Field and water surface survival self-rescue and mutual-rescue VR simulation training system and method

Abstract

The invention relates to a field and water surface survival self-rescue and mutual-rescue VR simulation training system and method, and the method comprises the steps: generating a physical parameter set through constructing a water body-vegetation-human body coupling control equation set, and carrying out the dynamic boundary mapping based on student movement data and equipment signals, and generating a dynamic boundary condition and a rescue tool constraint; calling physical parameters to realize multi-physical field coupling solution, and generating a physical field output set in combination with collision response; converting the physical field data into a tactile feedback waveform and a rendering instruction set by utilizing a vegetation stiffness matrix, and realizing cross-modal synchronous feedback of mechanical characteristics and visual special effects; and dynamically optimizing the physical parameter set based on training feedback data to form a training difficulty adaptive closed-loop system. According to the method, the problem of insufficient training effect migration caused by multi-physical field coupling distortion and environment response stiffness in the prior art is solved, and the physical authenticity of a virtual training scene and the adaptability of student actions are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of wilderness rescue training technology, specifically to a VR simulation training system and method for wilderness and water survival self-rescue and mutual rescue. Background Technology

[0002] Wilderness and water survival self-rescue and mutual rescue training is a crucial component of emergency rescue systems. By highly recreating dangerous scenarios in complex natural environments, it helps trainees master core skills such as water rescue, vegetation obstacle crossing, and collaborative rescue. Virtual reality (VR) technology, due to its repeatability and safety advantages, has become an important medium for this type of training. Its core lies in constructing a physically realistic interactive environment, allowing trainees to gain an immersive experience in scenarios involving the coupling of multiple physical fields, such as virtual water flow, vegetation deformation, and human movement.

[0003] However, existing VR simulation training methods have significant limitations in achieving multiphysics coupling: on the one hand, the physical models of water, vegetation, and human interaction often employ discrete calculations, leading to distortions in the coupling of key characteristics such as fluid resistance, vegetation flexibility, and human inertia. For example, rescue tools lack directionally sensitive mechanical feedback when colliding with vegetation. On the other hand, training systems struggle to dynamically adapt to differences in trainees' abilities. During training, physical parameters such as water viscosity and vegetation stiffness remain constant, making it impossible to optimize environmental responses in real time based on the trainees' movement precision. These shortcomings result in significant deviations between the simulated environment and real disaster scenarios, preventing trainees from obtaining accurate mechanical perception feedback and hindering the effective transfer of training results to real-world situations. Summary of the Invention

[0004] Based on this, the purpose of this invention is to provide a VR simulation training system and method for wilderness and water surface survival self-rescue and mutual rescue that can realize dynamic coupling of multiple physical fields and support adaptive training process.

[0005] The objective of this invention is achieved through the following solution:

[0006] In a first aspect, the present invention provides a VR simulation training method for survival, self-rescue, and mutual rescue in the wild and on water, comprising the following steps:

[0007] S1: Construct a set of coupled control equations for water, vegetation and human body based on pre-stored physical field modeling parameters, and generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0008] S2: Process the acquired student motion data and equipment signals, acquire user joint position data through VR motion capture equipment, calculate joint inertia weight vector based on human mass matrix in physical parameter set, perform weighted mapping and command parsing in combination with controller signals, and generate dynamic boundary conditions and rescue tool constraints.

[0009] S3: Perform multi-physics coupling solution for dynamic boundary conditions and rescue tool constraints, call the viscosity coefficient and vegetation stiffness matrix in the physical parameter set to calculate the force of fluid on vegetation, combine the rescue tool constraints to solve the vegetation deformation displacement and update the fluid velocity field, generate a penalty force when a collision is detected, and generate a physical field output set.

[0010] S4: Process the physical field output set, combine it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic transformation, generate a feedback set containing tactile feedback waveforms and rendering instruction set, and send it to the VR simulation training device.

[0011] S5: Based on the collected training feedback data, combined with the inertial weight vector and dynamic boundary conditions, weighted error calculation is performed to update and adjust the viscosity coefficient and vegetation stiffness matrix of the physical parameter set, and a new physical parameter set is generated.

[0012] In one embodiment, the physical field modeling parameters of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention include fluid dynamic parameters simulating water flow characteristics, plant mechanical parameters simulating plant mechanical responses, and human dynamic parameters simulating human movement characteristics. S1 includes:

[0013] S11: The pre-stored fluid dynamics parameters are processed to construct multi-field coupling equations. Based on the Navier-Stokes equations, the water density and viscosity coefficient are fused to establish a pressure-velocity coupling model and generate the water control equations.

[0014] S12: Perform hierarchical stiffness matching processing on the pre-stored plant mechanical parameters, query the stem-leaf stiffness ratio according to the plant type database and calculate the equivalent stiffness matrix to generate the vegetation flexible body dynamic equation.

[0015] S13: Perform joint dynamics optimization on the pre-stored human kinematic parameters, construct a diagonal mass matrix based on the joint range of motion constraints, integrate the water body control equation, the vegetation flexible body dynamics equation and the diagonal mass matrix to generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0016] In one embodiment, S2 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0017] S21: Obtain the motion data of the trainee to be trained, perform coordinate system normalization on the user's joint position data, transform the three-dimensional space coordinates to the virtual scene coordinate system, and generate standardized joint data.

[0018] S22: Perform inertial weight calculation on the standardized joint data, calculate the mass ratio of each joint based on the human body mass matrix in the physical parameter set and normalize it to generate a weight allocation vector;

[0019] S23: Jointly analyze the weight allocation vector and the acquired training device signal, call the neural network to fuse weighted joint data to generate boundary mapping relationship, generate dynamic boundary conditions through fully connected layer transformation, and simultaneously analyze the handle displacement signal to extract displacement vector and direction angle, and generate rescue tool constraints through inverse kinematics solution.

[0020] In one embodiment, S3 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0021] S31: Perform fluid particle initialization processing on dynamic boundary conditions, distribute smooth particles according to the boundary shape and set the initial velocity field to generate the fluid computational domain;

[0022] S32: Couple the solution of the fluid computation domain and the rescue tool constraint, call the viscosity coefficient in the physical parameter set to calculate the fluid force based on the vortex constraint, and combine the rescue tool constraint to calculate the vegetation deformation displacement considering the fluid-structure interaction, and generate the fluid-vegetation coupling intermediate solution.

[0023] S33: Perform collision simulation on the obtained locations of rescue tools and vegetation nodes, perform dynamic response processing on the collision detection results of tool-vegetation distance calculated by spatial discretization algorithm, generate a penalty force associated with stiffness matrix when the tool-vegetation distance is less than the adaptive distance threshold, and generate a physical field output set containing fluid force, vegetation deformation displacement and penalty force coupling by combining the intermediate solution of fluid-vegetation coupling.

[0024] In one embodiment, the formula for calculating the penalty force of a VR simulation training method for survival and mutual rescue in the wild and on the water provided by the present invention is as follows:

[0025]

[0026] in, For the purpose of punishment, Basic penalty coefficient, Here is the vegetation stiffness matrix. Let be the unit direction vector from which the tool points to the vegetation node. This is the distance attenuation factor.

[0027] In one embodiment, S4 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0028] S41: Perform stiffness-correlated Jacobian analysis on the vegetation deformation displacement, fluid velocity field and coupling force tensor of the physical field output set, calculate the stiffness-correlated Jacobian matrix using the vegetation stiffness matrix in the physical parameter set, and generate dynamic force-displacement gradient mapping relationship.

[0029] S42: Perform tactile waveform synthesis processing on the coupling force tensor and dynamic force-displacement gradient mapping relationship, and generate stiffness-adaptive tactile feedback waveforms through force mapping calculation;

[0030] S43: Perform cross-modal rendering processing on the vegetation deformation displacement and fluid velocity field of the physical field output set, update the skeleton skin mesh based on the stiffness-related deformation parameters, generate a vortex particle system based on the coupling relationship between fluid vorticity and stiffness matrix, perform particle density field and mesh topology avoidance optimization processing on the vortex particle system to generate a rendering instruction set, integrate the haptic feedback waveform and the rendering instruction set to generate a feedback set and send it to the VR simulation training device.

[0031] In one embodiment, the tactile feedback waveform of the VR simulation training method for survival and mutual aid in the wild and on the water provided by the present invention is calculated by the following formula:

[0032]

[0033] in, For tactile feedback waveforms, As the reference amplitude, The Jacobian matrix is ​​related to stiffness. This is the stiffness attenuation factor. Let be the norm of the vegetation stiffness matrix. The fundamental frequency for haptic feedback, This is the force mapping matrix.

