A civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity

Abstract

This invention discloses a civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity. The method divides the cabin into seating, aisle, and exit areas and sets independent target point sets. A two-stage screening technique is employed to determine effective action edges, avoiding abnormal repulsive forces at aisle entrances. Differentiated social force model parameters are established to address mechanical deadlock and lateral oscillation issues. A priority passage mechanism triggered by a waiting threshold is used to overcome the problem of permanent waiting. The kinetic energy is kept constant during area transitions through kinetic energy conservation and transformation, addressing the issue of acceleration lag during turns. An aisle boundary constraint protection mechanism is set up to correct positions and velocities exceeding the boundaries in real time, eliminating lateral oscillations. Verified through personnel evacuation tests under different operating conditions, the average error between the simulation results and the actual evacuation time is controlled within 10%, significantly improving the realism and reliability of civil aircraft emergency evacuation simulations and providing a scientific calculation tool for airworthiness certification and cabin safety assessment.

Description

Technical Field

[0001] This invention belongs to the field of emergency evacuation simulation technology, specifically relating to a civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity. Background Technology

[0002] In civil aircraft emergency evacuation scenarios, the aircraft cabin, as a typical densely populated enclosed space, is characterized by narrow passageways (typically ≤55cm wide), a limited number of exits, and a dense distribution of obstacles. This unique environment makes it difficult for full-scale physical experiments to comprehensively capture the microscopic characteristics of human movement (such as individual velocity distribution, exit selection behavior, and congestion formation mechanisms), and such experiments are costly, time-consuming, and pose safety hazards. Therefore, high-precision computer simulation models have become a key technical means for airworthiness certification and safety assessment.

[0003] Current mainstream personnel evacuation simulation models mainly include cellular automata, agent models (such as Pathfinder), data-driven models (such as SMCNN-LSTM), and social force models. Each type of model has its own advantages in practical applications, but also inherent limitations: Cellular automata models, by discretizing space into a grid and using local rules to simulate crowd movement, are suitable for analyzing the impact of macroscopic parameters such as passenger density and type distribution. However, their "jumping" movement mechanism struggles to accurately reflect the physical interactions and congestion dynamics in continuous space, leading to unnatural queuing sequences and spatial distribution deviations in high-density scenarios. Agent models, through continuous motion engines simulating individual avoidance and queuing behaviors, possess strong rule-expressive capabilities and engineering applicability. However, their underlying algorithms lack transparency, making it difficult to characterize psychologically driven behaviors such as panic and herd mentality. Furthermore, they are prone to irrational movement phenomena such as "fast is slow" in global path planning and high-density panic scenarios. Data-driven models can identify behavioral patterns such as competition and cooperation from trajectory data with high classification accuracy. However, they heavily rely on large amounts of labeled data, resulting in weak model interpretability and difficulty in generalizing to dynamic evolution prediction in unseen scenarios.

[0004] In contrast, the social force model, based on the physical framework of Newtonian mechanics, offers advantages in transparency and interpretability in describing individual movement and group phenomena. This model provides a clear physical explanation for individual movement and group behavior through self-driving forces, interpersonal forces, and the forces between people and the environment. In recent years, researchers both domestically and internationally have improved upon it from several directions, for example: For example, Chinese patents with publication numbers CN119670392A and CN118052050A introduce dynamic environmental factors (such as fire products, temperature, and CO concentration) as velocity influence coefficients; another example is Chinese patent CN115758695A, which integrates behavioral heterogeneity and psychological emotion transmission mechanisms; and Chinese patent CN118965810A enhances the model's ability to characterize behavior in complex environments by introducing a field of vision and improving the exit selection strategy.

[0005] However, existing research still faces the following problems when applying social force models to the extremely confined and highly constrained environment of a civil aircraft cabin: First, existing social force models lack specific mechanical mechanisms and parameter adaptations for the civil aircraft cabin environment. The parameters of existing models are mostly derived from open or conventional indoor scenes (such as classrooms and office buildings), and have not been systematically calibrated for structures such as dense seats and narrow aisles in civil aircraft cabins, resulting in non-physical phenomena such as "mechanical deadlock", "lateral oscillation" and "path crossing seats" in the simulation. Secondly, existing social force models still do not adequately characterize the heterogeneity of passenger behavior. Although some studies have introduced personality, emotion or behavioral heterogeneity, they mostly remain at the theoretical level and lack coupling verification with empirical data. Third, the modeling of exit selection and area transition behavior in existing social force models is crude. Traditional models fix the target point as the exit and ignore the phased path planning behavior of passengers between the seating area, aisle area and exit area. Finally, existing social force models lack parameter calibration and model validation based on measured data: the credibility and engineering authority of current models in the cabin environment still face challenges, making it difficult to meet the stringent requirements for the accuracy of simulation results in airworthiness certification.

