Multi-attribute and multi-stage microscopic simulation method and device based on pedestrian behavior characteristics

Abstract

The invention discloses a multi-attribute and multi-stage microscopic simulation method and device based on pedestrian behavior characteristics. The microcosmic simulation method comprises the steps that the pedestrian attribute and the initial position of a pedestrian agent are initialized, the pedestrian attribute comprises the initial speed, and the size of the initial speed is determined based on emotion driving characteristics; under the condition that the preset stage judgment condition is not met, each pedestrian agent is controlled to be in a normalization stage, and the real-time normal speed is determined in real time based on the target guiding characteristic; under the condition that the preset stage judgment condition is met and the simulation scene is a non-emergency scene, continuously controlling each pedestrian agent to be in a normalization stage; under the condition that the preset stage judgment condition is met and the simulation scene is an emergency scene, each pedestrian intelligent body is controlled to be in an evacuation stage, and the real-time evacuation speed is determined in real time based on the target guiding characteristic, the emotion driving characteristic, the physical strength guarantee characteristic and the state evaluation characteristic. Therefore, the accuracy of microscopic simulation is improved.

Description

Technical Field

[0001] This application relates to the field of microscopic simulation, and more specifically, to a multi-attribute, multi-stage microscopic simulation method and apparatus based on pedestrian behavior characteristics. Background Technology

[0002] Microscopic simulation is primarily used to simulate individual pedestrian behavior and their interactions within their environment, and it is widely applied in fields such as architecture, transportation, and urban planning. In architectural design, it can assess pedestrian flow in locations such as shopping malls, airports, and stadiums, optimizing passageway widths, staircase layouts, and ensuring buildings provide a good level of service for pedestrians. For urban planning, it can simulate pedestrian traffic in areas such as sidewalks, pedestrian streets, and plazas, helping to optimize traffic facilities and reduce congestion. Furthermore, in emergency evacuation analysis, it can incorporate disaster data such as fires and floods to simulate pedestrian evacuation processes under disaster conditions and evaluate the effectiveness of evacuation plans. Overall, microscopic simulation helps improve the safety, convenience, and efficiency of public spaces.

[0003] Currently, traditional models for microscopic simulation have some shortcomings. These models theoretically simulate particles, gases, or liquids, but lack the simulation of individual characteristics of pedestrians. Instead, they treat non-human substances such as particles as human beings through mechanical principles and rules, and the accuracy of the simulation needs to be improved.

[0004] Therefore, improving the accuracy of microscopic simulation has become a technical problem that needs to be solved in this field. Summary of the Invention

[0005] In view of this, this application proposes a multi-attribute, multi-stage microscopic simulation method and apparatus based on pedestrian behavior characteristics to improve the accuracy of microscopic simulation.

[0006] Firstly, this application provides a multi-attribute, multi-stage micro-simulation method based on pedestrian behavior characteristics. This micro-simulation method includes: initializing pedestrian attributes and initial positions for each pedestrian agent, wherein for each pedestrian agent, the pedestrian attributes include an initial speed, the magnitude of which is determined based on emotion-driven characteristics; controlling each pedestrian agent to be in a normalization stage if a preset stage judgment condition is not met, wherein for each pedestrian agent, its real-time normal speed in the normalization stage is determined in real-time based on goal-oriented characteristics; continuing to control each pedestrian agent to be in the normalization stage if the preset stage judgment condition is met and the simulation scenario is a non-emergency scenario; and controlling each pedestrian agent to be in an evacuation stage if the preset stage judgment condition is met and the simulation scenario is an emergency scenario, wherein for each pedestrian agent, its real-time evacuation speed in the evacuation stage is determined in real-time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics.

[0007] Optionally, for each pedestrian agent, the magnitude of the initial speed is determined based on emotion-driven characteristics, including: the magnitude of the initial speed is determined based on a first emotion influence coefficient of the pedestrian agent, the first emotion influence coefficient being related to the pedestrian agent's travel purpose.

[0008] Optionally, for each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics. This includes: the real-time evacuation speed is determined in real time based on the initial speed, the evacuation impact coefficient, and the evacuation phase planned path; the evacuation phase planned path is determined based on the pedestrian agent's phase-determined location and safe location; and the evacuation impact coefficient includes: a second emotion impact coefficient, physical strength value, and a state impact coefficient, wherein the second emotion impact coefficient is related to the pedestrian agent's level of panic.

[0009] Optionally, for any pedestrian agent, the second emotion influence coefficient is determined based on at least one of the following: the distance between the pedestrian agent and the hazard source, the pedestrian density around the pedestrian agent, and the time the pedestrian agent stays in the hazard space.

[0010] Optionally, when the pedestrian agent is in the normalization phase, the micro-simulation method further includes: adjusting the real-time normal speed and / or planned path of the pedestrian agent in the normalization phase based on a local collision avoidance algorithm to avoid collisions.

