Shopping center traffic space crowd trajectory prediction method and system based on artificial jellyfish search algorithm

Abstract

The invention provides a shopping center traffic space crowd trajectory prediction method and system based on an artificial jellyfish search algorithm. The method comprises the following steps: 1, collecting shopping center scene parameter information and crowd behavior data, and associating the shopping center scene parameter information with the crowd behavior data; step 2, determining decision variables including position coordinates and speed components of individual crowds, and taking a channel attraction parameter, a node steering probability parameter, a stair / escalator / elevator selection probability parameter and a speed adjustment parameter as influence factors of the ocean current direction; 3, initializing a jellyfish population, taking a population individual as a jellyfish individual, and setting a population scale and an initialization decision variable according to a certain rule; and 4, performing algorithm iteration by using an artificial jellyfish search algorithm to obtain final crowd trajectory prediction. According to the method, accurate prediction of the crowd trajectory in the traffic space of the shopping center is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of crowd trajectory analysis and prediction, and in particular relates to a method and system for predicting crowd trajectory in a shopping center traffic space based on an artificial jellyfish search algorithm. Background Art

[0002] In the operation and management of modern shopping malls, accurate prediction of crowd trajectories is crucial to improving operational efficiency, optimizing customer experience and ensuring safety. As shopping malls continue to expand in scale and become more complex in function, the flow trajectories of people in them are affected by the interaction of multiple factors. Among them, the layout of traffic space (including passages, nodes, stairs, elevators, their geometric features, signs, etc.) plays a key role in guiding the flow and behavior of people.

[0003] Traditional trajectory prediction methods have too many limitations. This simulation based on natural biological behavior enables the algorithm to search and optimize in a way that is more in line with natural laws when solving certain complex problems, providing new ideas and methods for problem solving. It can reduce production costs while achieving the best prediction results.

[0004] The artificial Jellyfish Search (JS) optimizer is a new optimization algorithm proposed by Zhou Ruisheng in 2020. It has the characteristics of strong optimization ability and fast convergence speed. The principle of the algorithm is to simulate the search behavior of jellyfish, which involves their following the ocean current, the movement in the jellyfish group (active movement and passive movement), the time control mechanism of switching between these movements, and the process of their convergence into jellyfish clusters. In the algorithm, there are the above two assumptions: jellyfish either follow the ocean current or move within the group, and the "time control mechanism" controls the conversion between these types of movement. Jellyfish move in the ocean to find food, and they are more likely to be attracted to places with a larger food supply. In the process of trajectory prediction, the layout of the traffic space plays a key guiding role in the flow and behavior of the crowd. In the existing crowd trajectory prediction, how to accurately reflect the flexibility and adaptability of the crowd trajectory under the guidance of the traffic space, there is still a lack of effective solutions. Summary of the invention

[0005] The purpose of the present invention is to solve the problems in the prior art and meet the actual needs. A method and system for predicting the trajectory of people in the traffic space of a shopping center based on an artificial jellyfish search algorithm is proposed. The present invention uses an artificial jellyfish search algorithm to accurately reflect the guidance of traffic layout on crowd behavior and improve prediction efficiency.

[0006] The present invention is implemented by the following technical scheme. The present invention proposes a method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm. The method comprises the following steps:

[0007] Step 1: Collect shopping mall scene parameter information and crowd behavior data, establish a special data fusion platform, associate shopping mall scene parameter information and crowd behavior data, so that crowd behavior data can correspond to specific scene parameters;

[0008] Step 2: Treat the individuals in the crowd as jellyfish individuals, determine the position coordinates (x, y) as the basic decision variables, and the individual's horizontal and vertical velocity components are (v x ,v y ), the initial value of the speed is determined according to the statistical average of the collected crowd behavior data; other decision variables are defined as channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、p l and p m and the speed adjustment parameter v adjust ;

[0009] Step 3: Initialize the jellyfish population and determine the population size N; the initial position coordinates (x, y) of the jellyfish individuals are randomly generated within the shopping mall, and the velocity component (v x ,v y ) is randomly initialized according to the speed range of the collected data, and other decision variables are initialized within the value range;

[0010] Step 4: Use the artificial jellyfish search algorithm to iterate the algorithm to obtain the final crowd trajectory prediction;

[0011] The prediction trajectory is specifically as follows: taking various factors that guide the flow of people in the shopping center traffic space as ocean currents, respectively calculating the ocean current direction vectors of channel i, node j, stair s, escalator l and straight ladder m; for each jellyfish individual n, calculating the attraction of the ocean current to individual n and the repulsive force between individual jellyfish Determine the time step Δt based on the characteristics of the data acquisition device; update the speed and position of the individual people according to the calculated attraction, repulsion and inertia weight, and obtain new speed and position coordinates, so as to simulate the movement trajectory of the people in the traffic space;

[0012] When updating the position, it is necessary to check whether the crowd individuals exceed the boundary range of the shopping mall, that is, the channel attraction parameters, node turning probability parameters, stairs, escalators, and elevator selection parameters are all within the interval [0,1]. If they exceed the boundary, the parameters need to be adjusted back to a reasonable range to ensure that the predicted trajectory meets the actual space constraints;

[0013] The iteration termination condition is set, that is, the position change of all jellyfish individuals within M consecutive iterations is less than a certain threshold ξ. When the termination condition is met, the iteration is stopped and the output is the jellyfish trajectory in continuous time, that is, the predicted crowd trajectory.