[0006] In summary, current research urgently needs to construct a social force model framework that accurately calibrates parameters, deeply integrates physical constraints and behavioral heterogeneity, and undergoes empirical verification for the high-constraint environment of civil aircraft cabins. This is not only a key means to improve the accuracy and engineering applicability of emergency evacuation simulations, but also a core issue in promoting its substantial supporting role in airworthiness certification. This invention addresses the aforementioned technological gap by proposing an improved social force model simulation method to systematically solve the adaptability and realism issues of existing technologies in cabin environments. Summary of the Invention

[0007] This invention aims to overcome the technical deficiencies of existing social force models in civil aircraft cabin emergency evacuation simulation applications, and provides a civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity. By optimizing the target point selection mechanism, obstacle action boundary determination method, and mechanical parameter system of the social force model, and by introducing intelligent decision-making modules and motion constraint mechanisms, this invention significantly improves the simulation accuracy of passenger motion states and behavioral decisions in confined cabin environments, making the simulation results more consistent with measured data, and providing a reliable computational tool for civil aircraft airworthiness certification and safety assessment.

[0008] To achieve the above objectives, the present invention provides the following technical solution: The aforementioned civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity includes the following steps: Step 1: Perform simulation initialization, including loading cabin geometry data and individual passenger attributes, setting target point sets for each area, and initializing timers; individual passenger attributes include position, desired speed, mass, and radius; Step 2: In each simulation step, based on the passenger's current position, a dynamic partition target point selection is performed to obtain the passenger's target point and desired direction of movement; the cabin space is divided into a seating area, a cabin aisle area, and an exit aisle area, and a target point set is preset for each area; the passenger's location is determined in real time based on the centroid coordinates, and a target point is selected from the target point set of the current location as the moving target. Step 3: In each simulation step, the effective obstacle boundary determination algorithm is used to determine the effective action edges that exert force on the passenger, the repulsive force between the passenger and the effective action edges is calculated, and the repulsive force between the passenger and the effective action edges is used as the basis for the calculation. Components perpendicular to the boundary direction As the repulsive force actually acting on the passenger; the effective obstacle boundary determination algorithm is divided into two stages. The first stage uses the passenger's centroid as the center and a preset radius to filter candidate boundaries. The second stage performs secondary filtering by the overlap relationship between the projection line segment of the passenger's centroid on the candidate boundary and the boundary itself. Step 4: In each simulation step, integrate the passengers' self-driving force. Repulsive force between passengers And the repulsive force that the boundary actually exerts on the passengers. The social force of each passenger is calculated, and passenger movement is controlled based on the social force; among which the passenger's self-driving force... Determine the direction of the force based on the desired direction of motion obtained in step 1; Step 5: During the simulation, check the waiting time of passengers at the boundary of the seating area. If it exceeds the set threshold, activate the priority passage mechanism to temporarily relieve the repulsive force of other passengers on that passenger and increase the passenger's expected speed. Step 6: During the simulation, when the passenger completes the area transition, the velocity direction is corrected based on the principle of energy conservation, so that the passenger's kinetic energy scalar value remains unchanged during the turn; Step 7: When the passenger's center of mass is within the cabin aisle area, monitor the spatial relationship between the passenger and the cabin aisle boundary in real time during each simulation step. When it is detected that the passenger's updated position exceeds the cabin aisle boundary, correct the passenger's position and adjust the velocity component. Step 8: Check if all passengers have reached the exit. If not, return to step 2 to continue the iteration.

[0009] In a further preferred embodiment, in step 2, the target point set in the seating area is located within the cabin aisle area, distributed in the cabin aisle area near the exit positions of each row of seats; the target point set in the cabin aisle area is located within the exit aisle area, distributed in the exit aisle area near the exit direction of the cabin aisle; and the target point set in the exit aisle area is located at the cabin exit position, distributed at each exit door.

[0010] Further optimization, step 2, is as follows: Step 2.1: Divide the cabin into regions and set the target point set for each region: Determine the boundary coordinates of each region based on the cabin layout data; Step 2.2: At each simulation time step, based on the passenger's centroid coordinates Determine its region; Step 2.3: If the passenger is in the seating area, calculate the Euclidean distance from the passenger to all target points in the seating area target point set, and select the closest point as the current target point; when the passenger's centroid enters the cabin aisle area, the target point automatically switches to the cabin aisle area target point set; when the passenger enters the exit aisle area, the target point switches to the exit aisle area target point set; during the selection process, if there are multiple equidistant minimum distance points, a random function is used to randomly select one from the minimum distance point set.

[0011] A further optimized solution, step 3, is as follows: Step 3.1: Using the passenger's centroid as the center and a preset radius as the effective range, traverse all obstacle boundaries and select boundary segments that are less than or equal to the preset radius from the passenger's centroid as a set of candidate effective edges; Step 3.2: For each candidate effective action edge, calculate the projection point of the passenger's centroid on the straight line where the candidate effective action edge is located, and extend the passenger radius along the direction of the straight line where the candidate effective action edge is located to determine the projection line segment; determine whether the projection line segment overlaps with the candidate effective action edge. If there is an overlap, the candidate effective action edge is determined to be an effective action edge; otherwise, it is discarded. Step 3.3: For the finally selected effective action edges, calculate the repulsive force in the social force model. And take its component perpendicular to the boundary direction. As a repulsive force that actually acts on passengers.