[0011] Secondly, this application also provides a multi-attribute, multi-stage micro-simulation device based on pedestrian behavior characteristics. The micro-simulation device includes: an initialization module for initializing pedestrian attributes and initial positions of pedestrian agents, wherein for each pedestrian agent, the pedestrian attributes include an initial speed, the magnitude of which is determined based on emotion-driven characteristics; and a processing module for: controlling each pedestrian agent to be in a normalization stage if a preset stage judgment condition is not met, wherein for each pedestrian agent, its real-time normal speed in the normalization stage is determined in real-time based on goal-oriented characteristics; continuing to control each pedestrian agent to be in the normalization stage if the preset stage judgment condition is met and the simulation scenario is a non-emergency scenario; and controlling each pedestrian agent to be in an evacuation stage if the preset stage judgment condition is met and the simulation scenario is an emergency scenario, wherein for each pedestrian agent, its real-time evacuation speed in the evacuation stage is determined in real-time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics.

[0012] Optionally, for each pedestrian agent, the magnitude of the initial speed is determined based on emotion-driven characteristics, including: the magnitude of the initial speed is determined based on a first emotion influence coefficient of the pedestrian agent, the first emotion influence coefficient being related to the pedestrian agent's travel purpose.

[0013] Optionally, for each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics. This includes: the real-time evacuation speed is determined in real time based on the initial speed, the evacuation impact coefficient, and the evacuation phase planned path; the evacuation phase planned path is determined based on the pedestrian agent's phase-determined location and safe location; and the evacuation impact coefficient includes: a second emotion impact coefficient, physical strength value, and a state impact coefficient, wherein the second emotion impact coefficient is related to the pedestrian agent's level of panic.

[0014] Optionally, for any pedestrian agent, the second emotion influence coefficient is determined based on at least one of the following: the distance between the pedestrian agent and the hazard source, the pedestrian density around the pedestrian agent, and the time the pedestrian agent stays in the hazard space.

[0015] Optionally, the processing module is further configured to: when the pedestrian agent is in the normalization phase, adjust the real-time normal speed and / or normal phase planned path of the pedestrian agent in the normalization phase based on a local collision avoidance algorithm in order to avoid collisions.

[0016] Thirdly, this application also provides a machine-readable storage medium storing instructions that cause a machine to perform the aforementioned microscopic simulation method.

[0017] Fourthly, this application also provides an electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the executable instructions to implement the microscopic simulation method described above.

[0018] According to the technical solution of this application, the pedestrian attributes and initial position of the pedestrian intelligent agent are initialized. For each pedestrian intelligent agent, the pedestrian attributes include an initial speed, the magnitude of which is determined based on emotion-driven characteristics. If the preset stage judgment condition is not met, each pedestrian intelligent agent is controlled to be in a normalization stage. For each pedestrian intelligent agent, its real-time normal speed in the normalization stage is determined in real time based on goal-oriented characteristics. If the preset stage judgment condition is met and the simulation scenario is a non-emergency scenario, each pedestrian intelligent agent continues to be controlled to be in the normalization stage. If the preset stage judgment condition is met and the simulation scenario is an emergency scenario, each pedestrian intelligent agent is controlled to be in an evacuation stage. For each pedestrian intelligent agent, its real-time evacuation speed in the evacuation stage is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics. In this way, pedestrian simulation is realized. Furthermore, pedestrian simulation takes into account the walking patterns of people under normal and emergency conditions. In normal conditions, pedestrians are in a normalization phase, while in emergency conditions, they are in an evacuation phase. Moreover, when determining the speed, pedestrian behavioral characteristics such as goal orientation, emotion-driven characteristics, physical strength preservation characteristics, and state assessment characteristics are taken into account. Therefore, individual human characteristics are incorporated into pedestrian simulation, and micro-simulation is performed based on these individual characteristics, thereby improving the accuracy of micro-simulation.

[0019] Other features and advantages of this application will be described in detail in the following detailed description section. Attached Figure Description

[0020] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application, and the illustrative embodiments and descriptions thereof are used to explain this application. In the drawings: Figure 1 A flowchart of a multi-stage microscopic simulation method incorporating multi-attribute-driven approaches according to a preferred embodiment of this application; Figure 2 This is a structural block diagram of a multi-stage microscopic simulation device integrating multi-attribute driving according to a preferred embodiment of this application. Detailed Implementation

[0021] The technical solution of this application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] In a first aspect, the embodiments of this application provide a microscopic simulation method.

[0023] Figure 1 This is a flowchart of a multi-stage microscopic simulation method incorporating multi-attribute-driven approaches according to a preferred embodiment of this application. Figure 1 As shown, this microscopic simulation method includes the following:

[0024] In step S10, the pedestrian attributes and initial position of the pedestrian agent are initialized. For each pedestrian agent, the pedestrian attributes include an initial velocity, the magnitude of which is determined based on emotion-driven characteristics. Emotion-driven characteristics refer to the influence of emotions.

[0025] Optionally, for each pedestrian agent, the magnitude of the initial velocity is determined based on emotion-driven characteristics, including: the magnitude of the initial velocity is determined based on a first emotion influence coefficient of the pedestrian agent, which is related to the pedestrian agent's travel purpose.

[0026] In this embodiment of the application, the pedestrian agent is constructed before initializing the pedestrian attributes and initial position.