[0014] Furthermore, a data fusion platform that associates shopping mall scene parameter information with crowd behavior data is established. The data fusion platform adopts a three-layer architecture, including a data collection layer, an edge computing layer, and a cloud fusion analysis layer;

[0015] Among them, the data collection layer is composed of Wi-Fi positioning devices and smart cameras, which are responsible for collecting scene parameter information and crowd behavior data of the shopping mall; the edge computing layer is distributed in various areas of the shopping mall, close to the data collection equipment, and is composed of multiple edge computing nodes. These nodes can perform real-time pre-processing and preliminary analysis of the collected data, filter out redundant information to obtain the required traffic space information and crowd behavior information; the cloud fusion analysis layer is located in the data center, which is used for deep integration, analysis and storage of data from the edge computing layer.

[0016] Furthermore, the channel attraction parameter The node's turning probability parameter p jk , the selection probability parameter p of stairs, escalators, and elevators s 、p l and p m The value range of is [0-1], and each parameter is affected by different factors;

[0017] Among them, the channel attraction parameter of each channel i is The factors affected: channel width i , signage i The weights of each factor are α and β respectively, and α+β=1. Among them, max(widht) and max(signage) are the maximum values ​​of all channel widths and signage perfection, respectively;

[0018] The turning probability parameter p for each possible turning direction k of each node j jk ,

[0019] For each staircase s, escalator l and elevator m, select the probability parameter ps 、p l and p m , are all affected by the convenience of location s / l / m Attraction of surrounding shops s / l / m Influence.

[0020] Furthermore, various factors that guide the flow of people in the shopping center traffic space are regarded as ocean currents. Depending on the influencing factors, the ocean currents have different direction vectors;

[0021] Among them, for channel i, according to its width and the vector corresponding to the identification factor And the corresponding weight coefficients α, β, the channel current direction vector is

[0022] For node j, according to the turning probability parameter p of each turning direction k, jk , and the environmental vector associated with each turn direction The ocean current direction vector of node j is Turn probability parameter p jk As a weight, it reflects the probability of different turning directions in the crowd's choice;

[0023] For stairs s, escalators l and elevators m, the vectors corresponding to their locations and surrounding store attraction factors are: And the weight coefficients δ and ε, where δ + ε = 1, calculate the ocean current direction vector

[0024] Furthermore, the attraction to each jellyfish is Includes channel attraction Node attractiveness Stairs / elevator or escalator / staircase attraction Target attraction

[0025] Among them, the channel attraction is: Among them, k 1 is the channel attraction coefficient, is the channel attraction weight parameter of the channel where the individual is located, is the current direction vector of the channel;

[0026] The node attraction is: Among them, k 2 is the node attraction coefficient, j represents the nodes around individual n, p nj is the probability that individual n is affected by node j, is the current direction vector of the node;

[0027] Stairs / elevators or escalators / ladders have the following attractions: Among them, k 3 is the attraction coefficient of stairs / elevators or escalators / elevators, s represents the stairs / elevators or escalators / elevators around individual n, is the probability that an individual chooses stairs / elevator or escalator / elevator, is the current direction vector of the stairs / elevator or escalator / staircase;

[0028] The target attraction is: Among them, k 4 is the target attraction coefficient, is the vector pointing from the individual position to the target position, is the distance from the individual to the target location;

[0029] The total attraction is:

[0030] Furthermore, for any two jellyfish individuals n and q, the repulsive force between the jellyfish individuals is For any two jellyfish individuals n and q, the repulsive force is proportional to the distance d between them. nq Inversely proportional, the repulsive factor is r, then in is the vector pointing from individual n to individual q. All repulsive forces interacting with individual n are superimposed, and the total repulsive force is

[0031] Furthermore, during the prediction process, the speed and movement position of the jellyfish in continuous time are updated based on the time step Δt;

[0032] The speed update formula is: in, is the new velocity of individual n at time t+Δt, is the old velocity of individual n at time t, w is the inertia weight, a is the attraction factor, r is the repulsion factor, is the attraction experienced by individual n at time t, is the repulsive force on individual n at time t;

[0033] The position update formula is: in, is the new position of individual n at time t+Δt, is the old position of individual n at time t.

[0034] The present invention also proposes a shopping center traffic space crowd trajectory prediction system based on an artificial jellyfish search algorithm, the system comprising:

[0035] Data collection module: collects shopping center scene parameter information and crowd behavior data, establishes a special data fusion platform, associates shopping center scene parameter information and crowd behavior data, and enables crowd behavior data to correspond to specific scene parameters;

[0036] Decision variable setting module: The individuals in the crowd are regarded as jellyfish individuals, and the position coordinates (x, y) are determined as the basic decision variables. The velocity components of the individuals in the horizontal and vertical directions are (v x ,v y ), the initial value of the speed is determined according to the statistical average of the collected crowd behavior data; other decision variables are defined as channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、p l and p m and the speed adjustment parameter v adjust ;

[0037] Initialization module: Initialize the jellyfish population and determine the population size N; the initial position coordinates (x, y) of the jellyfish individuals are randomly generated within the shopping mall, and the velocity component (v x ,v y ) is randomly initialized according to the speed range of the collected data, and other decision variables are initialized within the value range;

[0038] Trajectory prediction module: uses the artificial jellyfish search algorithm to iterate the algorithm to obtain the final crowd trajectory prediction;

[0039] The prediction trajectory is specifically as follows: taking various factors that guide the flow of people in the shopping center traffic space as ocean currents, respectively calculating the ocean current direction vectors of channel i, node j, stair s, escalator l and straight ladder m; for each jellyfish individual n, calculating the attraction of the ocean current to individual n and the repulsive force between individual jellyfish Determine the time step Δt based on the characteristics of the data acquisition device; update the speed and position of the individual people according to the calculated attraction, repulsion and inertia weight, and obtain new speed and position coordinates, so as to simulate the movement trajectory of the people in the traffic space;

[0040] When updating the position, it is necessary to check whether the crowd individuals exceed the boundary range of the shopping mall, that is, the channel attraction parameters, node turning probability parameters, stairs, escalators, and elevator selection parameters are all within the interval [0,1]. If they exceed the boundary, the parameters need to be adjusted back to a reasonable range to ensure that the predicted trajectory meets the actual space constraints;

[0041] The iteration termination condition is set, that is, the position change of all jellyfish individuals within M consecutive iterations is less than a certain threshold ξ. When the termination condition is met, the iteration is stopped and the output is the jellyfish trajectory in continuous time, that is, the predicted crowd trajectory.

