Group motion simulation method and system based on manual intervention and computer equipment

Abstract

The invention belongs to the technical field of group motion simulation, and particularly relates to a group motion simulation method and system based on manual intervention and computer equipment. Comprising the following steps: building a group motion scene, adding intelligent agents for simulating individuals in a group into the group motion scene, and controlling each intelligent agent to move so as to simulate the group motion condition to obtain a group motion simulation result; in the simulation process, when a certain agent which can be manually intervened in microscopic regulation and control has abnormal motion, the simulation time is adjusted to be before the abnormal motion, so that the agent moves according to a microscopic regulation and control instruction of a microscopic decision-making person. A manual control strategy is added in the simulation process through microscopic regulation and control, so that a simulation result is closer to human behaviors, and the flexibility and simulation quality of an algorithm are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of group motion simulation, and particularly relates to a group motion simulation method and system based on manual intervention, and a computer device. Background Art

[0002] Crowd simulation plays a crucial role in understanding and simulating crowd dynamics and individual behaviors. It is widely used in fields such as public safety, traffic planning, and disaster emergency. It can reproduce large-scale crowd movements at a low cost and infer future crowd movement trends, which is beneficial for decision-makers to intervene in advance and prevent accidents. However, the existing group simulation methods only use algorithms to simulate human behaviors for simulation. Limited by the inflexible logic of algorithm settings, since human behaviors have the characteristics of flexibility and variability, only using algorithm simulation to simulate human behaviors will result in a large difference between the simulation results and the actual crowd movement results, and cannot accurately predict crowd movements, resulting in the lack of authenticity and reference significance of the simulation results. Summary of the Invention

[0003] The purpose of the present invention is to provide a group motion simulation method and system based on manual intervention, and a computer device, so as to solve the technical problem of the lack of authenticity of group motion simulation results in the prior art.

[0004] To solve the above technical problem, the present invention provides a group motion simulation method based on manual intervention, including the following steps:

[0005] Build a group motion scenario, add agents for simulating individuals in the group motion scenario, and control the movement of each agent to simulate the group motion situation, so as to obtain a group motion simulation result;

[0006] During the simulation process, when a motion anomaly occurs in an agent that can be micro-controlled manually, adjust the simulation time to before the occurrence of the motion anomaly, and enable the agent to move according to the micro-control instructions of the micro-decision maker.

[0007] Further, the micro-control instruction is a control instruction issued by the micro-decision maker in the first-person virtual reality scenario substituted into the agent that needs to be micro-controlled.

[0008] Further, the operation of adjusting the simulation time to before the occurrence of the motion anomaly is an operation performed by the macro-decision maker in the mixed reality perspective.

[0009] Further, the process of controlling the movement of each agent includes:

[0010] 1) Match in the database storing the agent's movement trajectories based on the agent's current position and average movement direction, obtain the agent's current destination according to the matching result, and obtain the complete movement trajectories of each agent to the destination based on the current positions and current destinations of each agent;

[0011] 2) Repeat step 1) to obtain the group movement trajectory.

[0012] Further, the process of matching in the database storing the agent's movement trajectories based on the agent's current position and average movement direction and obtaining the agent's current destination includes: calculating the historical average movement direction of each agent movement trajectory in the database; calculating the angular deviation between the agent's average movement direction and the historical average movement direction of each agent movement trajectory, and the position deviation between the agent's current position and the current positions in each agent movement trajectory in the database, summing the two calculated deviations to obtain the total deviation, and taking the destination of the agent movement trajectory in the database with the smallest total deviation from the agent without a set destination as the destination of this agent.

[0013] Further, the group movement scenario includes a mixed reality movement scenario and the corresponding virtual reality movement scenario. The agent exists in both the mixed reality movement scenario and the virtual reality movement scenario, and the movement states in the two scenarios are the same.

[0014] The present invention is an improved invention. Its beneficial effects are as follows: When performing group simulation, the present invention first constructs a group movement scenario, then adds agents to the group movement and sets the movement conditions of each agent, performs simulation of the group movement according to the set movement conditions to obtain the group movement simulation result. During the simulation process, when an agent's movement is abnormal, the simulation time is adjusted to before the agent's movement abnormality occurs, and micro-control is performed on this agent. Micro-control is to control the agent's movement according to the micro-control instruction issued by the decision-maker. By adding an artificial control strategy during the simulation process through micro-control, the simulation result is closer to human behavior, thereby improving the flexibility and simulation quality of the algorithm.

[0015] To solve the above technical problems, the present invention also provides a computer device, including a processor, and the processor is used to implement the method steps as described in the method of group movement simulation based on manual intervention of the present invention when executing a computer program.