[0027] In the embodiments of this application, the number and type of pedestrian intelligent agents can be determined according to specific circumstances. For example, the number and type of pedestrian intelligent agents to be constructed can be determined based on the following:

[0028] Determine the simulation OD matrix, pedestrian categories, and the proportion of each pedestrian category for the simulation scene. The simulation OD matrix includes multiple OD pairs in the simulation scene and the number of pedestrians corresponding to each OD pair. For each OD pair, calculate the number of pedestrians in each category based on its corresponding number of pedestrians, pedestrian category division, and corresponding proportion.

[0029] For example, the simulation scenario is a subway hub under flood conditions, simulating the evacuation of passengers in the subway hub during a flood disaster. The simulation OD matrix is... The matrix is ​​shown in Table 1. For example, there are 1028 people traveling from the intercity to M6, and 561 people traveling from M6 to the intercity.

[0030] Table 1 Simulation OD Matrix of Pedestrians in Metro Hubs For example, the pedestrian composition ratio is shown in Table 2, which specifies that for each OD pair, pedestrians are divided into eight categories and the proportion of each category.

[0031] Table 2 Pedestrian Category Classification and Corresponding Proportions Optionally, in this embodiment, the initial speed of each pedestrian agent can be determined according to the following: For each pedestrian agent, an initial speed is determined based on the initial speed and the corresponding first emotion influence coefficient. For example, the initial speed is determined by adding the corresponding first emotion influence coefficient to the initial speed. Specifically, the initial speed is multiplied by the corresponding first emotion influence coefficient.

[0032] The initial speed is the result of the personalized pedestrian agent's initial speed and "emotion-driven" factors. Each pedestrian has their own travel purpose, such as commuting, going to school, seeking medical treatment, shopping, or leisure exercise. Different travel purposes result in different levels of tension and anxiety (urgency), which will have different effects on walking speed. An initial speed is assigned to the pedestrian agent, and then a first emotion influence coefficient is added. The first emotion influence coefficients corresponding to different travel purposes are shown in Table 3.

[0033] Table 3 In this embodiment, the initialization speed can be determined according to specific circumstances. For example, a speed range for pedestrian agents can be set, and speeds can be randomly assigned to pedestrian agents within this range. Optionally, in this embodiment, the initialization speed of each pedestrian agent can be determined based on the following:

[0034] For each pedestrian category, a speed distribution range is set. For all pedestrian agents in that category, the speed distribution range is used to determine the initial speed of each pedestrian agent in that category.

[0035] Furthermore, in the embodiments of this application, there are many ways to assign an initial position to each pedestrian agent.

[0036] For example, each pedestrian agent has a corresponding OD. For all pedestrian agents, all pedestrian agents can be placed around point O (starting point); or half can be placed around point O (starting point) and half around point D (ending point).

[0037] Alternatively, for each OD pair, determine the path from point O to point D, and for all pedestrian agents corresponding to that OD pair, take the coordinates (e.g., randomly select values) along the determined path and assign them to the pedestrian agents.

[0038] Optionally, in this embodiment, a map may be prepared first, and then the initial position of each pedestrian agent on the map may be determined. The map is used to describe the scene, such as using a grid map or a polygon map. The grid or polygon must have relevant attributes that enable the agent to identify height, distance, which grids or polygons can support the agent, which cannot, and so on.

[0039] Optionally, in this embodiment, each pedestrian agent may be assigned at least one of the following: gender, age, height, and weight. Pedestrian attributes also include at least one of gender, age, height, and weight. Specifically, for each pedestrian category, a corresponding distribution interval is set for each of the at least one of these attributes. For all pedestrian agents in that category, values ​​are taken within the distribution interval according to a normal distribution to determine the initial value of each pedestrian agent in that category. For example, as shown in Table 3, distribution intervals are set for gender, age, height, weight, and speed for each pedestrian category.

[0040] Table 3 Pedestrian Attribute Table In step S11, each pedestrian agent is controlled to be in the normalization phase. The normalization phase refers to a simulated environment that is not in a disaster environment; for example, a disaster environment could be a flood, fire, or similar situation. For each pedestrian agent, its real-time normal speed in the normalization phase is determined in real-time based on its goal-oriented characteristics. The goal-oriented characteristics of a pedestrian mean that each trip has a starting point and an ending point; the ending point is the destination and determines the pedestrian's direction. In the normalization phase, the pedestrian agent completes its journey according to the predetermined starting point and ending point.

[0041] Specifically, during the normalization phase, each pedestrian agent is controlled to move based on its corresponding real-time normal speed and the planned path for the normal stage. For each pedestrian agent, the real-time normal speed is determined in real-time based on its initial speed and the planned path for the normal stage, while the planned path for the normal stage is determined based on its initial position and destination. The destination of the pedestrian agent can be determined based on the OD in the simulation OD matrix. The real-time normal speed is determined in real-time as the simulation progresses; for example, it can be determined once every simulation duration. The real-time normal speed is the walking speed of the pedestrian agent during the normalization phase.

[0042] During the normalization phase, the pedestrian agent needs to start walking; therefore, the key to pedestrian simulation is the direction and speed values. The magnitude of the real-time normal speed is determined in real time based on the initial speed, and the direction of the real-time normal speed is determined based on the path planned during the normalization phase.