[0042] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm are implemented.

[0043] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm.

[0044] Beneficial effects of the present invention:

[0045] The present invention combines an innovative data fusion platform with an artificial jellyfish search algorithm to achieve accurate prediction of crowd trajectories in shopping center traffic spaces. The data fusion platform effectively integrates multi-source data to improve data quality and availability; the artificial jellyfish search algorithm fully considers traffic space guidance and the interaction between individuals in the crowd, and can accurately simulate changes in crowd behavior, providing a decision-making basis for shopping center operations and management, such as optimizing store layout, adjusting traffic facility operation strategies, and formulating precision marketing plans, thereby improving operational efficiency, optimizing customer experience, and ensuring safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0047] Figure 1 It is a flow chart of the method for predicting crowd trajectory in shopping center traffic space based on artificial jellyfish search algorithm according to the present invention;

[0048] Figure 2 It is a block diagram of the shopping center traffic space crowd trajectory prediction system based on the artificial jellyfish search algorithm described in the present invention. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] The present invention proposes a method and system for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm. The method comprises the following steps: step one: collecting shopping center scene parameter information and crowd behavior data, and associating the shopping center scene parameter information and crowd behavior data; step two: determining decision variables, including position coordinates and velocity components of individual crowd members, and taking channel attraction parameters, node turning probability parameters, stair / escalator / elevator selection probability parameters and speed adjustment parameters as influencing factors of the direction of ocean currents; step three: initializing a jellyfish population, taking individual crowd members as jellyfish individuals, setting the population size and initializing decision variables according to certain rules; step four: taking factors guiding the flow of crowds in the shopping center traffic space as ocean currents, calculating the ocean current direction vector and the attraction and repulsion forces on each jellyfish individual, updating the individual speed and position according to the inertia weight, attraction factor and repulsion factor, checking boundary conditions, and stopping iteration when the position changes of all jellyfish individuals within M consecutive iterations are less than a certain threshold value ξ, and finally obtaining a predicted trajectory generated by the artificial jellyfish search algorithm.

[0051] Specifically, combined Figure 1-Figure 2 The present invention proposes a method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm, the method comprising the following steps:

[0052] Step 1: Collect shopping mall scene parameter information and crowd behavior data, establish a special data fusion platform, and associate the shopping mall scene parameter information and crowd behavior data so that the crowd behavior data can correspond to specific scene parameters; the shopping mall scene parameter information is the traffic space information within the building, including channel information, stairs, escalators and elevators information and node information; the crowd behavior information includes pedestrian behavior information and trajectory feature data.

[0053] Step 2: Treat the individuals in the crowd as jellyfish individuals, determine the position coordinates (x, y) as the basic decision variables, and the individual's horizontal and vertical velocity components are (v x ,v y ), the initial value of the speed is determined according to the statistical average of the collected crowd behavior data; other decision variables are defined as channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、pl and p m and the speed adjustment parameter v adjust ;

[0054] Step 3: Initialize the jellyfish population and determine the population size N; the initial position coordinates (x, y) of the jellyfish individuals are randomly generated within the shopping mall, and the velocity component (v x ,v y ) is randomly initialized according to the speed range of the collected data, and other decision variables are initialized within the value range;

[0055] Step 4: Use the artificial jellyfish search algorithm to iterate the algorithm to obtain the final crowd trajectory prediction;

[0056] The prediction trajectory is specifically as follows: various factors guiding the flow of people in the traffic space of the shopping center are regarded as ocean currents, and the ocean current direction vectors of channel i, node j, stairs s, escalators l and straight ladders m are calculated respectively; for each jellyfish individual n, considering the guiding force of the traffic space on the flow of people and the mutual repulsion between the individuals of the crowd, and at the same time, in order to avoid excessive aggregation, the magnitude of the repulsive force is inversely proportional to the distance between the individuals of the crowd, so the attraction of the ocean current to the individual n is calculated. and the repulsive force between individual jellyfish The time step Δt is determined based on the characteristics of the data acquisition device; the speed and position of the crowd individuals are updated according to the calculated attraction, repulsion and inertia weight. The speed update formula comprehensively considers the influence of the previous speed, attraction and repulsion, so that the crowd individuals gradually adjust their direction and speed during the movement. The position update is calculated based on the updated speed to obtain new speed and position coordinates, thereby simulating the movement trajectory of the crowd in the traffic space;

[0057] When updating the position, it is necessary to check whether the crowd individuals exceed the boundary range of the shopping mall, that is, the channel attraction parameters, node turning probability parameters, stairs, escalators, and elevator selection parameters are all within the interval [0,1]. If they exceed the boundary, the parameters need to be adjusted back to a reasonable range to ensure that the predicted trajectory meets the actual space constraints;

[0058] The iteration termination condition is set, that is, the position change of all jellyfish individuals within M consecutive iterations is less than a certain threshold ξ. When the termination condition is met, the iteration is stopped and the output is the jellyfish trajectory in continuous time, that is, the predicted crowd trajectory.

[0059] Establish a data fusion platform that associates shopping mall scene parameter information with crowd behavior data. The data fusion platform adopts a three-layer architecture, including data collection layer, edge computing layer and cloud fusion analysis layer;

[0060] Among them, the data collection layer is composed of Wi-Fi positioning devices and smart cameras, which are responsible for collecting scene parameter information and crowd behavior data of the shopping mall; the edge computing layer is distributed in various areas of the shopping mall, close to the data collection equipment, and is composed of multiple edge computing nodes. These nodes can perform real-time pre-processing and preliminary analysis of the collected data, filter out redundant information to obtain the required traffic space information and crowd behavior information; the cloud fusion analysis layer is located in the data center, which is used for deep integration, analysis and storage of data from the edge computing layer.

