An emergency rescue simulation method and system, a computer device and a storage medium

Abstract

The application relates to an emergency rescue simulation method and system, computer equipment and a storage medium, and relates to the technical field of computer simulation.The method comprises the following steps: acquiring multi-modal data of a plurality of users participating in simulation; for each user, a cognitive-physiological state vector representing the internal state of the user is calculated in real time based on the corresponding multi-modal data of the user; the cognitive-physiological state vectors of all users are converged, analyzed and a group cognitive coordination spectrum vector representing the overall coordination state of the team is generated, which contains indexes such as decision synchronization rate and information sharing entropy; a disaster control decision module adaptively decides and generates disaster control actions according to the spectrum vector; and an environment updating module updates a virtual environment according to the disaster control actions. Through the closed-loop feedback of the team cognitive state and the disaster environment evolution, objective quantitative evaluation and adaptive targeted training of the team coordination efficiency are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer simulation, in particular to an emergency rescue simulation method and system, a computer device and a storage medium. BACKGROUND

[0002] Emergency rescue simulation technology, as an important means to improve the combat capability of emergency response personnel and optimize rescue plans, has been widely applied and developed. Current simulation technology is mostly focused on building high-fidelity virtual physical environments, which can use advanced physical engines to accurately simulate the physical evolution process of disaster scenarios such as fire spread, smoke diffusion, building structure collapse, etc. Users immerse themselves in the virtual reality (VR) or augmented reality (AR) devices to practice operation skills and tactical processes.

[0003] However, at present, digital and intelligent means are gradually introduced in the field of emergency management and disaster rescue, and virtual environments are constructed for accident simulation and rescue drills to improve the ability and efficiency of responding to emergencies. Existing emergency rescue simulation technology mainly relies on finite element analysis, fluid dynamics simulation, and artificial neural network algorithms, but there are problems such as large consumption of computing resources, difficulty in balancing simulation accuracy and real-time performance.

[0004] In addition, the existing technology also exposes its shortcomings in the evaluation of team coordination effectiveness. Currently, the evaluation of team drilling effectiveness mostly relies on the subjective observation of instructors outside the scene or post-mortem discussions after the drill. This evaluation method is not only highly subjective and difficult to standardize, but also severely lacks process-oriented and objective quantitative data support. For deep-seated coordination problems such as whether team decisions are synchronized, whether key information is effectively transferred among members, and whether negative emotions are spreading in the team, the existing technology cannot provide real-time and quantifiable indicators for measurement, making the evaluation of team coordination capability often remain at a vague and superficial level, and unable to accurately identify specific weaknesses in team collaboration.

[0005] Finally, due to the lack of real-time and accurate insight into team coordination status, the existing simulation system also appears to be rigid in training mode. The triggering and evolution of disaster events are often based on pre-set scripts or fixed difficulty gradients, and cannot dynamically and specifically generate challenges according to the specific weaknesses exposed by the training team during the drill. For example, the system cannot adaptively generate a scenario to test distributed communication capabilities when it identifies communication bottlenecks in the team. This training mode is difficult to achieve targeted reinforcement training for different teams and different weaknesses, thereby limiting the utilization efficiency of training resources and the final training effect to some extent. SUMMARY

[0006] The application aims to provide an emergency rescue simulation method, system, computer device and storage medium, and solve the technical problem that the existing emergency rescue simulation technology cannot quantize the internal state of personnel and team coordination performance in real time, and dynamically couple them with disaster evolution, resulting in deviation between the simulation process and the real scene and lack of targeted training.

[0007] In a first aspect, the application provides an emergency rescue simulation method, which is realized by the following steps:

[0008] First, the multi-modal data of multiple users participating in the simulation is obtained. Then, for each user, based on the corresponding multi-modal data, the cognitive-physiological state vector representing the internal state of the user is calculated in real time through a preset cognitive state conversion function.