[0034] In one embodiment, S5 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0035] S51: Perform kinematic feature extraction on the collected training feedback data, calculate the weighted position deviation by combining the inertial weight vector and dynamic boundary conditions, and generate a biomechanical error vector through feature fusion.

[0036] S52: Perform physical constraint parameter update processing on the biomechanical error vector, dynamically adjust the viscosity coefficient and vegetation stiffness matrix based on the error amplitude, and generate a subset of progressively optimized physical parameters;

[0037] S53: Perform stability verification on a subset of physical parameters, correct parameter mutations through conservation law tests, and integrate water density with the verified parameters to generate a new set of physical parameters.

[0038] Secondly, this invention provides a VR simulation training system for survival, self-rescue, and mutual aid in the wild and on water, which is equipped with the following modules:

[0039] The physical parameter construction module is used to construct a set of coupled control equations for water, vegetation and human body based on pre-stored physical field modeling parameters, and generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0040] The motion data processing module is used to process the acquired student motion data and equipment signals. It acquires user joint position data through VR motion capture equipment, calculates joint inertia weight vectors based on the human body mass matrix in the physical parameter set, and performs weighted mapping and command parsing in combination with the controller signals to generate dynamic boundary conditions and rescue tool constraints.

[0041] The multi-field coupling solution module is used to perform multi-physics coupling solution of dynamic boundary conditions and rescue tool constraints. It calls the viscosity coefficient and vegetation stiffness matrix in the physical parameter set to calculate the force of fluid on vegetation, combines the rescue tool constraints to solve the vegetation deformation displacement and update the fluid velocity field, generates a penalty force when a collision is detected, and generates a physical field output set.

[0042] The physics feedback conversion module is used to process the physical field output set, combine it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic conversion, generate a feedback set containing tactile feedback waveforms and rendering instruction sets, and send it to the VR simulation training device.

[0043] The parameter dynamic update module is used to calculate the weighted error based on the collected training feedback data, combined with the inertial weight vector and dynamic boundary conditions, to update and adjust the viscosity coefficient and vegetation stiffness matrix of the physical parameter set, and to generate a new physical parameter set.

[0044] In summary, the VR simulation training method for wilderness and water survival self-rescue and mutual rescue provided in this application constructs a multi-physics field coupled control equation set of water-vegetation-human body, which can realize the unified dynamic interaction modeling of fluid resistance, vegetation flexibility and human inertia, thus solving the physical distortion problem caused by traditional discrete models. Based on biomechanical characteristics, the joint inertia weight vector is calculated and combined with equipment signal analysis to achieve the collaborative generation of dynamic boundary conditions and rescue tool constraints, effectively overcoming the rigidity of environmental response. By solving the fluid force, vegetation deformation displacement and collision penalty force in real time, the method can achieve direction-sensitive physical field output, significantly improving the realism of the interaction between rescue tools and vegetation. The tactile waveform conversion of vegetation stiffness matrix and eddy current effect rendering are used to achieve cross-modal consistency between mechanical feedback and visual representation. Finally, relying on the closed-loop parameter optimization mechanism of training feedback data, the method can achieve the effect of physical environment self-adaptation to trainee ability, fundamentally solving the industry pain point of insufficient transferability of training effect. This whole process design ensures the physical realism of the virtual scene while constructing an adaptive system for dynamic evolution of training difficulty.

[0045] To better understand and implement this invention, the following detailed description is provided in conjunction with the accompanying drawings. Attached Figure Description

[0046] Figure 1 A flowchart illustrating a VR simulation training method for survival, self-rescue, and mutual aid in the wild and on water, provided in an embodiment of this application;

[0047] Figure 2 A schematic diagram of the process for generating a feedback set including haptic feedback waveforms and a rendering instruction set, provided for embodiments of this application;

[0048] Figure 3 This is a schematic diagram of the structure of a VR simulation training system for survival and mutual rescue in the wild and on the water, provided as another embodiment of this application. Detailed Implementation

[0049] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0051] In one embodiment, such as Figure 1 As shown, a VR simulation training method for survival, self-rescue, and mutual rescue in the wild and on water is provided. This embodiment illustrates the method using a terminal as an example. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0052] S1: Construct a set of coupled control equations for water, vegetation and human body based on pre-stored physical field modeling parameters, and generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0053] Specifically, this embodiment determines pre-stored physical field modeling parameters based on sampling data from real field and water surface scenes. Specifically, for water parameters, the system collects the density and dynamic viscosity coefficient of freshwater in different temperature ranges, and uses a rotational viscometer to measure viscosity data at different water temperatures, establishing a water temperature-viscosity mapping table. For vegetation parameters, the system uses a universal testing machine to test the elastic modulus and Poisson's ratio of stems or branches for common field vegetation types, and discretizes the vegetation into a multi-node mechanical model based on the finite element method, generating a corresponding vegetation stiffness matrix. For human body parameters, the system divides human body models according to weight and height ranges based on a human biomechanical database, uses the DH parameter method to construct a multi-rigid-body linkage model including key joints, calculates the moment of inertia of each joint using the mass distribution formula, and generates a human body mass matrix.

[0054] For example, the system establishes three types of core equations based on multiphysics coupling theory and achieves coupling correlation. The system uses the incompressible Navier-Stokes equations to describe water motion, which include water density, fluid velocity field, pressure, dynamic viscosity coefficient, the force exerted by vegetation on the fluid, and the disturbance force exerted by human movement on the fluid; the latter two are used as coupling terms. The system constructs vegetation deformation equations based on Hooke's law, which include the vegetation stiffness matrix, vegetation node deformation displacement vector, vegetation gravity vector, and the force vector exerted by the fluid on the vegetation. This force vector is derived from the calculation results of the fluid equations, achieving flow-vegetation coupling. Preferably, the system can use the Lagrange equations to describe human motion, which include the human mass matrix, joint angular displacement vector, damping matrix, joint stiffness matrix, human joint driving torque, and the force vector exerted by the fluid on the human body. This force vector is derived from the calculation results of the fluid equations, achieving flow-human coupling.

[0055] The system organizes the key parameters involved in the above equations into a structured set of physical parameters. The set includes a water body subset, a vegetation subset, and a human body subset. The water body subset includes density, dynamic viscosity coefficient, and fluid mesh resolution. The vegetation subset includes stiffness matrix, elastic modulus, Poisson's ratio, and number of nodes. The human body subset includes mass matrix, joint rotational inertia, and link mass distribution ratio.

[0056] S2: Process the acquired student motion data and equipment signals, obtain user joint position data through VR motion capture equipment, calculate joint inertia weight vector based on the human body mass matrix in the physical parameter set, perform weighted mapping and command parsing in combination with the controller signal, and generate dynamic boundary conditions and rescue tool constraints.

[0057] Specifically, this embodiment employs a dual-mode acquisition scheme combining optical capture and inertial measurement unit (IMU) to acquire student motion data and device signals. Specifically, the optical capture system deploys multiple cameras to acquire the three-dimensional coordinates of key human joints. The IMU, integrated into the VR controller and head-mounted display, acquires joint angular velocity and angular acceleration data. The system fuses these two types of data using Kalman filtering to eliminate errors caused by optical occlusion. The system also acquires trigger press depth, joystick displacement, and button status via the VR controller. The acquired device signals are transmitted in real-time to the data processing unit through a designated interface. The system calculates joint inertial weight vectors based on the human body mass matrix, extracts the diagonal elements corresponding to each joint in the human body mass matrix (i.e., joint rotational inertia), normalizes the rotational inertia, and calculates the ratio of each joint's rotational inertia to the sum of all joint rotational inertia to obtain the inertial weight of each joint, thereby generating the joint inertial weight vector.

[0058] Furthermore, the system performs weighted mapping between the handle signal and the inertial weight vector and parses the mapping results into instructions. The system maps trigger button press depth to the grip force of the rescue tool, calculating this by combining wrist joint inertial weight and the maximum grip force of the rescue tool. The system maps joystick displacement to the movement speed of the rescue tool, also calculating this by combining elbow joint inertial weight and the maximum movement speed of the rescue tool. The system converts the mapped physical quantities into machine-recognizable constraint instructions. Different physical quantities correspond to different constraint types: grip force corresponds to the fixed constraint between the rescue tool and the hand, and movement speed corresponds to the tool's trajectory constraint. The system constructs dynamic boundary conditions based on human motion data. It fits the three-dimensional coordinates of human joints into a continuous human surface mesh, which serves as the moving boundary for the fluid equations. The system sets a no-slip boundary condition for this moving boundary, ensuring that the fluid velocity on the human surface matches the velocity on the human surface. The system compares the human motion trajectory with the coordinates of vegetation nodes in real time. When the distance between the human and vegetation reaches a set threshold, the vegetation contact boundary is triggered, and the system calculates the boundary force of the contact boundary according to Hertzian contact theory.