[0061] The edge computing nodes and data collection equipment are connected through a wireless network to ensure that data can be quickly and stably transmitted to the edge computing layer. At the same time, edge computing nodes share and collaborate through the internal network. Nodes in different regions can collaborate with each other to analyze cross-regional crowd behavior. The edge computing layer and the cloud fusion analysis layer communicate through a high-speed broadband network and upload the processed data to the cloud.

[0062] A Bayesian network model is constructed to integrate traffic space information and human trajectory data information. The status (smooth, congested, maintenance, etc.) and attributes (such as width, length, carrying capacity, etc.) of traffic space elements (channels, nodes, stairs, elevators, escalators, etc.) are used as nodes, and pedestrian behaviors (walking, staying, choosing a path, etc.) and trajectory characteristics (speed, direction, stay time, etc.) are used as other nodes. The conditional probability relationship between nodes is determined based on prior knowledge and data statistical analysis. For example, if it is known that a channel is narrow and often congested, then the probability of pedestrians choosing this channel may be low, while the probability of choosing other alternative paths (such as adjacent wide channels) will increase. This relationship can be represented by a conditional probability table.

[0063] Shopping mall scene parameter information includes: channel information; information about stairs, escalators and elevators; and node information. Crowd behavior information includes: pedestrian behavior information and trajectory feature data. Among them, channel information: obtain the width data of all channels in the shopping mall and the identification information in the channels; information about stairs, escalators and elevators: for stairs, collect information such as their location, number, staircase width, surrounding space layout, etc.; for escalators and elevators, record their location, carrying capacity, speed, stop floors, entrance and exit locations, and surrounding space layout data; node information: determine the location and attributes of nodes such as intersections, confluences, and branch points between channels, including the number of channels connected by nodes, store types around nodes, and rest facilities. Pedestrian behavior information: walking, staying, and choosing a path; trajectory feature data: speed, direction, and dwell time.

[0064] Channel attraction parameters The node's turning probability parameter p jk , the selection probability parameter p of stairs, escalators, and elevatorss 、p l and p m The value range of is [0-1], and each parameter is affected by different factors;

[0065] Among them, the channel attraction parameter of each channel i is The factors affected: channel width i , signage i The weights of each factor are α and β respectively, and α+β=1. Among them, max(width) and max(signage) are the maximum values ​​of all channel widths and signage perfection, respectively;

[0066] The turning probability parameter p for each possible turning direction k of each node j jk ,

[0067] For each staircase s, escalator l and elevator m, select the probability parameter p s 、p l and p m , are all affected by the convenience of location s / l / m Attraction of surrounding shops s / l / m Influence.

[0068] The various factors that guide the flow of people in the shopping center traffic space are regarded as ocean currents. Depending on the influencing factors, the ocean currents have different direction vectors;

[0069] Among them, for channel i, according to its width and the vector corresponding to the identification factor And the corresponding weight coefficients α, β, the channel current direction vector is

[0070] For node j, according to the turning probability parameter p of each turning direction k, jk , and the environmental vector associated with each turn direction This vector comprehensively considers factors such as the width of the turning direction connecting channel, signs, and the attractiveness of surrounding shops. The ocean current direction vector of node j is Turn probability parameter p jk As a weight, it reflects the probability of different turning directions in the crowd's choice;

[0071] For stairs s, escalators l and elevators m, the vectors corresponding to their locations and surrounding store attraction factors are: And the weight coefficients δ and ε, where δ + ε = 1, calculate the ocean current direction vector

[0072] Each jellyfish individual is attracted by Includes channel attraction Node attractiveness Stairs / elevator or escalator / staircase attraction Target attractiveness

[0073] Among them, the channel attraction is: Among them, k 1 is the channel attraction coefficient, is the channel attraction weight parameter of the channel where the individual is located, is the current direction vector of the channel;

[0074] The node attraction is: Among them, k 2 is the node attraction coefficient, j represents the nodes around individual n, p nj is the probability that individual n is affected by node j, is the current direction vector of the node;

[0075] Stairs / elevators or escalators / ladders have the following attractions: Among them, k 3 is the attraction coefficient of stairs / elevators or escalators / elevators, s represents the stairs / elevators or escalators / elevators around individual n, is the probability that an individual chooses stairs / elevator or escalator / elevator, is the current direction vector of the stairs / elevator or escalator / staircase;

[0076] The target attraction is: Among them, k 4 is the target attraction coefficient, is the vector pointing from the individual position to the target position, is the distance from the individual to the target location;

[0077] The total attraction is:

[0078] For any two jellyfish individuals n and q, the repulsive force between the jellyfish individuals is For any two jellyfish individuals n and q, the repulsive force is proportional to the distance d between them. nq Inversely proportional, the repulsive factor is r, then in is the vector pointing from individual n to individual q. All repulsive forces interacting with individual n are superimposed, and the total repulsive force is

[0079] During the prediction process, the speed and movement position of the jellyfish in continuous time are updated based on the time step Δt;

[0080] The speed update formula is: in, is the new velocity of individual n at time t+Δt, is the old velocity of individual n at time t, w is the inertia weight, a is the attraction factor, r is the repulsion factor, is the attraction experienced by individual n at time t, is the repulsive force on individual n at time t;

[0081] The position update formula is: in, is the new position of individual n at time t+Δt, is the old position of individual n at time t.

[0082] Using the fitness function, F(v i )=w 1 E distance +w 2 E node +w 3 E flow .