[0009] After obtaining the individual cognitive-physiological state vectors of all users, the vectors are aggregated and calculated, and a group cognitive coordination spectrum vector representing the overall coordination state of the team is analyzed and generated.

[0010] Then, a disaster control decision module adaptively decides and generates a disaster control action according to the group cognitive coordination spectrum vector.

[0011] Finally, an environment update module updates the virtual environment state in the simulation according to the disaster control action, thereby completing a dynamic feedback loop.

[0012] As a specific implementation, the multi-modal data is obtained by collecting real-time physiological indicators of the user through a biological sensor worn by the user to form a physiological data stream, extracting stress factors of the microenvironment in which the user is located from the physical engine of the simulation system to form an environmental data stream, and recording the operation interaction log of the user in the virtual environment to form a behavior data stream. The cognitive-physiological state vector is a data structure that includes at least one or more of the following quantitative indicators: cognitive load level, physiological stress level, or panic index.

[0013] As a specific implementation, the group cognitive coordination spectrum vector is a data structure that includes at least one or more of the following coordination indicators: decision synchronization rate, information sharing entropy, or emotional contagion index. These indicators quantize the coordination performance of the team from different dimensions.

[0014] In a specific embodiment, the decision synchronization rate is represented by symbol η dsr , which is used to quantify the consistency of the cognitive state response of team members after receiving critical instructions. This index can be calculated by the following formula:

[0015]

[0016] wherein: Var(·) is a variance calculation function, ΔL cog,i is the change of the cognitive load level of the i-th user within a preset time window after receiving the critical instruction, is a preset maximum variance value for normalization, and N is the total number of users.

[0017] In another specific embodiment, the information sharing entropy, denoted by H info , is used to characterize the balance of the distribution of critical information among team members. This index can be calculated by the following formula:

[0018]

[0019] wherein: p i (t) is the probability proportion of critical information initiated or transmitted by the i-th user within a specific time window t; and N is the total number of users.

[0020] In yet another specific embodiment, the emotional contagion index, denoted by κ emo , is used to quantify the propagation intensity of negative emotions in the team network. This index is defined as the spectral radius of a correlation matrix R. The elements ρ ij in the correlation matrix R represent the correlation of the panic index time series of user i and user j, which can be calculated by the following formula:

[0021] wherein: ρ ij is the correlation of the panic index time series of user i and user j; Cov(·,·) is a covariance calculation function; L pan,i (t ′ ) and L pan,j (t ′ ) are the panic index time series of user i and user j within a past time window t ′ ; and are the standard deviations of the corresponding panic index time series, respectively.

[0022] In a preferred embodiment, the step of adaptively deciding and generating disaster control actions is specifically: obtaining, by the disaster control decision module, a global state, wherein the data structure of the global state includes the cognitive-physiological state vectors of all users, the group cognitive synergy frequency spectrum vector, and the physical world state representing the current virtual environment physical state; based on the global state, the module outputs the disaster control action through a preset decision model. Moreover, the disaster control action is generated for the team cognitive synergy weakness revealed by the group cognitive synergy frequency spectrum vector.

[0023] In another preferred embodiment, the method further comprises a down-regulation step: based on the cognitive-physiological state vector of each user, real-time regulating the behavioral capability of the user in the virtual environment. The behavioral capability includes the operation accuracy or the virtual field of view of the user. For example, when the physiological stress level or the cognitive load level of a user exceeds a preset threshold, the system processes the original input instruction of the user through a state-dependent conversion function to reduce the operation accuracy of the final action or narrow the virtual field of view.

[0024] The second aspect of the present application provides an emergency rescue simulation system, comprising:

[0025] a data acquisition module, configured to acquire multi-modal data of a plurality of users participating in the simulation;

[0026] an individual state solving module, connected to the data acquisition module, configured to, for each user, based on the multi-modal data corresponding to the user, real-time solve a cognitive-physiological state vector representing the internal state of the user;

[0027] a group spectrum analysis module, connected to the individual state solving module, configured to converge the cognitive-physiological state vectors of all users, analyze and generate a group cognitive coordination spectrum vector representing the overall coordination state of the team;

[0028] a disaster control decision module, connected to the group spectrum analysis module, configured to adaptively decide and generate a disaster control action according to the group cognitive coordination spectrum vector;

[0029] an environment updating module, connected to the disaster control decision module, configured to update the virtual environment in the simulation according to the disaster control action.

