A pedestrian motion trajectory evolution method and system based on a social force model
By integrating physical forces into the social force model, a pedestrian dynamics model was constructed and parameters were combined, which solved the problem of not considering collision factors in pedestrian movement and achieved more accurate dynamic analysis and force simulation of crowds.
Patent Information
- Application Number
- CN202411384899.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2025-12-26
- Estimated Expiration
- 2044-09-30
AI Technical Summary
Existing social force models fail to effectively consider pedestrian collision factors when describing pedestrian movement, resulting in a large deviation between the crowd distribution and the real scene, and the crowd compression deformation is not obvious, with unreasonable oscillation phenomena.
Integrating the physical forces of pedestrians with other individuals and obstacles into the social force model, a pedestrian dynamics model is constructed. By parametrically combining physical and social forces, the model reflects the force situation of pedestrians in different situations and updates the speed and position of pedestrians.
It effectively solves the problem of multiple collisions in crowds, and the dynamic analysis results are more consistent with the real situation, and can more accurately reflect the force situation of pedestrians in different situations.
Smart Images

Figure CN119167653B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of crowd disturbance, and in particular to a pedestrian motion trajectory evolution method and system based on a social force model. BACKGROUND
[0002] In recent years, with the increasing large-scale crowd activities in public places, the problem of crowd safety has become prominent. According to incomplete statistics, on average, thousands of people die in crowd gathering events every year in the world, and more people are injured. Therefore, the problem of crowd safety has attracted much attention, and a large number of researches have been carried out on the description, laws and guidance of group motion behavior. The unsafe state of the crowd is jointly constituted by a variety of factors, including external factors such as place type, space structure, organization and management, safety measures, and sudden events, and internal factors such as crowd type and crowd behavior characteristics. Many scholars have proposed a series of pedestrian flow models in the study of crowd behavior and characteristics, including cellular automaton model, magnetic field force model, queuing model, social force model and gas dynamics model. Among them, the social force model has attracted more and more attention.
[0003] The social force (SF) model assumes that individuals are affected by three forces during evacuation: driving force, force between people, and force between people and obstacles. In the actual movement process of pedestrians, it is impossible for pedestrians to maintain a constant speed state unchanged, but due to the driving force, the pedestrian will try to restore the original state as much as possible within a certain time. At the same time, the change of the speed of the pedestrian is also affected by the external influence and the distance from the adjacent pedestrians and the wall. In addition, compared with the normal movement of rational pedestrians, panic pedestrians become disordered and chaotic due to the spread of panic emotions, which can cause collisions and falls of themselves or surrounding pedestrians.
[0004] There are still some deficiencies in the current research: 1) the pedestrians in the traditional social force model do not consider the collision factor of the pedestrians, resulting in a large deviation between the crowd distribution and the real scene; 2) the crowd compression deformation is not obvious, there is a large gap between the pedestrians, and there is an unreasonable oscillation phenomenon. SUMMARY
[0005] The purpose of the present application is to overcome the defects of the prior art and provide a pedestrian motion trajectory evolution method and system based on a social force model, which realizes the dynamic analysis of pedestrians in different situations.
[0006] The purpose of the present application can be achieved by the following technical solutions:
[0007] A pedestrian motion trajectory evolution method based on a social force model, comprising the following steps:
[0008] The simulated environment was set up according to the real scene, and the initial position of the pedestrian was obtained from the video using the object detection algorithm;
[0009] The actual forces experienced by a pedestrian during collision and non-collision moments are obtained based on the pedestrian dynamics model. The pedestrian's velocity is updated based on the change of the pedestrian's velocity over time, and then the pedestrian's position is updated.
[0010] The pedestrian's trajectory is obtained by visualizing the pedestrian's position at each moment during the movement.
[0011] The process of constructing the pedestrian dynamics model is as follows:
[0012] Based on the social force model, the interaction forces between pedestrians and other individuals, and the interaction forces between pedestrians and walls, are obtained;
[0013] Social forces are constructed based on the interaction forces between pedestrians and other individuals and the interaction forces between pedestrians and walls; physical forces and the change of pedestrian speed over time are constructed based on the contact or collision effects between pedestrians and other individuals.
[0014] A pedestrian dynamics model is constructed based on the changes in pedestrian speed over time, physical forces, and social forces.
[0015] Furthermore, the pedestrian's speed changes over time as follows:
[0016]
[0017] In the formula, τ i This indicates that the pedestrian returns to their original speed within a specific time. Let m represent the velocity of pedestrian i. i Indicate the mass of pedestrian i. Let i represent the velocity of pedestrian i at time t. This represents the interaction force between pedestrian i and pedestrian j. This represents the repulsive force exerted by the wall on pedestrian i. Indicates the desired direction.