[0016] The present invention is an improved invention, and its beneficial effects are the same as those of the method of group movement simulation based on manual intervention of the present invention.

[0017] To solve the above technical problems, the present invention also provides a group motion simulation system based on manual intervention, including a manual intervention decision-making system and a simulation module. The manual intervention decision-making system is used to receive the micro-control instructions of the micro decision-makers and send them to the simulation module. The simulation module is used to implement the following steps:

[0018] Build a group motion scenario, add agents for simulating individuals in the group to the group motion scenario, set the motion conditions of the agents, and make each agent move according to the set motion conditions to simulate the group motion situation and obtain the group motion simulation result. During the simulation process, when a motion anomaly occurs in an agent that can be manually intervened and micro-controlled, adjust the simulation time to before the motion anomaly occurs, and control the motion of the agent according to the micro-control instructions of the micro decision-makers.

[0019] Further, the manual intervention decision-making system includes a virtual reality decision-making system for receiving the micro-control instructions of the micro decision-makers.

[0020] Further, the manual intervention decision-making system includes a mixed reality decision-making system for adjusting the simulation time to before the motion anomaly occurs when a motion anomaly occurs in an agent that can be manually intervened and micro-controlled.

[0021] The present invention is an improved invention, and its beneficial effects are the same as those of the group motion simulation method based on manual intervention of the present invention. Description of the Drawings

[0022] Figure 1 It is a schematic diagram of the virtual reality-based group motion simulation method of the system embodiment of the present invention;

[0023] Figure 2 It is an example diagram of the macro perspective of the system embodiment of the present invention;

[0024] Figure 3 It is an example diagram of the micro perspective of the system embodiment of the present invention. Detailed Embodiments

[0025] A method, system and computer device for group motion simulation based on manual intervention in the present invention. When performing group simulation, first, a group motion scenario is built. Then, agents are added to the group motion and the motion conditions of each agent are set. The simulation of the group motion is carried out according to the set motion conditions to obtain the group motion simulation result. During the simulation process, when an agent's motion is abnormal, the simulation time is adjusted to before the agent's motion becomes abnormal, and micro-control is performed on the agent. Micro-control is to control the agent's motion according to the micro-control instruction sent by the decision-maker. By adding a human control strategy during the simulation through micro-control, the simulation result is closer to human behavior, thus improving the flexibility and simulation quality of the algorithm.

[0026] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0027] System embodiment:

[0028] The group motion simulation system based on manual intervention in the present invention includes a manual intervention decision-making system and a simulation module. The manual intervention decision-making system is used to receive the micro-control instruction of the decision-maker and send it to the simulation module. The simulation module is used to implement the steps of the group motion simulation based on manual intervention in the present invention.

[0029] In this embodiment, the group motion simulation system based on manual intervention further includes a mixed reality motion scenario, a virtual reality motion scenario, physical agents and virtual agents. All agents in the virtual reality motion scenario are virtual agents, and there are corresponding physical agents or virtual agents in the mixed reality motion scenario. The motion states of the corresponding agents in the two scenarios are exactly the same at the same time. In this embodiment, the manual intervention decision-making system includes a virtual reality decision-making system (VR helmet) and a mixed reality decision-making system (MR helmet). The virtual reality decision-making system is used to receive the micro-control instruction of the decision-maker. The mixed reality decision-making system is used to adjust the simulation time to before the motion abnormality occurs when a micro-control-enabled agent has a motion abnormality. The mixed reality decision-making system is also used to complete operations such as scheduling resources, setting task goals or coordinating actions.

[0030] In this embodiment, the agents involved include the following four categories:

[0031] 1) Virtual - autonomous motion - pedestrian: Abbreviated as A pedestrian, which is controlled by the group motion simulation algorithm to generate its motion trajectory.

[0032] 2) Virtual - Decision Intervention - Pedestrian: Abbreviated as I Pedestrian. There are two types of control modules for the I Agent. One is controlled by the group motion simulation model, and the other is controlled by the micro - operator using virtual reality (VR) technology after the macro - operator issues a manual intervention instruction.

[0033] 3) Physical - Autonomous Movement - Car: Abbreviated as A Car. This car is a physical car that moves in the electronic sand table, and there is a corresponding digital avatar in the virtual scene. The A Car is controlled by the group motion simulation method to generate its movement trajectory.

[0034] 4) Physical - Decision Intervention - Car: Abbreviated as I Car. This car is a physical car that moves in the electronic sand table, and there is a corresponding digital avatar in the virtual scene. There are two types of control modules for the I Car. One is controlled by the group motion simulation model, and the other is controlled by the micro - operator using virtual reality (VR) technology after the macro - operator issues a manual intervention instruction.