[0030] The third aspect of the present application provides a computer device, comprising a processor and a memory, wherein the memory stores a computer program, and the computer program is executed by the processor to implement the method of the first aspect of the present application.

[0031] The fourth aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the method of the first aspect of the present application.

[0032] In summary, the present application includes at least one of the following beneficial technical effects:

[0033] 1. The application characterizes the conduction relationship of risk in the building space by constructing an abstract weighted space adjacency graph, and quantifies the risk by using algebraic operation based on the hazard correlation matrix. The evaluation method based on graph theory and matrix operation has much lower computational complexity than the physical simulation method, so it can realize near real-time evaluation and spread path prediction of the risk situation in the whole building without relying on large-scale computing clusters, effectively avoiding the problem of balancing simulation accuracy and real-time performance in traditional technology, and providing a lightweight and efficient decision support tool for emergency command and rescue exercise;

[0034] 2. The application can realize objective and quantitative evaluation of team coordination performance, by analyzing the cognitive-physiological state vector of all users in real time and calculating the group cognitive coordination spectrum vector containing specific indicators such as decision synchronization rate, information sharing entropy and emotional contagion index. The traditional exercise team coordination performance which is difficult to measure and fuzzy is converted into a set of traceable, analyzable structured data, which provides accurate data support for subsequent review, evaluation and capability improvement;

[0035] 3. The application realizes adaptive and targeted training of rescue team. The disaster control decision module can adaptively decide and generate disaster control actions corresponding to the team coordination short board revealed by the group cognitive coordination spectrum vector in real time. When the system identifies that the team information sharing is not smooth, it can generate communication failure events, instead of blindly increasing physical risks, so as to guide the team to strengthen training in specific ability dimensions and improve the training pertinence and efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a method flowchart of the application;

[0037] Figure 2 is a system architecture diagram of the application;

[0038] Figure 3 is a computer device structure schematic diagram of the application. DETAILED DESCRIPTION

[0039] The application is further described in detail in combination with the accompanying Figure 1

[0040] Embodiment: An emergency rescue simulation method, comprising the following steps:

[0041] S1, obtaining multi-modal data of a plurality of users participating in simulation;

[0042] ​In the specific implementation of the present application, this step is a basic data collection process that is carried out continuously during the simulation initialization phase and throughout the entire running period. The core of it is to construct a comprehensive data portrait for each user participating in the simulation, which contains the inherent characteristics and real-time dynamic reactions. The multi-modal data is specifically composed of static data and dynamic data.

[0043] Firstly, before the simulation task starts, the system will obtain or configure the static data of each user. This part of data constitutes the baseline of the user's ability and characteristics. Preferably, the static data is organized as a static profile vector. The vector provides personalized initial parameters or weight adjustment basis for the subsequent calculation of the cognitive-physiological state vector. The static profile vector can include: experience level: this is a quantitative indicator representing the user's familiarity with the emergency rescue field. For example, the number of times the user participates in related tasks, the total length of historical training, and other information can be used to assign values.

[0044] Psychological quality baseline: this is an indicator reflecting the user's psychological endurance and adjustment ability under normal stress. For example, the result of a standardized psychological assessment questionnaire can be used for pre-determination and input.

[0045] Professional skill level: this is an indicator representing the user's mastery of the professional skills of the role he plays. It can be set according to his historical examination results or certification level.