[0059] S3: Perform multi-physics coupling solution for dynamic boundary conditions and rescue tool constraints. Calculate the fluid's force on vegetation by calling the viscosity coefficient and vegetation stiffness matrix in the physical parameter set. Combine the rescue tool constraints to solve for vegetation deformation displacement and update the fluid velocity field. When a collision is detected, generate a penalty force and generate a physical field output set.

[0060] Specifically, the system can employ a combined finite volume method and finite element method framework to perform multi-physics coupling solutions for dynamic boundary conditions and rescue tool constraints. The system develops a fluid solver based on the OpenFOAM open-source library. The fluid solver uses a structured hexahedral mesh as the fluid mesh, sets a specific time step to meet the stability requirements of the CFL number, and uses the SIMPLE algorithm to achieve pressure-velocity coupling. The system also performs secondary development based on ABAQUS to obtain a vegetation solver. The system imports the vegetation stiffness matrix into the vegetation solver, uses the Newmark-β method to solve the vegetation deformation equation, and sets corresponding β and γ values ​​to ensure numerical stability. The time step of the vegetation solver is synchronized with that of the fluid solver. Preferably, the system can use a data interaction protocol to construct a coupling interface, through which real-time data transmission between the fluid solver and the vegetation solver is achieved, and a specific transmission frequency is set to ensure coupling timeliness.

[0061] Furthermore, the system calls the viscosity coefficient and vegetation stiffness matrix from the physical parameter set to calculate the fluid force on the vegetation. The system first extracts the velocity of the fluid grid where the vegetation node is located, and then uses the drag force formula to calculate the fluid force on a single vegetation node. The formula includes water density, drag coefficient, the vegetation projection area corresponding to the node, and the difference between the fluid velocity and the vegetation node's movement velocity. The system integrates the fluid force on all vegetation nodes to obtain the fluid force vector on the entire vegetation, and substitutes this vector into the vegetation elasticity equation. Combining the rescue tool constraints, the system solves for the vegetation deformation displacement and updates the fluid velocity field. The system substitutes the fluid force vector on the vegetation and the force corresponding to the rescue tool constraints into the vegetation elasticity equation to solve for the deformation displacement of each vegetation node and updates the vegetation geometric model based on this displacement. The system uses the immersion boundary method to embed the updated vegetation geometric model into the fluid grid. The velocity value of the fluid grid near the vegetation boundary is adjusted through the interpolation function. During the interpolation process, the vegetation boundary coordinates, fluid grid node coordinates, interpolation kernel function, and fluid velocity unaffected by vegetation are combined for calculation.

[0062] Furthermore, the system detects collisions and generates penalty forces when collisions are detected. The system employs a combination of axis-aligned bounding boxes and discrete collision detection for collision detection. First, the system compares the axis-aligned bounding boxes of the human body or rescue tool with the axis-aligned bounding boxes of vegetation or water body boundaries in real time. When the two types of bounding boxes intersect, the GJK algorithm is used to perform triangular facet-level collision detection to determine whether a real collision exists. When a collision is detected, the system uses a spring-damped penalty model to generate a penalty force. The penalty force calculation includes penalty stiffness, collision penetration depth, damping coefficient, and the difference between the velocity of the colliding body and the velocity of the collided body. The system substitutes the generated penalty force into the human body dynamics equation and the vegetation deformation equation to influence the human body's trajectory and enhance the vegetation deformation effect. The system organizes the results obtained from multi-physics coupling solutions into a set of physical field outputs. The output set includes fluid sub-outputs, vegetation sub-outputs, and human sub-outputs. The fluid sub-outputs include velocity field distribution, pressure field distribution, and fluid force vector. The vegetation sub-outputs include nodal deformation displacement vector, vegetation stress distribution, and collision penalty force. The human sub-outputs include joint force vector and human motion trajectory.

[0063] S4: Process the physical field output set, combine it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic transformation, generate a feedback set containing haptic feedback waveforms and rendering instruction set, and send it to the VR simulation training device.

[0064] Specifically, the system processes the physical field output set and combines it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic transformation to generate a feedback set. During the tactile feedback waveform transformation process, the system extracts the deformation displacement in the vegetation sub-output and calculates the reaction force of the vegetation on the human body or tool by combining it with the vegetation stiffness matrix; then, the reaction force is converted into a tactile feedback amplitude, and the transformation process adopts a linear mapping method; the system determines the feedback frequency according to the rate of change of the force, and selects a sine wave or square wave as the tactile feedback waveform according to the type of force change, with abrupt force corresponding to a specific waveform and gradual force corresponding to another specific waveform.

[0065] During the rendering instruction set conversion process, the system converts the velocity field output by the fluid sub-output into water surface ripple rendering parameters, where velocity is related to the amplitude and wavelength of the ripples; it converts vegetation deformation displacement into vegetation vertex update instructions, updating the vertex coordinates of the vegetation model at a specific frequency to match the refresh rate of the VR headset; and it converts collision penalty force into collision effect instructions, which include parameters such as particle number and particle velocity.

[0066] The system organizes the converted haptic feedback parameters and rendering instruction sets into a feedback set. The feedback set includes a haptic feedback subset, a rendering instruction subset, and synchronization instructions. The haptic feedback subset includes feedback force amplitude, feedback frequency, and waveform type. The rendering instruction subset includes water ripple parameters, vegetation vertex update coordinates, and collision effect parameters. The synchronization instructions include feedback delay compensation values ​​and frame synchronization signals.

[0067] The system sends the feedback set to the VR simulation training device and controls the device. The system uses the UDP protocol to transmit the feedback set and sets a specific transmission rate to ensure that the data transmission latency of each frame meets the requirements. The system is compatible with various VR simulation training devices. After receiving the haptic feedback waveform, the haptic feedback glove generates stress feedback through a built-in micro linear motor. After receiving the rendering command set, the VR headset updates the rendering of fluid, vegetation, and collision effects in the virtual scene. After receiving the collision penalty force signal, the force feedback handle generates vibration feedback through a built-in vibration motor. The vibration intensity is related to the amplitude of the penalty force. The system calibrates the transmission latency of the feedback set every certain number of frames and ensures that the time difference between haptic feedback and visual rendering meets the requirements by adjusting the latency compensation value in the synchronization command.

[0068] S5: Based on the collected training feedback data, combined with the inertial weight vector and dynamic boundary conditions, weighted error calculation is performed to update and adjust the viscosity coefficient and vegetation stiffness matrix of the physical parameter set, and a new physical parameter set is generated.

[0069] Specifically, the system collects training feedback data at a frequency consistent with the data collection frequency. The collected training feedback data includes objective and subjective data. Objective data includes operational accuracy data and equipment feedback data. Operational accuracy data includes the deviation between the actual position of the rescue tool and the target position and the time to complete the action. Equipment feedback data includes the feedback latency of the VR device and the error of haptic feedback force. Subjective data is collected through the interactive interface built into the VR headset. After completing each training task, the trainee rates the realism of the training scenario, and the system records this rate as subjective data.

[0070] Furthermore, the system combines inertia weight vectors and dynamic boundary conditions to calculate weighted errors. The system first calculates each error component: the position deviation weight error is obtained by multiplying the tool position deviation corresponding to each joint by the inertia weight of that joint and then summing the results; the feedback force error is calculated as the ratio of the absolute value of the tactile feedback force error to the target feedback force, and this ratio is normalized; the delay error is calculated as the ratio of the VR device's feedback delay to the maximum allowable delay, and this ratio is normalized; the subjective rating error is calculated as the ratio of the difference between 5 and the student's subjective rating to 5, and this ratio is normalized. The system multiplies each error component by its corresponding error weight coefficient and then sums the results to obtain the total weighted error. The error weight coefficients can be adjusted according to the training objectives.

[0071] Furthermore, the system updates and adjusts the viscosity coefficient and vegetation stiffness matrix of the physical parameter set based on the total weighted error. The system can use a PID control algorithm for parameter updates. For the viscosity coefficient, when the total weighted error is greater than a set threshold, the viscosity coefficient is increased, and the increase is determined by the original viscosity coefficient and the formula including the total weighted error. When the total weighted error is less than another set threshold, the viscosity coefficient is decreased, and the decrease is determined by the original viscosity coefficient and the formula including the total weighted error. The updated viscosity coefficient must be within the range determined based on real water body data. For the vegetation stiffness matrix, the system extracts the diagonal elements of the stiffness matrix, i.e., the stiffness coefficients of each node, and determines the updated stiffness coefficients by the original stiffness coefficients and the formula including the total weighted error and the vegetation stiffness adjustment coefficient. The system verifies the positive definiteness of the updated stiffness matrix through eigenvalue decomposition to ensure that the vegetation deformation equation has a unique solution. The verification criterion is that the minimum eigenvalue of the stiffness matrix satisfies the set conditions. The system integrates the updated viscosity coefficient and vegetation stiffness matrix with the unadjusted parameters in S1 to generate a new set of physical parameters. This new set of physical parameters is used for steps S1 to S4 of the next training round, forming a closed-loop adaptive mechanism of training-feedback-optimization.

