to each other. The connected component decomposition is performed on the acoustic connected graph, and each connected component is defined as a dominant transfer domain. The boundary between dominant transfer domains corresponds to the location of the sound shadow region. The dominant potential energy field is discontinuous between different dominant transfer domains. The maneuver constraint map construction module is used to obtain the density profile of the underwater combat area, extract the thermocline boundary and density cascade boundary from the density profile, and construct the maneuver constraint map within each advantage transfer domain. The effectiveness prediction module is used to input the acoustic connectivity graph and maneuver constraint graph into the combat effectiveness prediction model to obtain the intrinsic support of each spatial location for underwater combat operations. The potential energy field solution module is used to solve the Poisson equation with the intrinsic support as the potential energy value within each dominant transfer domain to obtain a block-continuous dominant potential energy field. The strategy generation module is used to generate action strategies for each combat entity using the gradient direction of the advantageous potential energy field as the movement direction. When the movement path of the combat entity crosses the boundary of the advantage transfer domain, a detection command or attack command is triggered at the crossing point. When the preset simulation termination condition is met, the simulation result containing all triggered commands is output. When the preset simulation termination condition is not met and the preset environment update time is reached, update the sound velocity profile and density profile, and return to the step of extracting the sound ray propagation path from the sound velocity profile.

[0016] The beneficial effects of this invention compared to existing technologies are as follows: This invention overcomes the fundamental technical defect of existing potential energy field methods that neglect the underwater acoustic shadow zone, leading to the disconnectivity of advantages. Existing technologies, such as the artificial potential field method and improved potential energy models, assume a globally continuous distribution of battlefield advantages. However, in the underwater combat environment, the acoustic shadow zone obstructs sound ray propagation, causing the advantage to be divided into multiple unconnected independent regions. This invention, for the first time, identifies and utilizes this physical law, dividing the combat space into advantage transfer domains through acoustic connectivity graph connectivity component decomposition. The potential energy field is solved block by block within each domain, achieving accurate modeling of the disconnectivity of underwater advantages. Based on this, this invention automatically triggers detection or attack commands at the boundaries of the advantage transfer domains, transforming physical boundaries into command trigger points. This significantly improves the consistency between simulation results and the real underwater physical environment, and the accuracy of commands triggered at the boundaries of the simulation system is significantly improved, providing a reliable physical model basis for underwater combat simulation.

[0017] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in this application.

[0018] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. 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 underwater combat situation simulation and deduction method in a complex marine hydrological environment according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the partitioning of the dominant transfer domain and the solution of the potential field under acoustic connectivity constraints in an embodiment of the present invention; Figure 3 This is an architecture diagram of an underwater combat situation simulation and deduction platform in a complex marine hydrological environment, as described in this embodiment of the invention. Detailed Implementation

[0020] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0021] like Figure 1 , Figure 2 , Figure 3 As shown, this invention provides an embodiment of an underwater combat situation simulation and deduction method for complex marine hydrological environments, comprising: Obtain sound velocity and density profiles for underwater combat zones; The sound ray propagation path is extracted from the sound speed profile, and an acoustic connectivity graph is constructed using spatial location pairs that have direct sound rays or single-reflection sound rays as edges. Identify the connected components in the acoustic connectivity graph, where the presence of the sound shadow region causes different connected components to be disconnected from each other. The acoustic connectivity graph is decomposed into connected components. Each connected component is defined as a dominant transfer domain. The boundary between dominant transfer domains corresponds to the location of the sound shadow zone. The dominant potential energy field is discontinuous between different dominant transfer domains. Extract the thermocline boundary and density clinch boundary from the density profile, and construct a kinematic constraint map within each dominant transfer domain. By inputting the acoustic connectivity graph and maneuver constraint graph into the combat effectiveness prediction model, the intrinsic support of each spatial location for underwater combat operations can be obtained. Within each dominant transfer domain, the Poisson equation is solved using the intrinsic support as the potential energy value to obtain a block-continuous dominant potential energy field. Using the gradient direction of the advantageous potential energy field as the direction of movement, an action strategy is generated for each combat entity. When the movement path of a combat entity crosses the boundary of the superiority transfer domain, a detection command or an attack command is triggered at the crossing point. When the preset simulation termination condition is met, the simulation result containing all triggered instructions is output. When the preset simulation termination condition is not met and the preset environment update time is reached, update the sound velocity profile and density profile, and return to the step of extracting the sound ray propagation path from the sound velocity profile.

[0022] In this embodiment, sound velocity and density profiles of the underwater combat zone are acquired. The underwater combat zone refers to the seawater space covered by the simulation, which is defined by longitude, latitude, and depth intervals. The sound velocity profile is the vertical distribution data sequence of sound velocity varying with depth at each horizontal grid point within the underwater combat zone. The density profile is the vertical distribution data sequence of seawater density varying with depth at each horizontal grid point within the underwater combat zone.

[0023] In this embodiment, sound ray propagation paths are extracted from the sound speed profile, and an acoustic connectivity graph is constructed using pairs of spatial locations where there are direct sound rays or first-reflection sound rays as edges. A sound ray propagation path refers to the curved trajectory followed by a sound wave emitted from a sound source as it propagates through seawater; this trajectory is calculated from the sound speed profile using ray tracing. The acoustic connectivity graph is an undirected graph structure where nodes represent discrete spatial locations within the underwater operational area, and edges represent the existence of direct sound rays or first-reflection sound rays between two spatial locations.

[0024] In this embodiment, connected components in the acoustic connectivity graph are identified, where the presence of shadow regions prevents different connected components from being interconnected. A shadow region is a spatial area where sound waves cannot reach due to the geometric shielding effect of sound ray bending or reflection from the seabed or sea surface. In the acoustic connectivity graph, the shadow region manifests as the boundary region between different connected components.

[0025] In this embodiment, the acoustic connectivity graph is decomposed into connected components, and each connected component is defined as a dominant transfer domain. The boundaries between dominant transfer domains correspond to the locations of the acoustic shadow zone, and the dominant potential energy field is discontinuous between different dominant transfer domains. The dominant potential energy field refers to a scalar function field defined within the underwater combat area. The value of this function at each spatial location represents the overall support level for carrying out underwater combat operations at that location.

[0026] In this embodiment, thermocline boundaries and density gradient boundaries are extracted from the density profile, and a kinematic constraint map is constructed within each dominant transfer domain. The thermocline boundary refers to the upper and lower boundaries of the depth range where the vertical temperature gradient of seawater exceeds a preset temperature gradient threshold. The density gradient boundary refers to the upper and lower boundaries of the depth range where the vertical density gradient of seawater exceeds a preset density gradient threshold. The kinematic constraint map is a label map marking the type of gradient region to which each spatial location belongs. Gradient region types include the upper mixing zone above both the thermocline and density gradient, the gradient transition zone between the thermocline and density gradient, and the deep isodense zone below both the thermocline and density gradient.

[0027] In this embodiment, the acoustic connectivity graph and maneuver constraint graph are input into the operational effectiveness prediction model to obtain the intrinsic support of each spatial location for underwater combat operations. The operational effectiveness prediction model is a three-layer backpropagation neural network. The input layer receives inputs including the number of neighboring nodes of the spatial location in the acoustic connectivity graph, a Boolean value indicating whether the spatial location is located in a shadow zone, the type encoding of the hierarchical region to which the spatial location belongs, and the horizontal and vertical distances between the spatial location and the target area. The output layer outputs the intrinsic support of the spatial location for sonar detection operations, submarine maneuvering operations, torpedo attack operations, and underwater acoustic communication operations. All four output values ​​are real numbers between 0 and 1. The network is trained using measured underwater exercise data. The input data for the training data comes from measured sound velocity and density profiles, and the output data comes from the actual effectiveness scores of various operations recorded during the exercises. Underwater combat operations include sonar detection operations, submarine maneuvering operations, torpedo attack operations, and underwater acoustic communication operations.

