105 results about "Social force model" patented technology

The invention discloses a crowd exceptional event detecting method based on an LBP (Local Binary Pattern) weighted social force model. The method comprises the following steps of: computing a light stream vector of a sampled point based on a block-matching method; extracting dynamic textures of the sampled point by a time-space domain local binary pattern, and performing spectral analysis of Fourier transform; computing social force of the sampled point based on LBP weighted social force model; performing histogram quantization on the social force and performing classification on the video sequence based on a support vector machine to detect the exceptional behavior. Through combination of light stream and LBP frequency spectrum, the method provided by the invention innovatively calculates the social force to detect the exceptional behavior of the crowd, thus avoiding background modeling, prospect detection and detection and track of target, improving robustness and reduce calculated amount, and being particularly fit for occassions with large density of the crowd and complicated environment.

The invention provides a fire emergency evacuation simulation method. On the basis of establishing a real three-dimensional fire scene, researches are made around a vector-field-based global motion planning algorithm to navigate the escape of a plurality of individuals in the fire scene in real time. The method comprises the following steps of calculating a vector-field-based safety field in real time by virtue of the algorithm, constructing a fire-scene-based crowd behavior model in combination of a social force model-based local collision avoiding algorithm, and finally providing fire escape guidance for each escape individual in the scene according to the fire-scene-based crowd behavior model. According to the fire emergency evacuation simulation method, a theoretical basis is provided, the calculation accuracy can be ensured, and verification application is developed under various experimental conditions; the method has the characteristics of low calculation complexity, high simulation effect trueness and the like, and the real-time emergency evacuation of hundreds of escape individuals can be simulated during practical application.

The present invention discloses a computer simulation system for large-size crowd evacuation and a method therefor. The system comprises: an evacuation scene model establishing unit, used for extracting a feature of an evacuation scene, so as to obtain a three-dimensional model of the evacuation scene; an evacuation scene global path planning unit, used for performing a global path plan on the evacuation scene according to the three-dimensional model of the evacuation scene, and calculating global paths of all entrances and exits in the evacuation scene; a crowd activity generation unit, used for performing crowd grouping according to the feature of the evacuation scene, a relationship among individuals and individual position information, so as to obtain a plurality of groups; and a crowd simulation unit, used for screening a leader of each group and calculating a real-time speed of an individual according to a modified social force model. In a simulation process, a moving direction of each group is decided by the leader, and the motion speed of the individual is adjusted in real time, so as to maintain travel consistency of each group and achieve behavior simulation of evacuated crowd.

The invention relates to a simulation method and system for crowd evacuation based on a multi-bee colony algorithm. The method comprises the steps of setting evacuation scene parameter information, creating an evacuation scene model and a character model, and importing the character model into the evacuation scene model, wherein the evacuation scene model serves as the environment space of the crowd evacuation, and the character model serves as an evacuation crowd; extracting semantic information of the evacuation scene model, setting evacuation crowd parameter information under an evacuation scene, and conducting crowd initialization according to the evacuation crowd parameter information; dividing the evacuation crowd into groups, selecting a leading bee out from each group, adopting the multi-bee colony algorithm and adding auxiliary population to plan macroscopic paths, adopting a social force model to guide the movement of microscopic population, obtaining a final crowd evacuation path to conduct the simulation of the crowd evacuation. Through the adoption of the method, the simulation of the crowd evacuation in complex scenes can be really and effectively achieved, the evacuation time is shortened, the utilization rate of exits is increased, and the method can provided help for real evacuation rehearsals.

The invention discloses a group evacuation simulation system and method by combining an artificial bee colony and a social force model. The method comprises the following steps: acquiring an evacuation scene parameter to construct an evacuation scene three-dimensional model; finding all exits of the evacuation scene in the three-dimensional model; dividing to-be-evacuated crow in the evacuation scene into a plurality of groups according to the individual-to-individual relation and the position from the exit, screening the individual closest to the exit position in each group as a leader of each group; using each exit of the evacuation scene as the food source, and the leader as the leader bee in the group, thereby establishing one-to-one mapping with each parameter in the artificial bee colony; under the leading of the leader of each group, executing a parallel artificial bee colony algorithm to dynamically plan a path to move to the exit; and if the leader arriving the corresponding exit, waiting at the exit until the individual is inexistent in each group, and ending the crow evacuation simulation. Through the adoption of the method disclosed by the invention, the simulation efficiency and the channel efficiency in the public place are improved, and the assistance is offered for the real evacuation drill.