[0043] The normal phase path planning refers to the walking path of the pedestrian agent during the normalization phase, determined based on the initial position and destination. Direction is a result of "goal-oriented" planning; macroscopically, global path planning is performed based on the initial position and destination to determine the normal phase planned path. For example, using... Algorithms, such as Dijkstra's algorithm, are used to find the (shortest) path for a pedestrian agent from its initial position to its destination on a map. Alternatively, other algorithms can be used to determine the planned path during the normal phase; this is not restricted. Optionally, in the embodiments of this application, collision avoidance functions, such as the RVO algorithm or ORCA algorithm, are added to the pedestrian agent at the micro level to avoid collisions with other pedestrian agents, and the planned path during the normal phase is fine-tuned locally.

[0044] In step S12, it is determined whether a preset stage determination condition is met. Specifically, the preset stage determination condition can be determined according to specific circumstances. If yes, then step S13 is executed; if no, then step S11 is executed. In addition, step S12 is executed once at regular intervals, and the length of the interval can be determined according to specific circumstances.

[0045] For example, the preset stage judgment condition can be that the simulation runtime reaches a preset duration value. Specifically, the simulation runtime is determined by accumulating time from the start of the simulation, and the determined simulation runtime is compared with the preset duration value to determine whether the preset stage judgment condition is met.

[0046] Alternatively, the preset stage judgment condition could be that the distribution of all pedestrian intelligent agents reaches an ideal distribution state. Specifically, the distribution of pedestrian intelligent agents could be visualized and displayed to the user, who could then judge whether the ideal distribution state has been reached and provide feedback on the judgment result. Based on the feedback judgment result, it could be determined whether the preset stage judgment condition has been met.

[0047] In step S13, it is determined whether the simulation scenario is an emergency scenario. If not, step S14 is executed; if yes, step S15 is executed. An emergency scenario is a simulation scenario in a disaster environment, such as a flood or fire. In this embodiment, whether the simulation scenario is an emergency scenario can be preset before simulation.

[0048] In step S14, each pedestrian agent is controlled to be in the normalization phase. Specifically, each pedestrian agent continues to be controlled to move based on its corresponding real-time normal speed and the planned path of the normalization phase.

[0049] In step S15, each pedestrian agent is controlled to be in the evacuation phase. For each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics. The evacuation phase refers to a simulated environment in a disaster environment, such as a flood or fire. During the evacuation phase, the pedestrian agent moves towards the nearest safe point or a predetermined safe point to complete the evacuation.

[0050] Physical fitness characteristics refer to the need to consider physical factors. Condition assessment characteristics refer to the potential harm and impact on pedestrians caused by disaster environments. For example, in a fire evacuation scenario, flames and smoke can injure or even kill pedestrians; while in a flood, wading through water can slow evacuation, cause slips, falls, or even drowning.

[0051] When the pedestrian agent is in the evacuation phase, the agent is controlled to continue moving based on its corresponding real-time evacuation speed and the planned evacuation path. The planned evacuation path is determined based on the pedestrian agent's phase-determined position and safe location. The phase-determined position refers to the position of the pedestrian agent when the simulation scenario is determined to be an emergency scenario. The real-time evacuation speed is determined in real time as the simulation progresses; for example, it can be determined once every simulation duration. The real-time evacuation speed is the walking speed of the pedestrian agent during the evacuation phase. The magnitude of the real-time evacuation speed is determined in real time based on the initial speed and the evacuation influence coefficient, and the direction of the real-time evacuation speed is determined based on the planned evacuation path.

[0052] Optionally, for each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics, and may include the following: The real-time evacuation speed is determined in real time based on the initial speed, evacuation impact coefficient, and evacuation phase planned path. The evacuation phase planned path is determined based on the pedestrian agent's phase-determined location and safe location. The evacuation impact coefficient includes: a second emotion impact coefficient, physical strength value, and a state impact coefficient. The second emotion impact coefficient is related to the pedestrian agent's level of panic.

[0053] During the evacuation phase, pedestrian agents walk or run in emergency scenarios, so the key to pedestrian simulation is still direction and speed, but there are more influencing factors.

[0054] The direction remains "goal-oriented," but the goal is a safe evacuation location. For example, the goal of all pedestrian agents is to reach Level 1, which is defined as a safe area under flood conditions. Global path planning is performed from the pedestrian agent's stage-determined location to the nearest point on Level 1 to determine the planned evacuation path for each stage. For example, using... The algorithm, such as Dijkstra's algorithm, is determined.

[0055] The real-time adjustment of speed is a comprehensive result of "emotional drive, physical strength guarantee, and status assessment".

[0056] "Emotion-driven" evacuation phase. In emergency situations, people experience anxiety and panic, leading them to walk or even run quickly. A second emotional influence coefficient is added to the initial speed, for example, multiplied; this second emotional influence coefficient is related to the degree of panic of the pedestrian agent.

[0057] There are many ways to determine the second emotional influence coefficient. For example, the simple coefficient method gives 1.6 for men and 1.76 for women.