[0012] In a further optimized scheme, in step 4, the error between the total time of the evacuation experiment in the simulation and the actual evacuation time is less than 10%. The social force model parameters are corrected in combination with the motion state of passengers in the simulation, and the repulsion force strength coefficient between passengers is determined to be 200N, the repulsion force strength coefficient between passengers and the seat boundary is 500N, and the repulsion force strength coefficient between passengers and the cabin boundary is 2000N.

[0013] A further optimized solution, step 5, is as follows: Step 5.1: Set a waiting timer for each passenger to record the time they stay at the boundary of the seating area. When the waiting time of a passenger at the boundary of the seating area exceeds a preset threshold, it is determined that the passenger is in a stagnant state. Step 5.2: Among the passengers who are stationary on both sides of the aisle, calculate the priority score of each passenger and select the individual with the highest priority score as the priority passage object; Step 5.3: For the selected priority passage object, temporarily release the repulsive force of other passengers on the priority passage object, retain its self-driving force and boundary repulsive force, and at the same time increase the expected speed of the priority passage object; Step 5.4: Once the priority passage object enters the cabin aisle, the normal mechanical action mode is restored, and the waiting timers for all relevant passengers are reset.

[0014] A further optimized solution involves taking the priority score in step 5. ,in m For passenger quality, The speed expected by passengers.

[0015] In a further optimized approach, step 5.3 increases the expected speed of priority passage objects by 15%-25%.

[0016] A further optimized solution, the specific process of step 6 is as follows: When a passenger has completely moved from the seating area into the aisle area, the longitudinal velocity component is set to the absolute value of the lateral velocity just before entering the aisle area multiplied by the direction indicator variable, while the lateral velocity component is cleared to zero. When a passenger has completely moved from the cabin aisle area to the exit aisle area, their lateral velocity component is set to the absolute value of their longitudinal velocity at the moment of entry multiplied by the direction indicator variable, while the longitudinal velocity component is cleared to zero.

[0017] A further optimized solution, step 7, is as follows: Assume the cabin aisle area starts from the lower left corner. and the top right corner Defined, passenger radius is In the current simulation step, the passenger's updated location is: The following tests and corrections were performed: If the passenger's updated horizontal coordinates The corrected passenger horizontal coordinates The passenger's lateral velocity component is set to zero, while the longitudinal velocity remains constant. If the passenger's updated horizontal coordinates The corrected passenger horizontal coordinates The passenger's lateral velocity component is set to zero, while the longitudinal velocity remains constant.

[0018] Beneficial effects This invention proposes a civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity. Through a series of improvements, including dynamic partitioning of target points, determination of effective action edges, differentiated adjustment of social force model parameters, passenger waiting timeout decision-making, conservation and conversion of motion energy, and aisle boundary constraint protection, it systematically solves the technical problems inherent in traditional social force models in the extremely confined environment of a civil aircraft cabin. These problems include distortions in area transition behavior, abnormal repulsive forces at aisle entrances, mechanical deadlock and lateral fluctuations, permanent waiting and congestion, delayed acceleration during turns, and lateral oscillations within the aisle. This significantly improves the simulation accuracy of passenger motion states and behavioral decisions in confined spaces. Verified through 24 evacuation tests under different conditions, the average error between the simulation results and the actual evacuation time is consistently controlled within 10%, fully demonstrating the model's reliability and engineering applicability. This provides a scientific and reliable computational tool for civil aircraft emergency evacuation simulation analysis, airworthiness certification, and cabin safety assessment.

[0019] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0020] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of dynamic partitioning and target point division according to an embodiment of the present invention; Figure 2 The diagram illustrates the effective edge screening method according to an embodiment of the present invention; (a) an unreasonable edge obstructing passengers from entering the passageway, and (b) an effective edge screening method. Figure 3This is a schematic diagram of the repulsion force calculation method between the occupant and the obstacle boundary according to an embodiment of the present invention; (a) Schematic diagram of the repulsion force calculation method when the occupant projection is inside the obstacle boundary; (b) Schematic diagram of the repulsion force calculation problem in the traditional social force model when the occupant projection partially overlaps with the obstacle boundary. Figure 4 This is a schematic diagram of the obstruction of passenger movement on both sides of the aisle according to an embodiment of the present invention; (a) schematic diagram of the force on the movement of passengers on both sides of the aisle when they are far apart, and (b) schematic diagram of the force on the movement of passengers on both sides of the aisle when they are close together. Figure 5 This is a schematic diagram illustrating the state where a passenger is unable to enter the aisle due to force balance according to an embodiment of the present invention. Figure 6 This is a schematic diagram illustrating the movement of a passenger turning into the aisle according to an embodiment of the present invention; Figure 7 This is a schematic diagram of the simulation modeling process in an embodiment of the present invention; Figure 8 The following is a schematic diagram of the experimental site and simulation in an embodiment of the present invention; (a) the size and layout of the experimental site, and (b) a snapshot of a simulation experiment. Figure 9 To compare the simulated and measured evacuation times. Detailed Implementation

[0021] The technical solutions provided by the present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are only for explaining the present invention and are not intended to limit the scope of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0022] See Figures 1 to 7 This invention provides a civil aircraft emergency evacuation simulation method based on an improved social force model that considers passenger heterogeneity. By optimizing the target point selection mechanism, obstacle action boundary determination method, and mechanical parameter system of the social force model, and by introducing an intelligent decision-making module and motion constraint mechanism, the simulation accuracy of passenger motion state and behavior decision-making in a narrow cabin environment is significantly improved.