[0083] Among them, w 1 、w 2 、w 3 is the weight coefficient, E distance is the trajectory distance error, E node is the node passing error, E flow is the flow error, which evaluates the degree of match between the predicted trajectory and the actual crowd behavior. The trajectory distance error is obtained by calculating the average Euclidean distance between the predicted trajectory points and the actual trajectory points; the node passing error statistics predicts the difference between the time when the crowd passes through the key node and the actual passing time; the flow error compares the difference between the predicted flow of different traffic space elements and the actual flow of people.

[0084] The algorithm terminates when the number of iterations reaches the maximum number of iterations or an individual with a fitness value less than the pre-set accuracy threshold appears in the population. The continuous number of iterations M in the iteration termination condition ranges from [50, 200], and the accuracy threshold ξ ranges from [0.01, 0.1]. The specific value can be adjusted and determined according to factors such as the scale of the shopping center and the complexity of crowd behavior. For large shopping centers with complex and changeable crowd behavior, the M value can be appropriately increased to between 150 and 200, while the ξ value can be reduced to between 0.01 and 0.05 to ensure the accuracy of the prediction results; for small shopping centers or shopping centers with relatively simple crowd behavior, the M value is between 50 and 100, and the ξ value is between 0.05 and 0.1, which improves the calculation efficiency while ensuring a certain prediction accuracy.

[0085] The present invention also proposes a shopping center traffic space crowd trajectory prediction system based on an artificial jellyfish search algorithm, the system comprising:

[0086] Data collection module: collects shopping mall scene parameter information and crowd behavior data, establishes a special data fusion platform, associates shopping mall scene parameter information and crowd behavior data, so that the crowd behavior data can correspond to specific scene parameters; the shopping mall scene parameter information is the traffic space information within the building, including channel information, stairs, escalators and elevators information and node information; the crowd behavior information includes pedestrian behavior information and trajectory feature data.

[0087] Decision variable setting module: The individuals in the crowd are regarded as jellyfish individuals, and the position coordinates (x, y) are determined as the basic decision variables. The velocity components of the individuals in the horizontal and vertical directions are (v x ,v y ), the initial value of the speed is determined according to the statistical average of the collected crowd behavior data; other decision variables are defined as channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、p l and p m and the speed adjustment parameter v adjust ;

[0088] Initialization module: Initialize the jellyfish population and determine the population size N; the initial position coordinates (x, y) of the jellyfish individuals are randomly generated within the shopping mall, and the velocity component (v x ,v y ) is randomly initialized according to the speed range of the collected data, and other decision variables are initialized within the value range;

[0089] Trajectory prediction module: uses the artificial jellyfish search algorithm to iterate the algorithm to obtain the final crowd trajectory prediction;

[0090] The prediction trajectory is specifically as follows: various factors guiding the flow of people in the traffic space of the shopping center are regarded as ocean currents, and the ocean current direction vectors of channel i, node j, stairs s, escalators l and straight ladders m are calculated respectively; for each jellyfish individual n, considering the guiding force of the traffic space on the flow of people and the mutual repulsion between the individuals of the crowd, and at the same time, in order to avoid excessive aggregation, the magnitude of the repulsive force is inversely proportional to the distance between the individuals of the crowd, so the attraction of the ocean current to the individual n is calculated. and the repulsive force between individual jellyfish The time step Δt is determined based on the characteristics of the data acquisition device; the speed and position of the crowd individuals are updated according to the calculated attraction, repulsion and inertia weight. The speed update formula comprehensively considers the influence of the previous speed, attraction and repulsion, so that the crowd individuals gradually adjust their direction and speed during the movement. The position update is calculated based on the updated speed to obtain new speed and position coordinates, thereby simulating the movement trajectory of the crowd in the traffic space;

[0091] When updating the position, it is necessary to check whether the crowd individuals exceed the boundary range of the shopping mall, that is, the channel attraction parameters, node turning probability parameters, stairs, escalators, and elevator selection parameters are all within the interval [0,1]. If they exceed the boundary, the parameters need to be adjusted back to a reasonable range to ensure that the predicted trajectory meets the actual space constraints;

[0092] The iteration termination condition is set, that is, the position change of all jellyfish individuals within M consecutive iterations is less than a certain threshold ξ. When the termination condition is met, the iteration is stopped and the output is the jellyfish trajectory in continuous time, that is, the predicted crowd trajectory.

[0093] Example

[0094] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0095] The artificial Jellyfish Search (JS) optimizer is a new optimization algorithm proposed by Zhou Ruisheng in 2020. It has the characteristics of strong optimization ability and fast convergence speed. The algorithm is based on simulating the search behavior of jellyfish, which involves their movement (active and passive movement) in the jellyfish group following the ocean current, the time control mechanism of switching between these movements, and the process of their aggregation into jellyfish clusters.

[0096] Combination Figure 1-2 The present invention proposes a method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm, the method comprising:

[0097] Step 1: Collect data and establish a dedicated data fusion platform to associate shopping center scene parameter information with crowd behavior data so that crowd behavior data can correspond to specific scene parameters;

[0098] Consider a large shopping mall, covering an area of ​​100,000 square meters, with a multi-story building structure, including many shops, restaurants, entertainment facilities, etc. During peak hours, the flow of people is large and the crowd behavior is complex and diverse.

[0099] Through the data collection layer of the data fusion platform, Wi-Fi positioning devices and smart cameras fully cover all areas of the shopping mall. The Wi-Fi positioning device monitors the signals of customer mobile devices connected to the network in real time to obtain customer location information. The Wi-Fi positioning accuracy is 3-5 meters; the smart camera collects images at a frame rate of 25 frames per second, covering key locations such as passages, stairs, elevators, store entrances, and rest areas.

