Search

CN-121997552-A - Data-driven uphill pedestrian biped gait and social force coupling simulation method

CN121997552ACN 121997552 ACN121997552 ACN 121997552ACN-121997552-A

Abstract

The invention provides a data-driven uphill pedestrian biped gait and social force coupling simulation method, which belongs to the technical field of pedestrian simulation and comprises the steps of step 1, collecting gait dynamics data, step 2, constructing a biped social force coupling model, step 21, simulating initialization, step 22, step triggering judgment, step 23, social force and gait parameter calculation, step 24, foot drop point planning and collision response, step 25, mass center position updating, and step 26, simulating circulation and termination. The invention aims to realize high-precision and high-fidelity simulation of pedestrian movement from microscopic gait to macroscopic flow rate under different gradient conditions through a complete technical chain from experiment to simulation.

Inventors

  • XIE WEI
  • Xuan Chenrui
  • TIAN BINGWEI
  • DI BAOFENG

Assignees

  • 四川大学

Dates

Publication Date
20260508
Application Date
20251224

Claims (10)

  1. 1. A data-driven uphill pedestrian biped gait and social force coupling simulation method is characterized by comprising the following steps: step 1, acquiring gait dynamics data of pedestrians under different uphill slopes; Step 2, constructing a biped social force coupling model based on gait dynamics data, wherein a simulation flow of the biped social force coupling model comprises the following steps: step 21, setting geometric parameters of a simulation scene, initializing pedestrian groups, setting total simulation duration and simulation time step length, and setting initial gait parameters and initial bipedal states for each pedestrian based on gait dynamics data; Step 22, traversing all pedestrians in each simulation time step, judging whether a walking condition is met based on a bipedal gait dynamics principle, if so, executing step 23 and step 24 to update gait parameters and bipedal states, otherwise, directly transferring to step 25 and continuing to use the historical gait parameters and the historical bipedal states of the current pedestrians; step 23, calculating social force and key gait parameters of pedestrians, wherein the key gait parameters are used as the latest gait parameters of the current time step; Step 24, planning a target foothold of the swing foot according to the latest gait parameters, detecting whether the target foothold collides with an obstacle or other pedestrians, triggering a deceleration strategy if the target foothold collides, re-planning a foothold until the collision is released, and then updating the two-foot state to obtain the latest two-foot state of the current time step; Step 25, if the pedestrian meets the stepping condition, calculating the advancing direction displacement and the lateral displacement of the pedestrian according to the latest gait parameter and the latest bipedal state, and if the pedestrian does not meet the stepping condition, calculating the advancing direction displacement and the lateral displacement of the pedestrian according to the historical gait parameter and the historical bipedal state of the current pedestrian, wherein the advancing direction displacement and the lateral displacement of the pedestrian are calculated according to the initial gait parameter and the initial bipedal state of the current pedestrian in initial simulation; Updating the centroid coordinate position of the pedestrian based on the forward direction displacement and the lateral displacement; And 26, judging whether the simulation time reaches the preset total simulation duration, if so, ending the simulation, otherwise, updating the simulation time, recording gait parameters and the two-foot state of the current time step as historical data, and returning to the step 22.
  2. 2. The data-driven uphill pedestrian biped gait and social force coupling simulation method of claim 1, further comprising the steps of 3, calibrating key parameters of a biped social force coupling model by adopting an optimization algorithm, and verifying microscopic gait reproducibility and macroscopic dynamics of the model in a multi-gradient scene.
  3. 3. The data-driven uphill pedestrian biped gait and social force coupling simulation method according to claim 1 is characterized in that step 1 of collecting gait dynamics data of pedestrians under different uphill gradients specifically comprises the steps of carrying out constant-speed walking experiments under preset different uphill gradients and collecting data in real time, wherein the gait dynamics data at least comprise the step length, the step time, the step frequency, the step width, the support phase duration, the swing phase duration, the double support phase duration, the ground reaction force statistic value and the step direction angle of each step of walking.
  4. 4. The data-driven uphill pedestrian bipedal gait and social force coupling simulation method of claim 1, wherein the setting of the initial gait parameters and the initial bipedal state for each pedestrian based on the gait dynamics data in step 21 specifically comprises: Setting initial gait parameters, namely setting an initial stepping direction angle along the ascending direction of a ramp, and referring to the average advancing direction statistical value of pedestrians under the same gradient, wherein the initial step length adopts the statistical average value of the step length of the pedestrians under the same gradient; The initial two-foot state setting comprises that the initial supporting foot position and the initial position of the mass center of the pedestrian are vertically corresponding in the ramp plane and accord with the natural position relation between the supporting foot and the mass center when a human body stands, the initial swinging foot position is set on the symmetrical side of the initial supporting foot based on the initial step width and the initial stepping direction angle, whether the initial two-foot position collides with an obstacle or other pedestrians is detected after setting, and if the initial two-foot position collides, the relative position of the initial supporting foot and the swinging foot is readjusted until the collision is released.