[0072] In summary, the VR simulation training method for wilderness and water survival self-rescue and mutual rescue provided in this application constructs a multi-physics field coupled control equation set of water-vegetation-human body, which can realize the unified dynamic interaction modeling of fluid resistance, vegetation flexibility and human inertia, thus solving the physical distortion problem caused by traditional discrete models. Based on biomechanical characteristics, the joint inertia weight vector is calculated and combined with equipment signal analysis to achieve the collaborative generation of dynamic boundary conditions and rescue tool constraints, effectively overcoming the rigidity of environmental response. By solving the fluid force, vegetation deformation displacement and collision penalty force in real time, the method can achieve direction-sensitive physical field output, significantly improving the realism of the interaction between rescue tools and vegetation. The tactile waveform conversion of vegetation stiffness matrix and eddy current effect rendering are used to achieve cross-modal consistency between mechanical feedback and visual representation. Finally, relying on the closed-loop parameter optimization mechanism of training feedback data, the method can achieve the effect of physical environment self-adaptation to trainee ability, fundamentally solving the industry pain point of insufficient transferability of training effect. This whole process design ensures the physical realism of the virtual scene while constructing an adaptive system for dynamic evolution of training difficulty.

[0073] In one embodiment, the physical field modeling parameters of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention include fluid dynamic parameters simulating water flow characteristics, plant mechanical parameters simulating plant mechanical responses, and human dynamic parameters simulating human movement characteristics. S1 includes:

[0074] S11: The pre-stored fluid dynamics parameters are processed to construct multi-field coupling equations. Based on the Navier-Stokes equations, the water density and viscosity coefficient are integrated to establish a pressure-velocity coupling model and generate the water body control equations.

[0075] Specifically, the system calls pre-stored fluid dynamics parameters, which are obtained based on sampling from real water scenes and include basic mechanical property data of water under different environmental conditions. The system is based on incompressible fluid mechanics theory and selects the Navier-Stokes equations as the core equation framework to describe the motion state of water. Preferably, the system performs formal adaptation of the Navier-Stokes equations, neglecting compressibility effects and retaining viscous and inertial terms for the flow characteristics of water in outdoor and surface survival scenarios, ensuring that the equations accurately reflect the velocity changes and pressure distribution of water during interactions with humans and vegetation.

[0076] Furthermore, the system incorporates pre-stored water density and viscosity coefficients into the equations to construct a pressure-velocity coupled model. During model construction, based on the fluid continuum assumption, the system uses water density as the core parameter of the inertial term, substituting it into the mass conservation equation to ensure the equations reflect the inertial effects of water motion. Simultaneously, the viscosity coefficient is substituted into the viscous term of the momentum conservation equation to describe the hindering effect of viscous forces within the water on flow. To achieve pressure and velocity coupling, the system employs the core logic of the SIMPLE algorithm to establish the correlation between the pressure correction equation and the velocity correction equation. Iterative calculations ensure that the pressure and velocity fields satisfy the mass and momentum conservation conditions, avoiding calculation errors caused by the separation of velocity and pressure.

[0077] Through the above processing steps, the system generates the water body control equations, which consist of two parts: a mass conservation equation and a momentum conservation equation. The mass conservation equation constrains the continuity of water flow, ensuring that the mass of water flowing into and out of a control volume per unit time is equal. The momentum conservation equation describes the relationship between water velocity changes and forces, including inertial force terms, viscous force terms, pressure gradient terms, and external force terms. The external force terms reserve a coupling interface for interaction with humans and vegetation, laying the foundation for subsequent multi-field coupled solutions. The system standardizes the generated water body control equations, clarifying the physical meaning and mathematical expression of each variable in the equations, ensuring that the equations can be directly called by subsequent solvers.

[0078] S12: Perform hierarchical stiffness matching processing on the pre-stored plant mechanical parameters, query the stem-leaf stiffness ratio according to the plant type database and calculate the equivalent stiffness matrix to generate the vegetation flexible body dynamic equation.

[0079] Specifically, the system extracts pre-stored plant mechanical parameters, which cover stem and leaf mechanical property data for different vegetation types, including basic stiffness data and deformation response characteristics data for various vegetation types. The system also accesses a plant type database, which stores classification information of common wild vegetation and corresponding stem-leaf structure ratios and mechanical property differences. Through vegetation type matching, the system queries the database for the stem-leaf stiffness ratio of the currently processed vegetation type to determine the weighting relationship between stems and leaves in mechanical load-bearing capacity.

[0080] Furthermore, the system performs hierarchical stiffness matching processing on pre-stored plant mechanical parameters. For the stem structure of the vegetation, the system treats it as the main load-bearing component and uses a beam element mechanical model for stiffness calculation. Based on pre-stored parameters such as stem elastic modulus and cross-sectional dimensions, the system calculates the bending and torsional stiffness of the stem using material mechanics formulas. For the leaf structure of the vegetation, the system treats it as an auxiliary flexible component and uses a membrane element mechanical model for stiffness calculation. Based on pre-stored parameters such as leaf thickness and elastic coefficient, the system calculates the flexible stiffness of the leaf in the in-plane and out-of-plane directions. The system weights and integrates the stem stiffness and leaf stiffness according to the retrieved stem-leaf stiffness ratio. Using the finite element discretization method, the entire vegetation is divided into several mechanical nodes. Based on the nodal force balance conditions, the system calculates the stiffness contribution value corresponding to each node, thereby generating the equivalent stiffness matrix of the entire vegetation.

[0081] Preferably, the system constructs the vegetation flexible body dynamic equations using the generated equivalent stiffness matrix as the core. These equations use the deformation displacement of vegetation nodes as the core variables. The left side of the equation contains a product term of the equivalent stiffness matrix and the deformation displacement, used to describe the elastic restoring force during vegetation deformation. The right side of the equation contains terms related to the vegetation's own gravity and external forces. The external force term has a reserved interface for interaction with water bodies and human bodies, allowing it to receive load data such as water body forces and human contact forces transmitted during subsequent coupled solution processes. The system adapts the boundary conditions of the vegetation flexible body dynamic equations, clarifies the constraint state of the nodes connecting the vegetation and the ground, and ensures that the equations accurately reflect the deformation response of the vegetation under stress, ultimately completing the standardized generation of the equations.

[0082] S13: Perform joint dynamics optimization on the pre-stored human kinematic parameters, construct a diagonal mass matrix based on the joint range of motion constraints, integrate the water body control equation, the vegetation flexible body dynamics equation and the diagonal mass matrix to generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0083] Specifically, the system reads pre-stored human kinematic parameters, which are constructed based on human biomechanical research data and include data related to the range of motion, rotational characteristics, and mass distribution of each joint. Based on joint range of motion constraints, the system defines the degrees of freedom of motion for each key joint, clarifying the rotational limits of each joint in three-dimensional space, such as the flexion-extension, abduction-adduction angle range of the hip joint, and the flexion-extension angle range of the elbow joint. Based on these constraints, the system optimizes the construction logic of the human mass matrix to ensure that the mass matrix matches the actual motion capabilities of the joints.

[0084] Furthermore, the system constructs a diagonal mass matrix, where the diagonal elements correspond to the rotational inertia of each key joint in the human body. During the calculation of rotational inertia, the system uses pre-stored human body mass distribution data, considering the mass proportion of the limb segment to which the joint belongs, and combining parameters such as the length and cross-sectional dimensions of the limb segment, to calculate the rotational inertia value of each joint using the rigid body rotational inertia calculation formula. Simultaneously, the system fine-tunes the calculated rotational inertia based on joint range of motion constraints. For example, for joints with a large range of motion, the rotational inertia calculation coefficient is appropriately adjusted to ensure that the rotational inertia reflects the inertial characteristics of the joint under different motion states. The system arranges the rotational inertia of each joint in order of joint number, forming a diagonal mass matrix, with off-diagonal elements set to zero to simplify subsequent dynamic calculations.