[0018] Furthermore, the interaction forces between the pedestrian and other individuals are as follows:
[0019]
[0020] In the formula, A i The parameter r represents the magnitude of the influence related to pedestrian i. ij d represents the actual distance between pedestrian i and pedestrian j. ij Indicates the set baseline or reference distance, B ia parameter representing the decay speed of the influence index function related to the pedestrian i, a parameter representing the directionality of the force, a parameter representing the change amount of the relative speed between the pedestrian i and the pedestrian j, a parameter representing the component direction of the force in the tangential direction, k and k represent larger constants;
[0021] The interaction force between the pedestrian and the wall is as follows:
[0022]
[0023] where r i represents the distance between the object i and the wall, d iw represents the set reference distance or reference distance, a parameter representing the directionality of the force, a parameter representing the speed of the pedestrian i, a parameter representing the component direction of the force in the tangential direction. Further, the combined force on the pedestrian is as follows:
[0024] F i = a · F i,j_phy + b · F i_SF
[0025] a, b ∈ (0, 1)
[0026] where F i,j_phy represents the physical force on the pedestrian i, represents the social force on the pedestrian i, a and b are parameters.
[0027] Further, the dynamic model of the pedestrian is as follows:
[0028]
[0029] a, b ∈ (0, 1)
[0030] where m i represents the mass of the pedestrian i, represents the speed of the pedestrian i at time t.
[0031] Further, the social force is as follows:
[0032]
[0033] where τ i represents the time for the pedestrian to recover to the original speed, represents the speed of the pedestrian i, m i represents the mass of the pedestrian i, represents the speed of the pedestrian i at time t, represents the interaction force between pedestrian i and pedestrian j, represents the repulsion force of the boundary or obstacle to pedestrian i, represents the desired direction.
[0034] Further, the physical force is as follows:
[0035]
[0036] wherein m represents the mass of the pedestrian, u j represents the set of pedestrians moving together with pedestrian i, v k,- and v i,- represents the speed of pedestrian k and pedestrian i before collision.
[0037] Further, the physical force includes the interaction force between the pedestrian and other individuals.
[0038] Further, the physical force and the social force are linearly combined by adjusting the parameters a and b, and the actual force between pedestrians in different situations is reflected according to the combined force received by the pedestrian.
[0039] According to another aspect of the present application, a pedestrian motion trajectory evolution system based on a social force model is provided, comprising:
[0040] An initial position acquisition module is configured to set up a simulation environment according to a real scene, and obtain the initial position of the pedestrian from a video by using a target detection algorithm.
[0041] A pedestrian position updating module is configured to obtain the actual force received by the pedestrian at the collision moment and the non-collision moment in the motion process according to a pedestrian dynamics model, update the speed of the pedestrian according to the change of the speed of the pedestrian with time, and further update the position of the pedestrian.
[0042] A motion trajectory generation module is configured to obtain the motion trajectory of the pedestrian according to the position of the pedestrian at all moments in the motion process.
[0043] The construction process of the pedestrian dynamics model is as follows:
[0044] The interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall are obtained based on the social force model.
[0045] The social force is constructed according to the interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall, and the physical force and the change of the speed of the pedestrian with time are constructed according to the contact or collision effect between the pedestrian and other individuals.
[0046] The pedestrian dynamics model is constructed according to the change of the speed of the pedestrian with time, the physical force and the social force.
[0047] Compared with the prior art, the present application has the following beneficial effects:
[0048] 1. The present application integrates the physical force of the pedestrian contact with other individuals and obstacles into the classic social force model, can describe multiple simultaneous contacts, and the physical force effectively solves the inevitable multiple collisions in the crowd, so that the analysis result of the dynamics of the crowd is more in line with the actual situation.
[0049] 2. The present application constructs a pedestrian dynamics model, linearly combines the physical force and the social force through parameters, simulates the stress under different situations by adjusting the parameters, and can more accurately reflect the actual stress between pedestrians under different situations. BRIEF DESCRIPTION OF DRAWINGS
[0050] Figure 1 A flowchart of a pedestrian motion trajectory evolution method based on a social force model is provided for the present application.
[0051] Figure 2 A crowd scene when an earthquake occurs in a certain place in the embodiment, wherein (2a) is a real scene image, (2b) is a simulated scene image, wherein the red, green and blue circles respectively represent panic, normal and calm pedestrians, the green rectangle represents a safe exit, and the yellow circle represents a danger source.