  5. 5. The data-driven uphill pedestrian biped gait and social force coupling simulation method according to claim 1, wherein the core basis for determining whether the step condition is satisfied based on the biped gait dynamics principle in step 22 is the relative relationship between the projection position of the pedestrian centroid on the ramp plane and the front end point position of the current supporting leg.
  6. 6. The data-driven uphill pedestrian bipedal gait and social force coupling simulation method of claim 1, wherein calculating the social force of the pedestrian in step 23 comprises calculating the self-driving force of the pedestrian, the interaction force between the pedestrian and other individuals and the interaction force between the pedestrian and the wall, wherein an adaptive relaxation time model is introduced during the calculation of the self-driving force of the pedestrian, and the adaptive relaxation time model comprehensively reflects the biomechanical load caused by uphill walking, and specifically comprises: correlating and quantifying the forward tilting degree of the trunk through a sine function of the gradient angle; Correlating and quantifying the non-linear increase in friction demand by an exponential form of the tangent function of the slope angle; Correlating the ratio of the actual weight of the pedestrian to the reference weight and quantifying the influence of the weight on the exercise response; and correlating and quantifying fatigue effects through the accumulated motion height of the pedestrian active walking stage, wherein the accumulated motion height is reset to a base line value when the instantaneous speed of the pedestrian is lower than a minimum motion speed threshold value, and the reference weight and the base line value are both determined based on experimental data statistics.
  7. 7. The data-driven uphill pedestrian biped gait and social force coupling simulation method according to claim 1 is characterized in that key gait parameters in step 23 comprise a stepping direction angle, a stepping length and a stepping width, wherein the stepping length is calculated by calculating logic that the stepping length is equal to the product of the current instantaneous speed of a pedestrian and a single step time, the single step time is influenced by acceleration and adaptive relaxation time, specifically, acceleration influencing factors and relaxation time influencing factors are overlapped based on experimental calibrated average single step time, the acceleration influencing factors are calculated by the ratio of an acceleration reference value to the current instantaneous acceleration of the pedestrian, the relaxation time influencing factors are calculated by the ratio of the current adaptive relaxation time to a relaxation time reference value, and the acceleration reference value and the relaxation time reference value are determined based on experimental data statistics.
  8. 8. The data-driven uphill pedestrian biped gait and social force coupling simulation method of claim 7, wherein step width setting logic in step 23 is that step width obeys standard normal distribution, the average value of the standard normal distribution is statistical average value of pedestrian step width under the same gradient, standard deviation is statistical standard deviation of pedestrian step width under the same gradient, specific value of step width is that the product of the standard deviation and standard normal distribution random variable is added and subtracted from the average value, and the standard normal distribution random variable obeys the distribution of 0 and 1 variance to simulate individual difference of different pedestrian step widths.
  9. 9. The method for simulating the coupling of the biped gait and the social force of the pedestrian on the uphill according to claim 1, wherein the step 24 is characterized in that the triggering deceleration strategy specifically comprises the steps of reducing the current instantaneous speed of the pedestrian to a preset safe speed interval, wherein the safe speed interval is set based on a low-speed adjustment state statistical value when the pedestrian walks on the uphill, shortening the step length by reducing the instantaneous speed, further adjusting the position of the target foothold of the swinging foot until no collision between the target foothold and an obstacle or other pedestrians is detected, and updating the biped state specifically comprises the steps of updating the original supporting foot to a new swinging foot, updating the target foothold of the swinging foot without collision after the planning to a new supporting foot position, and ensuring that the updated biped state has no overlapping in a ramp plane and accords with the natural gesture of biped walking.
  10. 10. The data-driven uphill pedestrian biped gait and social force coupling simulation method is characterized in that key parameters of a biped social force coupling model in the step 3 comprise a posture adaptation coefficient, a friction sensitivity coefficient, a quality influence coefficient and a fatigue sensitivity coefficient in a self-adaptive relaxation time model, an acceleration sensitivity coefficient and a relaxation time sensitivity coefficient in step length calculation, an optimization target of an optimization algorithm is to minimize the sum of squares of errors of a model simulation step length and an experiment acquisition step length, the range of values of all key parameters is limited to be between 0 and 1, microscopic gait reproducibility and macroscopic dynamics characteristics of a verification model specifically comprise consistency of microscopic level comparison simulation step length and experiment step length, a density-speed relationship obtained by macroscopic level comparison simulation is consistent with a basic graph of a hill pedestrian population dynamics experiment, and a verification scene covers all preset uphill slopes.