[0046] Secondly, after the simulation process starts, the system will continuously and real-time collect the dynamic data of each user. Dynamic data is a direct reflection of the user's real-time interaction with the virtual environment, mainly including the following three parallel data streams:

[0047] Physiological data stream

[0048] This data stream is obtained through wearable biosensor devices configured for each user. Preferably, such sensing devices can be integrated into the user's virtual reality (VR) head-mounted display, interactive handle, or dedicated physiological monitoring wristband, etc. devices to complete data collection without disturbing the user's immersive experience. This data stream provides a series of quantitative indicators that can objectively reflect the user's autonomic nervous system activity and physiological stress response, which can specifically include heart rate, heart rate variability, skin electrical response, etc. This physiological data stream is the most direct source of data for judging the user's internal physiological state changes.

[0049] The environmental data stream is dynamically extracted and generated by the physical simulation engine built in the simulation system according to the real-time position and state of each user in the virtual environment. It aims to quantify the objective stress exerted by the external environment in which the user is located. The parameters in this data stream can include: the environmental temperature, smoke concentration, environmental visibility, environmental noise decibel value, and the real-time distance between the user and the preset dangerous source, etc. This data stream provides a key environmental stress basis for subsequent calculation of the user's cognitive load and physiological stress level.

[0050] The behavior data stream is a structured log generated by recording all operation interaction events of the user in the virtual environment in real time. It reflects the user's explicit decision-making and execution ability under the driving of environmental information and internal state. This data stream can record in detail: the user's three-dimensional space movement path and instantaneous speed; the user's interaction frequency, duration and operation mode with virtual devices or tools; the start and end time, execution duration and success or failure of the user performing a specific task; and the frequency and duration of voice or text communication initiated through the system communication module, etc.

[0051] In summary, step S1 integrates the static profile vector of each user, and simultaneously collects the physiological data stream, environmental data stream and behavior data stream in real time, which together constitute a complete and dynamically updated multi-modal data set. This data set provides comprehensive and reliable data support for subsequent step S2 of precise and individual cognitive-physiological state vector calculation from four dimensions of user's inherent characteristics, internal physiological response, external environmental stress and explicit behavior performance.

[0052] S2, for each user, based on the corresponding multi-modal data, a cognitive-physiological state vector representing the internal state of the user is calculated in real time;

[0053] In the specific implementation of the present application, this step is performed in parallel by the individual state calculation module configured in the system and corresponding to each user. This design ensures that the state evaluation of each user is independent and real-time.

[0054] Specifically, the individual state calculation module receives the complete multi-modal data set related to the current user i from step S1. This data set includes the static profile vector of the user, as well as the real-time updated physiological data stream, environmental data stream and behavior data stream.

[0055] The core of the module is a preset cognitive state transition function, denoted by f c . The function of this function is to establish a nonlinear mapping relationship between the above multi-dimensional input data and the user's internal cognitive physiological state. The cognitive state transition function takes multi-modal data as input, and outputs the cognitive-physiological state vector C i(t). The functional relationship can be expressed as:

[0056] C i (t) = f c (D phys,i (t), D env,i (t), D beh,i (t), U i );

[0057] wherein: C i (t) represents the cognitive-physiological state vector of user i at time t; f c represents the cognitive state transition function; D phys,i (t) represents the physiological data stream of user i at time t; D env,i (t) represents the environmental data stream of user i at time t; D beh,i (t) represents the behavioral data stream of user i at time t; U i represents the static profile vector of user i.

[0058] Preferably, the cognitive state transition function f c can be implemented through at least one of the following technical paths:

[0059] A pre-trained machine learning model, such as a recurrent neural network or its variant long short-term memory network, can be employed to construct the function. The model learns the complex mapping pattern from input data to output state through supervised learning training on a historical data set containing multi-modal data and corresponding cognitive state labels.

[0060] A fuzzy logic system based on expert knowledge can be employed to construct the function. The system describes the relationship between different input data combinations and specific cognitive state levels by defining a series of rules.

