A high-rise building elevator management system and method based on a space-time information platform

Abstract

The application provides a high-rise building elevator management system and method based on a space-time information platform, which comprises: obtaining elevator state data of elevators in a building; using machine vision technology to monitor passenger flow in an elevator waiting hall and in a car in real time to generate passenger state data; based on a space-time information platform, generating a digital twin model of an elevator system according to the elevator state data; according to the passenger state data and the elevator state data, using a preset first algorithm to make intelligent decisions on elevator scheduling and outputting a first scheduling strategy; according to the first scheduling strategy, applying a preset second algorithm to simulate an elevator running path in the digital twin model and dynamically adjusting the elevator scheduling strategy to reduce passenger waiting time and energy consumption, thereby obtaining a second scheduling strategy; and managing the elevators in the building according to the second scheduling strategy. Through the combination of machine vision, digital twin technology and intelligent algorithms, the elevator scheduling efficiency and passenger experience can be effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent buildings, and particularly relates to a high-rise building elevator management system and method based on a space-time information platform. BACKGROUND

[0002] With the acceleration of urbanization and the increasing number of high-rise buildings, elevators as an important tool for vertical transportation have become increasingly important in people's daily lives. However, traditional elevator management methods usually rely on simple rules or fixed algorithms, which makes them difficult to accurately predict and respond to complex user request changes. This limitation leads to low efficiency of elevator operation, long waiting time for users, and other problems. SUMMARY

[0003] The present application is based on the above problems, and proposes a high-rise building elevator management system and method based on a space-time information platform. By combining machine vision, digital twin technology and intelligent algorithms, an innovative elevator management method is provided, which can effectively improve the scheduling efficiency of elevators and passenger experience.

[0004] Therefore, one aspect of the present application proposes a high-rise building elevator management method based on a space-time information platform, comprising:

[0005] Obtaining elevator state data of elevators in a building;

[0006] Using machine vision technology to monitor passenger flow in elevator waiting halls and in cars in real time, generating passenger state data;

[0007] Based on the space-time information platform, generating a digital twin model of the elevator system according to the elevator state data;

[0008] According to the passenger state data and the elevator state data, using a preset first algorithm to make intelligent decisions on elevator scheduling, outputting a first scheduling strategy;

[0009] According to the first scheduling strategy, applying a preset second algorithm to simulate elevator operation paths in the digital twin model, dynamically adjusting the elevator scheduling strategy to reduce passenger waiting time and energy consumption, obtaining a second scheduling strategy;

[0010] According to the second scheduling strategy, managing the elevators in the building.

[0011] Optionally, the step of using machine vision technology to monitor passenger flow in elevator waiting halls and in cars in real time, generating passenger state data, comprises:

[0012] Obtaining first image data of the waiting hall and second image data of the car;

[0013] In combination with a preset waiting-for-elevator passenger behavior model, the first image data is analyzed to obtain waiting-for-elevator passenger state data;

[0014] In combination with a preset in-elevator passenger behavior model, the second image data is analyzed to obtain in-elevator passenger state data;

[0015] After the waiting-for-elevator passenger state data and the in-elevator passenger state data are labeled, the passenger state data is obtained.

[0016] Optionally, the step of generating a digital twin model of the elevator system according to the elevator state data based on the space-time information platform comprises:

[0017] A three-dimensional scene coordinate system is established based on the space-time information platform, and the three-dimensional scene coordinate system comprises: establishing a global coordinate system with the center of the bottom floor of the building as the origin; establishing a local coordinate system with the center of the bottom of each elevator shaft as the origin; mapping and converting the global coordinate system and the local coordinate system;

[0018] A static model of the elevator system is constructed, comprising: constructing three-dimensional models of elevator components such as elevator cars, guide rails, counterweight systems, shafts, etc. in the three-dimensional scene coordinate system according to the elevator three-dimensional data and the elevator shaft three-dimensional data; calibrating the relative position relationship and motion constraint relationship between each elevator component; constructing an elevator component material library to assign corresponding material properties to each elevator component;

[0019] A dynamic model of the elevator system is constructed, comprising: establishing a physical motion model of the elevator system, including a car motion model, a traction system model, and a counterweight system model; establishing an environment model of the elevator system according to the elevator interior environment data, including a temperature field model, an airflow field model, and a noise field model; establishing a component state model of the elevator system according to the elevator accessory data, including the working state, the wear state, and the fault state of each component;

[0020] The static model and the dynamic model are fused to generate a digital twin model of the elevator system, comprising: establishing a corresponding relationship between physical quantities in the dynamic model and geometric features in the static model; according to real-time collected elevator state data, the digital twin model is updated in real time; a bidirectional data interaction mechanism between the digital twin model and the physical elevator system is established.

[0021] Optionally, the step of applying a preset first algorithm to make intelligent decisions on elevator dispatching according to the passenger state data and the elevator state data, and outputting a first dispatching strategy comprises:

[0022] The waiting-for-elevator passenger state data in the passenger state data is extracted to obtain waiting-for-elevator passenger state data of each floor;

[0023] Elevator car status data is generated based on the passenger status data inside the car and the elevator status data in the passenger status data.

[0024] Extract a first feature set from the passenger status data of each floor and the elevator car status data;

[0025] Based on the first feature set and the first algorithm, intelligent decision-making is made for elevator scheduling, and a first scheduling strategy is output.