[0085] Furthermore, the system integrates the water body control equation, the vegetation flexible body dynamics equation, and the currently constructed diagonal mass matrix. During integration, the system clarifies the coupling logic between the three equations: the external force term in the water body control equation is linked to the vegetation flexible body dynamics equation and the human body model corresponding to the diagonal mass matrix, used to transmit the forces exerted by the water on the vegetation and the human body; the external force term in the vegetation flexible body dynamics equation is linked to the human body model corresponding to the diagonal mass matrix, used to transmit the contact forces between the vegetation and the human body. Based on the integrated equation system, the system extracts core parameters, including water density and viscosity coefficient from the water body control equation, the equivalent stiffness matrix from the vegetation flexible body dynamics equation, and the human body mass matrix corresponding to the diagonal mass matrix. These parameters are then organized in a structured form to generate a set of physical parameters.

[0086] In one embodiment, S2 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0087] S21: Obtain the motion data of the trainee to be trained, perform coordinate system normalization on the user's joint position data, transform the three-dimensional space coordinates to the virtual scene coordinate system, and generate standardized joint data.

[0088] Specifically, the system acquires motion data from the trainee, focusing on the positional data of key joints throughout the trainee's body. The data source is continuous joint position information output by VR motion capture devices. Preferably, the system performs coordinate system normalization on the acquired user joint position data. This normalization process uses a preset reference point in the virtual scene as a benchmark to determine the origin and coordinate axis directions of the reference coordinate system. The reference point is selected as a fixed position in the virtual scene that is associated with the training scene structure, and the coordinate axis directions are consistent with the spatial dimensions of the virtual scene. The system calculates the transformation relationship between the coordinate system of the original joint position data and the reference coordinate system, eliminating deviations caused by differences in coordinate systems from different acquisition devices through a coordinate transformation algorithm, ensuring that all joint position data are consistent within the same reference frame.

[0089] After coordinate system normalization, the system transforms the processed 3D spatial coordinates to the virtual scene coordinate system. It calls the coordinate system parameters of the virtual scene, which include the spatial scale and coordinate mapping rules of the virtual scene. Based on these rules, a correspondence is established between the reference coordinate system and the virtual scene coordinate system. A linear transformation formula is then used to map the joint position data from the reference coordinate system to the virtual scene coordinate system. During the transformation process, the system performs continuity checks on the data, eliminating discrete data points caused by equipment acquisition errors to ensure that the transformed joint position data continuously reflects the movement trajectory of the trainee's joints. The system then standardizes the format of the transformed joint data, clarifying the data storage structure and field definitions. The standardized data for each joint includes the 3D coordinate values ​​in the virtual scene coordinate system and the data acquisition timestamp, ultimately generating standardized joint data to provide a unified format of input data for subsequent inertia weight calculations.

[0090] S22: Perform inertial weight calculation on the standardized joint data, calculate the mass ratio of each joint based on the human body mass matrix in the physical parameter set and normalize it to generate a weight allocation vector.

[0091] Specifically, the system calls the human body mass matrix from the physical parameter set. This matrix stores mass-related parameters corresponding to each key joint of the human body. The system extracts data from the human body mass matrix, extracting elements directly associated with each joint. These elements reflect the mass contribution characteristics of the corresponding joint. Based on the extracted elements, the system calculates the mass percentage of each joint. The calculation logic is to use the sum of the mass contributions of all joints as a benchmark, and then calculate the ratio of the mass contribution of each individual joint to the sum. This ratio is the mass percentage of the corresponding joint.

[0092] Furthermore, the system performs normalization processing on the calculated mass proportions of each joint. During this process, a linear transformation is used to adjust the mass proportions of all joints to the same numerical range, ensuring that the sum of the mass proportions of all joints meets the preset conditions and eliminating weight bias caused by excessive differences in the mass contributions of different joints. After normalization, the system sorts the normalized mass proportions of each joint, following the same sorting rules as the joint numbering order in the standardized joint data, ensuring a one-to-one correspondence between the mass proportion of each joint and the corresponding joint data. The system then integrates the sorted normalized mass proportions in sequence to form a vector data structure, where each element corresponds to the weight value of a joint, ultimately generating a weight allocation vector for subsequent joint analysis with the training device signals.

[0093] S23: Jointly analyze the weight allocation vector and the acquired training device signal, call the neural network to fuse weighted joint data to generate boundary mapping relationship, generate dynamic boundary conditions through fully connected layer transformation, and simultaneously analyze the handle displacement signal to extract displacement vector and direction angle, and generate rescue tool constraints through inverse kinematics solution.

[0094] Specifically, the system calls upon standardized joint data and weight allocation vectors, while simultaneously acquiring training device signals, including operation and status signals output by the VR controllers. The system first fuses the weight allocation vectors and standardized joint data, associating each weight value in the weight allocation vector with its corresponding standardized joint data to form weighted joint data. The association method involves establishing a mapping relationship between the standardized coordinate values ​​of each joint and its corresponding weight value.

[0095] Preferably, the system can invoke a preset neural network model, inputting weighted joint data into the neural network. This neural network is used to learn the correlation between joint data and the virtual scene boundary. The neural network extracts features from the weighted joint data through hidden layers, uncovering the potential relationship between joint movement and scene boundary interaction, and generating boundary mapping relationships. The system inputs the boundary mapping relationships into the fully connected layer of the neural network. The fully connected layer transforms the boundary mapping relationships into a parametric form that conforms to the physical rules of the virtual scene through linear transformation. These parameters include information such as the spatial range and motion characteristics of the boundary. Based on these parameters, the system generates dynamic boundary conditions, which are updated in real time according to the student's joint movements.

[0096] Simultaneously, the system analyzes the handle displacement signal from the acquired training device signals, extracting displacement-related parameters to obtain the displacement vector and orientation angle. The displacement vector reflects the range of handle movement in space, while the orientation angle reflects the spatial direction of the handle's movement. The system invokes an inverse kinematics algorithm, using the displacement vector and orientation angle as input, to calculate the spatial attitude and motion constraints of the rescue tool in the virtual scene. During the calculation, the structural characteristics of the rescue tool are considered to ensure that the inverse solution results conform to the actual motion laws of the tool. Based on the results of the inverse kinematics solution, the system extracts the motion constraint parameters of the rescue tool. These parameters include the tool's range of movement, rotation angle limitations, etc., ultimately generating rescue tool constraints for subsequent multiphysics coupling solutions.

[0097] In one embodiment, S3 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0098] S31: Perform fluid particle initialization processing on dynamic boundary conditions, distribute smooth particles according to the boundary shape and set the initial velocity field to generate the fluid computational domain.

[0099] Specifically, the system reads dynamic boundary conditions, which include boundary geometry data and boundary motion trend information in the virtual scene coordinate system. The system initializes the dynamic boundary conditions with fluid particles based on the smooth particle hydrodynamics method. First, it analyzes the geometric features of the boundary shape to determine the spatial extent of the fluid region. Then, it expands a predetermined region into the fluid based on the boundary surface, forming the distribution space of the fluid particles. Preferably, the system distributes smooth particles according to a preset particle spacing rule. During particle distribution, a kernel function is used to describe the interaction between particles. The core formula is:

[0100]

[0101] in, For kernel function values, Let be the distance between the two fluid particles. Let d be the smooth length and d be the spatial dimension. This formula ensures a smooth transition in particle interactions, preventing gaps or excessive density in particle distribution. The system sets the initial velocity field for the fluid particles, based on the initial conditions of the water body control equations. For simulating static water, the initial velocity vectors of all fluid particles are set to zero; for simulating flowing water, initial velocity vectors are assigned according to the preset flow direction and velocity. The system integrates the distributed fluid particles with the initial velocity field, recording the three-dimensional coordinates, initial velocity, mass, and other parameters of each fluid particle, forming a fluid computational domain containing particle set and velocity field data.

[0102] S32: Couple the solution processing of the fluid computation domain and the rescue tool constraint, call the viscosity coefficient in the physical parameter set to calculate the fluid force based on the vortex constraint, and combine the rescue tool constraint to calculate the vegetation deformation displacement considering the fluid-structure interaction, and generate the fluid-vegetation coupled intermediate solution.

[0103] Specifically, the system calls upon the viscosity coefficient from the physical parameter set to perform fluid force calculations based on vortex constraints in the fluid computational domain. Preferably, the system extracts the velocity vectors of each fluid particle in the fluid computational domain and calculates the vortex of each particle through a vector cross product operation. The core formula for vortex calculation is:

[0104]

[0105] in, The vorticity vector. For gradient operators, Let be the velocity vector of the fluid particles; this formula reflects the rotational characteristics of the fluid particles. The system then substitutes the viscosity coefficient into the eddy diffusion equation, which takes the form:

[0106]

[0107] in, For time, The viscosity coefficient is... The Laplace operator is used in this equation to calculate the vorticity diffusion coefficient in the fluid. Based on the diffusion coefficient and the relative vorticity difference between particles, the vorticity constraint force on each fluid particle is determined. The system reads the constraints of the rescue tool and substitutes them as fluid-structure interaction boundary conditions into the vegetation flexible body dynamics equation. It calculates the load of the fluid force on the vegetation nodes and the tool contact force, which together serve as the external force load for vegetation deformation. The system performs coupled iterative processing on the fluid force and vegetation deformation displacement. Based on the updated vegetation node positions, it adjusts the fluid particle distribution boundary and recalculates the fluid force until the changes in fluid force and vegetation deformation displacement are less than a set threshold, generating a fluid-vegetation coupled intermediate solution that includes the fluid velocity field, fluid force, and vegetation deformation displacement.