[0028] In one specific implementation of this embodiment, the combat effectiveness prediction model is a feedforward neural network. Its structure is as follows: the input layer contains 9 neurons, receiving the number of neighboring nodes of the spatial location in the acoustic connectivity graph, a Boolean value indicating whether the spatial location is located in a sound shadow zone, the encoding of the hierarchical region to which the spatial location belongs, and the horizontal and vertical distances between the spatial location and the target region. This is followed by a first hidden layer and a second hidden layer, each containing 64 neurons, both using the ReLU activation function. The output layer contains 4 neurons, using the Sigmoid activation function, outputting the intrinsic support of the spatial location for sonar detection operations, submarine maneuvering operations, torpedo attack operations, and underwater acoustic communication operations. The network uses the mean squared error loss function and the Adam optimizer, with an initial learning rate set to 0.001. During the training phase, no fewer than 10,000 samples are extracted from a historical underwater exercise database. Each sample contains a set of input feature vectors and corresponding real effectiveness label vectors annotated by military experts. Supervised learning is performed according to the above input-output format until the loss function converges.

[0029] In this embodiment, the Poisson equation is solved within each dominant transfer domain using the intrinsic support as the potential energy value to obtain a block-continuous dominant potential energy field. The specific process of solving the Poisson equation is as follows: within each dominant transfer domain, the intrinsic support value at each spatial location is used as the source term of the Poisson equation. The boundary of the dominant transfer domain is used as the Dirichlet boundary condition. The potential energy value on the boundary is set as the intrinsic support value of the adjacent spatial location inside the boundary. The Poisson equation is discretized into a system of linear equations using the finite difference method. The system of linear equations is solved using the multigrid method to obtain the potential energy value at each spatial location. This potential energy value constitutes a block-continuous dominant potential energy field.

[0030] In this embodiment, the gradient direction of the dominant potential energy field is used as the movement direction to generate an action strategy for each combat entity. The gradient direction of the dominant potential energy field refers to the direction in which the dominant potential energy field function changes most rapidly at each point in space, and this direction is calculated from the first-order partial derivative of the dominant potential energy field function with respect to spatial coordinates. The movement direction refers to the direction in which the combat entity moves from its current position towards the direction indicated by the gradient direction of the dominant potential energy field. A combat entity refers to a moving unit representing a submarine or an unmanned underwater vehicle in the simulation. The action strategy includes the sequence of movement paths for the combat entity and the movement speed at each point along the path.

[0031] In this embodiment, when the movement path of a combat entity crosses the boundary of an advantage transfer domain, a detection command or an attack command is triggered at the crossing point. The movement path of the combat entity refers to the spatial position sequence traversed by the combat entity from the start of the simulation to the current moment. The advantage transfer domain boundary refers to the interface between two adjacent advantage transfer domains, corresponding to the acoustic shadow zone location between two disconnected connected components in the acoustic connectivity graph. The crossing point refers to the spatial location of the combat entity when its movement path enters the second advantage transfer domain from the first advantage transfer domain. The detection command is the operational command that controls the combat entity to activate active or passive sonar for target search. The attack command is the operational command that controls the combat entity to launch torpedoes or release flares.

[0032] In this embodiment, when the preset simulation termination condition is met, the simulation result containing all triggered instructions is output. When the preset simulation termination condition is not met and the preset environment update time is reached, the sound velocity profile and density profile are updated, and the process returns to the step of extracting the sound ray propagation path from the sound velocity profile. The preset environment update time refers to the environmental field refresh time point pre-set according to the simulation time, such as the hourly refresh every simulation hour. Environmental dynamic evolution rules include internal wave propagation models, frontal movement models, or ocean current change models, used to calculate the updated sound velocity profile and density profile according to these rules. The preset simulation termination condition is when the simulation duration reaches 3600 seconds. The simulation results include the trigger time, trigger location, and command type of all triggered commands.

[0033] In another embodiment of the present invention, identifying connected components in an acoustic connectivity graph includes: Using each spatial location as a sound source point, a preset number of sound rays with different initial emission angles are emitted using a ray tracing algorithm; Based on whether each sound ray is reflected by the sea surface or the seabed during its propagation, the sound ray connectivity type between each spatial location pair is recorded. The sound ray connectivity types include direct connectivity, single-reflection connectivity, and no connectivity. When there is no direct connection and no single reflection connection between two spatial locations, these two spatial locations are marked as sound-shadow zone blocking relationships; All spatial pairs with sound shadow zone blocking relationships are considered as sound shadow zone blocking relationships. Based on the sound shadow zone blocking relationships, disconnected connected components in the acoustic connectivity graph are identified.

[0034] In this embodiment, the preset quantity refers to the total number of sound rays emitted from each sound source point, and the value of this quantity is 360.

[0035] In this embodiment, the initial emission angle refers to the angle between the sound ray and the horizontal direction when it originates from the sound source point, and the angle range is from -90 degrees to +90 degrees.

[0036] In this embodiment, using each spatial location as a sound source, a preset number of sound rays with different initial emission angles are emitted using a ray tracing algorithm. The ray tracing algorithm calculates the propagation direction and distance of the sound rays layer by layer based on the vertical layered structure of the sound velocity profile, until the sound rays propagate to the boundary of the underwater combat area or exceed the preset maximum propagation distance. Of the 360 ​​sound rays emitted from each spatial location as a sound source, each sound ray has a different initial emission angle, which is distributed at equal intervals from -90 degrees to +90 degrees.

[0037] In this embodiment, the acoustic connectivity type between each spatial location pair is recorded based on whether each sound ray undergoes reflection from the sea surface or the seabed during propagation. The acoustic connectivity types include direct connectivity, single-reflection connectivity, and no connectivity. When a sound ray emitted from the first spatial location reaches the second spatial location directly without any reflection, the acoustic connectivity type between the spatial location pair is recorded as direct connectivity. When a sound ray emitted from the first spatial location reaches the second spatial location after experiencing exactly one sea surface reflection or one seabed reflection, the acoustic connectivity type between the spatial location pair is recorded as single-reflection connectivity. When all sound rays emitted from the first spatial location fail to reach the second spatial location, the acoustic connectivity type between the spatial location pair is recorded as no connectivity.

[0038] In this embodiment, when there is no direct connection and no single-reflection connection between two spatial locations, these two spatial locations are marked as having a sound shadow zone barrier relationship. The sound shadow zone barrier relationship indicates that two spatial locations cannot reach each other through direct sound rays or single-reflection sound rays due to the existence of the sound shadow zone. This relationship is the physical reason why advantages cannot be continuously transferred in underwater combat environments.

[0039] In this embodiment, all spatial location pairs with acoustic shadow zone blocking relationships are aggregated to generate a set of dominant disconnected boundaries. This set of dominant disconnected boundaries records all spatial location pairs within the underwater operational area that are blocked by acoustic shadow zones. This set is used in subsequent steps to identify the boundary locations of connected components in the acoustic connectivity graph.

[0040] In this embodiment, all spatial location pairs with sound shadow zone barriers are considered as sound shadow zone barriers. Based on these barriers, disconnected connected components in the acoustic connectivity graph are identified. The specific identification process is as follows: starting from any unvisited spatial location node in the acoustic connectivity graph, a depth-first search is used to traverse all nodes with direct or single-reflection connectivity to that node, grouping these nodes into the same connected component. This process is repeated until all nodes in the acoustic connectivity graph have been visited, resulting in all disconnected connected components in the acoustic connectivity graph.