[0100] The collected scene parameter information includes: the width of the channel varies between 2-5 meters, the sign information includes direction signs, store signs, etc., and the completeness of the signs varies in different areas;

[0101] The locations of stairs and elevators are reasonably distributed, the width of the stairwell is about 2-3 meters, the elevator has a carrying capacity of 10-15 people, the speed is moderate, the stop floors cover all floors, the entrance and exit locations are coordinated with the surrounding passages and store layout, and the surrounding space layout takes into account the flow and evacuation needs of personnel;

[0102] The locations of nodes (channel intersections, confluence points, branch points) are clear, the number of connecting channels ranges from 2 to 4, the types of shops around the nodes are rich and diverse, and the rest facilities are reasonably configured.

[0103] In terms of crowd behavior data, pedestrians’ walking paths, stop locations and time, path selection conditions, and trajectory feature data with speeds between 0.5-1.5 m / s, frequent changes in direction, and stay times ranging from a few seconds to several minutes are obtained.

[0104] Step 2: Treat the individuals in the crowd as jellyfish individuals, and randomly initialize the position coordinates (x, y) to cover the entire shopping center area. According to the statistical average of the collected crowd behavior data, determine the individual's horizontal and vertical velocity components (v x ,v y ) initial values, for example, the horizontal velocity is 0.8 m / s and the vertical velocity is 0.3 m / s.

[0105] Define the channel attraction parameter as The turning probability parameter of the node is p jk , the probability parameter for choosing stairs, escalators, and elevators is p s 、p l and p m .

[0106] In the shopping mall scenario, the width of the main channel is 4 meters, the maximum width of all channels is 5 meters, the sign perfection is 0.8, the maximum sign perfection is 1, α = 0.6, β = 0.4, then the channel attraction parameter

[0107] Node turning probability parameter p jkAccording to historical data statistics, at a certain three-channel intersection, the probability of turning left is 0.35, the probability of going straight is 0.4, and the probability of turning right is 0.25;

[0108] The selection probability parameter p of stairs, escalators, and elevators s 、p l and p m It is determined based on the convenience of the location and the attractiveness of the surrounding shops. For example, the probability of selecting an elevator near a popular shop area is higher and is initialized to 0.4, while the probability of selecting stairs in a relatively remote location is lower and is initialized to 0.15. The value range of each parameter is between [0,1].

[0109] Step 3: Determine the population size N = 200 to fully simulate the diversity of people in a large shopping mall during peak hours. Other decision variables (channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、p l and p m , speed adjustment parameter v adjust ) is initialized within the value range.

[0110] Step 4: Treat various factors that guide the flow of people in the shopping center traffic space as ocean currents, calculate the ocean current direction vector of each influencing factor respectively, and the jellyfish individuals move actively (attraction) and passively (repulsion) with the ocean currents.

[0111] For channel i, the ocean current direction vector is

[0112] In the formula, the vectors corresponding to the width and identification factors are And the corresponding weight coefficients α, β.

[0113] The ocean current direction vector of node j is In the formula, the turning probability parameter p of each turning direction k is jk , and the environmental vector associated with each turn direction Here the turning probability parameter p jk As a weight, it reflects the probability of different turning directions in people's choices.

[0114] For stairs s, escalators l, and ladders m, the current direction vector

[0115] In the formula, the vectors corresponding to factors such as location and surrounding store attractiveness are And weight coefficients δ and ε, where δ+ε=1.

[0116] Each jellyfish individual is attracted by Includes channel attraction Node attractiveness Stairs / elevator or escalator / staircase attraction Target attraction

[0117] Among them, the channel attraction is: Among them, k 1 is the channel attraction coefficient, is the channel attraction weight parameter of the channel where the individual is located, is the current direction vector of the channel;

[0118] The node attraction is: Among them, k 2 is the node attraction coefficient, j represents the nodes around individual n, p nj is the probability that individual n is affected by node j, is the current direction vector of the node;

[0119] Stairs / elevators or escalators / ladders have the following attractions: Among them, k 3 is the attraction coefficient of stairs / elevators or escalators / elevators, s represents the stairs / elevators or escalators / elevators around individual n, is the probability that an individual chooses stairs / elevator or escalator / elevator, is the current direction vector of the stairs / elevator or escalator / staircase;

[0120] The target attraction is: Among them, k 4 is the target attraction coefficient, is the vector pointing from the individual position to the target position, is the distance from the individual to the target location;

[0121] The total attraction is:

[0122] During the movement, individual jellyfish n will continue to move with the ocean currents and adjust their movement speed and direction.

[0123] For any two jellyfish individuals n and q, the repulsive force between the jellyfish individuals is For any two jellyfish individuals n and q, the repulsive force is proportional to the distance d between them. nq Inversely proportional, the repulsive factor is r, then in is the vector pointing from individual n to individual q. All repulsive forces interacting with individual n are superimposed, and the total repulsive force is

[0124] During the prediction process, the speed and movement position of the jellyfish in continuous time are updated based on the time step Δt;

[0125] The speed update formula is: in, is the new velocity of individual n at time t+Δt, is the old velocity of individual n at time t, w is the inertia weight, a is the attraction factor, r is the repulsion factor, is the attraction experienced by individual n at time t, is the repulsive force on individual n at time t;

[0126] The position update formula is: in, is the new position of individual n at time t+Δt, is the old position of individual n at time t.

[0127] During the update process, ensure that the parameters are within a reasonable range. For example, the channel attraction weight should be between 0-1, the steering probability should also be between 0-1, etc. If the parameter is out of range, adjust it to return to a reasonable range.

[0128] Using the fitness function, F(v i )=w 1 E distance +w 2 E node +w 3 E flow .

[0129] Among them, w 1 、w 2 、w 3 is the weight coefficient, E distance is the trajectory distance error, E node is the node passing error, E flow is the flow error, which evaluates the degree of match between the predicted trajectory and the actual crowd behavior. The trajectory distance error is obtained by calculating the average Euclidean distance between the predicted trajectory points and the actual trajectory points; the node passing error statistics predicts the difference between the time when the crowd passes through the key node and the actual passing time; the flow error compares the difference between the predicted flow of different traffic space elements and the actual flow of people.