[0108] S33: Perform collision simulation on the obtained locations of rescue tools and vegetation nodes, perform dynamic response processing on the collision detection results of tool-vegetation distance calculated by spatial discretization algorithm, generate a penalty force associated with stiffness matrix when the tool-vegetation distance is less than the adaptive distance threshold, and generate a physical field output set containing fluid force, vegetation deformation displacement and penalty force coupling by combining the intermediate solution of fluid-vegetation coupling.

[0109] Specifically, the system acquires rescue tool location data and vegetation node location data. The rescue tool location data comes from the signal analysis results of the training equipment, and the vegetation node location data comes from the vegetation flexible body model. Preferably, the system can use a spatial discretization algorithm to process the two types of location data. In the algorithm, the virtual scene space is divided into discrete units using a hash function. The hash function formula is:

[0110]

[0111] in, The hash value of the spatial unit. For position coordinates, The formula, representing the number of units in the spatial dimension, is used to quickly filter tool-vegetation node pairs that may collide. The straight-line distance between the filtered node pairs is then calculated to obtain the collision detection result. The system performs dynamic response processing on the collision detection result, invoking an adaptive distance threshold. This threshold is adjusted based on the diagonal element values ​​of the vegetation stiffness matrix. If the tool-vegetation distance is less than the adaptive distance threshold, the system calculates the penalty force according to a preset formula:

[0112]

[0113] in, For the purpose of punishment, Basic penalty coefficient, Here is the vegetation stiffness matrix. Let be the unit direction vector from which the tool points to the vegetation node. The distance attenuation factor is used. The system integrates the penalty force with the intermediate solution of fluid-vegetation coupling, performs format standardization on the integrated data, and finally generates a set of physical field outputs that includes fluid forces, vegetation deformation and displacement, and the coupling of penalty forces.

[0114] In one embodiment, such as Figure 2 As shown, S4 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by this invention specifically includes the following steps:

[0115] S41: Perform stiffness-correlated Jacobian analysis on the vegetation deformation displacement, fluid velocity field and coupling force tensor of the physical field output set, calculate the stiffness-correlated Jacobian matrix using the vegetation stiffness matrix in the physical parameter set, and generate the dynamic force-displacement gradient mapping relationship.

[0116] Specifically, the system reads vegetation deformation displacement, fluid velocity field, and coupling force tensor from the physical field output set. It first performs format standardization on these three types of data, and then converts the vegetation deformation displacement into vector form based on the number of vegetation nodes in the physical parameter set. Based on the particle distribution characteristics of the fluid computational domain, the fluid velocity field is transformed into a velocity vector field. According to the rules governing the composition of force components, the coupling force tensor is converted into matrix form. This ensures that the three data dimensions meet the requirements of subsequent calculations. The system calls the vegetation stiffness matrix from the physical parameter set. Initiate the stiffness-related Jacobian analytical process, and calculate the stiffness-related Jacobian matrix through partial differential operations. The calculation follows the formula:

[0117]

[0118] in, The stiffness-related Jacobian matrix describes the linear relationship between the coupling force tensor and vegetation deformation displacement. For the coupling force tensor in the physical field output set, This is the vegetation deformation displacement vector.

[0119] Furthermore, the system... To verify its validity, the system uses matrix rank analysis to determine if it is full rank. If the rank is insufficient, the system calls the vegetation stiffness matrix. Correction The revised version passed verification and is based on [the system]. Generate a dynamic force-displacement gradient mapping relationship. This relationship is stored in a data table, with each row containing the vegetation deformation displacement increment. Each column contains the coupling force increment. And satisfy This provides a quantitative correlation between force and displacement for subsequent tactile waveform synthesis.

[0120] S42: Perform tactile waveform synthesis processing on the coupling force tensor and dynamic force-displacement gradient mapping relationship, and generate stiffness-adaptive tactile feedback waveforms through force mapping calculation.

[0121] Specifically, the system extracts the Jacobian matrix. Read the coupling force tensor from the physical field output set The system performs dimensional matching on the two types of data. A force mapping matrix is ​​constructed. , The dimension is determined by the coupling force tensor. The number of components and the dimension of the force vector that the haptic feedback device can receive are determined by... Will Convert to A consistent format ensures that the data can be used for waveform calculations.

[0122] Preferably, the system performs tactile waveform synthesis by calling the following preset formula:

[0123]

[0124] in, For tactile feedback waveforms, The reference amplitude is used to set the feedback base strength; This is the Jacobian matrix for stiffness correlation, which correlates vegetation stiffness with changes in feedback force. This is the force mapping matrix, used to adapt the dimensions of the coupled force tensor. This is the stiffness attenuation factor, which adjusts the effect of stiffness on the waveform attenuation rate. The norm of the vegetation stiffness matrix is ​​calculated from the vegetation stiffness matrix. The fundamental frequency of the waveform is set as the haptic feedback fundamental frequency. The system determines the calculation time step according to the VR device refresh cycle and calculates time-by-time. After the calculation is completed, the waveform data is smoothed using the moving average method to generate a stiffness-adaptive tactile feedback waveform.

[0125] S43: Perform cross-modal rendering processing on the vegetation deformation displacement and fluid velocity field of the physical field output set, update the skeleton skin mesh based on the stiffness-related deformation parameters, generate a vortex particle system based on the coupling relationship between fluid vorticity and stiffness matrix, perform particle density field and mesh topology avoidance optimization processing on the vortex particle system to generate a rendering instruction set, integrate the haptic feedback waveform and the rendering instruction set to generate a feedback set and send it to the VR simulation training device.

[0126] Specifically, the system reads the vegetation deformation displacement from the physical field output set. Jacobian matrix based on stiffness correlation Extract the deformation parameters associated with stiffness, following the formula:

[0127]

[0128] in, These are stiffness-related deformation parameters, reflecting the influence of vegetation stiffness on deformation morphology. The system will... The input skeleton skin mesh update process adjusts the skeleton joint angles of the vegetation model and updates the vertex coordinates of the skin mesh according to the vertex weight allocation rules pre-calculated in the vegetation model binding stage, so that the vegetation visualization deformation and stiffness characteristics are consistent.

[0129] The system reads the fluid velocity field from the physical field output set. Calculate fluid vorticity The calculation follows the formula:

[0130]

[0131] in, For curl operator, This is the fluid vorticity vector, reflecting the intensity and direction of fluid rotation. The system will... Coupled with the vegetation stiffness matrix norm K, a vortex particle system is generated. The particle generation location is set to a region with large vorticity, and the initial velocity direction of the particles is consistent with the vorticity direction. The system calculates the distance between the particles and the vertices of the vegetation skin mesh, optimizes the particle position to avoid visual penetration, and then integrates the vegetation skin mesh update instructions and vortex particle system parameters to generate a rendering instruction set. Finally, it integrates the haptic feedback waveform to generate a feedback set, which is sent to the VR simulation training device through a data transmission protocol.

[0132] In one embodiment, S5 of the VR simulation training method for survival, self-rescue, and mutual aid in the wild and on the water provided by the present invention specifically includes the following steps:

[0133] S51: Perform kinematic feature extraction on the collected training feedback data, calculate the weighted position deviation by combining the inertial weight vector and dynamic boundary conditions, and generate a biomechanical error vector through feature fusion.

[0134] Specifically, after acquiring the collected training feedback data, the system initiates a kinematic feature extraction and processing flow. The training feedback data includes the trainee's actual joint position data, motion trajectory data, and movement timing data. The system first extracts kinematic features such as joint position deviation and motion velocity deviation from the data, where the joint position deviation is the difference between the trainee's actual joint position and the standard joint position. Preferably, the system calls the previously generated inertial weight vector. With dynamic boundary conditions, a weighted position deviation calculation model is constructed, where the weighted position deviation... The calculation formula is:

[0135]

[0136] in, This represents the vector dot product operation. This is the vector representing the actual joint position of the trainee. The standard joint position vector is defined by dynamic boundary conditions. The system integrates features such as weighted position deviation and motion velocity deviation into a unified biomechanical error vector through a feature fusion algorithm. This vector can quantify the degree of deviation between the trainee's movements and the standard movements, providing an error basis for subsequent adjustments to physical parameters.