[0026] Optionally, the second algorithm is an ant colony algorithm; the step of applying a preset second algorithm to simulate the elevator running path in the digital twin model according to the first scheduling strategy, dynamically adjusting the elevator scheduling strategy to reduce passenger waiting time and energy consumption, and obtaining the second scheduling strategy includes:

[0027] Constructing a graph structure model for an elevator scheduling scenario includes: mapping floor nodes, elevator locations, and target floors as vertices of a graph; mapping possible elevator running paths as edges of a graph; and assigning initial pheromone concentrations to the edges according to the first scheduling strategy.

[0028] Initialize the second algorithm parameters, including: setting the number of ants n to match the current number of scheduled tasks; setting the maximum number of iterations max_iter; setting the pheromone evaporation coefficient ρ1; and setting the state transition probability calculation parameters α1 and β1.

[0029] Construct a fitness function, including: calculating the average passenger waiting time; calculating the total energy consumption of the elevator system; setting waiting time weight w1 and energy consumption weight w2 according to the actual scheduling scenario; defining fitness as a weighted sum: f1 = w1 × waiting time + w2 × energy consumption;

[0030] The execution path search and optimization process includes: randomly assigning an initial node to each ant in the digital twin model; calculating the selection probability of the next node according to the state transition probability formula; and calculating the pheromone concentration τ based on the pheromone concentration τ. ij and heuristic information η ij Calculate the state transition probability p ij = (τ ij ^α1 × η ij ^β1) / Σ(τ ij ^α1 × η ij ^β1); Update global pheromone concentration: pheromone volatilization τ ij =(1-ρ1)×τ ij ;Pheromone increase Δτ ij = Q / Lk (Q is a constant, and Lk is the fitness value of path k);

[0031] Dynamically adjust the optimization strategy: monitor real-time state changes in the digital twin model; dynamically adjust heuristic information based on state changes; adaptively adjust the values ​​of α1 and β1 parameters; update the local optimum and global optimum in each iteration;

[0032] Generate a second scheduling strategy: select the globally optimal solution as the final scheduling scheme; convert the optimized path into specific scheduling instructions; and output detailed scheduling execution steps.

[0033] Optionally, the first algorithm is a fuzzy neural network algorithm; the step of making intelligent decisions on elevator scheduling based on the first feature set and the first algorithm, and outputting a first scheduling strategy, includes:

[0034] Preprocess the first feature set;

[0035] The first algorithm is obtained by constructing a fuzzy neural network structure, including: input layer configuration: setting the number of neurons to be consistent with the feature dimension, and defining the domain range of the input variables; fuzzy layer design: designing a membership function for each input variable and determining the number of fuzzy rules; inference layer implementation: establishing a fuzzy rule base and setting rule weights; output layer construction: defining the set of output variables and designing a defuzzification method.

[0036] Input the preprocessed first feature set into the first algorithm;

[0037] The fuzzification process includes: calculating the membership degrees of the input variables; and activating relevant fuzzy rules.

[0038] Implementing fuzzy reasoning includes: performing rule-based reasoning; aggregating the outputs of multiple rules to generate a comprehensive decision result;

[0039] The defuzzification process includes: applying the centroid method to calculate the sharp values ​​of the output variables and generate specific scheduling instructions; and output mapping transformation to map the sharp values ​​to scheduling actions and generate the execution sequence.

[0040] Generate the first scheduling strategy, including: output scheduling priority: assign service order to each elevator and determine the target floor sequence; generate execution plan: plan the operation path and set time nodes; and formulate emergency plan.

[0041] Optionally, the first algorithm is a fuzzy neural network algorithm; before the step of making intelligent decisions on elevator scheduling using a preset first algorithm based on the passenger status data and the elevator status data, and outputting a first scheduling strategy, the algorithm further includes a step of optimizing the first algorithm using a third algorithm, specifically:

[0042] Initializing the parameter optimization space includes: determining the fuzzy neural network parameters to be optimized, i.e., the membership function parameters of fuzzy rules, the connection weights of neural networks, and the threshold parameters of neurons; setting the value range and constraint conditions of each parameter; and establishing a parameter coding scheme to discretize the continuous parameter space;

[0043] Constructing an optimization objective function includes: designing evaluation indexes, i.e., scheduling decision accuracy, system response time, and calculation resource consumption; defining the objective function: f2 = q1 × accuracy + q2 × (1 / response time) + q3 × (1 / resource consumption); and setting the weights q1, q2, and q3 of each index;

[0044] Configuring a third algorithm includes: initializing an ant colony, i.e., adapting the number of ants m to the parameter dimension and randomly allocating initial parameter combinations; setting algorithm control parameters, i.e., pheromone evaporation coefficient ρ2, local search probability p_local, and global search probability p_global; and defining state transition rules, i.e., selecting parameter values based on a roulette strategy and considering the constraint relationship between parameters;

[0045] Performing a parameter optimization process includes: evaluating the current parameter combination, i.e., testing the performance of the fuzzy neural network on a validation dataset and calculating the objective function value; updating the local optimal solution, i.e., recording the optimal parameter combination found by each ant and updating the local pheromone concentration; updating the global optimal solution, i.e., comparing and updating the global optimal parameter combination and updating the global pheromone concentration; and applying an adaptive mechanism, i.e., dynamically adjusting the search step and adjusting the control parameters according to the optimization progress;

[0046] Implementing convergence control includes: setting termination conditions, i.e., reaching the maximum number of iterations, the objective function value converging, and the parameter change being less than a threshold; and applying an early stopping mechanism, i.e., monitoring the optimization effect and avoiding overfitting;

[0047] Updating the fuzzy neural network includes: updating the network using the optimized parameters, i.e., reconstructing the membership function, updating the connection weights, and adjusting the threshold parameters; and verifying the optimization effect, i.e., evaluating the performance on a test dataset and comparing the changes in indexes before and after optimization.