[0052] Figure 3 A motion trajectory of a pedestrian collision in the embodiment, wherein brown and green represent two different pedestrians, the green rectangle represents a safe exit, and the yellow circle represents a danger source. DETAILED DESCRIPTION
[0053] The present application will be described in detail below in combination with the drawings and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present application, and gives a detailed implementation manner and specific operation process, but the protection scope of the present application is not limited to the following embodiments.
[0054] Embodiment 1
[0055] The present embodiment provides a pedestrian motion trajectory evolution method based on a social force model, as shown in Figure 1 , comprising the following steps:
[0056] S1, set up a simulation environment according to the real scene, and obtain the initial position of the pedestrian from the video by using a target detection algorithm.
[0057] The present embodiment selects a panic crowd escape scene in a certain market when an earthquake occurs in a certain place, as shown in Figure 2 , the real scene image is as shown in Figure 2 (2a).
[0058] Before the earthquake, all pedestrians are in normal state, at t = 3 seconds, the earthquake caused panic in the crowd, pedestrians from the alley entrance out to avoid the collapse of houses. Due to panic, pedestrians fleeing from the exit of varying degrees of contact and collision.
[0059] First according to the real scene set simulation environment, set the safety exit, the source of danger is located at point (644, 83), as shown in (2b) of Figure 2 The initial position of the pedestrian is obtained from the monitoring video by using the target detection algorithm, and the initial speed of all pedestrians is set to 0.
[0060] S2, according to the pedestrian dynamics model, the actual force of the pedestrian in the process of collision and non-collision is obtained, the speed of the pedestrian is updated according to the change of the speed of the pedestrian with time, and then the position of the pedestrian is updated.
[0061] The construction process of the pedestrian dynamics model is as follows:
[0062] The interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall are obtained based on the social force model;
[0063] The social force is constructed according to the interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall, and the physical force and the change of the speed of the pedestrian with time are constructed according to the contact or collision effect between the pedestrian and other individuals;
[0064] The pedestrian dynamics model is constructed according to the change of the speed of the pedestrian with time, the physical force and the social force.
[0065] The interaction force between the pedestrian and other individuals is as follows:
[0066]
[0067] In the formula, A i represents the size of the influence related to the pedestrian i, exp is used to describe the nonlinear relationship of the force with the distance, r ij represents the actual distance between the pedestrian i and the pedestrian j, and d ij represents the set reference distance or reference distance, B i represents the parameter of the decay rate of the influence exponential function related to the pedestrian i, represents the directionality of the force, which points to the pedestrian j, represents the change of the relative speed between the pedestrian i and the pedestrian j, represents the component direction of the force in the tangential direction, and k and k represent larger constants; r ij -d ijrepresents the difference between the actual distance and the reference distance, reflecting the relative position change between the two pedestrians, g(x) represents the coefficient of the force, and d ij > r ij The function g(x) is equal to 0, otherwise g(x) is equal to the parameter (x).
[0068] The interaction force between the pedestrian and the wall is as follows:
[0069]
[0070] In the formula, r i represents the distance between object i and the wall, d iw represents the set reference distance or reference distance, represents the directionality of the force, pointing to the wall w, represents the speed of pedestrian i, represents the component direction of the force in the tangential direction, r i -d iw represents the difference between the actual distance and the reference distance, reflecting the relative position change between the two pedestrians, g(x) represents the coefficient of the force, and d
[0071] The speed of the pedestrian changes over time as follows:
[0072]
[0073] In the formula, τ i represents the speed of the pedestrian returning to the original speed at a certain time, represents the speed of pedestrian i, m i represents the mass of pedestrian i, represents the speed of pedestrian i at time t, represents the interaction force between pedestrian i and pedestrian j, represents the repulsive force of the wall on pedestrian i, represents the desired direction.
[0074] The combined force on the pedestrian is as follows:
[0075] F i = a·F i,j_phy +b·F i_SF
[0076] a, b ∈ (0, 1)
[0077] In the formula, F i,j_phy represents the physical force on pedestrian i, represents the social force on pedestrian i, and a and b are parameters.
[0078] The dynamics model of the pedestrian is as follows:
[0079]
[0080] a, b e (0, 1)
[0081] where m i denotes the mass of pedestrian i, denotes the velocity of pedestrian i at time t.
[0082] The physical force and social force are linearly combined by adjusting parameters a and b, and the actual force between pedestrians in different situations is reflected according to the combined force received by the pedestrians. At the moment of collision, a = 1, b = 0, the movement of the pedestrian is completely affected by the collision, and the effect of the social force is 0; while at the moment of non-collision, a = 0, b = 1, the movement of the pedestrian is completely affected by the social force.