The invention discloses a simulation system and method of an evacuation crowd behavior based on grid-density-relation. The system comprises an evacuation scene model building unit, an evacuation scene global path planning unit, a crowd activity generation unit and a crowd simulation unit, wherein the evacuation scene model building unit is used for extracting characteristics of an evacuation scene according to structural information of the evacuation scene to obtain a three-dimensional model of the evacuation scene; the evacuation scene global path planning unit used for performing global path planning on the evacuation scene according to the three-dimensional model of the evacuation scene, and calculating global paths of all entrances in the evacuation scene; the crowd activity generation unit used for performing crowd grouping according to the quantity of all entrances, a relationship between individuals in the crowd and individual location information in the evacuation scene; and the crowd simulation unit used for calculating the real-time speed of individuals in a simulation process according to a social force model, and in the simulation process, correcting the moving speed of each individual in the same group in real time, so as to keep the advancing consistency of each group, and realize behavior simulation of the evacuation crowd.

The invention discloses a medium scale crowd abnormal behavior detection method based on causal network analysis. The interaction among individuals is expressed by using a cognitive concept and social force model, the association among individuals is estimated according to Granger test, the causal cognitive complex network is constructed according to the communication among individuals, and the crowd behavior is expressed and analyzed on the macro by analyzing the function parameter of the network while the individual microcosmic information is remained, thus the abnormal behavior of the medium scale crowd can be detected and judged effectively.

The present invention relates to a crowd grouping evacuation simulation method and system based on a half DBSCAN clustering algorithm. The method comprises: creating an evacuation scene model and a crowd individual model, and initializing evacuation scene information and crowd individual information, so as to obtain initialization parameters; according to the initialization parameters, dividing a crowd into multiple groups by using a half DBSCAN algorithm, so as to obtain a crowd grouping result; and using the crowd grouping result as an initial state of crowd motion, and making a local path plan by using a social force model, so as to obtain a motion path of each individual in an evacuation scene and perform crowd evacuation simulation. According to the method, the half DBSCAN clustering algorithm is combined with the social force model to complete crowd evacuation simulation in a complex scene, so that crowd evacuation simulation in the complex scene is really and effectively realized, the accuracy of crowd evacuation simulation is improved, and the grouping phenomenon in the evacuation process is demonstrated vividly.

The invention provides a street-crossing pedestrian trajectory prediction method and system based on a social force model, and a medium. The method comprises the following steps: a basic data acquisition step: selecting a pedestrian-vehicle mixed zebra crossing area for early investigation, shooting a pedestrian walking video of the area, and performing video processing and image processing on thepedestrian walking video to obtain pedestrian walking feature data and pedestrian-vehicle interaction scene data; and a model parameter calibration step: preprocessing the obtained pedestrian walkingcharacteristic data and the obtained pedestrian-vehicle interaction scene data, and calibrating parameters of a logistic regression model and a social force model. According to the method, the improved social force model is utilized, the individual difference of pedestrians is fully considered, the pedestrian track prediction of the crossing street is applied to the field of automatic driving vehicle decision making, the driving safety of the automatic driving vehicle in the pedestrian-vehicle mixed zebra crossing area can be improved, the delay rate of the vehicle is reduced, and the road traffic capacity is improved.

The invention discloses an unmanned aerial vehicle path planning method based on a transfer learning strategy deep Q-network (DQN). The method comprises the steps of: firstly carrying out the modelingand description of a dynamic environment where a UAV is located through adoption of a grid method, and building a state space model and an action space model of the UAV; secondly, initializing network parameters of a DQN and the current state of the unmanned aerial vehicle; training the DQN by adopting a return mechanism based on a social force model under the static environment model to obtain anetwork weight and an optimal action value; migrating the network weight and the optimal action value obtained by training in the static environment to the dynamic environment by using transfer learning, and continuing neural network training to obtain an action to be executed by the UAV; and finally, calculating the position of the unmanned aerial vehicle at the current moment to achieve path planning of the unmanned aerial vehicle in the dynamic environment. According to the method, the problems of the low DQN training convergence rate, the non-ideal path planning and the low success rate when the unmanned aerial vehicle performs path planning in a dynamic environment are effectively solved.