[0048] Optionally, the method for constructing the waiting passenger behavior model includes:

[0049] Collecting basic behavior data includes: acquiring image sequence data, i.e., recording the motion trajectory of waiting passengers, capturing posture change information, collecting facial expression data, and collecting gesture action data; collecting operation behavior data, i.e., recording the key operation sequence, acquiring card swiping information, and collecting voice instruction data; and collecting environmental parameters, i.e., recording the crowding degree of the waiting area, monitoring the environmental noise level, and collecting lighting condition data;

[0050] Extract behavior features from basic behavior data, including: processing spatio-temporal features, i.e. analyzing motion trajectory features, extracting position distribution features, calculating velocity and acceleration features; identifying interaction features, i.e. extracting human-computer interaction patterns, analyzing crowd interaction features, identifying group behavior patterns; analyzing emotional features, i.e. identifying facial expression changes, analyzing body language features, evaluating emotional states;

[0051] Establish a behavior classification system according to the extracted behavior features, including: defining basic behavior types, i.e. normal waiting behavior, emergency travel behavior, special demand behavior; constructing composite behavior patterns, i.e. identifying behavior sequence associations, extracting behavior combination features, establishing behavior transition rules; designing abnormal behavior recognition, i.e. defining abnormal behavior features, establishing early warning trigger mechanisms, developing response strategies;

[0052] Construct a probabilistic state transition model according to the behavior features and the behavior classification system, including: establishing a state space, i.e. defining a set of behavior states, designing a state feature vector, determining state transition conditions; calculating transition probabilities, i.e. based on historical data statistics, applying Bayesian inference, updating probability matrices; realizing state prediction, i.e. predicting next state probabilities, evaluating prediction reliability, dynamically adjusting prediction models;

[0053] Realize model self-adaptation optimization to obtain a waiting passenger behavior model, including: designing a feedback mechanism, i.e. collecting prediction error data, analyzing error distribution features, adjusting model parameters; performing online learning, i.e. updating behavior feature libraries, optimizing classification rules, improving prediction accuracy; realizing scene adaptation, i.e. identifying scene changes, adjusting model parameters, updating prediction strategies.

[0054] Optionally, the method for constructing the in-car passenger behavior model comprises:

[0055] Collect in-car multi-modal data, including: acquiring visual data, i.e. collecting passenger body posture data, recording passenger position distribution data, identifying face orientation information; collecting audio data, i.e. recording voice command information, collecting environmental sound features, identifying abnormal sound events; recording sensor data, i.e. collecting weight change data, monitoring vibration information, obtaining temperature and humidity parameters;

[0056] Extract passenger features from the collected in-car multi-modal data, including: analyzing spatial features, i.e. calculating passenger density distribution, extracting station position pattern features, identifying movement path features; identifying interaction behavior, i.e. analyzing human-computer interaction patterns, extracting interpersonal interaction features, identifying group behavior patterns; evaluating emotional states, i.e. analyzing facial expression changes, identifying body language features, evaluating tension level indicators;

[0057] According to the passenger characteristics, a behavior classification model is established, including: defining basic behavior categories, i.e., normal elevator riding behavior, emergency situation behavior, and special demand behavior; constructing complex behavior patterns, i.e., identifying behavior sequence rules, extracting behavior combination features, and establishing behavior transition rules; designing an abnormality detection mechanism, i.e., defining abnormal behavior features, establishing early warning thresholds, and formulating emergency response strategies;

[0058] In combination with passenger characteristics and classification models, an in-car passenger behavior model is obtained, including: constructing a state transition network, i.e., defining a state space, calculating transition probabilities, and predicting the next state; applying time series prediction, i.e., analyzing historical behavior sequences, predicting future behavior trends, and evaluating prediction reliability; and performing multi-modal fusion, i.e., integrating multi-source data information, optimizing prediction results, and improving prediction accuracy;

[0059] A safety monitoring mechanism is designed for the in-car passenger behavior model, including: implementing real-time monitoring, i.e., monitoring congestion levels, detecting abnormal behavior, and assessing safety risks; establishing an early warning system, i.e., setting multi-level early warning thresholds, defining early warning trigger conditions, and formulating early warning response strategies; and executing emergency handling, i.e., identifying emergency situations, starting emergency plans, and recording event logs.

[0060] Another aspect of the present application provides a high-rise building elevator management system based on a space-time information platform, for executing a high-rise building elevator management method based on a space-time information platform, including: a server, a communication module, and an elevator;

[0061] The server is configured to:

[0062] Obtain elevator state data of elevators in a building;

[0063] Real-time monitoring of passenger flow in elevator waiting halls and cars using machine vision technology generates passenger state data;

[0064] Based on a space-time information platform, a digital twin model of the elevator system is generated based on the elevator state data;

[0065] According to the passenger state data and the elevator state data, a first algorithm is used to make intelligent decisions on elevator dispatching, and a first dispatching strategy is output;

[0066] According to the first dispatching strategy, a second algorithm is applied to simulate elevator operation paths in the digital twin model, dynamically adjust the elevator dispatching strategy to reduce passenger waiting time and energy consumption, and obtain a second dispatching strategy;

[0067] According to the second dispatching strategy, the elevators in the building are managed.