[0061] By executing the cognitive state transition function, the system generates a cognitive-physiological state vector C i (t) for each user in real time. The vector is a multi-dimensional data structure that contains deep internal state indicators that cannot be directly measured by external devices. Preferably, the vector contains at least the following dimension quantitative indicators:

[0062] C i (t) = [L cog,i (t), L str,i (t), L pan,i (t)];

[0063] wherein:

[0064] Cognitive load level (L cog,i(t)): This index aims to quantify the mental resources consumed by the user's brain when performing the current task. Its calculation process will take into account factors such as task complexity (e.g., the frequency of multi-target operations), communication data volume, and interference information (e.g., high-decibel noise) in the environmental data stream.

[0065] Physiological stress level (L str,i (t)): This index aims to quantify the degree of tension and vigilance of the user's physiological system triggered by perceived threats or stress. Its calculation mainly relies on changes in indicators in the physiological data stream, such as significant increases in heart rate, decreases in heart rate variability, and increases in galvanic skin response, combined with weighting of dangerous source information in the environmental data stream.

[0066] Panic index (L pan,i (t)): This index aims to assess the user's tendency to make irrational decisions or lose control of behavior. In addition to relying on extreme physiological stress indicators, it will focus on analyzing abnormal patterns in the behavior data stream, such as aimless wandering, repetitive error operations, or communication interruptions.

[0067] In summary, step S2 successfully converts the multi-source, heterogeneous raw data collected in step S1 into a set of standardized key indicators that can represent the user's internal cognitive and psychological state, through the individual state calculation module and cognitive state transition function independently configured for each user. The cognitive-physiological state vector set {C1(t), C2(t),..., C N (t)} output by this step provides the necessary and high-quality data basis for step S3 to perform macro analysis of the team's overall collaborative state.

[0068] S3, gather all users' cognitive-physiological state vectors, analyze and generate a group cognitive synergy spectrum vector representing the team's overall collaborative state;

[0069] In the specific implementation of the present application, this step is performed by a group spectrum analysis module within the system. This module serves as a centralized computing unit, with its data input connected to the output of all users' individual state calculation modules to real-time gather the cognitive-physiological state vector set {C1(t), C2(t),..., C N (t)} of all users at time t.

[0070] The core function of this module is to perform statistical and dynamic analysis on the gathered individual state data through a series of pre-set mathematical models and algorithms, thereby calculating and generating a structured data that can multi-dimensionally and quantitatively describe the current overall collaborative effectiveness of the team, i.e., the group cognitive synergy spectrum vector, denoted by symbol S team (t).

[0071] Preferably, the group cognitive synergy spectrum vector contains a series of synergy indicators, which reveal the internal operation state of the team from different dimensions. In a specific embodiment, the vector can contain at least one or more of the following synergy indicators: first, a key synergy indicator is the decision synchronization rate, denoted by η dsr . This indicator aims to quantify the consistency of the response speed and intensity of each member at the cognitive level when the team receives a unified key instruction. A high synchronization rate reflects that the team has good instruction understanding and execution consistency. The calculation process is as follows: the system monitors the communication data stream, and when a key instruction is identified to be issued at time t c , it tracks and records the change amount ΔL cog,i of the cognitive load level L cog,i of each user i in a very short preset time window Δt that follows. Then, by calculating the variance of the change amount of all users and performing normalization processing, the decision synchronization rate is obtained. The calculation formula can be expressed as:

[0072]

[0073] where: η dsr (t c ) represents the decision synchronization rate at time t c when the instruction is issued; Var(·) represents the variance calculation function; ΔL cog,i (t c , Δt) represents the change amount of the cognitive load level of the i-th user in the time window from t c to t c + Δt; Var max represents a preset maximum variance value for normalization, which is a fixed parameter set according to historical data or expert experience; N represents the total number of users.