[0023] This embodiment details the various steps of the civil aircraft emergency evacuation simulation method provided by the present invention, such as... Figure 7 As shown, the complete simulation process in this embodiment includes the following steps: Step 1: Perform simulation initialization, including loading cabin geometry data and individual passenger attributes, setting target point sets for each area, and initializing timers; individual passenger attributes include position, desired speed, mass, radius, etc.

[0024] Step 2: In each simulation step, based on the passenger's current position, perform dynamic partition target point selection to obtain the passenger's target point and desired direction of movement.

[0025] Traditional social force models treat the cabin exit as a fixed target point, with the direction of the passenger's self-driving force pointing from their current location to the target point, and the direction of the line connecting the passenger to the selected exit as the desired velocity direction; however, simulations have revealed that, for example... Figure 1 As shown, on the one hand, due to the obstruction of seats, passengers in the three rows of seats on the left and right sides of the cabin cannot cross the seats, making it inappropriate to directly set the target point as the cabin exit. On the other hand, for passengers in the three middle rows, when they move into the aisle area, due to the relatively small initial lateral velocity and the continuous horizontal self-driving force, they may not accelerate sufficiently before crossing the aisle to enter the other side of the seating area, which is inconsistent with reality. This method of selecting a target point by using the cabin exit as a fixed target point causes passengers to have a desired direction towards the exit while still in the seating area, resulting in non-physical movements such as "straight-through-seat movement".

[0026] To address this issue, the present invention considers the passenger's evacuation sequence: first entering the cabin aisle, then the exit aisle, and finally exiting through the aircraft exit. Therefore, based on this actual evacuation path, the cabin is divided into three continuous functional areas: the seating area, the cabin aisle area, and the exit aisle area, as described below. Figure 1 The blue, orange, and green dashed boxes indicate the target point sets for each area: the seating area target point set (red dots), the cabin aisle area target point set (blue dots), and the exit aisle area target point set (green dots).

[0027] like Figure 1 As shown, the target points in the seating area are located within the cabin aisle area, specifically distributed in the aisle area near the exits of each row of seats. When passengers rise from their seats, these red dots serve as primary targets, guiding them from their seats into the aisle. The placement of these red dots in the aisle area, rather than the seating area, ensures that passengers must enter the aisle before proceeding, preventing the unreasonable behavior of directly crossing seats. The target points in the aisle area are located within the exit aisle area, specifically distributed in the exit aisle area near the exit direction. The blue dots serve as secondary targets, guiding passengers longitudinally along the cabin aisle to gradually approach the exit aisle area. The placement of these blue dots in the exit aisle area ensures that passengers must enter the exit aisle before reaching the exit, preventing the non-physical behavior of prematurely turning towards the exit. The target points in the exit aisle area are located at the cabin exits, specifically distributed at each exit door. The green dots serve as the final targets, guiding passengers directly from the exit aisle area to the exit door to complete the evacuation.

[0028] The specific steps are as follows: Step 2.1: Divide the area and set the target point set for each area: Based on the cabin layout data, determine the boundary coordinates of each area. Taking a twin-aisle aircraft as an example, the seating area is the area between the rows of seats, the aisle area is the longitudinal aisle, and the exit area is the area near the cabin door.

[0029] Step 2.2: Region Determination: At each simulation time step, based on the passenger centroid coordinates... Determine its region.

[0030] Step 2.3: Dynamic Target Point Selection: If the passenger is located in the seating area, calculate the Euclidean distance from the passenger to all target points in the seating area target point set, and select the closest point as the current target point. When the passenger's center of mass enters the cabin aisle area, the target point automatically switches to the cabin aisle area target point set; when the passenger enters the exit aisle area, the target point switches to the exit aisle area target point set. During the selection process, if multiple equidistant minimum distance points exist, a random function is used to randomly select one from the minimum distance point set to eliminate path uniformity. After obtaining the target point, the direction from the passenger to the target point is taken as the desired motion direction, i.e., the direction of the passenger's self-driving force.

[0031] This mechanism ensures that passengers' desired direction always points to the target point in the current area, avoiding unreasonable behavior of directly heading to the exit across areas, and making the passenger's movement trajectory conform to the actual path of "seat → aisle → exit".

[0032] Step 3: In each simulation step, the effective obstacle boundary determination algorithm is used to determine the effective action edge that exerts a force on the passenger, and the repulsion force between the passenger and the effective action edge is calculated.