[0068] The high-rise building elevator management method based on a space-time information platform provided by the application comprises the following steps: obtaining elevator state data of elevators in a building; using machine vision technology to monitor passenger flow in an elevator waiting hall and in a car in real time to generate passenger state data; generating a digital twin model of an elevator system based on a space-time information platform according to the elevator state data; using a preset first algorithm to make intelligent decisions on elevator scheduling according to the passenger state data and the elevator state data, and outputting a first scheduling strategy; simulating an elevator running path in the digital twin model according to the first scheduling strategy by using a preset second algorithm, dynamically adjusting the elevator scheduling strategy to reduce passenger waiting time and energy consumption, and obtaining a second scheduling strategy; and managing the elevators in the building according to the second scheduling strategy. The application provides an innovative elevator management method by combining machine vision, digital twin technology and intelligent algorithms, and can effectively improve the scheduling efficiency of elevators and passenger experience. BRIEF DESCRIPTION OF DRAWINGS

[0069] Fig. 1 is a flowchart of the high-rise building elevator management method based on a space-time information platform provided by an embodiment of the application;

[0070] Fig. 2 is a schematic block diagram of the high-rise building elevator management system based on a space-time information platform provided by an embodiment of the application. DETAILED DESCRIPTION

[0071] In order to more clearly understand the above-mentioned purposes, features and advantages of the application, the application will be further described in detail below with reference to the drawings and specific embodiments. It should be noted that the embodiments of the application and the features in the embodiments can be combined with each other without conflict.

[0072] In the following description, many specific details are set forth in order to provide a thorough understanding of the application, but the application can also be implemented in other ways different from those described herein, and therefore, the scope of protection of the application is not limited by the specific embodiments disclosed below.

[0073] The terms "first", "second", and the like in the specification and claims of the application and the above-mentioned drawings are used to distinguish different objects, and are not used to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or device.

[0074] Reference to“an embodiment” herein means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase“in one embodiment” in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily all

[0075] Some embodiments of the application will now be described with reference to the following drawings. Figs. 1-2

[0076] As shown in Fig. 1 , one embodiment of the application provides a high-rise building elevator management method based on a space-time information platform, comprising:

[0077] Obtaining elevator state data of elevators in the building (including elevator three-dimensional data, elevator shaft three-dimensional data, elevator internal environment data, elevator accessory data, etc.);

[0078] It can be understood that a laser scanner can be used to perform three-dimensional scanning on the elevator and its shaft to obtain accurate spatial data; laser scanning data and building information model (BIM) software are combined to generate a three-dimensional model of the elevator and its shaft; sensors (such as ultrasonic sensors, infrared sensors, etc.) are installed in the elevator shaft to monitor the state and structural changes of the shaft in real time; a camera is used to take images of the elevator shaft, and three-dimensional information of the shaft is extracted through image processing algorithms; temperature, humidity, and illumination environment sensors are installed in the elevator car to monitor the environmental parameters inside the elevator in real time; air quality sensors are used to monitor the air quality inside the elevator, including carbon dioxide, volatile organic compounds, and other indicators; intelligent sensors are installed on the elevator's auxiliary equipment (such as door control systems, control panels, etc.) to monitor their working state and performance in real time; data from various sensors is aggregated through a data integration platform to form elevator accessory state data; a data management system is established to integrate and analyze the obtained elevator state data to form a visual elevator state report; digital twin technology is used to combine real-time data with three-dimensional models to reflect the running state of the elevator and environmental changes in real time. Through the above methods, various state data of elevators in the building can be comprehensively obtained to provide a basis for intelligent management and scheduling of elevators.

[0079] Real-time monitoring of passenger flow in the elevator waiting hall and the car using machine vision technology to generate passenger state data;

[0080] Based on the space-time information platform, generating a digital twin model of the elevator system according to the elevator state data;

[0081] ​According to the passenger state data and the elevator state data, a first scheduling strategy is output by applying a preset first algorithm to make an intelligent decision on elevator scheduling;

[0082] According to the first scheduling strategy, a second scheduling strategy is obtained by applying a preset second algorithm to simulate an elevator running path in the digital twin model and dynamically adjusting the elevator scheduling strategy to reduce passenger waiting time and energy consumption.

[0083] According to the second scheduling strategy, the elevators in the building are managed.

[0084] The present application provides an innovative elevator management method by combining machine vision, digital twin technology and intelligent algorithms, which can effectively improve the scheduling efficiency of elevators and passenger experience.

[0085] In some possible embodiments of the present application, the step of generating passenger state data by using machine vision technology to monitor passenger flow in the elevator waiting hall and the car in real time comprises:

[0086] Obtaining first image data of the waiting hall and second image data of the car;

[0087] Analyzing the first image data in combination with a preset waiting passenger behavior model to obtain waiting passenger state data;

[0088] In this step, the waiting passenger input data (such as pressing the up / down key, inputting the target floor number, swiping the elevator card, etc.) and the waiting passenger key feature data (such as face data, key behavior data, key item data, etc.) in the first image data are identified; according to the waiting passenger input data, the waiting passenger key feature data and the waiting passenger behavior model, the waiting passenger state data (such as whether waiting for an elevator, whether the waiting passenger is going up or down, the destination floor, age, gender, body type, health status, item data, etc.) is obtained.

[0089] Analyzing the second image data in combination with a preset in-car passenger behavior model to obtain in-car passenger state data;

[0090] In this step, the in-car passenger input data (such as pressing the floor key, inputting the target floor number, speaking the floor number, swiping the elevator card, etc.) and the in-car passenger key feature data (such as face data, key behavior data, key item data, etc.) in the second image data are identified; according to the in-car passenger input data, the in-car passenger key feature data and the in-car passenger behavior model, the in-car passenger state data (such as the destination floor, age, gender, body type, health status, item data, etc.) is obtained.

[0091] The waiting-for-elevator passenger state data and the in-car passenger state data are labeled as the passenger state data.