The invention discloses a metro station people emergency evacuation simulation method and system. The method comprises the steps of 1, building a metro station emergency evacuation simulation scene, and initializing people to be evacuated; 2, simulating perception for surroundings and mental and physiological status of the people to be evacuated; 3, according to the perception result of the surroundings and mental and physiological status of the people to be evacuated, building a social force model including the mental and physiological status of the people to be evacuated; 4, according to the social force model including the mental and physiological status of the people to be evacuated, building a motion path and motion speed decision model of the people to be evacuated; 5, performing simulating on the motion path and motion speed decision model of the people to be evacuated according to the time step method, and obtaining the motion status and the final simulation result of the people to be evacuated at each simulation moment. The emergency evacuation measure formulation and implementation reasonability and effectiveness can be improved.

The invention discloses a social force-based intersection crossing type construction area traffic flow model establishing method. The method comprises the following steps of: 1, establishing a signalized intersection construction area traffic flow model; 2, carrying out data acquisition on a signalized intersection construction area; 3, building a simulation platform to simulate a real construction area traffic flow; 4, calibrating a signalized intersection construction area traffic flow mode parameter by utilizing the data acquired in the step 2; and 5, verifying the signalized intersection construction area traffic flow model by utilizing the data acquired in the step 2 and the simulation result. According to the method, the method is capable of better reflecting the heterogeneity of traffic flows. Compared with traditional micro traffic flow models, the social force-based model is more suitable for traffic flows without lane division. According to the method, an existing social force model is improved firstly, then the new model is calibrated, and the calibrated model is utilized to simulate and analyze traffic flows.

The invention relates to a pedestrian evacuation simulation method and system adopting an artificial neural network-based Q-Learning algorithm. The method comprises the steps that to-be-evacuated pedestrians are divided into a plurality of groups according to initial related parameters, a leader is selected in each group, and the rest of the to-be-evacuated pedestrians are taken as followers; in each group, the leader preferentially learns and selects an optimal path obtained by performing global planning on evacuation paths through the artificial neural network-based Q-Learning algorithm, and the followers calculate a resultant force of the followers, the pedestrians in the group, the pedestrians between the groups, and environments according to a social force model, avoid barriers and follow the leader; and the to-be-evacuated pedestrians are all evacuated. According to the method and the system, global path planning is carried out in combination with the advantages of reinforcement learning and an artificial neural network; the deficiencies of pure reinforcement learning are made up for; a bottom layer matches with the social force model to guide a movement; and quick and effective way finding and relatively real evacuation of crowds can be realized.

The invention provides a panic crowd escape simulation method and is to simulate crowd escape characteristics under the panic condition truly. The method uses a social force model to carry out crowd modeling; according to the characteristics of the panic crowd, a new interaction force is invented, and a character model is introduced; and behavior characteristics of agents with different characters are simulated by adjusting the crowd simulation parameters. In the aspect of escape route planning technology, uniform grid technology is invented, so that Agent is allowed to sense the ambient environment more accurately, and route selection is various. In route planning, for the case that under a real scene, crowd visual fields are always blocked by barriers and the like, the partial visual field technology is invented to enable the crowd to only regard points in the visual field range as target points for navigation, thereby guaranteeing authenticity of the simulation.

The invention provides a fused automatic driving automobile street-crossing pedestrian trajectory prediction method and system, and the method comprises the steps: obtaining street-crossing pedestrianmotion state information, pedestrian individual feature information and vehicle motion state information according to a vehicle-mounted sensor fusion algorithm; calibrating parameters in the social force model according to the state data obtained by the vehicle-mounted sensor; training structural weight and offset parameters of an LSTM model according to the state data obtained by the vehicle-mounted sensor; utilizing the social force model and the LSTM model to respectively predict the motion trails of pedestrians crossing the street; importing the motion trail predicted by the model and theactual trail truth value of the pedestrian crossing the street into a Stacking fusion model, and training a structure weight; and outputting the optimal prediction track of the pedestrian crossing the street within a first preset time length in the future by utilizing the Stacking fusion model. According to the method, the Stacking algorithm is used for fusing the social force model and the LSTMmodel, the effect of reducing variance and deviation is achieved, and therefore it is guaranteed that the predicted trajectory is closer to the actual trajectory of the pedestrian.