[0041] In another embodiment of the invention, constructing a maneuver constraint graph within each advantage transfer domain includes: Calculate the temperature gradient of the density profile along the vertical direction, and mark the continuous depth interval where the temperature gradient exceeds the preset temperature gradient threshold as a thermocline. The spatial region located above the thermocline and above the density climax is marked as the upper mixing region; the depth region located below the thermocline and below the density climax is marked as the deep isodense region; and the spatial region located between the thermocline and the density climax is marked as the climax transition region. A maneuver constraint map is generated using the region type to which each spatial location belongs as the maneuver constraint information.

[0042] In this embodiment, the density gradient of the density profile is calculated along the vertical direction. The density gradient is the ratio of the density difference between two adjacent depth layers to the depth difference. For each spatial location, the density gradient between every two adjacent depth layers is calculated sequentially in ascending order of depth.

[0043] In this embodiment, the preset density gradient threshold refers to a pre-set density gradient critical value, which takes the value of... .

[0044] In this embodiment, consecutive depth intervals where the density gradient exceeds a preset density gradient threshold are marked as density jump layers. When the density gradients of multiple consecutive depth layers all exceed... At that time, these depth layers are merged into a density jump interval, and the upper and lower boundary depths of the interval are recorded.

[0045] In this embodiment, the temperature gradient of the density profile is calculated along the vertical direction. The temperature gradient is the ratio of the temperature difference between two adjacent depth layers to the depth difference. For each spatial location, the temperature gradient between every two adjacent depth layers is calculated sequentially in ascending order of depth.

[0046] In this embodiment, the preset temperature gradient threshold refers to a pre-set temperature gradient critical value, which is 0.02 degrees Celsius / meter.

[0047] In this embodiment, continuous depth intervals with temperature gradients exceeding a preset temperature gradient threshold are marked as thermocline layers. When the temperature gradients of multiple consecutive depth layers all exceed 0.02 degrees Celsius / meter, these depth layers are merged into a single thermocline layer interval, and the upper and lower boundary depths of this interval are recorded.

[0048] In this embodiment, the spatial region located above the thermocline and above the density cascade is designated as the upper mixing zone. The upper mixing zone refers to the seawater layer between the sea surface and the shallower of the upper boundaries of the thermocline and density cascade, where temperature and density are vertically uniformly distributed.

[0049] In this embodiment, the depth range located below the thermocline and below the density cascade is designated as the deep isodense zone. The deep isodense zone refers to the seawater layer between the deeper of the lower boundaries of the thermocline and the density cascade and the seabed, where temperature and density change slowly with depth.

[0050] In this embodiment, the spatial region located between the thermocline and the density cascade is designated as the cascade transition zone. The cascade transition zone refers to the area where the thermocline and density cascade overlap, and within this region, temperature and density change drastically with depth, significantly affecting submarine maneuvering and underwater acoustic propagation.

[0051] In this embodiment, a maneuver constraint map is generated using the region type to which each spatial location belongs as maneuver constraint information. The maneuver constraint map is a label matrix with the same size as the underwater combat area spatial grid. The value of each element in the matrix represents the region type to which the corresponding spatial location belongs. Region types include three types: upper mixed zone, interlayer transition zone, and deep equal-density zone.

[0052] In another embodiment of the invention, obtaining the intrinsic support for underwater combat operations at each spatial location includes: The dominant transit domain identifier of each spatial location in the acoustic connectivity graph is used as a constraint condition for the intrinsic support calculation. For sonar detection operations, based on the judgment that there is a direct sound ray or a single reflected sound ray reaching the target area from the current spatial location, and the judgment that the current spatial location is located below the thermocline, a nonlinear weighting function is used to calculate the intrinsic support of each spatial location for the sonar detection operation. For submarine maneuvering operations, based on the judgment results of whether the current spatial position is located in the transition zone between different layers and whether the current spatial position is located in the sound shadow zone, the intrinsic support degree of each spatial position for maneuvering stealth is calculated using the fuzzy membership function. For torpedo attack operations, the intrinsic support of each spatial location for the torpedo attack operation is calculated based on the judgment results of the location of the sound convergence zone from the current spatial location to the target area, and the judgment results of whether the current spatial location has a preset maneuver space range. For underwater acoustic communication operations, based on the judgment results of whether there is a stable acoustic ray channel between the current spatial location and the communication target, and the judgment results of whether there is multipath interference, the signal-to-noise ratio prediction model is used to calculate the intrinsic support of each spatial location for the underwater acoustic communication operation. The calculated intrinsic support is grouped and stored according to the dominant transit domain identifier to ensure that the intrinsic support between different dominant transit domains does not participate in cross-domain coupling calculation.

[0053] In this embodiment, the dominant transit domain identifier of each spatial location in the acoustic connectivity graph is used as a constraint for intrinsic support calculation. The dominant transit domain identifier is a unique number of the dominant transit domain to which each spatial location belongs, generated during the connected component decomposition step. During intrinsic support calculation, coupling calculations are performed only between spatial locations within the same dominant transit domain; the intrinsic support of spatial locations in different dominant transit domains does not affect each other.

[0054] In this embodiment, sonar detection operations refer to the process by which a combat entity uses active or passive sonar to search for, detect, and track underwater targets.

[0055] In this embodiment, the target area refers to the preset spatial range in which the enemy combat entity may be located during the simulation, and this range is jointly defined by the longitude range, latitude range and depth range.

[0056] In this embodiment, for sonar detection operations, based on the determination of whether a direct sound ray or a first-reflection sound ray reaches the target area from the current spatial location, and the determination of whether the current spatial location is located below the thermocline, a nonlinear weighting function is used to calculate the intrinsic support of each spatial location for the sonar detection operation. The nonlinear weighting function adopts the form of a Sigmoid function, with inputs being Boolean values ​​for the presence of a direct sound ray or a first-reflection sound ray (1 for presence, 0 for absence) and Boolean values ​​for whether the location is below the thermocline (1 for yes, 0 for no). The output value is a real number ranging from 0 to 1. When a direct sound ray or a first-reflection sound ray exists and the location is below the thermocline, the output value is close to 1; when no sound ray exists and the location is above the thermocline, the output value is close to 0.

[0057] In this embodiment, for submarine maneuvering, based on the judgment results of whether the current spatial position is located in the transition zone and whether the current spatial position is located in the sound shadow zone, a fuzzy membership function is used to calculate the intrinsic support degree of each spatial position for maneuver stealth. The fuzzy membership function adopts the form of a triangular membership function. The inputs are Boolean values ​​for whether the submarine is located in the transition zone (1 for yes, 0 for no) and whether it is located in the sound shadow zone (1 for yes, 0 for no). The output value is a real number between 0 and 1. When the submarine is located in both the transition zone and the sound shadow zone, the output value is 1; when the submarine is located in neither the transition zone nor the sound shadow zone, the output value is 0.

[0058] In this embodiment, for torpedo attack operations, the intrinsic support of each spatial location for the torpedo attack operation is calculated based on the judgment result of the location of the acoustic convergence zone from the current spatial location to the target area, and the judgment result of whether the current spatial location has a preset maneuvering space range. The judgment result of the location of the acoustic convergence zone from the current spatial location to the target area refers to a Boolean value indicating whether the current spatial location is within the acoustic convergence zone. The judgment result is "yes" when the current spatial location is within the acoustic convergence zone from which the sound emitted from that location converges after reflection from the sea surface or seabed to the target area. The preset maneuvering space range refers to the minimum horizontal radius of movement required for the combat entity to complete the torpedo launch, which is 250 meters. The judgment result of whether the current spatial location has a preset maneuvering space range refers to a Boolean value indicating whether there is sufficient space for the combat entity to maneuver within a circular area centered on the current spatial location and with a radius of 250 meters. The judgment result is "yes" when there are no obstacles in this area and the water depth meets the maneuvering requirements. The intrinsic support of the torpedo attack operation is a weighted sum of the two judgment results mentioned above. The Boolean value located in the sound convergence zone has a weight of 0.6, and the Boolean value with a preset maneuvering space range has a weight of 0.4, so that the intrinsic support continuously takes values ​​between 0 and 1.