[0033] In traditional social force models, the repulsive force of obstacles on passengers is calculated based solely on the single boundary closest to the passenger's center of mass, neglecting the situation where passengers may be affected by multiple obstacle boundaries simultaneously in a confined cabin environment. Therefore, this invention proposes a two-stage effective obstacle boundary determination method, as follows: Step 3.1: Distance Filtering: With the passenger's center of mass as the center and a preset radius To determine the effective range, all obstacle boundaries (including seat boundaries and cabin walls) are traversed, and those located at a distance less than or equal to the passenger's center of gravity are selected. The boundary segments are used as the set of candidate effective edges. This step ensures that only neighboring boundaries that have a significant impact on passengers are considered.

[0034] Step 3.2: Projection Overlap Filtering: After distance filtering in step 3.1, some candidate boundaries may still generate unreasonable repulsive forces, such as... Figure 2As shown in (a), the squares represent three adjacent seats. Assuming the passenger moves to the right, when the passenger approaches the aisle, the right boundary of the rightmost seat (red line segment) becomes an effective edge, creating a lateral repulsive force in the opposite direction of movement, hindering the passenger from entering the aisle. To solve this problem, this invention further performs a secondary screening based on the overlap between the passenger's projection onto the candidate boundary line and the candidate boundary itself. When the projection coincides with the candidate boundary (partially or completely), the candidate boundary is retained. Conversely, when the projection does not coincide with the candidate boundary, the candidate boundary is eliminated.

[0035] The specific approach is as follows: Figure 2 (b) Assuming AB is a candidate boundary after distance filtering in step 3.1, first calculate the projection point O′ of the passenger centroid O on the line AB; then, along the AB direction, with the passenger radius as half the width, determine the endpoints C′ and D′ of the projection line segment; determine whether C′ or D′ is on the line segment AB. If at least one point is on the line segment AB, then retain the candidate boundary as a valid edge; otherwise, discard it.

[0036] After the above two-stage screening, the final set of effective action edges is obtained.

[0037] Step 3.3: For the finally selected effective action edges, calculate the repulsive force in the social force model. And take its component perpendicular to the boundary direction. As a repulsive force actually acting on passengers (see...) Figure 3 This design avoids creating unnecessary lateral forces on passengers at the boundary, ensuring a smooth passage for them.

[0038] Traditional social force models, when calculating the repulsive force between passengers and the boundary, only use the shortest distance from the passenger's center of mass to the boundary line segment as input: in For passengers i Boundary segment w The repulsive force, The coefficient of repulsion between the passenger and the boundary. For passenger radius, The shortest distance from the passenger's center of mass to the boundary line segment. This is the coefficient representing the range of the repulsive force between the passenger and the boundary. Let be the unit normal vector pointing from the boundary line segment to the passenger.

[0039] Generally speaking, the direction of force is as follows: Figure 3 As shown in (a), however, there is a partial overlap between the passenger projection and the boundary line segment, such as... Figure 3As shown in (b), if the calculation is still performed according to the definition, the shortest distance from the passenger's center of mass to the boundary line segment is the distance from the passenger's center of mass to the right end of the boundary line segment. The direction of the repulsive force is not perpendicular to the boundary line segment, which will cause the boundary line segment to generate a lateral component force on the passenger, pushing the passenger away from the range of action of the boundary line segment, which is inconsistent with reality.

[0040] Therefore, the repulsive force between passengers and the boundary is taken from the social force model. Components perpendicular to the boundary direction As the repulsive force actually acting on passengers, the direction of the force is always perpendicular to the boundary to avoid generating unreasonable lateral components. This method effectively eliminates abnormal repulsive forces at the passageway entrance, ensuring smooth passenger access to the passageway.

[0041] Step 4: In each simulation step, integrate the passengers' self-driving force. Repulsive force between passengers And the repulsive force that the boundary actually exerts on the passengers. The social force of each passenger is calculated, and passenger movement is controlled based on this social force. For each passenger... i Its current simulation steps t social forces for: Among them, the passenger's self-driving force The direction of the force is determined based on the desired direction of motion obtained in step 1; while the repulsive force between passengers... According to the formula Sure, For passengers i Passengers j The repulsive force, The coefficient of repulsion between passengers. For passengers i With passengers j The sum of the radii, For passengers i With passengers j The distance between the centers of mass, This is the coefficient representing the range of repulsive force between passengers. For passengers j Pointing to passengers i The unit normal vector.

[0042] In traditional social force models, both the repulsion force coefficients between passengers and between passengers and boundaries use uniform preset parameters. However, in actual simulations, it was found that these uniform preset parameters lead to excessive repulsion forces between passengers at the aisle entrance in the cabin environment (see...). Figure 4), and the lateral fluctuations caused by the seat boundaries. Figure 4 As shown in (a), assume passengers 1 and 2 walk towards the aisle simultaneously. Taking passenger 1 as an example, he is subject to a self-driving force to the right and a repulsive force from passenger 2, which is moving to the left. As the lateral distance between them decreases, the repulsive force increases. When the repulsive force exceeds the self-driving force, passenger 1 decelerates and then moves to the left, unable to enter the aisle. Figure 4 (b)). For passengers in the seating and aisle areas, the repulsive force at the seat boundary causes significant lateral fluctuations in the occupants, deviating from their actual movement trajectory.