[0092] In the embodiment, the target detection algorithm is used to identify passenger individuals in the first image data and / or the second image data, determine passenger number, posture (such as standing, sitting, walking, etc.), gender, body shape, age, health condition, personal belongings, etc., track the motion trajectory of the passenger through the tracking algorithm, and determine the passenger state data (such as whether waiting for the elevator, whether entering or exiting the car, the destination floor of the in-car passenger, whether the waiting-for-elevator passenger is going up or down, etc.) of the waiting-for-elevator passenger and / or the in-car passenger in combination with the preset waiting-for-elevator passenger behavior model and / or the in-car passenger behavior model.

[0093] In the embodiment, the waiting-for-elevator passenger state data includes but is not limited to the number of people going up, the number of people going down, age, gender, body shape, health condition, personal belongings data, destination floor, etc., and the in-car passenger state data includes but is not limited to the number of people, age, gender, body shape, health condition, personal belongings data, destination floor, etc. In the embodiment, the machine vision technology is used to obtain passenger image information through image acquisition sensors in the waiting hall and the car, analyze and process the image, identify the position, posture, behavior, etc. of the passenger, and thus accurately count the number of waiting passengers and the number of in-car passengers and other passenger state data that can indicate the intention of the passenger.

[0094] The scheme of the embodiment can accurately count the passenger state data in the waiting hall and the car through the machine vision technology, thus providing important decision support for elevator dispatching, upgrading the traditional simple passenger detection to all-around state analysis, improving the understanding of the intention of the passenger through the introduction of the behavior model, and supporting more intelligent dispatching decisions through very comprehensive feature recognition.

[0095] In some possible embodiments of the application, the following can also be included: according to the distribution of the passenger in the waiting hall and the car, in combination with the floor layout and the running direction of the elevator, the flow direction and the flow change trend of the passenger are predicted, thus providing more accurate decision basis for the elevator dispatching; the passenger flow state information is updated in real time, when a significant change in the passenger flow is detected, the running mode and the dispatching strategy of the elevator are timely adjusted (for example, the running frequency of the elevator is increased or the partition running strategy is adjusted during the peak period); the behavior of the waiting-for-elevator / in-car passenger is analyzed, whether the passenger is normally waiting for the elevator or has abnormal behavior (such as jumping in the car, prying the door, etc.) is judged, and an alarm signal should be timely sent to the elevator dispatching center and the relevant management department for the abnormal behavior.

[0096] In some possible embodiments of the application, the step of generating the digital twin model of the elevator system according to the elevator state data includes:

[0097] establish a global coordinate system with the center of the building bottom floor as the origin; map and convert the global coordinate system and the local coordinate system;

[0098] construct a static model of the elevator system, including: constructing three-dimensional models of elevator components such as elevator cars, guide rails, counterweight systems, and shafts in the three-dimensional scene coordinate system according to the elevator three-dimensional data and elevator shaft three-dimensional data; calibrating the relative position relationship and motion constraint relationship between each elevator component; constructing an elevator component material library to assign corresponding material properties to each elevator component;

[0099] construct a dynamic model of the elevator system, including: establishing a physical motion model of the elevator system, including a car motion model, a traction system model, and a counterweight system model; establishing an environment model of the elevator system according to the elevator internal environment data, including a temperature field model, an airflow field model, and a noise field model; establishing a component state model of the elevator system according to the elevator accessory data, including the working state, wear state, and fault state of each component;

[0100] fuse the static model and the dynamic model to generate a digital twin model of the elevator system, including: establishing a corresponding relationship between physical quantities in the dynamic model and geometric features in the static model; updating the digital twin model in real time according to real-time collected elevator state data; establishing a bidirectional data interaction mechanism between the digital twin model and the physical elevator system;

[0101] The digital twin model further includes the following functional modules:

[0102] state monitoring module: real-time monitoring and display of the running state of the elevator system;

[0103] fault diagnosis module: fault diagnosis based on historical data analysis and real-time data comparison;

[0104] performance evaluation module: evaluation of the running efficiency and service quality of the elevator system;

[0105] prediction and early warning module: prediction of possible faults and performance degradation.

[0106] The scheme of the embodiment realizes accurate spatial positioning of the elevator system by establishing a unified three-dimensional scene coordinate system; facilitates multi-elevator collaborative scheduling through mapping of the global coordinate system and the local coordinate system, realizes accurate three-dimensional modeling and spatial relationship expression of components of the elevator system; realizes real-time simulation and prediction of the elevator motion state, can reflect changes in the environment and component state of the elevator system, supports dynamic performance evaluation of the elevator system; provides comprehensive state monitoring and fault diagnosis capabilities; supports predictive maintenance, discovers potential problems in advance; provides reliable data support and simulation verification environment for elevator scheduling decisions; breaks through the limitations of static management in traditional elevator management, realizes an intelligent management mode combining virtual and real, and provides data support for optimization and upgrading of the elevator system.

[0107] In some possible embodiments of the application, the step of making intelligent decisions on elevator scheduling according to the passenger state data and the elevator state data and outputting a first scheduling strategy by using a preset first algorithm comprises:

[0108] extracting each floor waiting passenger state data (including up waiting passenger state data (such as destination floor, age, gender, body shape, health status, and personal item data) and down waiting passenger state data (such as destination floor, age, gender, body shape, health status, and personal item data)) from the waiting passenger state data in the passenger state data;

[0109] generating elevator car state data (including elevator car running state data (such as current floor, current running direction, request frequency of each floor, current car environment parameters, and current running parameters) and the number of passengers in the car (such as the total number of passengers in the car, the number of passengers getting off at each destination floor, age, gender, body shape, health status, and personal item data)) according to the in-car passenger state data in the passenger state data and the elevator state data;

[0110] extracting a first feature set from the each floor waiting passenger state data and the elevator car state data;

[0111] making intelligent decisions on elevator scheduling according to the first feature set and the first algorithm and outputting a first scheduling strategy.