[0059] In this embodiment, for underwater acoustic communication operations, based on the judgment results of whether a stable acoustic ray channel exists between the current spatial location and the communication target, and the judgment results of whether multipath interference exists, a signal-to-noise ratio prediction model is used to calculate the intrinsic support of each spatial location for the underwater acoustic communication operation. The communication target refers to the spatial location of the receiver conducting underwater acoustic communication with the current combat entity in the simulation. A stable acoustic ray channel refers to a direct acoustic ray or a single-reflection acoustic ray existing between the current spatial location and the communication target, where the acoustic energy attenuation along the ray propagation path is less than a preset attenuation threshold. The judgment result of whether a stable acoustic ray channel exists between the current spatial location and the communication target is a Boolean value indicating the existence of such a channel. The judgment result of whether multipath interference exists refers to a Boolean value indicating whether there are two or more acoustic rays with different propagation paths from the current spatial location to the communication target, and the arrival time difference of each path is less than a preset time difference threshold. The signal-to-noise ratio (SNR) prediction model is a three-layer backpropagation neural network. The input layer receives Boolean values ​​indicating the existence of a stable acoustic channel, the existence of multipath interference, and the horizontal and vertical distances between the current spatial location and the communication target. The output layer outputs the predicted SNR value, ranging from 0 to 30 dB (excluding 0). The network is trained using measured underwater acoustic communication data. The intrinsic support of the underwater acoustic communication action is the normalized result of dividing the predicted SNR value by 30 dB. When the predicted SNR is less than 0 dB, the intrinsic support of the underwater acoustic communication action is set to 0.

[0060] In this embodiment, underwater acoustic communication operations refer to the process by which combat entities exchange data, commands, or status information through underwater acoustic channels.

[0061] In this embodiment, the calculated intrinsic support is grouped and stored according to the dominant transitive domain identifier to ensure that the intrinsic support between different dominant transitive domains does not participate in cross-domain coupling calculations. Specifically, a dictionary data structure is established with the dominant transitive domain identifier as the key, and each key corresponds to the intrinsic support matrix of all spatial locations within that dominant transitive domain. During the subsequent Poisson equation solution, each dominant transitive domain independently uses its own intrinsic support matrix, and different dominant transitive domains do not exchange data. The dominant transitive domain identifier is a unique number for each dominant transitive domain, generated during the connected component decomposition step, and its value ranges from 1 to the total number of dominant transitive domains.

[0062] In another embodiment of the present invention, the method for identifying crossing points in the boundary of the advantage transfer domain includes: The acoustic connected graph is traversed using either breadth-first search or depth-first search to identify all maximal connected subgraphs. Each maximally connected subgraph is mapped to a dominant transit region; Record the boundary positions between different dominant transmission domains, and the boundary positions correspond to the acoustic shadow region in the acoustic connectivity graph; The boundary location of each dominant transfer domain is overlaid with underwater topographic data, and the mutation rate of intrinsic support on both sides of the boundary is calculated. When the mutation rate exceeds the preset mutation threshold, the corresponding boundary segment is marked as a crossing point.

[0063] In this embodiment, breadth-first search or depth-first search is used to traverse the acoustic connected graph to identify all maximal connected subgraphs. Breadth-first search starts from any unvisited node in the acoustic connected graph, first visiting all its neighboring nodes, then visiting the neighbors of those neighbors in turn, expanding outwards layer by layer until no further expansion is possible. All visited nodes constitute a connected subgraph. Depth-first search starts from any unvisited node in the acoustic connected graph, visiting nodes as deeply as possible along a path until no further depth is possible, then backtracking to the previous node and continuing to explore other branches. All visited nodes constitute a connected subgraph. This process is repeated until all nodes in the acoustic connected graph have been visited. Among all the resulting connected subgraphs, if a connected subgraph is not contained within any other connected subgraph, then that connected subgraph is a maximal connected subgraph.

[0064] In this embodiment, each maximal connected subgraph corresponds to a dominant transit region. Each maximal connected subgraph corresponds to a unique dominant transit region, and the number of dominant transit regions is equal to the number of maximal connected subgraphs. The set of spatial locations contained in each dominant transit region is the set of spatial locations corresponding to all nodes in that maximal connected subgraph.

[0065] In this embodiment, the boundary positions between different dominant transfer domains are recorded, and these boundary positions correspond to the sound shadow region in the acoustic connectivity graph. For any two adjacent dominant transfer domains, the boundary node pairs between their corresponding maximal connected subgraphs in the acoustic connectivity graph are extracted, and the set of midpoints of the spatial positions corresponding to these boundary node pairs is recorded as the boundary position between the two dominant transfer domains. This boundary position corresponds to the location of the sound shadow region in physical space.

[0066] In this embodiment, underwater topographic data refers to seabed depth grid data of the underwater combat area, which is acquired by a multibeam echo sounder system with a grid resolution of 500 meters by 500 meters.

[0067] In this embodiment, the boundary locations of each dominant transfer domain are overlaid with underwater topographic data, and the abrupt change rate of intrinsic support on both sides of the boundary is calculated. For each boundary location, a preset distance (500 meters) is extended vertically from the boundary location into the first and second dominant transfer domains. The intrinsic support values ​​of all spatial locations along the extension path are extracted. The absolute value of the difference between the average intrinsic support on the first and second dominant transfer domain sides is calculated, and this difference is divided by the preset distance to obtain the abrupt change rate. The unit of the abrupt change rate is per meter.

[0068] In this embodiment, the preset mutation threshold refers to the critical value of the intrinsic support mutation rate that is set in advance, and the critical value is 0.01 per meter.

[0069] In this embodiment, the corresponding boundary segment is marked as a command trigger point. When the mutation rate at the boundary location exceeds 0.01 per meter, the boundary location is marked as a command trigger point. A command trigger point indicates that at this boundary location, the advantage changes drastically between two advantage transfer domains, and triggering a detection or attack command when a combat entity crosses this boundary has high tactical value.

[0070] In another embodiment of the present invention, solving the Poisson equation within each dominance transfer domain using the intrinsic support as the potential energy value to obtain a block-continuous dominance potential energy field includes: Within each dominance transfer domain, the intrinsic support value of each spatial location is used as the source term of the Poisson equation; The boundary of the dominant transfer domain is used as the first type of boundary condition, and the potential energy value on the boundary is set as the intrinsic support value of the adjacent spatial location inside the boundary. The Poisson equation is discretized into a system of linear equations using the finite difference method. The multigrid method is used to solve the linear equations to obtain a block-continuous dominant potential energy field.

[0071] In this embodiment, within each advantage transfer domain, the intrinsic support value of each spatial location is used as the source term of the Poisson equation. The Poisson equation states that the Laplace operator of the potential energy field equals the negative source term. Within each advantage transfer domain, the calculated intrinsic support value at each spatial location is substituted into the source term position of the Poisson equation, establishing a mathematical relationship between the advantage potential energy field to be solved and the intrinsic support at that location. The advantage potential energy field is a dimensionless mathematical evaluation field used to characterize the relative advantage distribution in the operational space.

[0072] In this embodiment, the boundary of the dominant transfer domain is used as the first type of boundary condition, and the potential energy value on the boundary is set as the intrinsic support value of the adjacent spatial location inside the boundary. The first type of boundary condition means that the value of the potential field is directly specified on the boundary of the solution domain. For the boundary location of the dominant transfer domain, the intrinsic support value of the spatial location at a position one grid step vertically extended into the dominant transfer domain from the boundary location is used as the potential energy value at that boundary location.

[0073] In this embodiment, the potential energy value at the boundary refers to the value of the dominant potential energy field to be solved at the boundary of the dominant transfer domain. The potential energy value at the boundary is not obtained by solving the Poisson equation, but is directly set as a known condition.