Description

Data-driven uphill pedestrian biped gait and social force coupling simulation method Technical Field The invention relates to the technical field of pedestrian simulation, in particular to a data-driven uphill pedestrian bipedal gait and social force coupling simulation method. Background The urban construction is promoted to increase the number of the vertical buildings, the slopes (such as subway stations, malls and stadium slopes) become common facilities in public spaces, the pedestrians are influenced by gravity to generate backward tilting moment when ascending, the pedestrians need to maintain balance by adjusting the gait (such as trunk forward tilting and step length adjustment), the problems of increasing physiological load and increasing out-of-control risks in the scenes exist, and even group safety accidents (such as pear Tai court slope stepping events) can be possibly induced, so that accurate ascending pedestrian simulation technology is needed to support safety assessment. Defects of the traditional technology: modeling limitation is that a mainstream social force model simplifies pedestrians into particles, key biomechanical characteristics such as step length, step frequency, trunk gesture and the like during ascending cannot be represented, and deviation from real walking behavior is large; The traditional model is mostly designed in a level ground scene, the nonlinear influence of the gradient on the expected speed and the motion adjustment capability of pedestrians is not considered, and systematic deviation exists between macroscopic speed, flow and other output and a real ramp scene; The parameter reliability is low, the model parameter is set by depending on experience, driving and verification of real biomechanical experimental data are lacked, biomechanical rationality is insufficient, and self-adaptive behavior of pedestrians for coping with gradient cannot be reproduced. Disclosure of Invention The invention provides a data-driven uphill pedestrian biped gait and social force coupling simulation method, which aims to realize high-precision and high-fidelity simulation of pedestrian movement from microscopic gait to macroscopic flow rate under different gradient conditions through a complete technology chain from experiment to simulation. In order to achieve the above purpose, the invention adopts the following technical scheme: a data-driven uphill pedestrian biped gait and social force coupling simulation method comprises the following steps: step 1, acquiring gait dynamics data of pedestrians under different uphill slopes; Step 2, constructing a biped social force coupling model based on gait dynamics data, wherein a simulation flow of the biped social force coupling model comprises the following steps: step 21, setting geometric parameters of a simulation scene, initializing pedestrian groups, setting total simulation duration and simulation time step length, and setting initial gait parameters and initial bipedal states for each pedestrian based on gait dynamics data; Step 22, traversing all pedestrians in each simulation time step, judging whether a walking condition is met based on a bipedal gait dynamics principle, if so, executing step 23 and step 24 to update gait parameters and bipedal states, otherwise, directly transferring to step 25 and continuing to use the historical gait parameters and the historical bipedal states of the current pedestrians; step 23, calculating social force and key gait parameters of pedestrians, wherein the key gait parameters are used as the latest gait parameters of the current time step; Step 24, planning a target foothold of the swing foot according to the latest gait parameters, detecting whether the target foothold collides with an obstacle or other pedestrians, triggering a deceleration strategy if the target foothold collides, re-planning a foothold until the collision is released, and then updating the two-foot state to obtain the latest two-foot state of the current time step; Step 25, if the pedestrian meets the stepping condition, calculating the advancing direction displacement and the lateral displacement of the pedestrian according to the latest gait parameter and the latest bipedal state, and if the pedestrian does not meet the stepping condition, calculating the advancing direction displacement and the lateral displacement of the pedestrian according to the historical gait parameter and the historical bipedal state of the current pedestrian, wherein the advancing direction displacement and the lateral displacement of the pedestrian are calculated according to the initial gait parameter and the initial bipedal state of the current pedestrian in initial simulation; Updating the centroid coordinate position of the pedestrian based on the forward direction displacement and the lateral displacement; And 26, judging whether the simulation time reaches the preset total simulation duration, if so, ending the simulation, otherwise, up