[0130] The algorithm terminates when the number of iterations reaches the maximum number of iterations or when an individual with a fitness value less than the preset accuracy threshold appears in the population. The number of continuous iterations M = 200, the threshold ξ = 0.03, during the iteration process, the position changes of all jellyfish individuals are continuously monitored, and the iteration is stopped when the termination condition is met. The jellyfish trajectory in continuous time is output, that is, the predicted population trajectory.

[0131] The present invention is a method for predicting the trajectory of a crowd in a shopping center traffic space based on an artificial jellyfish search algorithm. The artificial jellyfish search algorithm simulates the search behavior of jellyfish in the ocean, including following ocean currents, moving in groups, and a time control mechanism for switching movement types. This bionic characteristic enables the algorithm to better simulate the behavioral changes of the crowd in a complex traffic space when dealing with the problem of predicting the trajectory of the crowd in the shopping center. The movement of the crowd in the shopping center is affected by many factors, just as jellyfish are affected by ocean currents, food distribution and other factors. In the shopping center, the crowd will flow naturally along the channel (similar to the ocean current channel) and will change direction or speed due to factors such as store distribution (similar to food sources) and sign guidance (similar to clues in the marine environment). The artificial jellyfish search algorithm can naturally capture these behavioral characteristics, so that the prediction result is more in line with the actual movement mode of the crowd. The artificial jellyfish search algorithm itself has the characteristics of strong optimization ability and fast convergence speed. When predicting the trajectory of the crowd, it can quickly find a better solution, reducing computing time and resource consumption. Compared with some traditional trajectory prediction algorithms, it does not require a large number of iterative calculations to achieve a better prediction effect.

[0132] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm are implemented.

[0133] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the method for predicting crowd trajectories in a shopping center traffic space based on an artificial jellyfish search algorithm.

[0134] The memory in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct RAM bus RAM (DRRAM). It should be noted that the memory of the method described in the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0135] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a high-density digital video disc (DVD)), or a semiconductor medium (eg, a solid state disc (SSD)).

[0136] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software. The steps of the method disclosed in conjunction with the embodiment of the present application can be directly embodied as a hardware processor for execution, or a combination of hardware and software modules in a processor for execution. The software module can be located in a storage medium mature in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in conjunction with its hardware. To avoid repetition, it is not described in detail here.

[0137] It should be noted that the processor in the embodiment of the present application can be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiment can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiment of the present application can be directly embodied as a hardware decoding processor to perform, or the hardware and software modules in the decoding processor can be combined and performed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0138] The above is a detailed introduction to the method and system for predicting crowd trajectories in shopping center traffic spaces based on an artificial jellyfish search algorithm proposed in the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scopes. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for predicting crowd trajectories in shopping center traffic space based on artificial jellyfish search algorithm, characterized in that: The method comprises the following steps: Step 1: Collect shopping mall scene parameter information and crowd behavior data, establish a special data fusion platform, associate shopping mall scene parameter information and crowd behavior data, so that crowd behavior data can correspond to specific scene parameters; Step 2: Treat the individuals in the crowd as jellyfish individuals, determine the position coordinates (x, y) as the basic decision variables, and the individual's horizontal and vertical velocity components are (v x ,v y ), the initial value of the speed is determined according to the statistical average of the collected crowd behavior data; other decision variables are defined as channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、p l and p m and the speed adjustment parameter v adjust ; Step 3: Initialize the jellyfish population and determine the population size N; the initial position coordinates (x, y) of the jellyfish individuals are randomly generated within the shopping mall, and the velocity component (v x ,v y ) is randomly initialized according to the speed range of the collected data, and other decision variables are initialized within the value range; Step 4: Use the artificial jellyfish search algorithm to iterate the algorithm to obtain the final crowd trajectory prediction; The prediction trajectory is specifically as follows: taking various factors that guide the flow of people in the shopping center traffic space as ocean currents, respectively calculating the ocean current direction vectors of channel i, node j, stair s, escalator l and straight ladder m; for each jellyfish individual n, calculating the attraction of the ocean current to individual n and the repulsive force between individual jellyfish Determine the time step Δt based on the characteristics of the data acquisition device; update the speed and position of the individual people according to the calculated attraction, repulsion and inertia weight, and obtain new speed and position coordinates, so as to simulate the movement trajectory of the people in the traffic space; When updating the position, it is necessary to check whether the crowd individuals exceed the boundary range of the shopping mall, that is, the channel attraction parameters, node turning probability parameters, stairs, escalators, and elevator selection parameters are all within the interval [0,1]. If they exceed the boundary, the parameters need to be adjusted back to a reasonable range to ensure that the predicted trajectory meets the actual space constraints; The iteration termination condition is set, that is, the position change of all jellyfish individuals within M consecutive iterations is less than a certain threshold ξ. When the termination condition is met, the iteration is stopped and the output is the jellyfish trajectory in continuous time, that is, the predicted crowd trajectory.

2. The method according to claim 1, characterized in that Establish a data fusion platform that associates shopping mall scene parameter information with crowd behavior data. The data fusion platform adopts a three-layer architecture, including data collection layer, edge computing layer and cloud fusion analysis layer; Among them, the data collection layer is composed of Wi-Fi positioning devices and smart cameras, which are responsible for collecting scene parameter information and crowd behavior data of the shopping mall; the edge computing layer is distributed in various areas of the shopping mall, close to the data collection equipment, and is composed of multiple edge computing nodes. These nodes can perform real-time pre-processing and preliminary analysis of the collected data, filter out redundant information to obtain the required traffic space information and crowd behavior information; the cloud fusion analysis layer is located in the data center, which is used for deep integration, analysis and storage of data from the edge computing layer.