[0074] In this embodiment, the adjacent spatial position inside the boundary refers to the spatial position reached after moving vertically one grid step from the boundary position of the dominance transfer domain into the dominance transfer domain. The grid step is determined when dividing the spatial grid of the underwater combat area, with a horizontal grid step of 500 meters and a vertical grid step of 10 meters.

[0075] In this embodiment, the intrinsic support value of adjacent spatial locations inside the boundary refers to the intrinsic support value calculated at the spatial location reached after moving vertically one grid step from the boundary location of the dominance transfer domain into the dominance transfer domain.

[0076] In this embodiment, the Poisson equation is discretized into a system of linear equations using the finite difference method. The finite difference method approximates the derivative of a function at a given spatial location by using the difference in function values ​​at that point. Specifically, the second-order partial derivative at each spatial location is represented by a weighted sum of the differences in potential energy values ​​between that location and its neighboring locations, transforming the Poisson equation into an algebraic equation concerning the potential energy value at each spatial location. After establishing algebraic equations for all spatial locations, a system of linear equations is formed, with the potential energy value at each spatial location being the unknown. The number of equations in this system equals the total number of spatial locations within the dominance transfer domain.

[0077] In this embodiment, a multigrid method is used to solve the linear equations to obtain a block-continuous dominant potential energy field. The multigrid method refers to iteratively solving the linear equations on grids of varying coarseness. First, several iterations are performed on the original fine grid to smooth the error. Then, the residuals are constrained to a coarser grid to solve for the correction, which is then extended back to the fine grid for correction. This process is repeated until the linear equations converge. After solving, a potential energy value is obtained for each spatial location, and the potential energy values ​​at all spatial locations constitute the dominant potential energy field. Since each dominant transfer domain solves the Poisson equation independently, the potential energy fields between different dominant transfer domains are independent and discontinuous at the boundaries, hence the term "block-continuous dominant potential energy field." Because the Poisson equation for each dominant transfer domain is solved independently, and its Dirichlet boundary conditions are determined by the eigenvalue support within its respective domain, the potential energy values ​​calculated on both sides of the common boundary of adjacent dominant transfer domains are usually unequal, thus mathematically creating a discontinuity in the potential energy field at the boundaries of the dominant transfer domains. This discontinuity precisely reflects the physical barrier effect of the sound shadow zone on the transmission of combat advantages.

[0078] In another embodiment of the present invention, triggering a probe command or attack command at a crossing point includes: Record the first superiority transfer domain identifier before the combat entity crosses the superiority transfer domain boundary and the second superiority transfer domain identifier after crossing it. Extract the first eigensupport of the nearest position to the boundary within the first advantage transfer domain and the second eigensupport of the nearest position to the boundary within the second advantage transfer domain from the hydro-operation coupling tensor. Calculate the ratio of the first intrinsic support to the second intrinsic support; When the ratio exceeds the preset detection threshold, it is determined that the area has been crossed, triggering an active sonar detection command. When the ratio is lower than the preset concealment threshold, it is determined that a passive listening command or an evasion maneuver command will be triggered after crossing the area. When the ratio is between the preset concealment threshold and the preset detection threshold, it is determined that the area has been crossed and a regular detection command is triggered.

[0079] In this embodiment, the first advantage transfer domain identifier before the combat entity crosses the boundary of the advantage transfer domain and the second advantage transfer domain identifier after the crossover are recorded. When the combat entity moves from the current spatial location to the next spatial location, if the advantage transfer domain identifier of the current spatial location is different from the advantage transfer domain identifier of the next spatial location, a crossover is determined to have occurred. The advantage transfer domain identifier before the crossover is recorded as the first advantage transfer domain identifier, and the advantage transfer domain identifier after the crossover is recorded as the second advantage transfer domain identifier.

[0080] In this embodiment, the hydro-operation coupling tensor refers to a three-dimensional data structure that arranges the intrinsic support of each spatial location for each underwater combat action according to the spatial location dimension, the combat action type dimension, and the support value dimension. The size of the spatial location dimension is equal to the total number of discrete spatial locations within the underwater combat area, the size of the combat action type dimension is equal to 4, and the size of the support value dimension is equal to 1.

[0081] In this embodiment, the first intrinsic support of the nearest position to the boundary within the first advantage transfer domain and the second intrinsic support of the nearest position to the boundary within the second advantage transfer domain are extracted from the hydro-operation coupling tensor. The nearest position to the boundary refers to the spatial position reached after moving vertically one grid step from the boundary position into the advantage transfer domain. The first intrinsic support refers to the intrinsic support value extracted from the hydro-operation coupling tensor at the nearest position to the boundary within the first advantage transfer domain. The second intrinsic support refers to the intrinsic support value extracted from the hydro-operation coupling tensor at the nearest position to the boundary within the second advantage transfer domain.

[0082] In actual numerical calculations, when the second intrinsic support is lower than a preset minimum value (e.g., 0.01), an active sonar detection command is directly triggered.

[0083] In this embodiment, the preset detection threshold refers to a pre-set critical value for the intrinsic support ratio that triggers an active sonar detection command, and this critical value is 0.8. When the ratio of the first intrinsic support to the second intrinsic support exceeds 0.8, it is determined that an active sonar detection command needs to be triggered.

[0084] In this embodiment, the preset concealment threshold refers to a pre-set critical value for the intrinsic support ratio that triggers a passive eavesdropping command or an evasion maneuver command, and this critical value is 0.2. When the ratio of the first intrinsic support to the second intrinsic support is lower than 0.2, it is determined that a passive eavesdropping command or an evasion maneuver command needs to be triggered.

[0085] In this embodiment, the active sonar detection command refers to the operational command that controls the combat entity to emit sonar pulse signals and receive echoes to actively detect targets. The active sonar detection command includes three parameters: transmission frequency, transmission pulse width, and transmission power. The transmission frequency is set to 3000 Hz, the transmission pulse width is set to 0.1 seconds, and the transmission power is set to 1000 watts.

[0086] In this embodiment, the passive listening command refers to the operational command that controls the combat entity to shut down the active sonar transmitter and listen to the target by receiving acoustic signals from the environment solely through passive sonar. The passive listening command includes two parameters: the listening frequency band and the gain. The listening frequency band ranges from 100 Hz to 5000 Hz, and the gain is 30 dB.

[0087] In this embodiment, the evasion maneuver command refers to the operational command that controls the combat entity to change its course, depth, or speed to avoid enemy detection or attack. The evasion maneuver command includes three parameters: course change, depth change, and speed change. The course change is ±90 degrees, the depth change is ±20 meters, and the speed change is ±3 knots.

[0088] In this embodiment, the conventional detection command refers to the command that controls the combat entity to perform detection operations with default parameters. This command is between the active sonar detection command and the passive listening command. The conventional detection command adopts a compromise parameter between the active sonar detection command and the passive listening command, with a transmission frequency of 5000 Hz, a transmission pulse width of 0.05 seconds, a transmission power of 500 watts, and the passive listening channel is activated simultaneously.

[0089] To prevent combat entities from repeatedly triggering commands by crossing back and forth near the boundary of the advantage transfer domain, a delay mechanism can be introduced into the above decision logic: after a combat entity crosses the boundary, it must remain in the area after crossing for more than a preset dwell time threshold (e.g., 30 seconds) before the corresponding detection or attack command is officially triggered; if it returns to the original advantage transfer domain within the dwell time threshold, no command is triggered.