[0137] S52: Perform physical constraint parameter update processing on the biomechanical error vector, dynamically adjust the viscosity coefficient and vegetation stiffness matrix based on the error amplitude, and generate a subset of progressively optimized physical parameters.

[0138] Specifically, the system receives the biomechanical error vector. Then, the magnitude of the error vector can be calculated using the L2 norm. The system determines the direction and magnitude of parameter adjustment based on the error amplitude: when the error amplitude is large, the parameter adjustment magnitude is increased to accelerate the adaptation speed; when the error amplitude is small, the adjustment magnitude is decreased to ensure parameter stability. The system targets the viscosity coefficient within the set of physical parameters. Adjustment formulas are constructed for the vegetation stiffness matrix K, the viscosity coefficient adjustment formula, and the vegetation stiffness matrix adjustment formula, respectively:

[0139]

[0140]

[0141] in, The adjusted viscosity coefficient. The viscosity coefficient before adjustment. This is the adjusted vegetation stiffness matrix. The original vegetation stiffness matrix is ​​shown below. The error coefficient is used to correlate the error magnitude with the adjustment range. The system completes parameter adjustment using the above formula, generating a subset of progressively optimized physical parameters that includes the new viscosity coefficient and the new vegetation stiffness matrix.

[0142] S53: Perform stability verification on a subset of physical parameters, correct parameter mutations through conservation law tests, and integrate water density with the verified parameters to generate a new set of physical parameters.

[0143] Specifically, the system performs stability verification on the generated subset of physical parameters. The core of this process is to ensure the stability of the physical field simulation after parameter adjustment through conservation law checks, prioritizing momentum and energy conservation checks. The momentum conservation check formula is as follows:

[0144]

[0145] in, This represents the total change in momentum of the system. Let be the mass of each object in the physical field, such as a water element, a vegetation unit, or a human limb. This represents the change in velocity of each object. When the total change in momentum exceeds a preset reasonable range, the system determines that there is a sudden change in parameters and corrects the parameters using a linear interpolation algorithm to bring the change in momentum back to a reasonable range. After completing the stability verification, the system calls the water density from the previous set of physical parameters. Compare it with the verified viscosity coefficient Vegetation stiffness matrix By integrating and incorporating the human body mass matrix, a new set of physical parameters is generated. This set can be used for physical field modeling of the next training scenario, enabling adaptive optimization of the training process.

[0146] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0147] Based on the same inventive concept, this application also provides a VR simulation training system for implementing the aforementioned VR simulation training method for wilderness and water surface survival, self-rescue, and mutual aid. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the VR simulation training system for wilderness and water surface survival, self-rescue, and mutual aid provided below can be found in the limitations of the VR simulation training method for wilderness and water surface survival, self-rescue, and mutual aid described above, and will not be repeated here.

[0148] Preferably, such as Figure 3 As shown, this invention provides a VR simulation training system 600 for wilderness and water surface survival self-rescue and mutual rescue, which is configured with the following modules:

[0149] The physical parameter construction module 610 is used to construct a set of coupled control equations for water-vegetation-human body based on pre-stored physical field modeling parameters, and generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0150] The motion data processing module 620 is used to process the acquired student motion data and equipment signals. It acquires user joint position data through VR motion capture equipment, calculates joint inertia weight vectors based on the human body mass matrix in the physical parameter set, and performs weighted mapping and command parsing in combination with the controller signals to generate dynamic boundary conditions and rescue tool constraints.

[0151] The multi-field coupling solution module 630 is used to perform multi-physics coupling solution of dynamic boundary conditions and rescue tool constraints. It calls the viscosity coefficient and vegetation stiffness matrix in the physical parameter set to calculate the force of fluid on vegetation, combines the rescue tool constraints to solve the vegetation deformation displacement and update the fluid velocity field, generates a penalty force when a collision is detected, and generates a physical field output set.

[0152] The physics feedback conversion module 640 is used to process the physical field output set, combine it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic conversion, generate a feedback set containing tactile feedback waveforms and rendering instruction sets and send it to the VR simulation training device.

[0153] The parameter dynamic update module 650 is used to perform weighted error calculation based on the collected training feedback data, combined with the inertial weight vector and dynamic boundary conditions, to update and adjust the viscosity coefficient and vegetation stiffness matrix of the physical parameter set, and to generate a new physical parameter set.

[0154] Preferably, the physical parameter construction module 610 provided in this application is configured with the following units:

[0155] The water body equation construction unit is used to construct multi-field coupling equations for pre-stored fluid dynamic parameters. Based on the Navier-Stokes equation, it integrates water density and viscosity coefficient to establish a pressure-velocity coupling model and generate water body control equations.

[0156] The vegetation equation construction unit is used to perform hierarchical stiffness matching processing on the pre-stored plant mechanical parameters, query the stem-leaf stiffness ratio according to the plant type database and calculate the equivalent stiffness matrix to generate the vegetation flexible body dynamic equation.

[0157] The parameter set integration unit is used to perform joint dynamics optimization on pre-stored human kinematic parameters. Based on the joint range of motion constraints, a diagonal mass matrix is ​​constructed. The water body control equation, vegetation flexible body dynamics equation and diagonal mass matrix are integrated to generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

[0158] Preferably, the motion data processing module 620 provided in this application is configured with the following units:

[0159] The joint coordinate normalization unit is used to acquire the motion data of the trainee to be trained, perform coordinate system normalization processing on the user's joint position data, transform the three-dimensional space coordinates to the virtual scene coordinate system, and generate standardized joint data.

[0160] The inertia weight calculation unit is used to perform inertia weight calculation on standardized joint data. It calculates the mass ratio of each joint based on the human body mass matrix in the physical parameter set and normalizes it to generate a weight allocation vector.

[0161] The signal joint analysis unit is used to jointly analyze the weight allocation vector and the acquired training device signals, call the neural network to fuse weighted joint data to generate boundary mapping relationships, generate dynamic boundary conditions through fully connected layer transformation, and simultaneously analyze the handle displacement signal to extract displacement vector and orientation angle, and generate rescue tool constraints through inverse kinematics solution.

[0162] Preferably, the multi-field coupling solution module 630 provided in this application is configured with the following units:

[0163] The fluid particle initialization unit is used to initialize the fluid particles for dynamic boundary conditions. It distributes smooth particles according to the boundary shape and sets the initial velocity field to generate the fluid computational domain.

[0164] The fluid-vegetation coupling solution unit is used to perform coupled solution processing of the fluid computation domain and the rescue tool constraints. It calls the viscosity coefficient in the physical parameter set to calculate the fluid force based on vorticity constraints, and combines the rescue tool constraints to calculate the vegetation deformation displacement considering fluid-structure interaction, generating a fluid-vegetation coupling intermediate solution.

[0165] The collision response processing unit is used to simulate collisions between the acquired rescue tool locations and vegetation node locations, perform dynamic response processing on the collision detection results of the tool-vegetation distance calculated by the spatial discretization algorithm, generate a penalty force associated with the stiffness matrix when the tool-vegetation distance is less than the adaptive distance threshold, and generate a physical field output set containing the coupling of fluid force, vegetation deformation displacement and penalty force by combining the intermediate solution of fluid-vegetation coupling.

[0166] Preferably, the physical feedback conversion module 640 provided in this application is configured with the following units:

[0167] The stiffness Jacobian analytical unit is used to perform stiffness-correlated Jacobian analytical processing on the vegetation deformation displacement, fluid velocity field and coupling force tensor of the physical field output set. It uses the vegetation stiffness matrix in the physical parameter set to calculate the stiffness-correlated Jacobian matrix and generate dynamic force-displacement gradient mapping relationship.

[0168] The tactile waveform synthesis unit is used to perform tactile waveform synthesis processing on the coupling force tensor and the dynamic force-displacement gradient mapping relationship, and to generate stiffness-adaptive tactile feedback waveforms through force mapping calculation.

[0169] The cross-modal rendering unit is used to perform cross-modal rendering processing on vegetation deformation displacement and fluid velocity field of the physical field output set. It updates the skeleton skin mesh based on the deformation parameters related to stiffness, generates vortex particle system according to the coupling relationship between fluid vorticity and stiffness matrix, performs particle density field and mesh topology avoidance optimization processing on vortex particle system to generate rendering instruction set, integrates haptic feedback waveform and rendering instruction set to generate feedback set and sends it to VR simulation training device.