[0112] In an embodiment, the first algorithm is a fuzzy neural network algorithm; the first algorithm comprises an input layer, a first rule set, a first function, and an output layer; the input parameters received by the input layer include the current number of waiting passengers, the number of passengers in the elevator car, the floor request frequency, and the current state of the elevator, etc.; the first rule set defines a plurality of rules, such as “if the number of waiting passengers is large and the number of passengers in the car is small, then the elevator is preferentially dispatched”; the first function uses a triangular or trapezoidal membership function to quantify the fuzziness of the input parameters; and the output layer outputs a dispatching strategy (including but not limited to: dispatching priority and recommended target floor of the elevator, etc.).

[0113] The scheme of the embodiment intelligently adjusts the dispatching strategy of the elevator according to real-time passenger demand and elevator state, improves the service efficiency of the elevator, reduces the waiting time of passengers, and improves the satisfaction of passengers; can quickly respond to different passenger demands and elevator states according to real-time data changes, and ensure the flexibility and adaptability of the elevator system; the elevator management method based on the space-time information platform and combined with the application of intelligent algorithms can effectively improve the management efficiency and service quality of the elevator in a high-rise building; not only considers traditional passenger number and floor, but also contains passenger characteristics; quantitatively processes uncertainty through fuzziness; and makes decisions based on real-time state data.

[0114] In some possible embodiments of the application, the second algorithm is an ant colony algorithm; and the step of applying a preset second algorithm in the digital twin model to simulate the elevator running path according to the first dispatching strategy, dynamically adjusting the elevator dispatching strategy to reduce passenger waiting time and energy consumption, and obtaining a second dispatching strategy, comprises:

[0115] A graph structure model of the elevator dispatching scene is constructed, including: mapping the floor nodes, elevator positions, and target floors as vertices of the graph; mapping possible elevator running paths as edges of the graph; and assigning an initial pheromone concentration to the edges according to the first dispatching strategy;

[0116] Second algorithm parameters are initialized, including: setting the number of ants n to be adapted to the current dispatching task; setting the maximum number of iterations max_iter; setting the pheromone evaporation coefficient ρ1; and setting the state transition probability calculation parameters α1 and β1;

[0117] A fitness function is constructed, including: calculating the average waiting time of passengers; calculating the total energy consumption of the elevator system; setting the waiting time weight w1 and the energy consumption weight w2 according to the actual dispatching scene; and defining the fitness as a weighted sum: f1=w1×waiting time+w2×energy consumption;

[0118] A path search and optimization process is performed, including: randomly assigning an initial node to each ant in the digital twin model; calculating the selection probability of the next node according to the state transition probability formula: calculating the selection probability of the next node according to the pheromone concentration τ ij and heuristic information ηij (related to path length, elevator load) Calculate state transition probability p ij = (τ ij ^α1 × η ij ^β1) / Σ(τ ij ^α1 × η ij ^β1); Update global pheromone concentration: Pheromone evaporation τ ij = (1-ρ1) × τ ij ; Pheromone increase Δτ ij = Q / Lk (Q is a constant, Lk is the fitness value of path k);

[0119] Dynamic adjustment optimization strategy: Monitor real-time state changes in the digital twin model; dynamically adjust heuristic information according to state changes; adaptively adjust α1, β1 parameter values; update local optimal solution and global optimal solution in each iteration;

[0120] Generate the second scheduling strategy: Select the global optimal solution as the final scheduling scheme; convert the optimized path into specific scheduling instructions; output detailed scheduling execution steps.

[0121] In this embodiment, the constituent factors of heuristic information: path length related factors (such as the distance from the current floor to the target floor of the elevator, the number of stopping floors in the path, the number of turns (the number of times the running direction is changed) in the path) and elevator load related factors (such as the current car load rate, the number of passengers expected to get on and off in the path, the remaining capacity of the car, etc.). The calculation formula of heuristic information: η ij = 1 / (w11 × distance + w21 × load rate + w31 × turn number); where w11, w21, w31 are weight coefficients; distance, load rate, and turn number need to be normalized; η ij The larger the value, the more priority the path has. Dynamic adjustment of heuristic information includes: adjusting weights according to different time periods (paying more attention to load balancing during peak hours and paying more attention to energy efficiency during flat peak hours) and adjusting according to special situations (reducing load weight when special passengers, increasing distance weight in emergency situations, and reducing speed weight when elderly people take the elevator). Synergistic effect of heuristic information and pheromone: state transition probability calculation p ij = (τ ij ^α1 × η ij ^β1) / Σ(τ ij ^α1 × η ij ^β1); τ ij is the pheromone concentration, α1 is the importance factor of pheromone, and β1 is the importance factor of heuristic information. This scheme can optimize the rationality of path selection, improve the real-time performance of scheduling decision, enhance the adaptability of the algorithm, and achieve multi-objective trade-off scheduling optimization.

[0122] The scheme of the embodiment can significantly reduce the average waiting time of passengers, reduce the overall energy consumption of the elevator system, and improve the efficiency of elevator scheduling; can quickly respond to changes in the scheduling scene, has good dynamic optimization capability, and avoids local optimal solution; the traditional ant colony algorithm is deeply integrated with the elevator scheduling scene, a dynamic parameter adjustment mechanism is realized, and more accurate scheduling optimization capability is provided.