[0058] Alternatively, a quantitative method can be used to determine the second emotional impact coefficient based on at least one of the following: the distance between the pedestrian agent and the hazard source (e.g., flood), the pedestrian density around the pedestrian agent, and the time the pedestrian agent spends in the hazardous space (e.g., the time spent in a wading area). Emotional drivers during evacuation can be adjusted in real-time by quantifying the degree of panic, surpassing simple coefficient methods (e.g., 1.6 for men, 1.76 for women) and employing multi-factor dynamic calculations to more accurately reflect psychological reactions in emergency scenarios.

[0059] Second emotional influence coefficient k panic It is a function of distance, density, and wading time, as shown in the following formula: k panic =1+α×f distance +β×f density +γ×f water_time .

[0060] Among them, f distance It is a distance function to the hazard source (such as a flood): f distance =max(0,1-d / d max ).

[0061] d is the current distance between the pedestrian agent and the hazard source (e.g., the edge of a flood). max It is the maximum influence distance (e.g., 20 meters; beyond this distance, the pedestrian agent considers it to have no influence), when d≤d max It will take effect at that time.

[0062] f density It is the pedestrian density function: f density=ρ / ρ max ρ is the local density (people / square meter). ma x represents the maximum density (e.g., 5 people per square meter), and high density exacerbates panic.

[0063] f water_time It is a function of wading time: f water_time =min(1,t / t max ), where t is the cumulative time a pedestrian spends in a dangerous area (e.g., a wading area). max It is the maximum time threshold (e.g., 30 seconds), where prolonged exposure to a dangerous area (e.g., prolonged wading) increases panic.

[0064] α, β, and γ are all weighting parameters, calibrated based on empirical data to reflect the relative importance of each factor. For example, α=0.5, β=0.3, γ=0.2.

[0065] In emergency situations, people panic and walk quickly, which rapidly depletes their energy and significantly impacts their chosen speed. Therefore, the issue of "energy conservation" needs to be considered. "Energy conservation" primarily considers the decrease in speed due to pedestrian fatigue over time.

[0066] Physical fitness tracking simulates the dynamic impact of fatigue on speed during the evacuation phase. Through physical fitness tracking and consumption mechanisms, it achieves more reasonable speed adjustments. Specifically, the impact of physical fitness on real-time evacuation speed can be considered based on the following:

[0067] Physical strength initialization. Each pedestrian agent's physical strength is initialized based on their age. For example, E initial =E base ×(1-0.01×age offset), where E base It is a basic physical strength (e.g., 100 units for adults and 60 units for the elderly), with an age offset of max(0, age-30) to simulate age-related decline.

[0068] Physical exertion. Physical exertion is dynamically consumed over time and with increasing speed. The exhaustion rate is calculated using the formula: dE / dt = -c × v 2 Where c is the consumption coefficient (e.g., 0.05); v is the current velocity (m / s), and the squared term simulates the accelerated fatigue of high-intensity exercise.

[0069] Speed ​​correction. Current speed is adjusted based on stamina percentage: v = v base ×min(1,E / E threshold ), where E threshold This is the stamina threshold (e.g., 30 units); E is the current stamina value. When E <E threshold At that point, the velocity decreases linearly. Furthermore, a nonlinear correction can be introduced to simulate extreme fatigue: v = v base×(0.7+0.3×E / E initial When E < 50.

[0070] Emergency scenarios, besides being physically demanding, can also cause injury. Examples include the high temperatures and toxic fumes of a fire. In flood scenarios, water affects speed and a person's condition. For instance, pedestrians in the water may slip or fall, drown, or even die. Therefore, a "condition assessment" is necessary.

[0071] State assessment simulates the harm and state changes of pedestrians caused by environmental hazards (such as water and fire) during the evacuation phase. It influences speed in real time through a state machine model, introducing continuous and multi-dimensional state transitions into the model. Specifically, the impact of state on real-time evacuation speed can be considered based on the following:

[0072] State variables: The pedestrian agent's state includes normal, wading, slipping, injured, drowning, etc. State transitions are based on environmental detection, which determines the pedestrian agent's current state.

[0073] 1) Wading state. Triggered when water depth h > 0.1 meters, velocity is affected by attenuation coefficient k. water Impact: k water =max(0.3,1-0.5×h). Where h is the water depth (meters).

[0074] 2) Slipping condition. Based on the probability model, the probability of slipping is p. slip p slip =p0×(1+α slip ×h). Where p0 is the base probability (e.g., 0.1), α slip This is the water depth coefficient (e.g., 0.5). If you slip, the state changes to slipping, and the velocity v=0 for Δt=3 seconds (simulated recovery time).

[0075] 3) Drowning state. When water depth h > 1.0 × height, and time t > t_drowning. drown If triggered at 10 seconds, the speed will be permanently reduced to 0, indicating death.

[0076] 3) Accumulated damage. In scenarios such as fire, the damage value D accumulates over time: D = ∫δ harm dt, where δ harm This is the injury rate (e.g., a smoke inhalation rate of 0.1 units / second). When D > D max (Determined based on the pedestrian's health condition), the condition changes to injury, and the speed is permanently reduced (e.g., multiplied by 0.5).