[0074] In addition, another important synergy indicator is the information sharing entropy, denoted by H info . This indicator aims to evaluate the balance of the distribution of key information among all members by analyzing the communication pattern within the team. The calculation of this indicator is based on the statistics of effective communication events in the behavior data stream. Specifically, the system counts the number of key information initiated or successfully transmitted by each user i within a certain time window, and calculates the probability proportion p i (t) in the total amount of information in the team. The information sharing entropy is calculated by the Shannon entropy formula. This indicator helps to identify possible information bottlenecks or information overload problems in team communication. The calculation formula can be expressed as:

[0075]

[0076] where: Hinfo (t) represents the information sharing entropy at time t; p i (t) represents the probability of critical information initiated or transmitted by the i-th user within a specific time window; N represents the total number of users.

[0077] Furthermore, a synergy indicator reflecting the emotional stability of the team is the emotional contagion index, denoted by symbol κ emo . This indicator aims to quantify the intensity and risk of the spread of negative emotions (represented by the panic index in the cognitive-physiological state vector) in the social network composed of team members. The calculation process is as follows: first, obtain the time series data of the panic index L pan of any two users i and j within a past time window t'. Then, calculate the Pearson correlation coefficient ρ ij of the two time series. Repeat this calculation for all possible pairs of users in the team, and construct an N x N correlation matrix R(t) = [ρ ij (t)].

[0078] The emotional contagion index is defined as the spectral radius (the largest eigenvalue) of this correlation matrix, which can effectively capture the amplification effect of the strongest emotional transmission path in the network. The formula for calculating the correlation coefficient ρ ij is as follows:

[0079]

[0080] where: ρ ij (t) represents the correlation of the panic index time series of user i and user j at time t; Cov(·,·) represents the covariance calculation function; L pan,i (t ′ ) and L pan,j (t ′ ) represent the panic index time series of user i and user j within a past time window t ′ (length τ); σ Lpan,i and σ Lpan,j represent the standard deviations of the corresponding panic index time series.

[0081] In summary, step S3 uses the group spectrum analysis module to effectively refine and aggregate the massive, microscopic individual cognitive data generated in step S2 into a set of core indicators that can macroscopically and multidimensionally describe the team synergy state. The group cognitive synergy spectrum vector S team (t) output by this step provides a clear and quantitative snapshot of the team state for the disaster control decision module in step S4, enabling it to make more targeted intervention decisions that match the current cognitive weaknesses of the team.

[0082] S4, according to the group cognitive synergy spectrum vector, adaptively deciding and generating a disaster control action;

[0083] In the specific implementation of the present application, this step is performed by a disaster control decision module configured in the system. The function of the module is that the generated disaster event is no longer a preset, fixed script, but a real-time, dynamic response according to the current performance of the team.

[0084] Specifically, the decision of the disaster control decision module is not only based on the group cognitive synergy spectrum vector output in step S3, but also based on a more comprehensive global state. The global state is represented by symbol S(t), which integrates the core information of the entire simulation system at time t, and its composition can be represented as:

[0085] S(t)=({C1(t),...,C N (t)},S team (t),W(t));

[0086] Where: {C1(t),...,C N (t)} represents the set of individual cognitive-physiological state vectors of all users generated by step S2; S team (t) represents the group cognitive synergy spectrum vector representing the overall team synergy state generated by step S3; W(t) represents a physical world state vector representing the current virtual environment physical state, which can include fire spread range, remaining passable path, building structure stability, etc. Information provided by the physical simulation engine of the system.

[0087] By obtaining the global state S(t), the disaster control decision module can simultaneously perceive the internal state of each micro individual, the synergy efficiency of the macro team, and the physical reality of the objective environment under a unified data framework.

[0088] Inside the module, its decision-making process is performed by a preset decision strategy function, denoted by symbol π. The function of the strategy function is to map the currently obtained global state S(t) to a specific disaster control action A g (t) to be executed in the virtual environment. Its functional relationship can be represented as:

[0089] A g (t)=π(S(t));

[0090] Where: A g (t) represents the disaster control action generated at time t; π represents the decision strategy function; S(t) represents the global state at the current time.

[0091] Preferably, the decision strategy function π can be obtained by offline training methods such as deep reinforcement learning.