[0035] The above four types of agents can be divided into two categories according to whether they can perform decision intervention (micro - regulation), namely, agents that can be micro - regulated (including I Pedestrians and I Cars) and agents that cannot be micro - regulated (including A Pedestrians and A Cars). As other implementation methods, the agents can also only include agents that can be micro - regulated.

[0036] In this embodiment, the obstacles are controlled by the macro - decision - maker. The generation position, generation time, and movement trajectory can be determined either during the scene initialization process or during the simulation process. The obstacle can also be a static obstacle, and only the generation position and generation time need to be determined for the static obstacle.

[0037] In this embodiment, since the car is a physical car, a corresponding electronic sand table is required. Using the physical car and the electronic sand table to simulate vehicle - type agents can make the simulation process more intuitive and vivid, facilitating the display of the simulation process. As other implementation methods, the A Car and the I Car can also use completely virtual cars, that is, digital avatars in the virtual scene, without using physical cars and electronic sand tables, reducing the simulation cost.

[0038] At the same time, since the carts and sand tables in the system are physical entities, the human intervention decision-making system in this embodiment includes an MR helmet and a VR helmet. The MR helmet is worn by the macro decision-maker to provide the electronic sand table function with a global perspective, observe the global situation map, agent distribution, and dynamic movement information in the virtual-real fusion environment, facilitating the macro decision-maker to make macro decisions; the VR helmet is worn by the micro decision-maker to provide an immersive first-person operation environment, supporting the local task execution of the micro decision-maker. The micro decision-maker wears the VR helmet to perceive the scene details in real time from the virtual reality perspective of the agent and issue micro operation instructions to complete specific operation tasks.

[0039] As another implementation, when both types of cart agents are virtual agents and there is no sand table, the human intervention decision-making system can include two VR helmets, and the macro decision-maker also wears the VR helmet to make macro decisions.

[0040] The group motion simulation method based on human intervention of the present invention, as Figure 1 described, includes the following steps:

[0041] Step 1: Initialize the scene of group motion.

[0042] Before performing the simulation, it is necessary to set the scene of group motion according to the needs of the simulation first.

[0043] Specifically, first build the group motion scene, and place pedestrian A, pedestrian I, obstacles, cart A, and cart I in the built scene. Cart A and pedestrian A are controlled by the group motion simulation method, and pedestrian I and cart I are also controlled by the group motion simulation method before the decision intervention. After the placement, set the positions and movement trajectories of the obstacles. In this embodiment, there are pedestrians, carts, and obstacles in the scene to be simulated. The agents added to the scene include pedestrians and carts, and in addition to the agents, obstacles are also set. As another implementation, according to the needs of the simulation scene, the agents added can also only include any one of pedestrians and carts, and it can be selected whether to add obstacles according to the situation in the simulation scene.

[0044] The group motion simulation method of the present invention performs matching in the database for storing the motion paths of agents to obtain the short-term destinations of the agents, and obtains the complete motion trajectories of the agents to the destinations based on the short-term destinations, current positions, current speeds, and obstacle positions of all agents.

[0045] The group motion simulation method of this embodiment selects the data and model hybrid driven approach (DAMHDA). The destination determination method based on "current position - average motion direction" of this method can calculate the short-term movement target of the agent. Then, the model-driven method is used to generate the complete motion trajectory of the agent. The specific method is as follows:

[0046] First, obtain the candidate destination of the agent according to the destinations of the agents in the database that are close to the average motion direction of the controlled agent: Calculate the current position and average motion direction of the controlled agent. Based on this information, search for a matching individual in the database that is most similar to its position and average motion direction, and use the destinations of these matching individuals as the destination of the controlled agent. The calculation formula for the average motion direction of the agent is as follows:

[0047]

[0048] In the formula, (dirx t (i), diry t (i)) represents the average movement direction of the i-th agent, t is the current time, x t is the abscissa of the current position, x k is the abscissa of the position at time k, y t is the ordinate of the current position, y k is the ordinate of the current position.

[0049] The method for matching the current position and direction of the controlled agent with the motion trajectories of the agents in the database is as follows:

[0050]

[0051] In the formula, f is the matching result, dis(i) is the deviation between the current position of the controlled agent and the current position of the matching agent, dir_x_cen and dir_y_cen represent the current average motion direction of the research object, cos is used to measure the angular deviation of the average motion direction, and g(i) is the total deviation. Since the value of cos is only used to measure the angular deviation, not the actual angular deviation, and is inversely proportional to the angle in the range of 0 - π, the larger the cos value, the smaller the deviation angle. Therefore, there is a minus sign between the two deviations.