[0077] Therefore, the influence of the state on the real-time evacuation speed is that the real-time evacuation speed is the product of the initial velocity and the state influence coefficient: v = v base ×k stateAmong them, k state This is the state influence coefficient. When the state is normal, k state k is 1; when the state is wading, k state For k water When the state is slipping, k state k is 0; when the state is injured, k state It is 0.5; when the state is drowning, k state It is 0.

[0078] In the embodiments of this application, when the real-time evacuation speed of the pedestrian intelligent agent is only affected by the physical strength value, the real-time evacuation speed can be calculated according to the above calculation formula; when the real-time evacuation speed of the pedestrian intelligent agent is affected by the physical strength value and other parameters, the speed considering the influence of physical strength can be calculated first according to the above calculation formula, and then the influence of other parameters on the speed can be considered to calculate the real-time evacuation speed.

[0079] Optionally, in this embodiment, when the pedestrian agent is in the normalization phase, the micro-simulation method further includes: adjusting the real-time normal speed and / or planned path of the pedestrian agent in the normalization phase based on a local collision avoidance algorithm to avoid collisions. Specifically, the local collision avoidance algorithm can be the RVO algorithm or the ORCA algorithm.

[0080] In the embodiments of this application, during the normalization phase, the real-time speed correction does not consider the influence of factors such as "physical strength protection," but only collision avoidance. Specifically, the real-time speed correction adjusts the speed in real time due to local collision avoidance (such as using the ORCA algorithm) and the influence of surrounding density.

[0081] During the normalization phase, real-time speed adjustments are based on a dynamic process of local environmental interactions (such as collision avoidance) to make the speed more closely match real pedestrian behavior.

[0082] Optionally, in embodiments of this application, when the simulation scenario is an emergency scenario, the real-time evacuation speed and / or planned evacuation path can be adjusted based on a local collision avoidance algorithm to avoid collisions. Specifically, the local collision avoidance algorithm can be the RVO algorithm or the ORCA algorithm.

[0083] Optionally, in the embodiments of this application, when the simulation scenario is an emergency scenario, the global path planning can take into account the dynamics of the environment. Specifically, map attributes change in real time during the evacuation phase (such as flood spread), and the path planning needs to be updated periodically (once per second) to avoid dangerous areas. The ingenuity lies in the use of real-time data fusion, making the simulation more realistic.

[0084] When performing local collision avoidance, the behavioral parameters are adjusted: During the evacuation phase, panic causes pedestrians to be more aggressive. Therefore, the time horizon τ (i.e., collision avoidance prediction time) in the ORCA algorithm is shortened from 2-5 seconds in the normal phase to 0.5-1 seconds to simulate more urgent avoidance behavior.

[0085] In the embodiments of this application, the rules of each agent are executed at a fixed period, that is, the process is advanced according to the simulation step size.

[0086] The key feature of this application is that it endows the intelligent agent with observed human characteristics, making it a true pedestrian simulation (the core point), rather than attributing the intelligent agent with observed and summarized characteristics and laws of particles or other non-human matter. At different stages of the simulation, the pedestrian intelligent agent exhibits different attributes and methods.

[0087] In this embodiment, the microscopic pedestrian simulation model is divided into three stages: initialization, normalization, and evacuation. The normalization simulation includes two stages: initialization and normalization, while the emergency simulation includes three stages: initialization, normalization, and evacuation.

[0088] Initialization Phase. In this phase, attributes are assigned to each pedestrian agent and integrated with the map. Based on the given pedestrian attribute allocation method and principles, attributes such as gender, age, height, and weight are assigned to the agent. Based on the given initial coordinate allocation method and principles, a map coordinate is selected and assigned to the agent to determine its initial position. Pedestrians do not move during this phase.

[0089] Normalization Phase. Pedestrian simulation typically takes place during peak hours, when a large number of pedestrians are scattered across the scene due to their individual travel purposes and preferences. After the initialization phase, the pedestrian agents' positions are either relatively concentrated or relatively regular, differing somewhat from real-world scenarios. Therefore, a simulation period is required starting from the initialization phase. After several frames, the pedestrian agents with different speeds and paths will disperse. When all agents are in reasonable positions within the scene, making the simulation scenario at the current time closely resemble the real-world scenario, the normalization phase is complete. Normal pedestrian simulation continues after the normalization phase, ending once the simulation objective is achieved.

[0090] Evacuation Phase. This is a special phase, occurring only in pedestrian simulations under emergency conditions, such as evacuation during floods or fires. During this phase, all agents must change their destinations to ensure safe evacuation. This phase may require assessing the impact of the disaster environment (flood, fire) on pedestrian evacuation and the resulting harm to pedestrians.

[0091] Regarding the attributes and methods that intelligent agents should possess to make them more human-like, this application argues that human travel and walking behavior is not driven by social forces, magnetism, or the diffusion patterns of gases or liquids, but rather by "goal orientation, emotion-driven, physical strength-based, and state-assessment-based" factors.

[0092] 1) Goal-oriented. Every pedestrian journey has a starting point and an ending point. The ending point is the goal of the journey and determines the direction of the pedestrian. During normal operations, pedestrians complete their journeys according to the predetermined starting point and ending point. During evacuation, pedestrians move to the nearest safe point to complete the evacuation.