[0092] The core of this step lies in its "adaptive" nature, i.e. the generated disaster regulation action A g (t) is directly related to the team cognitive synergy weakness revealed by the group cognitive synergy vector S team (t). For example:

[0093] If the decision module receives S team (t) for a long time, the decision synchronization rate η dsr continuously falls below the preset threshold, indicating that the team is hesitant or inconsistent in executing instructions. At this time, the decision strategy π may generate a non-fatal challenge that requires precise cooperation among team members within a short period of time, such as a rescue passage gate that requires multiple people to operate simultaneously.

[0094] If the information sharing entropy H team (t) in S info is too low, indicating the emergence of information silos in the team. At this time, the decision strategy π may not increase physical risk, but generate an event simulating the temporary failure of a key communication node, in order to prompt the team to establish new, distributed communication links.

[0095] Step S4 takes the team's cognitive synergy state as input through a disaster regulation decision module, and outputs a disaster regulation action with a clear purpose through a strategic decision process. This step ensures that the evolution of the simulation environment is no longer random or preset, but becomes a dynamic tool that can purposefully detect, expose and train the team's cognitive synergy capabilities. The disaster regulation action A g (t) output by this step will serve as a direct instruction for the environment update in step S5.

[0096] S5, update the virtual environment in the simulation according to the disaster regulation action.

[0097] In the specific implementation of the present application, this step is performed by an environment update module. This module serves as a bridge between the decision layer and the physical simulation layer in the system architecture. Its input end is connected to the output end of the disaster regulation decision module in step S4 to receive the disaster regulation action A g (t) generated by the decision.

[0098] Specifically, the disaster regulation action A g(t) is not a simple trigger signal, but a structured data instruction containing explicit execution parameters. Preferably, the instruction can be a data object containing at least fields of action type, action target, action location, intensity parameter, and duration, so as to ensure that the executed environmental update is accurate and controllable.

[0099] The core task of the environmental update module after receiving the disaster control action instruction is to parse the instruction and convert it into a series of specific function calls or parameter modifications to the underlying physical simulation engine. The physical simulation engine is a basic component responsible for simulating all physical laws in the virtual environment.

[0100] In order to further illustrate the implementation of this step, the specific execution process of the environmental update module is illustrated as follows in combination with different disaster control actions that may be generated in step S4:

[0101] If step S4 generates a disaster control action aimed at enhancing team collaborative operation because the team's decision synchronization rate η dsr is too low.

[0102] If step S4 generates a disaster control action aimed at interfering with a key communication node because the team's information sharing entropy H info is abnormal.

[0103] If step S4 generates a disaster control action aimed at increasing environmental stress to test the team's emotional stability because the team's emotional contagion index κ emo is rapidly rising.

[0104] This step, as the final execution end of the method of the present application, accurately materializes the upper-level decision into the dynamic evolution of the virtual environment through the environmental update module, ensuring that the simulation challenge is always closely associated with the real-time cognitive coordination state of the team.

[0105] In combination with the attached Figure 2 , another embodiment of the present application provides an emergency rescue simulation system, comprising:

[0106] a data acquisition module for acquiring multi-modal data of a plurality of users participating in the simulation;

[0107] an individual state solving module for, for each user, based on the corresponding multi-modal data, solving a cognitive-physiological state vector representing the user's internal state in real time;

[0108] a group spectrum analysis module for aggregating cognitive-physiological state vectors of all users, analyzing and generating a group cognitive coordination spectrum vector representing the overall coordination state of the team;

[0109] The disaster regulation decision module is configured to adaptively decide and generate a disaster regulation action according to the group cognition collaborative spectrum vector; and the environment updating module is configured to update the virtual environment in the simulation according to the disaster regulation action.

[0110] The system of the embodiment can be used to execute the method embodiments described above, and has similar principles and technical effects, which will not be described here.

[0111] The accompanying drawings are referred to in the description of the embodiments. Figure 3 The application also provides a computer device, comprising a processor and a memory, the memory storing a computer program executable by the processor, and the computer program is executed by the processor to perform the method described above.