[0052] Obtain the current destination of the controlled agent according to the matching result. Repeatedly use the above method to obtain the current destinations of all agents. According to the current positions, speeds, and destinations of each agent in the scene, combined with the positions of the obstacles at the current moment, obtain the complete motion trajectory of the agent to the destination. The formula is as follows:

[0053] Pn +1 = f(P n , Des, V, Obs)

[0054] wherein, P n represents the position information of all individuals at the nth time step, Des represents the destinations of all individuals, V represents the speeds of all individuals, including the average speed and the maximum speed. These speeds are calculated based on the historical trajectory information. Obs represents the position information of all obstacles in the scene at the nth time step, which is obtained according to the set obstacle movement trajectory.

[0055] Repeat the above destination matching and position calculation methods, continuously adjust the destinations and movement conditions of all agents, so as to obtain the group movement trajectory.

[0056] Step 2: Macro decision control.

[0057] Perform macro decision control according to the operation results of the group movement simulation model.

[0058] In this embodiment, the macro decision maker wearing the MR helmet observes the global state in the MR (mixed reality) mode and adjusts the states of the agents and the operation parameters of the simulation model. For example, when the macro decision maker finds that the movement trajectory of the I pedestrian or the I car is abnormal from the mixed reality perspective, the running time of the simulation model can be adjusted to before the abnormality occurs, and a micro decision instruction can also be issued to enable the micro decision maker wearing the VR helmet to view the pedestrian or car with abnormal movement trajectory from the first-person perspective and immerse into the virtual scene to control the pedestrian or car in the first person. The macro perspective of the MR helmet is as Figure 2 shown, and the first-person perspective of the VR helmet is as Figure 3 shown.

[0059] Step 3: Micro decision control.

[0060] After the macro decision is completed, run the simulation model for simulation. During the simulation process, the agents that do not receive micro decision instructions use the group movement simulation model to control their movement. For the agents that need to adopt micro decisions, the micro decision maker substitutes into the first-person perspective of the agent that needs to be micro-controlled to manually control the agent and make the agent move towards the destination.

[0061] Through micro decisions and macro decisions, the present invention adds manual control during the simulation process, making the simulation process affected by humans and no longer completely limited to the fixed simulation model, enhancing the flexibility of the simulation, being closer to human behavior, and thus enhancing the authenticity of the simulation results.

[0062] Step 4: Record each simulation process to obtain the group movement simulation results.

[0063] Record all the data generated during each simulation process to obtain the group motion simulation results, including the positions, movement directions, speeds, and accelerations of each agent at each moment. In addition, the macro decision instructions of the macro decision maker will also be recorded.

[0064] Step Five: Repeat the simulation.

[0065] In this embodiment, after one simulation is completed, the recorded simulation process data can be called to perform another simulation based on the results of the previous simulation. When performing the re - simulation, the macro decision maker can choose to increase the number of agents until the simulation results meet the set requirements, such as reaching a certain number of simulation times or the simulation results being close to the real situation.

[0066] Method Embodiment:

[0067] A method for group motion simulation based on manual intervention according to the present invention is as described in the method for group motion simulation based on manual intervention in the system embodiment. The specific process, principle, and beneficial effects of this method have been described in detail in the method embodiment, and will not be elaborated in this embodiment.

[0068] Computer Device Embodiment:

[0069] A computer device according to the present invention includes a processor, and the processor is used to implement the steps of the method for group motion simulation based on manual intervention as described in the system embodiment when executing a computer program. The specific process, principle, and beneficial effects of this method have been described in detail in the method embodiment, and will not be elaborated in this embodiment.

[0070] Among them, the processor can be a processor such as an MCU or an FPGA.

[0071] For a method, system, and computer device for group motion simulation based on manual intervention according to the present invention, when an abnormality occurs in the movement of an agent during the simulation process, the time is adjusted to before the abnormality occurs in the agent's movement, providing a time basis for manual micro - regulation correction. Then, micro - regulation is used, and the micro - decision maker substitutes into the first - person perspective of the agent with abnormal movement that can be micro - regulated to manually control it, making the control of the agent completely close to manual control, thereby improving the flexibility of the simulation and the quality of the simulation results.

[0072] Furthermore, the present invention obtains the current destination of the agent by matching in the database, and then obtains the complete movement trajectory of the agent from its current position to the destination according to the current position and destination of the agent.