[0090] In another embodiment of the present invention, generating an action strategy for each combat entity, using the gradient direction of the advantageous potential energy field as the direction of movement, includes: Within each advantage transfer domain, multiple candidate movement paths are generated by integrating along the negative gradient direction of the advantage potential field. Calculate the second derivative of the potential energy gradient at each point on each candidate movement path, solve the potential energy curvature based on the second derivative of the potential energy gradient at each point, and identify points whose potential energy curvature exceeds a preset curvature threshold as maneuver adjustment points. When there is a bifurcation point, calculate the matching degree between the potential energy gradient in each bifurcation direction and the maneuver constraint diagram, and select the direction with the highest matching degree as the main movement direction. When the paths of multiple combat entities intersect at the same spatial location, the path allocation order is determined based on the mission priority of each combat entity and the potential energy value of its dominant potential energy field at the corresponding intersecting spatial location. Based on the main direction of movement and the order of path allocation, generate the action strategy for each combat entity.

[0091] In this embodiment, the negative gradient direction of the dominant potential field refers to the direction in which the function value of the dominant potential field function decreases the fastest at each point in space. This direction is calculated by the first-order partial derivative of the dominant potential field function with respect to spatial coordinates to obtain the direction of the negative gradient vector.

[0092] In this embodiment, within each advantage transfer domain, multiple candidate movement paths are generated by integration along the negative gradient direction of the advantageous potential energy field. Integration refers to gradually accumulating displacement from the current position of the combat entity along the negative gradient direction of the advantageous potential energy field. Each step involves moving half a grid step, with half a grid step being 250 meters in the horizontal direction and 5 meters in the vertical direction. This process is repeated until the boundary of the advantage transfer domain is reached or the potential energy gradient magnitude is less than a preset gradient magnitude threshold, resulting in a continuous trajectory from the current position to the stopping position. Positions offset by half a grid step from each of the eight adjacent horizontal directions, one vertical upward direction, and one vertical downward direction (ten directions in total) of the combat entity's current position are used as alternative starting points. The integration process is repeated for each alternative starting point to generate multiple candidate movement paths.

[0093] In this embodiment, the second derivative of the potential energy gradient at each point on each candidate movement path is calculated. Based on the second derivative of the potential energy gradient at each point, the potential energy curvature is solved, and points whose potential energy curvature exceeds a preset curvature threshold are identified as maneuver adjustment points. The second derivative of the potential energy gradient refers to the rate of change of the dominant potential energy field gradient direction along the path. For each point on the path, the difference between the unit vector of the dominant potential energy field gradient direction at that point and the unit vector of the dominant potential energy field gradient direction at the next adjacent point is calculated. The magnitude of this difference is divided by the path length between the two points to obtain an approximate value of the second derivative of the potential energy gradient.

[0094] Potential energy curvature refers to the rate of change of the gradient direction of the dominant potential energy field along a path, and is equal to the absolute value of the second derivative of the potential energy gradient. Potential energy curvature reflects the degree of bending of the dominant potential energy field; the greater the curvature, the more drastic the directional change of the dominant potential energy field at that location.

[0095] The preset curvature threshold refers to a pre-defined critical value for the potential energy curvature, which is 0.1 per meter. When the potential energy curvature at a point on the path exceeds 0.1 per meter, that point is marked as a maneuvering adjustment point.

[0096] In this embodiment, a bifurcation point refers to a location in a candidate movement path where two or more branches appear. When two candidate movement paths start from the same location and then move in different directions during subsequent movements, that starting location is identified as a bifurcation point.

[0097] In this embodiment, the matching degree between the potential energy gradient in each bifurcation direction and the maneuver constraint diagram is calculated. The bifurcation direction refers to the initial movement direction vector of each branch path at the bifurcation point. The matching degree is the dot product between the unit vector in the bifurcation direction and the unit vector of the allowed movement direction at that bifurcation point in the maneuver constraint diagram. The allowed movement direction is determined according to the region type label of the bifurcation point in the maneuver constraint diagram according to the following rules: when the region type label is an upper-layer mixed region or a deep-layer iso-dense region, the allowed movement direction is all horizontal directions; when the region type label is a layer transition region, the allowed movement direction is only the horizontal direction parallel to the layer extension direction. The dot product value ranges from -1 to 1; a larger dot product value indicates a higher consistency between the bifurcation direction and the allowed movement direction. The allowed movement directions in the maneuver constraint diagram include eight horizontal directions and two vertical directions, totaling ten directions, each represented by a unit vector. The dot product value ranges from -1 to 1; a larger dot product value indicates a higher consistency between the bifurcation direction and the allowed movement direction in the maneuver constraint diagram.

[0098] In this embodiment, the convergence of the paths of multiple combat entities at the same spatial location means that the movement paths of two or more combat entities pass through the same spatial location. This spatial location is called the convergence point.

[0099] In this embodiment, the path allocation order is determined based on the mission priority of each combat entity and the potential energy value of its dominant potential energy field at the corresponding intersection point. Mission priority refers to the pre-defined importance level of each combat entity, with a value ranging from 1 to 10; a higher value indicates a higher priority. The potential energy value of the dominant potential energy field at the intersection point refers to the potential energy value extracted from the dominant potential energy field at that intersection point, with a value ranging from 0 to 1. The rule for determining the path allocation order is as follows: first, tasks are sorted from highest to lowest priority; when mission priorities are the same, they are sorted from highest to lowest potential energy value of the dominant potential energy field. The sorting result is the order in which each combat entity passes through the intersection point.

[0100] In this embodiment, an action strategy for each combat entity is generated based on the primary direction of movement and the path allocation order. The primary direction of movement refers to the direction with the highest matching degree selected at the bifurcation point. The action strategy is a data structure that includes the combat entity's identifier, the sequence of movement paths from the current position to the boundary of the advantage transfer domain, the movement speed of each point on the movement path, the primary direction of movement selected at the bifurcation point, the path allocation order at the intersection point, and the heading adjustment amount at the maneuver adjustment point. The movement speed is set to 3 knots, and the heading adjustment amount is set to ±15 degrees.

[0101] In another embodiment of the present invention, obtaining the sound velocity profile and density profile of the underwater combat zone includes: Collect satellite remote sensing data of sea surface temperature, Argo buoy temperature, salinity and depth profile data, real-time data from underwater sensor network, and historical hydrological observation data; All collected data are interpolated to the same spatial grid. For each spatial grid point, calculate the temperature gradient in the vertical direction between the satellite remote sensing sea surface temperature data and the real-time data from the underwater sensor network; When the temperature gradient exceeds the preset gradient threshold, the inversion parameters of the satellite remote sensing sea surface temperature data are calibrated in reverse using the real-time data of the underwater sensor network. The calibrated satellite remote sensing sea surface temperature data is fused with Argo buoy temperature, salinity, and depth profile data to generate a three-dimensional temperature, salinity, and depth field. Sound velocity profiles and density profiles were extracted from a three-dimensional thermo-salinity field.

[0102] In this embodiment, the satellite remote sensing sea surface temperature data refers to the sea surface temperature grid data obtained by the infrared radiometer on the satellite. The spatial resolution of this data is 1000 meters by 1000 meters, and the temporal resolution is once a day.

[0103] In this embodiment, Argo buoy temperature, salinity, and depth profile data refers to the temperature and salinity vertical profile data obtained from buoys in the global Argo ocean observation network. Each buoy returns a temperature, salinity, and depth profile from the sea surface to a depth of 2000 meters every 10 days, with a vertical resolution of 5 meters.

[0104] In this embodiment, the real-time data of the underwater sensor network refers to the temperature, salinity, and depth data collected in real time by fixed sensor nodes deployed in the underwater combat area. Each sensor node collects data once every 10 seconds, and the horizontal spacing between the sensor nodes is 10,000 meters and the vertical spacing is 50 meters.

[0105] In this embodiment, historical hydrological observation data refers to temperature, salinity, and depth observation records of the same area over the past ten years obtained from the marine observation database, including ship observation data, submersible observation data, and moored buoy observation data.