[0043] Therefore, this invention aims to ensure that the error between the total evacuation time in the simulation and the actual evacuation time is less than 10%. By adjusting the social force model parameters based on the passenger movement states in the simulation, the invention ultimately determines the repulsion force strength coefficient between passengers to be 200N, the repulsion force strength coefficient between passengers and the seat boundary to be 500N, and the repulsion force strength coefficient between passengers and the cabin boundary to be 2000N. By establishing differentiated social force model parameters, the invention more realistically reflects passenger behavior during actual evacuation.

[0044] Step 5: During the simulation, the waiting time of passengers at the boundary of the seating area is checked according to the timer. If it exceeds the set threshold, the priority passage mechanism is activated to temporarily relieve the repulsive force of other passengers on the passenger and increase the passenger's expected speed.

[0045] During actual simulation, we found that the following problems occur when passengers move near the cabin aisle: (1) Due to the narrow cabin aisle, the combined force of passengers in the aisle and those on the other side may cause passengers in the seating area to reach a force balance, thus preventing them from entering the aisle. For example Figure 5 The force analysis of the passengers showed that all three passengers eventually reached a force balance and stopped moving.

[0046] (2) When passengers are crowded in the aisle, it will cause passengers in the seating areas on both sides of the aisle to be unable to squeeze into the aisle for a long time, which is inconsistent with the actual situation.

[0047] To address the issues of passengers stalling at the aisle entrance due to force imbalance, and passenger congestion in the aisle causing delays for seated passengers, this invention proposes a priority passage mechanism based on a waiting time threshold. The specific implementation steps are as follows: Step 5.1: Stagnation state detection: A waiting timer is set for each passenger to record the time they spend at the boundary of the seating area. When the waiting time of a passenger at the boundary of the seating area (the position where they are about to enter the cabin aisle) exceeds a preset threshold (set to 3 seconds in this embodiment), it is determined that the passenger is in a state of stagnation due to force balance or aisle congestion.

[0048] Step 5.2: Priority Determination: Calculate the priority score for each passenger who is stationary on both sides of the aisle. ,in m For passenger quality, Based on passengers' expected speeds, the individual with the highest priority score is selected for priority passage. This rule is based on the assumption that passengers with greater mass and higher expected speeds are more likely to squeeze into the aisle.

[0049] Step 5.3: For the selected priority passage object, temporarily release the repulsive force of other passengers on the priority passage object, retain its self-driving force and boundary repulsive force, so that it can break through the force balance state and enter the cabin aisle. At the same time, increase the expected speed of the priority passage object by 15%-25% to ensure that it can quickly pass through the congestion area.

[0050] Step 5.4: Once the priority passage object enters the cabin aisle, the normal mechanical action mode is restored, and the waiting timers for all relevant passengers are reset.

[0051] This mechanism effectively solves the problems of passenger congestion caused by force balance in narrow spaces and passenger overcrowding in cabin aisles, making the simulation process more realistic.

[0052] Step 6: During the simulation, when the passenger completes the area transition, the velocity direction is corrected based on the principle of energy conservation, so that the passenger's kinetic energy scalar value remains unchanged during the turn.

[0053] During the simulation, it was found that passengers experienced a buildup phenomenon when turning, due to their inability to change speed and direction in time. For example... Figure 6 As shown, three passengers enter the aisle area sequentially under the combined action of self-driving force and interaction force. When passenger 1 is fully inside the aisle, he begins to accelerate along the aisle direction under the action of longitudinal self-driving force (along the aisle direction). However, due to the constraint of the seat boundary, the passenger's initial longitudinal velocity is close to zero, resulting in a relatively slow longitudinal acceleration process. At the same time, due to the blocking effect of this passenger, the movement of subsequent passengers is hindered, and they may even come to a standstill.

[0054] To solve the above technical problems, the present invention adopts a velocity correction method based on the principle of energy conservation: When a passenger has completely moved from the seating area into the aisle area, that is, in two adjacent simulation steps, the passenger's centroid coordinate moves from the seating area into the aisle area, then the longitudinal velocity component (i.e. the velocity component along the aisle direction) is set to the absolute value of the lateral velocity (i.e. the velocity component perpendicular to the aisle direction) before entering the aisle area multiplied by the direction indicator variable, and the lateral velocity component is cleared to zero. When a passenger has completely moved from the cabin aisle area to the exit aisle area, that is, in two adjacent simulation steps, the passenger's centroid coordinates move from the cabin aisle area to the exit aisle area. The passenger's lateral velocity component (i.e., the velocity component perpendicular to the cabin aisle direction) is set to the absolute value of the longitudinal velocity (i.e., along the cabin aisle direction) at the moment of entry multiplied by the direction indicator variable, and the longitudinal velocity component is cleared to zero.

[0055] This correction method ensures that the kinetic energy scalar value of passengers remains unchanged during area transitions, avoids motion discontinuity caused by sudden changes in velocity direction, and makes the steering process smooth and natural, which is more consistent with the actual situation.