[0112] The application also provides a storage medium, which stores a computer program, and the computer program is executed by the processor to perform the method described above.

[0113] The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read-only memory (EEPROM), an erasable programmable read-only memory (EPROM), a programmable read-only memory (PROM), a read-only memory (ROM), a magnetic storage, a flash memory, a magnetic disk or an optical disk.

[0114] The embodiments of the specific implementation are the preferred embodiments of the application, and are not intended to limit the protection scope of the application, wherein the same parts are denoted by the same reference numerals. Therefore, any equivalent changes made according to the structure, shape, principle of the application should be covered within the protection scope of the application.

Claims

1. An emergency rescue simulation method, characterized in that, Includes the following steps: Acquire multimodal data from multiple users participating in the simulation; For each user, a cognitive-physiological state vector representing the user's internal state is calculated in real time based on the corresponding multimodal data. By aggregating the cognitive-physiological state vectors of all users, we can analyze and generate a group cognitive-coordination spectrum vector that represents the overall collaborative state of the team. Based on the aforementioned group cognitive collaborative spectrum vector, an adaptive decision is made and a disaster control action is generated; The virtual environment in the simulation is updated based on the disaster control actions. The group cognitive collaboration spectrum vector includes at least one of the following collaboration indicators: Decision synchronization rate, information sharing entropy, or emotion contagion index; The emotion contagion index is the spectral radius of a correlation matrix, and the elements ρ of the correlation matrix are... ij Calculated using the following formula: Where: ρ ij Let L represent the correlation between the time series of panic indexes of users i and j; Cov(·,·) is the covariance calculation function; L pan,i (t′) and L pan,j (t′) represent the time series of the panic index for users i and j within the past time window t′, respectively; and These represent the standard deviations of the corresponding time series of the panic index.

2. The emergency rescue simulation method according to claim 1, characterized in that, The acquisition of multimodal data includes: Acquire physiological data streams collected via biosensors; Obtain an environmental data stream extracted from the virtual environment, representing the micro-environmental pressure in which the user is located; Acquire behavioral data streams that record user virtual operation logs; The cognitive-physiological state vector includes at least one of the following: cognitive load level, physiological stress level, or panic index.

3. The emergency rescue simulation method according to claim 1, characterized in that, The decision synchronization rate is calculated using the following formula: Where: Var(·) is the variance calculation function, ΔL cog,i Let be the change in the cognitive load level of the i-th user within a preset time window after receiving the key instruction. The preset maximum variance value is used for normalization, and N is the total number of users.

4. The emergency rescue simulation method according to claim 1, characterized in that, Also includes: Based on each user's cognitive-physiological state vector, the user's behavioral capabilities in the virtual environment are adjusted in real time, wherein the behavioral capabilities include the user's operational accuracy or virtual field of vision.

5. The emergency rescue simulation method according to claim 1, characterized in that, The adaptive decision-making and generation of a disaster control action specifically includes: Obtain a global state, which includes the cognitive-physiological state vectors of all users, the collective cognitive collaboration spectrum vector, and the physical world state representing the current physical state of the virtual environment; The disaster control action is generated based on the global state decision.

6. An emergency rescue simulation system, comprising an emergency rescue simulation method according to any one of claims 1-5, characterized in that, include: The data acquisition module is used to acquire multimodal data from multiple users participating in the simulation; The individual state calculation module is used to calculate, in real time, a cognitive-physiological state vector representing the user's internal state based on the corresponding multimodal data for each user. The group spectrum analysis module is used to aggregate the cognitive-physiological state vectors of all users, analyze and generate a group cognitive collaboration spectrum vector that represents the overall collaborative state of the team. The disaster control decision-making module is used to adaptively make decisions and generate disaster control actions based on the collective cognitive collaborative spectrum vector. The environment update module is used to update the virtual environment in the simulation based on the disaster control actions.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-5.

8. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-5.

Images