[0093] 2) Emotion-driven. Pedestrian walking speed is influenced by individual emotions. For example, commuters to work or school are usually more anxious and walk faster, while shoppers walk at a moderate pace, and those taking leisurely walks or exercising walk at a slower pace. During evacuation, pedestrians facing danger may experience panic, leading to a significant increase in speed, or even running.

[0094] 3) Physical Strength Assurance. During the normalization phase, physical strength is usually not a concern. However, it is crucial to consider during the evacuation phase, especially for evacuations of large spaces. Physical strength is closely related to the speed at which pedestrians evacuate.

[0095] 4) Status assessment. The evacuation phase may cause harm to pedestrians. For example, in a fire scenario, the flames and smoke can injure or even kill pedestrians; while in a flood, wading through water can slow down the evacuation, cause people to slip, fall, or even drown.

[0096] This application provides a human-centered pedestrian simulation model, encompassing three stages: initialization, normalization, and evacuation. This model embodies the characteristics of pedestrians: goal-oriented, emotion-driven, physical well-being, and state assessment. By treating the intelligent agent as an individual with independent characteristics and simulating these three stages (initialization, normalization, and evacuation), the model fully considers factors such as goal orientation, emotion-driven behavior, physical well-being, and state assessment, making the intelligent agent more closely resemble the behavioral characteristics of real pedestrians. This solves the problem of traditional models simulating intelligent agents as non-human entities.

[0097] The human-centered pedestrian simulation model provided in this application breaks through traditional thinking by treating pedestrians as individuals with independent consciousness and emotions. It simulates pedestrians' purposeful behavior by introducing a goal-oriented mechanism; considers the impact of emotions on pedestrian decision-making and actions by incorporating emotional driving factors; and integrates physical strength elements to simulate the constraints of physical strength in pedestrian movement, making the simulation model more human-centered and realistic. This invention is expected to propel pedestrian simulation technology to new heights, possessing significant innovative value and broad application prospects.

[0098] This application is based on an intelligent agent model to simulate pedestrian travel behavior according to human characteristics. This makes the results of micro-simulation closer to the behavioral characteristics of crowds in reality. At the same time, the model is easier to understand and easier to apply to the observation and research results of pedestrians in reality.

[0099] Secondly, this application also provides a multi-stage microscopic simulation device that integrates multiple attribute-driven approaches.

[0100] Figure 2 This is a structural block diagram of a multi-stage microscopic simulation device integrating multi-attribute driving according to a preferred embodiment of this application. Figure 2 As shown, the micro-simulation device includes an initialization module 10 and a processing module 20.

[0101] The initialization module 10 is used to initialize the pedestrian attributes and initial position of the pedestrian agent. For each pedestrian agent, the pedestrian attributes include an initial velocity, the magnitude of which is determined based on the emotion-driven characteristics.

[0102] The processing module 20 is used to control each pedestrian agent to be in the normalization stage when the preset stage judgment conditions are not met. For each pedestrian agent, its real-time normal speed in the normalization stage is determined in real time based on the goal-oriented characteristics. If the preset stage judgment conditions are met and the simulation scenario is a non-emergency scenario, the processing module 20 continues to control each pedestrian agent to be in the normalization stage. If the preset stage judgment conditions are met and the simulation scenario is an emergency scenario, the processing module 20 controls each pedestrian agent to be in the evacuation stage. For each pedestrian agent, its real-time evacuation speed in the evacuation stage is determined in real time based on the goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics.

[0103] Optionally, for each pedestrian agent, the magnitude of the initial velocity is determined based on emotion-driven characteristics, including: the magnitude of the initial velocity is determined based on a first emotion influence coefficient of the pedestrian agent, which is related to the pedestrian agent's travel purpose.

[0104] Optionally, for each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics, and state assessment characteristics. This includes: the real-time evacuation speed is determined in real time based on the initial speed, the evacuation impact coefficient, and the evacuation phase planned path; the evacuation phase planned path is determined based on the pedestrian agent's phase judgment position and safe location; and the evacuation impact coefficient includes: a second emotion impact coefficient, physical strength value, and a state impact coefficient, wherein the second emotion impact coefficient is related to the pedestrian agent's level of panic.

[0105] Optionally, for any pedestrian agent, the second emotion influence coefficient is determined based on at least one of the following: the distance between the pedestrian agent and the hazard source, the pedestrian density around the pedestrian agent, and the time the pedestrian agent stays in the hazard space.

[0106] Optionally, the processing module is also used to: when the pedestrian agent is in the normalization phase, adjust the real-time normal speed and / or planned path of the pedestrian agent in the normalization phase based on the local collision avoidance algorithm to avoid collisions.

[0107] The specific working principle of the microscopic simulation device provided in this application embodiment is similar to the specific working principle and benefits of the microscopic simulation method provided in this application embodiment, and will not be repeated here.

[0108] Thirdly, this application also provides a machine-readable storage medium storing instructions that cause a machine to perform the aforementioned microscopic simulation method.