[0056] Step 7: When the passenger's center of mass is within the cabin aisle area, monitor the spatial relationship between the passenger and the cabin aisle boundary in real time during each simulation step. When it is detected that the passenger's updated position exceeds the cabin aisle boundary, correct the passenger's position and adjust the velocity component.

[0057] During the simulation, for passengers whose center of mass is located within the cabin aisle area, stable passenger movement within the aisle is required to prevent abnormal positional regression due to seat boundary repulsion or inter-occupant interactions, such as being squeezed back into the seat area. To this end, this invention employs an aisle area boundary constraint mechanism. This mechanism, through dynamic position and velocity correction, ensures that passengers remain within a reasonable range after entering the cabin aisle area. The specific process is as follows: Assume the cabin aisle area starts from the lower left corner. and the top right corner Defined, passenger radius is In the current simulation step, the passenger's updated location is: The following tests and corrections were performed: If the passenger's updated horizontal coordinates This indicates that the passenger has exceeded the left boundary, so the corrected passenger horizontal coordinates are... The passenger's lateral velocity component is set to zero while the longitudinal velocity remains constant to avoid continuous oscillations caused by repulsive forces. If the passenger's updated horizontal coordinates This indicates that the passenger is outside the right boundary, so the corrected passenger horizontal coordinates are... The passenger's lateral velocity component is set to zero, while the longitudinal velocity remains constant.

[0058] This mechanism ensures that once passengers enter the aisle, their lateral movement is strictly confined within the aisle boundaries, preventing non-physical backlash or oscillations caused by repulsive forces. By synchronously zeroing the lateral velocity component, lateral fluctuations of occupants within the aisle are effectively eliminated, making the simulation process more consistent with real emergency evacuation scenarios.

[0059] Step 8: Check if all passengers have reached the exit. If not, return to step 2 to continue the iteration.

[0060] To verify the effectiveness of the model, this invention constructed an emergency evacuation experimental platform and simulated a 3-3 single-aisle cabin environment in classroom 331 of Chengzi Building at Northwestern Polytechnical University. Figure 8 As shown, the experimental site dimensions were: seat section length 875cm, aisle width 65cm, and front and rear door width 88cm. There were 42 participants (22 males and 22 females), aged 24-28, height 158-190cm, and shoulder width 34-48cm.

[0061] The design employs a full factorial approach, including: Seating layout: three rows on the left side of the aisle, three rows on the right side, and three rows on both sides; Exit status: Front door only, or both front and rear doors are open; Scenario modes: Fire, hard landing; Gender ratio: Male:Female = 7:3 or 3:7.

[0062] There are a total of 24 operating conditions.

[0063] The simulation parameters are set as follows: Expected speed: 1.6-1.8 m / s for men and 1.5-1.7 m / s for women (uniform distribution); Weight: Men 65-75kg, women 50-60kg; Radius: 0.22-0.24m for males, 0.18-0.20m for females; The parameters of the social force model are set according to the method in step 4.

[0064] Five simulations were performed for each set of working conditions, and the average evacuation time was compared with the actual measured time. Figure 9 The error analysis results for the first 12 working conditions are presented (the error distribution is similar for the latter 12 conditions with the male-to-female ratio reversed). (By...) Figure 9 As can be seen, the average error for all operating conditions is within 10%, indicating that the method of the present invention can accurately reproduce the real evacuation process.

[0065] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention without departing from the principles and spirit of the present invention.

Claims

1. A simulation method for emergency evacuation of civil aircraft based on an improved social force model, considering passenger heterogeneity, characterized in that: Includes the following steps: Step 1: Perform simulation initialization, including loading cabin geometry data and individual passenger attributes, setting target point sets for each area, and initializing timers; individual passenger attributes include position, desired speed, mass, and radius; Step 2: In each simulation step, based on the passenger's current position, a dynamic partition target point selection is performed to obtain the passenger's target point and desired direction of movement; the cabin space is divided into a seating area, a cabin aisle area, and an exit aisle area, and a target point set is preset for each area; the passenger's location is determined in real time based on the centroid coordinates, and a target point is selected from the target point set of the current location as the moving target. Step 3: In each simulation step, the effective obstacle boundary determination algorithm is used to determine the effective action edges that exert force on the passenger, the repulsive force between the passenger and the effective action edges is calculated, and the repulsive force between the passenger and the effective action edges is used as the basis for the calculation. Components perpendicular to the boundary direction As the repulsive force actually acting on the passenger; the effective obstacle boundary determination algorithm is divided into two stages. The first stage uses the passenger's centroid as the center and a preset radius to filter candidate boundaries. The second stage performs secondary filtering by the overlap relationship between the projection line segment of the passenger's centroid on the candidate boundary and the boundary itself. Step 4: In each simulation step, integrate the passengers' self-driving force. Repulsive force between passengers And the repulsive force that the boundary actually exerts on the passengers. The social force of each passenger is calculated, and passenger movement is controlled based on the social force; among which the passenger's self-driving force... Determine the direction of the force based on the desired direction of motion obtained in step 1; Step 5: During the simulation, check the waiting time of passengers at the boundary of the seating area. If it exceeds the set threshold, activate the priority passage mechanism to temporarily relieve the repulsive force of other passengers on that passenger and increase the passenger's expected speed. Step 6: During the simulation, when the passenger completes the area transition, the velocity direction is corrected based on the principle of energy conservation, so that the passenger's kinetic energy scalar value remains unchanged during the turn; Step 7: When the passenger's center of mass is within the cabin aisle area, monitor the spatial relationship between the passenger and the cabin aisle boundary in real time during each simulation step. When it is detected that the passenger's updated position exceeds the cabin aisle boundary, correct the passenger's position and adjust the velocity component. Step 8: Check if all passengers have reached the exit. If not, return to step 2 to continue the iteration.

2. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: In step 2, the target point set for the seating area is located within the cabin aisle area, distributed in the cabin aisle area near the exit positions of each row of seats; the target point set for the cabin aisle area is located within the exit aisle area, distributed in the exit aisle area near the exit direction of the cabin aisle; and the target point set for the exit aisle area is located at the cabin exit position, distributed at each exit door.

3. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: Step 2 is as follows: Step 2.1: Divide the cabin into regions and set the target point set for each region: Determine the boundary coordinates of each region based on the cabin layout data; Step 2.2: At each simulation time step, based on the passenger's centroid coordinates Determine its region; Step 2.3: If the passenger is located in the seating area, calculate the Euclidean distance from the passenger to all target points in the target point set of the seating area, and select the closest point as the current target point; When the passenger's center of mass enters the cabin aisle area, the target point automatically switches to the target point set of the cabin aisle area; when the passenger enters the exit aisle area, the target point switches to the target point set of the exit aisle area; during the selection process, if there are multiple equidistant minimum distance points, a random function is used to randomly select one from the minimum distance point set.

4. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: The specific process of step 3 is as follows: Step 3.1: Using the passenger's centroid as the center and a preset radius as the effective range, traverse all obstacle boundaries and select boundary segments that are less than or equal to the preset radius from the passenger's centroid as a set of candidate effective edges; Step 3.2: For each candidate effective action edge, calculate the projection point of the passenger's centroid on the straight line where the candidate effective action edge is located, and extend the passenger radius along the direction of the straight line where the candidate effective action edge is located to determine the projection line segment; determine whether the projection line segment overlaps with the candidate effective action edge. If there is an overlap, the candidate effective action edge is determined to be an effective action edge; otherwise, it is discarded. Step 3.3: For the finally selected effective action edges, calculate the repulsive force in the social force model. And take its component perpendicular to the boundary direction. As a repulsive force that actually acts on passengers.

5. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: In step 4, with the goal of ensuring that the error between the total time of the evacuation experiment in the simulation and the actual evacuation time is less than 10%, the parameters of the social force model are corrected in combination with the motion state of the passengers in the simulation. The repulsion force strength coefficient between passengers is determined to be 200N, the repulsion force strength coefficient between passengers and the seat boundary is 500N, and the repulsion force strength coefficient between passengers and the cabin boundary is 2000N.

6. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: Step 5 is as follows: Step 5.1: Set a waiting timer for each passenger to record the time they stay at the boundary of the seating area. When the waiting time of a passenger at the boundary of the seating area exceeds a preset threshold, it is determined that the passenger is in a stagnant state. Step 5.2: Among the passengers who are stationary on both sides of the aisle, calculate the priority score of each passenger and select the individual with the highest priority score as the priority passage object; Step 5.3: For the selected priority passage object, temporarily release the repulsive force of other passengers on the priority passage object, retain its self-driving force and boundary repulsive force, and at the same time increase the expected speed of the priority passage object; Step 5.4: Once the priority passage object enters the cabin aisle, the normal mechanical action mode is restored, and the waiting timers for all relevant passengers are reset.

7. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 6, characterized in that: Priority score in step 5 ,in m For passenger quality, The speed expected by passengers.

8. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 6, characterized in that: In step 5.3, the expected speed of priority passage objects is increased by 15%-25%.

9. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: The specific process of step 6 is as follows: When a passenger has completely moved from the seating area into the aisle area, the longitudinal velocity component is set to the absolute value of the lateral velocity just before entering the aisle area multiplied by the direction indicator variable, while the lateral velocity component is cleared to zero. When a passenger has completely moved from the cabin aisle area to the exit aisle area, their lateral velocity component is set to the absolute value of their longitudinal velocity at the moment of entry multiplied by the direction indicator variable, while the longitudinal velocity component is cleared to zero.

10. The civil aircraft emergency evacuation simulation method based on an improved social force model considering passenger heterogeneity as described in claim 1, characterized in that: The specific process of step 7 is as follows: Assume the cabin aisle area starts from the lower left corner. and the top right corner Defined, passenger radius is In the current simulation step, the passenger's updated location is: The following tests and corrections were performed: If the passenger's updated horizontal coordinates The corrected passenger horizontal coordinates The passenger's lateral velocity component is set to zero, while the longitudinal velocity remains constant. If the passenger's updated horizontal coordinates The corrected passenger horizontal coordinates The passenger's lateral velocity component is set to zero, while the longitudinal velocity remains constant.

Images