The invention discloses a pedestrian trajectory prediction method which integrates a social force model and a Kalman filter. The Kalman filter is divided into two parts: a time update part and a measurement update part. Adaptive mutation particle swarm optimization algorithm is used to identify the parameters of social force model by setting fitness function. The predicted pedestrian trajectory issimulated in the step 2, the pedestrian position value at the next time is calculated according to the Kalman time update formula in the step 1, and the prior estimate value shown in the descriptionis finally obtained; According to the updated formula of Kalman measurement, the measurement value Zk of pedestrian's current position is calculated, and the optimal estimation value is obtained by combining the prior estimation value shown in the description. The error threshold psi is set to judge the error between the predicted position of the social force model and the optimal estimated value,and the error is corrected to complete the trajectory prediction. It can predict the trajectory accurately when pedestrians take the initiative to avoid turning and walking in a straight line and effectively reduce the error between the actual trajectory and the predicted trajectory so as to meet the required prediction requirements.

The invention discloses a crowd evacuation simulation method and system based on an evacuation path set. The method includes the steps: extracting characteristics of an evacuation scene to obtain a two-dimensional model of the evacuation scene; acquiring all available evacuation paths based on the two-dimensional model of the evacuation scene and a social force model fusing obstacle avoiding strategies to obtain the evacuation path set; performing discretization operation on all paths in the evacuation path set, and calculating crowd distribution degree and block degree information in real time to obtain evaluation values of each individual in crown for the all paths in the evacuation path set; building a roulette according to the evaluation values of each individual in the crown for the all paths in the evacuation path set, and selecting an optimum evacuation path by the aid of pseudo random selection strategies based on the roulette and stable factors to achieve simulation of crowd evacuation processes. According to the method, evacuation strategies are changed in real time by a simulation individual according to current evacuation environments, and safety evacuation plans of sudden events can be scientifically guided.

The invention discloses a bus passenger boarding and alighting behavior simulation method based on a social force model, which comprises the steps of shooting a bus passenger flow video, acquiring a boarding and alighting behavior characteristic, performing square net division on a scene, performing simulation setup on an arrival rule of a bus, setting generation, distribution and movement rules of passengers, introducing a neighbor domain searching algorithm to express a passenger stress characteristic by using a vectorial force, introducing a visual field blocking judgment rule to correct a social acting force, and introducing a collision avoidance judgment rule to judge collision possibility in advance. The method introduces a pedestrian flow simulation concept, treats each bus passenger as a micro individual, achieves calculation of passenger boarding and alighting time and evaluation of boarding and alighting efficiency via quantification and modeling of micro behavior parameters of the bus passengers, and provides assistance for effective decision-making assistance bases and optimization of an urban bus trip.

The invention discloses an instance-based learning multi-Agent cooperation crowd evacuation simulation method and device. Sufficient path information samples are acquired through real videos and simulation, and original path information is extracted. Scene information is extracted according to the scene, and region segmentation is performed on the scene information so as to construct a scene information base. The knowledge path is segmented in a way of being corresponding to the scene region segmentation situation, and an original local path information base is constructed. Pedestrians are matched with the scene information database, the groups are divided according to the similarity and each group is guided by the guide Agent. The guide Agent is matched with the local path information base, and an approximation function is established and the local path is ensured to be optimal. Scene local path information exchange is realized by using the information interaction cooperation function in the multi-Agent system. The isomorphic instance is constructed by using interaction information through analogical learning and path selection is optimized. The path information drives the guide Agent, and the low-level motion of the Agent of the same group is driven by the social force model until all the pedestrians in the scene are evacuated to the safe location and the process is ended.

The invention relates to an abnormal crowd behavior detection algorithm based on optical flow computation, comprising the following steps: first, extracting the characteristics of crowd movement using an optical flow particulate vector field; then, calculating the interaction force between optical flow particulates based on a social force model; and finally, making a histogram entropy analysis of the interaction force to realize behavior judgment. Through the algorithm, an abnormal crowd behavior can be effectively judged and located. By calculating the interaction force in a social force model corresponding to a scene video and making a histogram entropy analysis, rapid and reliable abnormal crowd behavior detection is realized. The algorithm can be widely used in the field of video monitoring.