[0170] Preferably, the parameter dynamic update module 650 provided in this application is configured with the following units:

[0171] The motion feature extraction unit is used to extract kinematic features from the collected training feedback data, calculate the weighted position deviation by combining the inertial weight vector and dynamic boundary conditions, and generate a biomechanical error vector through feature fusion.

[0172] The parameter dynamic adjustment unit is used to update the physical constraint parameters of the biomechanical error vector. Based on the error amplitude, it dynamically adjusts the viscosity coefficient and vegetation stiffness matrix to generate a progressively optimized subset of physical parameters.

[0173] The parameter verification and integration unit is used to perform stability verification on a subset of physical parameters, correct parameter mutations through conservation law tests, and integrate water density with the verified parameters to generate a new set of physical parameters.

[0174] In one embodiment, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described VR simulation training method for survival and mutual rescue in the wild and on the water.

[0175] In one embodiment, this application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described VR simulation training method for survival and mutual rescue in the wild and on the water.

[0176] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0177] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0178] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A VR simulation training method for survival, self-rescue, and mutual aid in the wild and on water, characterized in that, Includes the following steps: S1: Construct a set of coupled control equations for water, vegetation and human body based on pre-stored physical field modeling parameters, and generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix. S2: Acquire the student's motion data and device signals, obtain the user's joint position data through VR motion capture device, calculate the joint inertia weight vector based on the human body mass matrix in the physical parameter set, perform weighted mapping and command parsing in combination with the handle signal, and generate dynamic boundary conditions and rescue tool constraints. S3: Perform multi-physics coupling solution on the dynamic boundary conditions and the rescue tool constraints, call the viscosity coefficient and vegetation stiffness matrix in the physical parameter set to calculate the fluid force on the vegetation, combine the rescue tool constraints to solve the vegetation deformation displacement and update the fluid velocity field, generate a penalty force when a collision is detected, and generate a physical field output set. S4: Process the physical field output set, combine it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic transformation, generate a feedback set containing tactile feedback waveforms and rendering instruction set, and send it to the VR simulation training device; S5: Based on the collected training feedback data, combined with the inertial weight vector and dynamic boundary conditions, a weighted error calculation is performed to update and adjust the viscosity coefficient and vegetation stiffness matrix of the physical parameter set, and a new physical parameter set is generated.

2. The method according to claim 1, characterized in that, The physical field modeling parameters include fluid dynamics parameters simulating water flow characteristics, plant mechanical parameters simulating plant mechanical responses, and human dynamics parameters simulating human movement characteristics. S1 includes: S11: The pre-stored fluid dynamics parameters are processed to construct multi-field coupling equations. Based on the Navier-Stokes equations, the water density and viscosity coefficient are fused to establish a pressure-velocity coupling model and generate the water control equations. S12: Perform hierarchical stiffness matching processing on the pre-stored plant mechanical parameters, query the stem-leaf stiffness ratio according to the plant type database and calculate the equivalent stiffness matrix to generate the vegetation flexible body dynamic equation. S13: Perform joint dynamics optimization on the pre-stored human kinematic parameters, construct a diagonal mass matrix based on the joint range of motion constraints, integrate the water body control equation, the vegetation flexible body dynamics equation and the diagonal mass matrix to generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix.

3. The method according to claim 1, characterized in that, S2 includes: S21: Obtain the motion data of the trainee to be trained, perform coordinate system normalization on the user's joint position data, transform the three-dimensional space coordinates to the virtual scene coordinate system, and generate standardized joint data. S22: Perform inertial weight calculation processing on the standardized joint data, calculate the mass ratio of each joint based on the human body mass matrix in the physical parameter set and normalize it to generate a weight allocation vector; S23: Jointly analyze the weight allocation vector and the acquired training device signal, call the neural network to fuse weighted joint data to generate boundary mapping relationship, generate dynamic boundary conditions through fully connected layer transformation, and simultaneously analyze the handle displacement signal to extract displacement vector and direction angle, and generate rescue tool constraints through inverse kinematics solution.

4. The method according to claim 1, characterized in that, S3 includes: S31: Perform fluid particle initialization processing on the dynamic boundary conditions, distribute smooth particles according to the boundary shape and set the initial velocity field to generate the fluid computational domain; S32: Perform coupled solution processing on the fluid computation domain and the rescue tool constraints, call the viscosity coefficient in the physical parameter set to calculate the fluid force based on vorticity constraints, and combine the rescue tool constraints to calculate the vegetation deformation displacement considering fluid-structure interaction, and generate a fluid-vegetation coupled intermediate solution; S33: Perform collision simulation on the obtained rescue tool location and vegetation node location, perform dynamic response processing on the collision detection result of the tool-vegetation distance calculated by the spatial discretization algorithm, generate a penalty force associated with the stiffness matrix when the tool-vegetation distance is less than the adaptive distance threshold, and generate a physical field output set containing fluid force, vegetation deformation displacement and penalty force coupling by combining the fluid-vegetation coupling intermediate solution.

5. The method according to claim 1, characterized in that, The formula for calculating the penalty force is: in, For the purpose of punishment, Basic penalty coefficient, Here is the vegetation stiffness matrix. Let be the unit direction vector from which the tool points to the vegetation node. This is the distance attenuation factor.

6. The method according to claim 1, characterized in that, S4 includes: S41: Perform stiffness-correlated Jacobian analysis on the vegetation deformation displacement, fluid velocity field and coupling force tensor of the physical field output set, calculate the stiffness-correlated Jacobian matrix using the vegetation stiffness matrix in the physical parameter set, and generate dynamic force-displacement gradient mapping relationship. S42: Perform tactile waveform synthesis processing on the coupling force tensor and dynamic force-displacement gradient mapping relationship, and generate a stiffness-adaptive tactile feedback waveform through force mapping calculation; S43: Perform cross-modal rendering processing on the vegetation deformation displacement and fluid velocity field of the physical field output set, update the skeleton skin mesh based on the stiffness-related deformation parameters, generate a vortex particle system based on the coupling relationship between fluid vorticity and stiffness matrix, perform particle density field and mesh topology avoidance optimization processing on the vortex particle system to generate a rendering instruction set, integrate the haptic feedback waveform and the rendering instruction set to generate a feedback set and send it to the VR simulation training device.

7. The method according to claim 1, characterized in that, The tactile feedback waveform is calculated using the following formula: in, For tactile feedback waveforms, As the reference amplitude, The Jacobian matrix is ​​related to stiffness. This is the stiffness attenuation factor. Let be the norm of the vegetation stiffness matrix. The fundamental frequency for haptic feedback, This is the force mapping matrix.

8. The method according to any one of claims 1-7, characterized in that, S5 includes: S51: Perform kinematic feature extraction processing on the collected training feedback data, calculate the weighted position deviation by combining the inertial weight vector and dynamic boundary conditions, and generate a biomechanical error vector through feature fusion; S52: Perform physical constraint parameter update processing on the biomechanical error vector, dynamically adjust the viscosity coefficient and vegetation stiffness matrix based on the error amplitude, and generate a progressively optimized subset of physical parameters; S53: Perform stability verification on the subset of physical parameters, correct parameter mutations through conservation law tests, and integrate water density with the verified parameters to generate a new set of physical parameters.

9. A VR simulation training system for survival, self-rescue, and mutual aid in the wild and on water, characterized in that, The system includes: The physical parameter construction module is used to construct a set of coupled control equations for water, vegetation and human body based on pre-stored physical field modeling parameters, and generate a set of physical parameters including water density, viscosity coefficient, vegetation stiffness matrix and human body mass matrix. The motion data processing module is used to process the acquired student motion data and equipment signals. It acquires user joint position data through VR motion capture equipment, calculates joint inertia weight vectors based on the human body mass matrix in the physical parameter set, and performs weighted mapping and command parsing in combination with the controller signals to generate dynamic boundary conditions and rescue tool constraints. The multi-field coupling solution module is used to perform multi-physics coupling solution of the dynamic boundary conditions and the rescue tool constraints. It calls the viscosity coefficient and vegetation stiffness matrix in the physical parameter set to calculate the force of the fluid on the vegetation, combines the rescue tool constraints to solve the vegetation deformation displacement and update the fluid velocity field, generates a penalty force when a collision is detected, and generates a physical field output set. The physical feedback conversion module is used to process the physical field output set, combine it with the vegetation stiffness matrix in the physical parameter set to perform physical characteristic conversion, generate a feedback set containing tactile feedback waveforms and rendering instruction sets, and send it to the VR simulation training device. The parameter dynamic update module is used to perform weighted error calculation based on the collected training feedback data, the inertial weight vector and dynamic boundary conditions, update and adjust the viscosity coefficient and vegetation stiffness matrix of the physical parameter set, and generate a new physical parameter set.

Images