[0083] The social force is not directly generated by physical contact, but is based on the psychological state and behavior rules of pedestrians, etc. The social force is as follows:
[0084]
[0085] where τ i denotes the velocity of pedestrian i at time t, denotes the velocity of pedestrian i, m i denotes the mass of pedestrian i, denotes the velocity of pedestrian i at time t, denotes the interaction force between pedestrian i and pedestrian j, denotes the repulsive force of the boundary or obstacle to pedestrian i, denotes the desired direction.
[0086] The physical force is composed of the interaction force between objects in contact. It reflects the direct physical interaction between pedestrians. The physical force is as follows:
[0087]
[0088] where m represents the mass of the pedestrian, u j represents the set of pedestrians moving with pedestrian j, v k,- and v i,- represent the velocities of pedestrian k and pedestrian i before collision.
[0089] When a collision occurs between pedestrians, the action force and the reaction force generated by the collision of pedestrians are measured, and the data such as the velocity and mass of the colliding pedestrians are recorded. In this embodiment, a certain collision occurring in the simulation scene is selected as an example, and the number of pedestrians designed is 2, marked as i and j.
[0090] The pedestrians are regarded as rigid discs, and the physical force between pedestrian i and pedestrian j is as follows:
[0091]
[0092] The speed of the pedestrian is updated according to the change of the speed of the pedestrian with time, and the position of the pedestrian is updated.
[0093] S3, the pedestrian motion trajectory is obtained by visualizing the position of the pedestrian at each moment in the movement process.
[0094] The position of the pedestrian at each moment in the whole movement process is visualized, as shown in Figure 3 Brown represents pedestrian i, and green represents pedestrian j. Since pedestrian i and pedestrian j have collided, they have been affected by F i,j_phy , the trajectory of the collided pedestrian j deviates from the original forward direction at a small angle, but the pedestrian still travels in the direction of the destination after the collision. This is consistent with the real situation.
[0095] Embodiment 2
[0096] The embodiment provides a pedestrian motion trajectory evolution system based on a social force model, comprising:
[0097] An initial position acquisition module is configured to simulate an environment according to a real scene, and obtain the initial position of the pedestrian from a video by using a target detection algorithm;
[0098] A pedestrian position updating module is configured to obtain the actual force acting on the pedestrian at the collision moment and the non-collision moment in the movement process according to a pedestrian dynamics model, update the speed of the pedestrian according to the change of the speed of the pedestrian with time, and then update the position of the pedestrian;
[0099] A motion trajectory generation module is configured to obtain the pedestrian motion trajectory according to the position of the pedestrian at all moments in the movement process;
[0100] The construction process of the pedestrian dynamics model is as follows:
[0101] The interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall are obtained based on the social force model;
[0102] The social force is constructed according to the interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall, and the physical force and the change of the speed of the pedestrian with time are constructed according to the contact or collision action between the pedestrian and other individuals;
[0103] The pedestrian dynamics model is constructed according to the change of the speed of the pedestrian with time, the physical force and the social force.
[0104] The rest is the same as in embodiment 1.
[0105] The preferred embodiments of the present application have been described above in detail. It should be understood that modifications and variations to the present application can be affected by those skilled in the art without departing from the scope of the application. Accordingly, it is intended that all possible modifications and alterations be included within the scope of the present application as defined by the following claims.
Claims
1. A pedestrian motion trajectory evolution method based on a social force model, characterized in that, The method comprises the following steps: An initial position of the pedestrian is obtained from a video according to a simulation environment set according to a real scene and a target detection algorithm; Actual forces on the pedestrian at a collision time and a non-collision time in a movement process of the pedestrian are obtained according to a pedestrian dynamics model, a speed of the pedestrian is updated according to a speed change of the pedestrian over time, and then a position of the pedestrian is updated; A movement trajectory of the pedestrian is obtained through visual processing according to the position of the pedestrian at each time in the movement process of the pedestrian; The construction process of the pedestrian dynamics model is as follows: Interaction forces between the pedestrian and other individuals and interaction forces between the pedestrian and a wall are obtained based on a social force model; A social force is constructed according to the interaction forces between the pedestrian and other individuals and the interaction forces between the pedestrian and the wall, and a physical force and a speed change of the pedestrian over time are constructed according to a contact or collision action between the pedestrian and other individuals, and a combined force on the pedestrian is as follows: wherein, represents a physical force experienced by a pedestrian, represents a social force experienced by a pedestrian, and are parameters, and by adjusting the parameters and linearly combining the physical force and the social force, reflecting the actual force experienced by the pedestrian under different situations according to the combined force experienced by the pedestrian, wherein the physical force is as follows: wherein, m represents the mass of the pedestrian, represents the set of pedestrians moving with the pedestrian j, and represents the pedestrian k and the pedestrian i velocity before the collision; The pedestrian dynamics model is constructed according to the speed change of the pedestrian over time, the physical force and the social force, and the dynamics model of the pedestrian is as follows: wherein denotes the mass of the pedestrian , denotes the velocity of the pedestrian at the time instant t.