The invention relates to a crowd evacuation simulation method and system based on video data driving. The method comprises the following steps that: obtaining the video record of an appointed area, and utilizing a KLT (Kanade-Lucas-Tomasi) tracking algorithm to extract the initial coordinate, the movement speed, the exit position and the movement path of crow evacuation from the video record; and according to the initial coordinate and the movement speed, initializing an individual to be evacuated, carrying out interleaving access on a relative position between the initial coordinate of each individual and an exit and the movement path to obtain a shortest path from the initial coordinate of each individual to the exit, adopting a social force model to calculate the speed and the direction of the individual in real time on the basis of the above initial coordinate, movement speed and shortest path in real time so as to realize the simulation of a crowd evacuation behavior. The initial setting and the evacuation leading path of simulated evacuation is from a real radio, and the simulated evacuation can be more truly realized so as to provide help for evacuation maneuvering.

The invention discloses an imitation learning social navigation method based on a feature map fused with pedestrian information. According to the imitation learning social navigation method, a robot is guided to imitate the movement habits of experts by introducing an imitation learning method, the navigation method conforming to the social specifications is planned, the planning efficiency is improved, the problem that the robot is locked is solved, and the robot is helped to be better integrated into a man-machine co-integration environment. The method comprises the steps of: acquiring a time sequence motion state of a pedestrian through pedestrian detection and tracking in a sequence RGB image and three-dimensional point cloud alignment; then, in combination with two-dimensional laser data and a social force model, acquiring a local feature map marked with pedestrian dynamic information; and finally, establishing a deep network with the local feature map, a current speed of the robot and a relative position of a target as input and a robot control instruction as output, training with expert teaching data as supervision, and acquiring a navigation strategy conforming to social specifications.

The invention discloses a building disaster evacuation simulation method based on an evacuation object social force evolution game model, which comprises the following steps: establishing an evacuation object social force model; Based on game theory, dividing the behavior of people at the exit into two kinds: cooperative behavior and competitive behavior, then, establishimg an evolutionary game model based on evacuation object social force model, and reproducing the cooperative and competitive behaviors between people by cellular automata, obtaining and saving the resultant force of social forces received by all evacuated objects according to the received related parameters, obtaining a time step by utilizing the maximum resultant force value therein, and determining the expected speed ofthe evacuated objects in the time period and the expected position after the time period; When the evacuation object confirms the advance, updating the position of the evacuation object and the speedthereof; When not advancing, only updating the speed of the evacuated object is ed; finishing the process when the evacuation object completes the evacuation in the current building; Otherwise, executing the process in a loop.

The invention discloses a social-force-model-based monitoring system, which comprises a data acquisition device and a preprocessing device, wherein the data acquisition device is used for shooting and identifying the two-dimensional plane position information of pedestrians and fixed objects; and the preprocessing device is used for computing behavioral characteristic values of the pedestrians and rejecting abnormal data. The monitoring system further comprises an analysis device, an alarming device and a human-computer interaction device, wherein the analysis device is used for calibrating contact repelling force strength and contact friction force strength and comparing the contact repelling force strength and the contact friction force strength with preset contact repelling force strength and preset contact friction force strength to determine whether to output an alarming signal or not; the alarming device is used for giving an alarm according to the alarming signal; and the human-computer interaction device is used for interaction between the monitoring system and a worker. The system combines video monitoring and computer mode identification effectively based on a social force model, monitors collective behaviors with simple operations, and maximally solves social security problems caused by abnormal collective behaviors.

The present invention discloses a video driven group behavior evacuation simulation method and apparatus. The method comprises: receiving a real scene database, tracking a video crowd in the databaseto obtain individual motion trajectories, calculating the dissimilarity between individuals to obtain small groups, and calculating the dissimilarity between small groups as subsequent initializationinformation; generating a social force model based on the simulation scene and the character model, and initializing the scene and the individual information of the crowd; and performing local path planning, and calculating a motion path of each individual in the each scene to reach the selected exit. According to the technical scheme disclosed by the present invention, the group phenomenon can beeffectively and realistically simulated, real reproduction of crowd evacuation under unexpected events in real scenes can be implemented, and the technical scheme becomes an important tool for predicting the evacuation time and evacuation status of populations distributed in different population densities and groups on the occasion that the population trajectory cannot be measured in the case ofthe emergency situation.

The invention discloses a method for detecting and predicting the passenger flow in an urban rail transit passage. The method includes the steps that the passenger flow S1 at the entrance and exit of the passage is collected, the relation between the outflow amount of the passenger flow at the two ends of the passage and density of the passenger flow in the passage is built on the basis of a social force model, an empirical value S2 between pedestrian density and the outflow amount is obtained through calculation, a black box model of the to-be-detected passage is built, and passenger flow density information in the passage at the current moment and passenger flow density information S3 at the next moment are calculated by means of the empirical value and the passenger flow at the entrance and exit of the passage at the current moment. The invention further discloses a system for detecting and predicting the passenger flow in the urban rail transit passage. By means of the scheme, safety hidden hazards caused due to excessively large passenger flow density can be effectively avoided. The detection and prediction problem of the passenger flow of an urban rail transit large-curvature passage can be effectively solved, and innovativeness, practicality and scientific research value are high.