[0106] In this embodiment, all collected data are uniformly interpolated to the same spatial grid. The interpolation uses an inverse distance weighted method, with a horizontal resolution of 500 meters by 500 meters and a vertical resolution of 10 meters for the target spatial grid. For each target grid point, the four closest observation data points are selected, and a weighted average is calculated using the reciprocal of the square of the distance from each observation data point to the grid point as the weight. This average is then used as the interpolation result for that grid point. For data from different sources, the priority order is satellite remote sensing data, Argo buoy data, underwater sensor network data, and historical observation data. When high-priority data is available, it is used first for interpolation.

[0107] In this embodiment, for each spatial grid point, the temperature gradient in the vertical direction between satellite remote sensing sea surface temperature data and real-time data from the underwater sensor network is calculated. For each spatial grid point, the sea surface temperature value for that point is extracted from the satellite remote sensing sea surface temperature data, and the temperature value of the first valid observation point below the sea surface is extracted from the real-time data of the underwater sensor network. The difference between this temperature value and the sea surface temperature value is calculated and divided by the depth difference between the two to obtain the surface temperature gradient estimated from the satellite remote sensing data. Simultaneously, the sea surface temperature value and the temperature value of the first valid observation point below the sea surface are extracted from the real-time data of the underwater sensor network, and the measured surface temperature gradient from the underwater sensor data is calculated using the same method. The temperature gradient is equal to the absolute value of the difference between the two temperature gradients.

[0108] In this embodiment, the preset gradient threshold refers to the critical value of the pre-set temperature gradient, which is 0.01 degrees Celsius per meter.

[0109] In this embodiment, when the temperature gradient exceeds a preset gradient threshold, the inversion parameters of the satellite remote sensing sea surface temperature data are calibrated using real-time data from the underwater sensor network. Reverse calibration refers to adjusting the atmospheric correction coefficient and sea surface emissivity parameters in the satellite remote sensing sea surface temperature inversion algorithm. This is done iteratively using the gradient descent method, with each iteration adjusting the step size by 10% of the current difference value, until the temperature gradient decreases to below 0.01 degrees Celsius per meter or the number of iterations reaches 100.

[0110] In this embodiment, the inversion parameters of satellite remote sensing sea surface temperature data include atmospheric correction coefficients and sea surface emissivity. Atmospheric correction coefficients are used to eliminate the effects of atmospheric absorption and scattering of infrared radiation, and their values ​​range from 0.8 to 1.2. Sea surface emissivity is used to characterize the sea surface's radiative capacity in the infrared band, and its values ​​range from 0.95 to 0.99.

[0111] In this embodiment, the calibrated satellite remote sensing sea surface temperature data refers to the satellite remote sensing sea surface temperature grid data that has been recalculated after reverse calibration.

[0112] In this embodiment, calibrated satellite-sensed sea surface temperature data is fused with Argo buoy temperature-salinity-depth (TST) profile data to generate a three-dimensional TST field. The fusion employs an optimal interpolation method, using the calibrated satellite-sensed sea surface temperature data as the surface constraint and the Argo buoy TST profile data as the vertical constraint to construct the three-dimensional temperature and salinity fields. The specific steps of the optimal interpolation method are as follows: First, based on the Argo buoy TST profile data, a background field is generated using the Kriging interpolation method. Then, using the calibrated satellite-sensed sea surface temperature data as the observation value, the difference between the observation value and the background field is calculated. This difference is then distributed to each depth layer according to spatial correlation weights, ultimately obtaining the three-dimensional temperature and salinity fields. The spatial correlation weights are in the form of a Gaussian function, with a horizontal correlation scale of 50 kilometers and a vertical correlation scale of 100 meters.

[0113] In this embodiment, sound velocity and density profiles are extracted from a three-dimensional temperature-salinity field. The sound velocity profile is calculated using the Mackenzie nine-term formula, and the density profile is calculated using the seawater state equation (using the UNESCO 1983 seawater state equation). For each spatial grid point, using temperature, salinity, and depth as inputs, the sound velocity and density values ​​are calculated layer by layer along the vertical direction to obtain the sound velocity and density profiles for that point.

[0114] This invention provides an embodiment of an underwater combat situation simulation platform for complex marine hydrological environments, comprising: The acoustic connectivity graph construction module is used to obtain the sound speed profile of the underwater combat area, extract the sound ray propagation path from the sound speed profile, and construct the acoustic connectivity graph with the spatial location pairs of the sound ray that have direct access or first reflection as edges. The connected component decomposition module is used to identify connected components in the acoustic connected graph. Due to the existence of the sound shadow region, different connected components are not connected to each other. The connected component decomposition is performed on the acoustic connected graph, and each connected component is defined as a dominant transfer domain. The boundary between dominant transfer domains corresponds to the location of the sound shadow region. The dominant potential energy field is discontinuous between different dominant transfer domains. The maneuver constraint map construction module is used to obtain the density profile of the underwater combat area, extract the thermocline boundary and density cascade boundary from the density profile, and construct the maneuver constraint map within each advantage transfer domain. The effectiveness prediction module is used to input the acoustic connectivity graph and maneuver constraint graph into the combat effectiveness prediction model to obtain the intrinsic support of each spatial location for underwater combat operations. The potential energy field solution module is used to solve the Poisson equation with the intrinsic support as the potential energy value within each dominant transfer domain to obtain a block-continuous dominant potential energy field. The strategy generation module is used to generate action strategies for each combat entity using the gradient direction of the advantageous potential energy field as the movement direction. When the movement path of the combat entity crosses the boundary of the advantage transfer domain, a detection command or attack command is triggered at the crossing point. When the preset simulation termination condition is met, the simulation result containing all triggered commands is output. When the preset simulation termination condition is not met and the preset environment update time is reached, update the sound velocity profile and density profile, and return to the step of extracting the sound ray propagation path from the sound velocity profile.

[0115] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for simulating and extrapolating underwater combat situations in complex marine hydrological environments, characterized in that, include: Obtain sound velocity and density profiles for underwater combat zones; The sound ray propagation path is extracted from the sound speed profile, and an acoustic connectivity graph is constructed using spatial location pairs that have direct sound rays or single-reflection sound rays as edges. Identify the connected components in the acoustic connectivity graph, where the presence of the sound shadow region causes different connected components to be disconnected from each other. The acoustic connectivity graph is decomposed into connected components. Each connected component is defined as a dominant transfer domain. The boundary between dominant transfer domains corresponds to the location of the sound shadow zone. The dominant potential energy field is discontinuous between different dominant transfer domains. Extract the thermocline boundary and density clinch boundary from the density profile, and construct a kinematic constraint map within each dominant transfer domain. By inputting the acoustic connectivity graph and maneuver constraint graph into the combat effectiveness prediction model, the intrinsic support of each spatial location for underwater combat operations can be obtained. Within each dominant transfer domain, the Poisson equation is solved using the intrinsic support as the potential energy value to obtain a block-continuous dominant potential energy field. Using the gradient direction of the advantageous potential energy field as the direction of movement, an action strategy is generated for each combat entity. When the movement path of a combat entity crosses the boundary of the superiority transfer domain, a detection command or an attack command is triggered at the crossing point. When the preset simulation termination condition is met, the simulation result containing all triggered instructions is output. When the preset simulation termination condition is not met and the preset environment update time is reached, update the sound velocity profile and density profile, and return to the step of extracting the sound ray propagation path from the sound velocity profile.

2. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Identifying connected components in an acoustic connectivity graph includes: Using each spatial location as a sound source point, a preset number of sound rays with different initial emission angles are emitted using a ray tracing algorithm; Based on whether each sound ray is reflected by the sea surface or the seabed during its propagation, the sound ray connectivity type between each spatial location pair is recorded. The sound ray connectivity types include direct connectivity, single-reflection connectivity, and no connectivity. When there is no direct connection and no single reflection connection between two spatial locations, these two spatial locations are marked as sound-shadow zone blocking relationships; All spatial pairs with sound shadow zone blocking relationships are considered as sound shadow zone blocking relationships. Based on the sound shadow zone blocking relationships, disconnected connected components in the acoustic connectivity graph are identified.

3. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Constructing a maneuver constraint graph within each advantage transfer domain includes: Calculate the temperature gradient of the density profile along the vertical direction, and mark the continuous depth interval where the temperature gradient exceeds the preset temperature gradient threshold as a thermocline. The spatial region located above the thermocline and above the density climax is marked as the upper mixing region; the depth region located below the thermocline and below the density climax is marked as the deep isodense region; and the spatial region located between the thermocline and the density climax is marked as the climax transition region. A maneuver constraint map is generated using the region type to which each spatial location belongs as the maneuver constraint information.

4. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Obtaining the intrinsic support for underwater operations at each spatial location includes: The dominant transit domain identifier of each spatial location in the acoustic connectivity graph is used as a constraint condition for the intrinsic support calculation. For sonar detection operations, based on the judgment that there is a direct sound ray or a single reflected sound ray reaching the target area from the current spatial location, and the judgment that the current spatial location is located below the thermocline, a nonlinear weighting function is used to calculate the intrinsic support of each spatial location for the sonar detection operation. For submarine maneuvering operations, based on the judgment results of whether the current spatial position is located in the transition zone between different layers and whether the current spatial position is located in the sound shadow zone, the intrinsic support degree of each spatial position for maneuvering stealth is calculated using the fuzzy membership function. For torpedo attack operations, the intrinsic support of each spatial location for the torpedo attack operation is calculated based on the judgment results of the location of the sound convergence zone from the current spatial location to the target area, and the judgment results of whether the current spatial location has a preset maneuver space range. For underwater acoustic communication operations, based on the judgment results of whether there is a stable acoustic ray channel between the current spatial location and the communication target, and the judgment results of whether there is multipath interference, the signal-to-noise ratio prediction model is used to calculate the intrinsic support of each spatial location for the underwater acoustic communication operation. The calculated intrinsic support is grouped and stored according to the dominant transit domain identifier to ensure that the intrinsic support between different dominant transit domains does not participate in cross-domain coupling calculation.

5. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Methods for identifying crossing points in the boundary of the advantage transfer domain include: The acoustic connected graph is traversed using either breadth-first search or depth-first search to identify all maximal connected subgraphs. Each maximally connected subgraph is mapped to a dominant transit region; Record the boundary positions between different dominant transmission domains, and the boundary positions correspond to the acoustic shadow region in the acoustic connectivity graph; The boundary location of each dominant transfer domain is overlaid with underwater topographic data, and the mutation rate of intrinsic support on both sides of the boundary is calculated. When the mutation rate exceeds the preset mutation threshold, the corresponding boundary segment is marked as a crossing point.

6. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Within each dominance transfer domain, the Poisson equation is solved using the intrinsic support as the potential energy value to obtain a block-continuous dominance potential energy field, including: Within each dominance transfer domain, the intrinsic support value of each spatial location is used as the source term of the Poisson equation; The boundary of the dominant transfer domain is used as the first type of boundary condition, and the potential energy value on the boundary is set as the intrinsic support value of the adjacent spatial location inside the boundary. The Poisson equation is discretized into a system of linear equations using the finite difference method. The multigrid method is used to solve the linear equations to obtain a block-continuous dominant potential energy field.

7. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Triggering probe or attack commands at the crossing point includes: Record the first superiority transfer domain identifier before the combat entity crosses the superiority transfer domain boundary and the second superiority transfer domain identifier after crossing it. Extract the first eigensupport of the nearest position to the boundary within the first advantage transfer domain and the second eigensupport of the nearest position to the boundary within the second advantage transfer domain from the hydro-operation coupling tensor. Calculate the ratio of the first intrinsic support to the second intrinsic support; When the ratio exceeds the preset detection threshold, it is determined that the area has been crossed, triggering an active sonar detection command. When the ratio is lower than the preset concealment threshold, it is determined that a passive listening command or an evasion maneuver command will be triggered after crossing the area. When the ratio is between the preset concealment threshold and the preset detection threshold, it is determined that the area has been crossed and a regular detection command is triggered.

8. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Using the gradient direction of the advantageous potential energy field as the direction of movement, the action strategy generated for each combat entity includes: Within each advantage transfer domain, multiple candidate movement paths are generated by integrating along the negative gradient direction of the advantage potential field. Calculate the second derivative of the potential energy gradient at each point on each candidate movement path, solve the potential energy curvature based on the second derivative of the potential energy gradient at each point, and identify points whose potential energy curvature exceeds a preset curvature threshold as maneuver adjustment points. When there is a bifurcation point, calculate the matching degree between the potential energy gradient in each bifurcation direction and the maneuver constraint diagram, and select the direction with the highest matching degree as the main movement direction. When the paths of multiple combat entities intersect at the same spatial location, the path allocation order is determined based on the mission priority of each combat entity and the potential energy value of its dominant potential energy field at the corresponding intersecting spatial location. Based on the main direction of movement and the order of path allocation, generate the action strategy for each combat entity.

9. The underwater combat situation simulation and deduction method under complex marine hydrological environments according to claim 1, characterized in that, Obtaining sound velocity and density profiles for underwater operational areas includes: Collect satellite remote sensing data of sea surface temperature, Argo buoy temperature, salinity and depth profile data, real-time data from underwater sensor network, and historical hydrological observation data; All collected data are interpolated to the same spatial grid. For each spatial grid point, calculate the temperature gradient in the vertical direction between the satellite remote sensing sea surface temperature data and the real-time data from the underwater sensor network; When the temperature gradient exceeds the preset gradient threshold, the inversion parameters of the satellite remote sensing sea surface temperature data are calibrated in reverse using the real-time data of the underwater sensor network. The calibrated satellite remote sensing sea surface temperature data is fused with Argo buoy temperature, salinity, and depth profile data to generate a three-dimensional temperature, salinity, and depth field. Sound velocity profiles and density profiles were extracted from a three-dimensional thermo-salinity field.

10. An underwater combat situation simulation platform for complex marine hydrological environments, characterized in that, include: The acoustic connectivity graph construction module is used to obtain the sound speed profile of the underwater combat area, extract the sound ray propagation path from the sound speed profile, and construct the acoustic connectivity graph with the spatial location pairs of the sound ray that have direct access or first reflection as edges. The connected component decomposition module is used to identify connected components in the acoustic connected graph. Due to the existence of the sound shadow region, different connected components are not connected to each other. The connected component decomposition is performed on the acoustic connected graph, and each connected component is defined as a dominant transfer domain. The boundary between dominant transfer domains corresponds to the location of the sound shadow region. The dominant potential energy field is discontinuous between different dominant transfer domains. The maneuver constraint map construction module is used to obtain the density profile of the underwater combat area, extract the thermocline boundary and density cascade boundary from the density profile, and construct the maneuver constraint map within each advantage transfer domain. The effectiveness prediction module is used to input the acoustic connectivity graph and maneuver constraint graph into the combat effectiveness prediction model to obtain the intrinsic support of each spatial location for underwater combat operations. The potential energy field solution module is used to solve the Poisson equation with the intrinsic support as the potential energy value within each dominant transfer domain to obtain a block-continuous dominant potential energy field. The strategy generation module is used to generate action strategies for each combat entity using the gradient direction of the advantageous potential energy field as the movement direction. When the movement path of the combat entity crosses the boundary of the advantage transfer domain, a detection command or attack command is triggered at the crossing point. When the preset simulation termination condition is met, the simulation result containing all triggered commands is output. When the preset simulation termination condition is not met and the preset environment update time is reached, update the sound velocity profile and density profile, and return to the step of extracting the sound ray propagation path from the sound velocity profile.

Images