3. The method according to claim 1, characterized in that Channel attraction parameters The node's turning probability parameter p jk , the selection probability parameter p of stairs, escalators, and elevators s 、p l and p m The value range of is [0-1], and each parameter is affected by different factors; Among them, the channel attraction parameter of each channel i is The factors affected: channel width i , signage i The weights of each factor are α and β respectively, and α+β=1. Among them, max(width) and max(signage) are the maximum values ​​of all channel widths and signage perfection, respectively; The turning probability parameter p for each possible turning direction k of each node j jk , For each staircase s, escalator l and elevator m, select the probability parameter p s 、p l and p m , are all affected by the convenience of location s / l / m Attraction of surrounding shops s / l / m Influence.

4. The method according to claim 1, characterized in that: The various factors that guide the flow of people in the shopping center traffic space are regarded as ocean currents. Depending on the influencing factors, the ocean currents have different direction vectors; Among them, for channel i, according to its width and the vector corresponding to the identification factor And the corresponding weight coefficients α, β, the channel current direction vector is For node j, according to the turning probability parameter p of each turning direction k, jk , and the environmental vector associated with each turn direction The ocean current direction vector of node j is Turn probability parameter p jk As a weight, it reflects the probability of different turning directions in the crowd's choice; For stairs s, escalators l and elevators m, the vectors corresponding to their locations and surrounding store attraction factors are: And the weight coefficients δ and ε, where δ + ε = 1, calculate the ocean current direction vector 5. The method according to claim 1, characterized in that Each jellyfish individual is attracted by Includes channel attraction Node attractiveness Stairs / elevator or escalator / staircase attraction Target attraction Among them, the channel attraction is: Among them, k1 is the channel attraction coefficient, is the channel attraction weight parameter of the channel where the individual is located, is the current direction vector of the channel; The node attraction is: Among them, k2 is the node attraction coefficient, j represents the nodes around individual n, and p nj is the probability that individual n is affected by node j, is the current direction vector of the node; Stairs / elevators or escalators / ladders have the following attractions: Among them, k3 is the attraction coefficient of stairs / elevators or escalators / elevators, s represents the stairs / elevators or escalators / elevators around individual n, is the probability that an individual chooses stairs / elevator or escalator / elevator, is the current direction vector of the stairs / elevator or escalator / staircase; The target attraction is: Among them, k4 is the target attraction coefficient, is the vector pointing from the individual position to the target position, is the distance from the individual to the target location; The total attraction is:

6. The method according to claim 1, characterized in that For any two jellyfish individuals n and q, the repulsive force between the jellyfish individuals is For any two jellyfish individuals n and q, the repulsive force is proportional to the distance d between them. nq Inversely proportional, the repulsive factor is r, then in is the vector pointing from individual n to individual q. All repulsive forces interacting with individual n are superimposed, and the total repulsive force is 7. The method according to claim 1, characterized in that During the prediction process, the speed and movement position of the jellyfish in continuous time are updated based on the time step Δt; The speed update formula is: in, is the new velocity of individual n at time t+Δt, is the old velocity of individual n at time t, w is the inertia weight, a is the attraction factor, r is the repulsion factor, is the attraction experienced by individual n at time t, is the repulsive force on individual n at time t; The position update formula is: in, is the new position of individual n at time t+Δt, is the old position of individual n at time t.

8. A crowd trajectory prediction system for shopping center traffic space based on artificial jellyfish search algorithm, characterized in that: The system comprises: Data collection module: collects shopping center scene parameter information and crowd behavior data, establishes a special data fusion platform, associates shopping center scene parameter information and crowd behavior data, and enables crowd behavior data to correspond to specific scene parameters; Decision variable setting module: The individuals in the crowd are regarded as jellyfish individuals, and the position coordinates (x, y) are determined as the basic decision variables. The velocity components of the individuals in the horizontal and vertical directions are (v x ,v y ), the initial value of the speed is determined according to the statistical average of the collected crowd behavior data; other decision variables are defined as channel attraction parameters Node turning probability parameter p jk , Stairs, escalators, elevators selection parameters p s 、p l and p m and the speed adjustment parameter v adjust ; Initialization module: Initialize the jellyfish population and determine the population size N; the initial position coordinates (x, y) of the jellyfish individuals are randomly generated within the shopping mall, and the velocity component (v x ,v y ) is randomly initialized according to the speed range of the collected data, and other decision variables are initialized within the value range; Trajectory prediction module: uses the artificial jellyfish search algorithm to iterate the algorithm to obtain the final crowd trajectory prediction; The prediction trajectory is specifically as follows: taking various factors that guide the flow of people in the shopping center traffic space as ocean currents, respectively calculating the ocean current direction vectors of channel i, node j, stair s, escalator l and straight ladder m; for each jellyfish individual n, calculating the attraction of the ocean current to individual n and the repulsive force between individual jellyfish Determine the time step Δt based on the characteristics of the data acquisition device; update the speed and position of the individual people according to the calculated attraction, repulsion and inertia weight, and obtain new speed and position coordinates, so as to simulate the movement trajectory of the people in the traffic space; When updating the position, it is necessary to check whether the crowd individuals exceed the boundary range of the shopping mall, that is, the channel attraction parameters, node turning probability parameters, stairs, escalators, and elevator selection parameters are all within the interval [0,1]. If they exceed the boundary, the parameters need to be adjusted back to a reasonable range to ensure that the predicted trajectory meets the actual space constraints; The iteration termination condition is set, that is, the position change of all jellyfish individuals within M consecutive iterations is less than a certain threshold ξ. When the termination condition is met, the iteration is stopped and the output is the jellyfish trajectory in continuous time, that is, the predicted crowd trajectory.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium for storing computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images