[0123] In some possible embodiments of the application, the first algorithm is a fuzzy neural network algorithm; the step of making intelligent decisions on elevator scheduling according to the first feature set and the first algorithm and outputting a first scheduling strategy comprises:

[0124] The pre-processing of the first feature set comprises: feature standardization processing (normalizing numerical value type features and encoding category type features); feature combination construction (generating time window features and constructing feature interaction items); feature importance evaluation (calculating feature weights and screening key features);

[0125] The first algorithm is constructed by constructing a fuzzy neural network structure, comprising: input layer configuration (setting the number of neurons to be consistent with the feature dimension, and defining the domain range of the input variable); fuzzy layer design (designing a membership function for each input variable and determining the number of fuzzy rules); inference layer implementation (establishing a fuzzy rule base and setting rule weights); output layer construction (defining an output variable set and designing a defuzzification method);

[0126] The pre-processed first feature set is input into the first algorithm;

[0127] The defuzzification process comprises: calculating the membership degree of the input variable (applying a Gaussian membership function to generate a fuzzy feature vector); and activating relevant fuzzy rules (matching rule conditions and calculating rule trigger intensity);

[0128] The fuzzy inference comprises: performing rule inference (applying a Mamdani inference mechanism to calculate rule output); and rule combination (aggregating the outputs of multiple rules to generate a comprehensive decision result);

[0129] The defuzzification is performed, comprising: applying the barycenter method (calculating the clear value of the output variable to generate a specific scheduling instruction); and output mapping conversion (mapping the clear value to a scheduling action to generate an execution time sequence);

[0130] The first scheduling strategy is generated, comprising: outputting scheduling priority (assigning service order to each elevator and determining the target floor sequence); generating an execution plan (planning a running path and setting a time node); and formulating an emergency plan (considering unexpected situations and pre-selecting a solution).

[0131] In the embodiment, the fuzzy neural network realizes intelligent scheduling of the elevator through steps of fuzzification, rule reasoning, learning optimization and intelligent decision making, so as to improve the transportation efficiency and service quality of the elevator.

[0132] In some possible embodiments of the application, the first algorithm is a fuzzy neural network algorithm; before the step of outputting the first scheduling strategy by using the preset first algorithm to make intelligent decisions on the elevator scheduling according to the passenger state data and the elevator state data, the step further includes a step of optimizing the first algorithm by using a third algorithm, specifically:

[0133] Initializing a parameter optimization space includes: determining fuzzy neural network parameters to be optimized, including membership function parameters of fuzzy rules, connection weights of neural networks, and threshold parameters of neurons; setting value ranges and constraint conditions of each parameter; and discretizing a continuous parameter space by establishing a parameter coding scheme;

[0134] Building an optimization objective function includes: designing evaluation indexes, including scheduling decision accuracy, system response time, and calculation resource consumption; defining an objective function: f2 = q1 x accuracy + q2 x (1 / response time) + q3 x (1 / resource consumption); and setting weights q1, q2 and q3 of each index;

[0135] Configuring the third algorithm includes: initializing an ant colony, including adapting the number of ants m to the parameter dimension and randomly allocating initial parameter combinations; setting algorithm control parameters, including pheromone evaporation coefficient p2, local search probability p_local, and global search probability p_global; and defining state transition rules, including selecting parameter values based on a roulette strategy and considering constraint relationships between parameters;

[0136] Performing a parameter optimization process includes: evaluating a current parameter combination, including testing fuzzy neural network performance on a verification data set and calculating an objective function value; updating a local optimal solution, including recording an optimal parameter combination found by each ant and updating local pheromone concentration; updating a global optimal solution, including comparing and updating a global optimal parameter combination and updating global pheromone concentration; and applying an adaptive mechanism, including dynamically adjusting a search step and adjusting control parameters according to optimization progress;

[0137] Implementing convergence control includes: setting termination conditions, including reaching a maximum number of iterations, converging an objective function value, and changing parameters by less than a threshold value; and applying an early stopping mechanism, including monitoring optimization effects and avoiding overfitting;

[0138] Updating the fuzzy neural network includes: updating the network using optimized parameters, including reconstructing membership functions, updating connection weights, and adjusting threshold parameters; and verifying optimization effects, including evaluating performance on a test data set and comparing changes in indexes before and after optimization.

[0139] The scheme of the embodiment can improve the decision accuracy of the fuzzy neural network, reduce system response time and reduce the consumption of computing resources; can realize dynamic optimization of parameters, improve the robustness of the algorithm, and enhance the adaptability to environmental changes; can accelerate the convergence speed of parameter optimization, avoid falling into local optimum, and reduce the computational overhead of the optimization process; can provide a reliable parameter optimization scheme to ensure the stability of the optimization result and support online optimization and update.

[0140] In some possible embodiments of the application, the method for constructing the waiting passenger behavior model comprises:

[0141] Collecting basic behavior data, including: acquiring image sequence data, i.e. recording the motion trajectory of the waiting passenger, capturing posture change information, collecting facial expression data and gesture action data; collecting operation behavior data, i.e. recording the key operation sequence, acquiring card swiping information, collecting voice instruction data; collecting environmental parameters, i.e. recording the crowdedness of the waiting area, monitoring the environmental noise level, collecting lighting condition data;

[0142] Extracting behavior features from the basic behavior data, including: processing spatiotemporal features, i.e. analyzing motion trajectory features, extracting position distribution features, calculating speed and acceleration features; recognizing interaction features, i.e. extracting human-computer interaction modes, analyzing crowd interaction features, recognizing group behavior modes; analyzing emotional features, i.e. recognizing facial expression changes, analyzing body language features, and evaluating emotional states;