[0073] Furthermore, there are also physical agents in the mixed reality scenario of the present invention, and corresponding virtual agents exist in the virtual simulation scenario. The physical agents and the corresponding virtual agents have the same motion state, which makes the motion of the agents more intuitively displayed, facilitating the observation of whether manual intervention is required, enabling timely manual intervention, reducing the time of abnormal agent motion, and improving the simulation efficiency.

[0074] Through the above method, a group motion simulation method and system, and a computer device based on manual intervention of the present invention can obtain more accurate group motion simulation results, and these simulation results can be applied to fields such as traffic command and layout design. For example, in the field of traffic command, when the simulation result indicates that there will be traffic congestion in the scenario, traffic commanders can be assigned in advance for guidance and dredging, thereby reducing traffic congestion. Also, the path selection strategy of the agents can be optimized according to the motion paths of the agents after manual intervention. In the field of layout design, the completed building layout design can be used as a template to build a group motion scenario. By observing the motion simulation results of the agents in the scenario, it can be judged whether there is crowd congestion, and thus whether the layout setting is reasonable.

Claims

1. A group motion simulation method based on manual intervention, characterized in that: The following steps are involved: Build a group motion scene, add agents to the scene to simulate individuals in the group, control the movement of each agent to simulate the group motion, and obtain the group motion simulation result; During the simulation process, when an intelligent agent that can be micro-regulated by human intervention has abnormal movement, the simulation time is adjusted to before the abnormal movement occurs, so that the intelligent agent moves according to the micro-regulation instructions of the micro-decision maker.

2. The group motion simulation method based on manual intervention according to claim 1, characterized in that: The micro-control instructions are control instructions issued by the micro-decision maker into the first-person perspective virtual reality scene of the intelligent body that requires micro-control.

3. The group motion simulation method based on manual intervention according to claim 1, characterized in that: The operation of adjusting the simulation time to before the motion anomaly occurs is performed by the macro decision maker under the mixed reality perspective.

4. The group motion simulation method based on manual intervention according to claim 1, characterized in that: The process of controlling the movement of each agent includes: 1) According to the current position and average movement direction of the agent, a match is performed in a database storing the movement trajectory of the agent, and the current destination of the agent is obtained according to the matching result. According to the current position and current destination of each agent, the complete movement trajectory of each agent to the destination is obtained; 2) Repeat step 1) to obtain the group motion trajectory.

5. The group motion simulation method based on manual intervention according to claim 4, characterized in that: A match is performed in a database storing the movement trajectory of the agent based on the current position and average movement direction of the agent, and the process of obtaining the current destination of the agent based on the matching result includes: calculating the historical average movement direction of each agent movement trajectory in the database; calculating the angular deviation between the average movement direction of the agent and the historical average movement direction of each agent movement trajectory, as well as the position deviation between the current position of the agent and the current position in each agent movement trajectory in the database, summing the two calculated deviations to obtain the total deviation, and taking the destination of the agent movement trajectory in the database that has the smallest total deviation with the agent without a set destination as the destination of the agent.

6. The group motion simulation method based on manual intervention according to claim 1, characterized in that: The group motion scene includes a mixed reality motion scene and a corresponding virtual reality motion scene. The intelligent agent exists in both the mixed reality motion scene and the virtual reality motion scene, and the motion state in the two scenes is the same.

7. A computer device comprising a processor, characterized in that: The processor is used to implement the steps of the group motion simulation method based on manual intervention as described in any one of claims 1-6 when executing the computer program.

8. A group motion simulation system based on human intervention, characterized in that: It includes a human intervention decision-making system and a simulation module. The human intervention decision-making system is used to receive micro-control instructions from micro-decision-makers and send them to the simulation module. The simulation module is used to implement the following steps: Build a group motion scene, add intelligent agents for simulating individuals in the group to the group motion scene, set the motion conditions of the intelligent agents, make each intelligent agent move according to the set motion conditions to simulate the group motion conditions, and obtain the group motion simulation results; during the simulation process, when an intelligent agent that can be manually intervened in micro-control has abnormal motion, adjust the simulation time to before the abnormal motion occurs, and control the movement of the intelligent agent according to the micro-control instructions of the micro-decision maker.

9. The group motion simulation system based on manual intervention according to claim 8, characterized in that: The human intervention decision-making system includes a virtual reality decision-making system for receiving micro-control instructions from micro-decision makers.

10. The group motion simulation system based on manual intervention according to claim 8, characterized in that: The human intervention decision system includes a mixed reality decision system that adjusts the simulation time to the time before the movement abnormality occurs when a certain intelligent body that can be micro-regulated by human intervention occurs abnormal movement.

Images