[0109] Fourthly, this application also provides an electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the executable instructions to implement the microscopic simulation method described above.

[0110] The preferred embodiments of this application have been described in detail above. However, this application is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this application, various simple modifications can be made to the technical solution of this application, and these simple modifications all fall within the protection scope of this application.

[0111] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this application will not describe the various possible combinations separately.

[0112] Furthermore, various different implementations of this application can be combined in any way, as long as they do not violate the spirit of this application, they should also be regarded as the content disclosed in this application.

Claims

1. A multi-attribute, multi-stage microscopic simulation method based on pedestrian behavior characteristics, characterized in that, This microscopic simulation method includes: Initialize the pedestrian attributes and initial position of the pedestrian agent, wherein, for each pedestrian agent, the pedestrian attributes include an initial velocity, the magnitude of which is determined based on emotion-driven characteristics; If the preset stage judgment conditions are not met, each pedestrian intelligent agent is controlled to be in the normalization stage, wherein, for each pedestrian intelligent agent, its real-time normal speed in the normalization stage is determined in real time based on the target guidance characteristics. If the preset stage judgment conditions are met and the simulation scenario is a non-emergency scenario, continue to control each pedestrian agent to be in the normalization stage. When the preset stage determination conditions are met and the simulation scenario is an emergency scenario, each pedestrian agent is controlled to be in the evacuation stage. For each pedestrian agent, its real-time evacuation speed in the evacuation stage is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics and state assessment characteristics.

2. The microscopic simulation method according to claim 1, characterized in that, For each pedestrian agent, the magnitude of the initial velocity is determined based on emotion-driven characteristics, including: The magnitude of the initial speed is determined based on a first emotion influence coefficient of the pedestrian agent, which is related to the pedestrian agent's travel purpose.

3. The microscopic simulation method according to claim 1, characterized in that, For each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength preservation characteristics, and state assessment characteristics, including: The real-time evacuation speed is determined in real time based on the initial speed, the evacuation impact coefficient, and the evacuation phase planning path. The evacuation phase planning path is determined based on the pedestrian agent's phase judgment position and safe location. The evacuation impact coefficient includes: a second emotion impact coefficient, physical strength value, and a state impact coefficient. The second emotion impact coefficient is related to the degree of panic of the pedestrian agent.

4. The microscopic simulation method according to claim 3, characterized in that, For any pedestrian agent, the second emotion influence coefficient is determined based on at least one of the following: the distance between the pedestrian agent and the hazard source, the pedestrian density around the pedestrian agent, and the time the pedestrian agent stays in the hazard space.

5. The microscopic simulation method according to claim 1, characterized in that, When the pedestrian agent is in the normalization phase, the microscopic simulation method further includes: Based on a local collision avoidance algorithm, the real-time normal speed and / or planned path of the pedestrian agent in the normalization phase are adjusted to avoid collisions.

6. A multi-attribute, multi-stage microscopic simulation device based on pedestrian behavior characteristics, characterized in that, The microscopic simulation device includes: An initialization module is used to initialize the pedestrian attributes and initial position of the pedestrian agent. For each pedestrian agent, the pedestrian attributes include an initial velocity, the magnitude of which is determined based on emotion-driven characteristics. Processing module, used for: If the preset stage judgment conditions are not met, each pedestrian intelligent agent is controlled to be in the normalization stage, wherein, for each pedestrian intelligent agent, its real-time normal speed in the normalization stage is determined in real time based on the target guidance characteristics. If the preset stage judgment conditions are met and the simulation scenario is a non-emergency scenario, continue to control each pedestrian agent to be in the normalization stage. When the preset stage determination conditions are met and the simulation scenario is an emergency scenario, each pedestrian agent is controlled to be in the evacuation stage. For each pedestrian agent, its real-time evacuation speed in the evacuation stage is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength protection characteristics and state assessment characteristics.

7. The microscopic simulation device according to claim 6, characterized in that, For each pedestrian agent, the magnitude of the initial velocity is determined based on emotion-driven characteristics, including: The magnitude of the initial speed is determined based on a first emotion influence coefficient of the pedestrian agent, which is related to the pedestrian agent's travel purpose.

8. The microscopic simulation device according to claim 6, characterized in that, For each pedestrian agent, its real-time evacuation speed during the evacuation phase is determined in real time based on goal-oriented characteristics, emotion-driven characteristics, physical strength preservation characteristics, and state assessment characteristics, including: The real-time evacuation speed is determined in real time based on the initial speed, the evacuation impact coefficient, and the evacuation phase planning path. The evacuation phase planning path is determined based on the pedestrian agent's phase judgment position and safe location. The evacuation impact coefficient includes: a second emotion impact coefficient, physical strength value, and a state impact coefficient. The second emotion impact coefficient is related to the degree of panic of the pedestrian agent.

9. A machine-readable storage medium, characterized in that, The machine-readable storage medium stores instructions for causing the machine to perform the microscopic simulation method according to any one of claims 1-5.

10. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store the processor's executable instructions; The processor is configured to read the executable instructions from the memory and execute the executable instructions to implement the microscopic simulation method according to any one of claims 1-5.

Images