The invention provides a cat swarm algorithm-based crowd evacuation simulation method and system, a medium and equipment. The method comprises the following steps: setting evacuation scene parameter information, creating an evacuation scene model and a person model, and importing the person model into the evacuation scene model; semantic information of the evacuation scene model is extracted, evacuation crowd parameter information is set in a corresponding evacuation scene, and crowd initialization is carried out according to the evacuation crowd parameter information; macroscopic path planning is performed by adopting a cat swarm algorithm, microscopic crowd movement guidance is performed by adopting a social force model, a final crowd evacuation path is obtained, and crowd evacuation simulation is performed; the advantages of a macroscopic evacuation simulation method and a microcosmic evacuation simulation method are combined, the defects of the two methods in large-scale crowd evacuation simulation are overcome, crowd evacuation can be quickly and effectively carried out, evacuation time is shortened, the utilization rate of an exit is increased, and help can be provided for real evacuation drilling.

The invention discloses a hospital acoustic environment simulation system based on a social force model. The system comprises a behavioral model and an acoustic model, wherein the behavioral model comprises a sound production spacing module, a duration module, a sound production position module and a sound production intensity module; the acoustic model comprises an outdoor noise module and an indoor sound field module; the system can defines each room in a hospital as an interactive Agent; when a certain sound source in the Agent is activated, firstly the sound production spacing module, the duration module, the sound production position module and the sound production intensity module in the behavioral model constitute basic attributes of a sound source sound production behavior; then sound transmission is simulated through the outdoor noise module and the indoor sound field module in the acoustic model, so that the initial sound pressure level in a receiving point position is obtained through computation; then the final sound pressure level in the receiving point position is determined by comparing the initial sound pressure level with a background noise sound pressure level and judging the quantity of sound production sound sources at identical time; finally, the plane distribution situation of an indoor sound field is computed through an interpolation method according to plane positions of all receiving points in the Agent and the final sound pressure levels of the receiving points.

The invention discloses a tabu search bee colony algorithm based crowd evacuation simulating method and system. The tabu search bee colony algorithm based crowd evacuation simulating method comprisesthe following steps: setting evacuation scenario parameter information and person model information, establishing an evacuation scenario model and a person model, and leading the person model into theevacuation scenario model; setting initial parameters of a bee colony algorithm based on tabu search; performing macro evacuation simulation: initializing initial parameters, performing macro path planning by adopting the bee colony algorithm based on tabu search, searching an initial path from an initial point to a target point, wherein the path consists of a plurality of points; performing micro evacuation simulation, wherein on the basis of the initial path, persons inevitably collide when number of evacuated crowd is greater than a set threshold value, micro crowd movement guidance is performed by adopting a social force model, collision-free evacuation movement is generated, and a final crowd evacuation path is obtained for performing crowd evacuation simulation; guiding out and storing evacuated person number, time for evacuation and the final crowd evacuation path for movement display and contrastive analysis used for crowd evacuation.

The invention relates to the field of crowd evacuation simulation, in particular to a crowd evacuation simulation method based on scene semantic information under a complex indoor structure, which comprises the following steps: firstly, carrying out target detection to obtain a pedestrian i coordinate and an initial speed vi, and initializing a simulation scene; determining a danger source, establishing a danger field for direct transmission of the danger signal by taking the danger source as a starting point, and establishing an information field for indirect transmission; dividing the pedestrian into different states according to the difference of the information obtained from the dangerous field or the information field, and correcting the expected speed of the pedestrian; adopting a pre-ERRT algorithm model to establish pedestrian evacuation path navigation, optimizing the evacuation direction in the pedestrian evacuation navigation path, and obtaining the final path navigation direction; introducing the affinity degree between pedestrians and pedestrian density information, correcting a social force model under a complex indoor structure, calculating the pedestrian acceleration, and performing weighted summation to obtain the direction and size of final speed variation. According to the method, the scene semantic information of a complex indoor structure can be obtained inreal time, and the crowd evacuation efficiency is improved.