2. The method of claim 1, wherein, The speed change of the pedestrian over time is as follows: wherein represents the time for the pedestrian to recover to the original speed, represents the speed of the pedestrian , represents the mass of the pedestrian , represents the speed of the pedestrian at the time , represents the interaction force between the pedestrian and the pedestrian , represents the repulsive force of the wall to the pedestrian , (t) represents the desired direction.
3. The method of claim 1, wherein the method further comprises: The interaction forces between the pedestrian and other individuals are as follows: In the formula, a parameter indicating the magnitude of influence related to the pedestrian, a parameter indicating the actual distance between the pedestrian and the vehicle, a parameter indicating the actual distance between the pedestrian and the vehicle, a parameter indicating the set reference distance or reference distance, a parameter indicating the decay speed of the influence exponential function related to the pedestrian, a parameter indicating the directionality of the force, a parameter indicating the change amount of the relative speed between the person and the pedestrian, a parameter indicating the change amount of the relative speed between the person and the pedestrian, a parameter indicating the change amount of the relative speed between the person and the pedestrian, a parameter indicating the component direction of the force in the tangential direction, a parameter indicating the component direction of the force in the tangential direction, a parameter indicating the component direction of the force in the tangential direction, a parameter indicating the component direction of the force in the tangential direction, The interaction forces between the pedestrian and the wall are as follows: wherein distance between the object distance between the object distance between the object directionality of the force pedestrian speed of the pedestrian direction of the tangential component of the force 4. The method of claim 1, wherein, The social force is as follows: wherein, denotes the time for the pedestrian to recover to the original velocity, denotes the velocity of the pedestrian , denotes the mass of the pedestrian , denotes the velocity of the pedestrian at the time , denotes the interaction force between the pedestrian and the pedestrian , denotes the repulsive force of the boundary or obstacle on the pedestrian , (t) denotes the desired direction.
5. The method of claim 1, wherein the method further comprises: The physical force comprises the interaction forces between the pedestrian and other individuals. 6.A pedestrian motion trajectory evolution system based on a social force model, characterized in that, The method comprises the following steps: An initial position of the pedestrian is obtained from a video according to a simulation environment set according to a real scene and a target detection algorithm; Actual forces on the pedestrian at a collision time and a non-collision time in a movement process of the pedestrian are obtained according to a pedestrian dynamics model, a speed of the pedestrian is updated according to a speed change of the pedestrian over time, and then a position of the pedestrian is updated; A movement trajectory of the pedestrian is obtained through visual processing according to the position of the pedestrian at each time in the movement process of the pedestrian; The construction process of the pedestrian dynamics model is as follows: Interaction forces between the pedestrian and other individuals and interaction forces between the pedestrian and a wall are obtained based on a social force model; A social force is constructed according to the interaction forces between the pedestrian and other individuals and the interaction forces between the pedestrian and the wall, and a physical force and a speed change of the pedestrian over time are constructed according to a contact or collision action between the pedestrian and other individuals, and a combined force on the pedestrian is as follows: wherein represents a physical force experienced by a pedestrian, represents a social force experienced by a pedestrian, and are parameters, and by adjusting the parameters and linearly combining the physical force and the social force, reflecting the actual force experienced by the pedestrian under different situations according to the combined force experienced by the pedestrian, wherein the physical force is as follows: wherein, m representing the mass of a pedestrian, representing the set of pedestrians moving with pedestrian j, and representing the pedestrian k and the pedestrian i the speed before the collision; The pedestrian dynamics model is constructed according to the speed change of the pedestrian over time, the physical force and the social force, and the dynamics model of the pedestrian is as follows: wherein denotes the mass of the pedestrian , denotes the velocity of the pedestrian at the time instant.
Citation Information
Patent Citations
Crowd simulation method for social force model based on relative speed
CN110276123A
Crowd emergency evacuation simulation method based on social force model and A star algorithm
CN116523160A