[0143] Establishing a behavior classification system according to the extracted behavior features, including: defining basic behavior types, i.e. normal waiting behavior, emergency travel behavior, and special demand behavior; constructing composite behavior modes, i.e. recognizing behavior sequence association, extracting behavior combination features, and establishing behavior transition rules; designing abnormal behavior recognition, i.e. defining abnormal behavior features, establishing an early warning trigger mechanism, and formulating response strategies;

[0144] Constructing a probabilistic state transition model according to the behavior features and the behavior classification system, including: establishing a state space, i.e. defining a behavior state set, designing a state feature vector, and determining state transition conditions; calculating transition probabilities, i.e. based on historical data statistics, applying Bayesian inference, and updating a probability matrix; realizing state prediction, i.e. predicting the next state probability, evaluating the prediction reliability, and dynamically adjusting the prediction model;

[0145] It can be understood that by constructing a probability state transition model (state space and transition probability), the next possible behavior of the passenger can be predicted, dynamic prediction of the passenger behavior sequence is realized, early planning and optimization of elevator dispatching are supported; the randomness and uncertainty of the passenger behavior can be processed by constructing the probability state transition model, the state transition probability is updated through Bayesian inference, and the reliability and robustness of the prediction result are improved; potential rules and patterns of the passenger behavior can be found, typical behavior sequences and transition paths are identified, and in-depth understanding of the passenger behavior is supported; probability basis is provided for elevator dispatching decision, active service based on prediction is supported, resource allocation and dispatching efficiency are optimized; behaviors deviating from the normal state transition mode are identified, potential abnormal situations are found in time, early warning and intervention mechanisms are provided; the transition probability is updated through continuous observation, the dynamic changes of the passenger behavior mode are adapted, and the generalization ability of the model is improved.

[0146] Model self-adaptive optimization is realized, and a waiting passenger behavior model is obtained, including: designing a feedback mechanism, that is, collecting prediction error data, analyzing error distribution characteristics, and adjusting model parameters; performing online learning, that is, updating a behavior feature library, optimizing classification rules, and improving prediction accuracy; realizing scene adaptation, that is, identifying scene changes, adjusting model parameters, and updating prediction strategies.

[0147] The scheme of the embodiment can improve behavior recognition accuracy, enhance behavior prediction capability, and realize timely discovery of abnormal behaviors; can quickly respond to scene changes, realize dynamic optimization of the model, and improve prediction reliability; can reduce behavior recognition delay, speed up response to abnormal behaviors, and improve decision support efficiency; can adapt to different population characteristics, support multi-scene applications, and have transfer learning capability.

[0148] In some possible embodiments of the application, the method for constructing the in-car passenger behavior model comprises:

[0149] Collecting in-car multi-modal data comprises: acquiring visual data, that is, collecting passenger body posture data, recording passenger position distribution data, and identifying face orientation information; collecting audio data, that is, recording voice instruction information, collecting environmental sound characteristics, and identifying abnormal sound events; recording sensor data, that is, collecting weight change data, monitoring vibration information, and acquiring temperature and humidity parameters;

[0150] Extracting passenger features from the collected in-car multi-modal data comprises: analyzing spatial features, that is, calculating passenger density distribution, extracting station mode features, and identifying moving path features; identifying interaction behaviors, that is, analyzing human-computer interaction mode, extracting interpersonal interaction features, and identifying group behavior mode; evaluating emotional state, that is, analyzing facial expression changes, identifying body language features, and evaluating tension level indicators;

[0151] According to the passenger characteristics, a behavior classification model is established, including: defining basic behavior categories, i.e. normal elevator riding behavior, emergency situation behavior, and special demand behavior; constructing complex behavior patterns, i.e. identifying behavior sequence rules, extracting behavior combination features, and establishing behavior transition rules; designing an abnormality detection mechanism, i.e. defining abnormal behavior features, establishing early warning thresholds, and formulating emergency response strategies;

[0152] In combination with passenger characteristics and classification models, an in-car passenger behavior model is obtained, including: constructing a state transition network, i.e. defining a state space, calculating transition probabilities, and predicting the next state; applying time series prediction, i.e. analyzing historical behavior sequences, predicting future behavior trends, and evaluating prediction reliability; and performing multi-modal fusion, i.e. integrating multi-source data information, optimizing prediction results, and improving prediction accuracy;

[0153] A safety monitoring mechanism is designed for the in-car passenger behavior model, including: implementing real-time monitoring, i.e. monitoring congestion levels, detecting abnormal behavior, and assessing safety risks; establishing an early warning system, i.e. setting multi-level early warning thresholds, defining early warning trigger conditions, and formulating early warning response strategies; and executing emergency handling, i.e. identifying emergency situations, starting emergency plans, and recording event logs.

[0154] The scheme of the embodiment can improve behavior recognition accuracy, enhance abnormality detection capability, and achieve accurate state prediction; can timely discover safety hazards, quickly respond to emergency situations, and improve operational safety; can optimize passenger experience, provide personalized services, and enhance service intelligence; and can improve operational efficiency, reduce management costs, and support intelligent decision-making.

[0155] See Fig. 2 Another embodiment of the present application provides a high-rise building elevator management system based on a space-time information platform, which is used to execute a high-rise building elevator management method based on a space-time information platform, including: a server, a communication module, and an elevator;

[0156] The server is configured to:

[0157] Obtain elevator state data of elevators in a building;

[0158] Generate passenger state data by using machine vision technology to monitor passenger flow in elevator waiting halls and cars in real time;

[0159] Based on a space-time information platform, generate a digital twin model of an elevator system according to the elevator state data;

[0160] According to the passenger state data and the elevator state data, use a preset first algorithm to make intelligent decisions on elevator dispatching, and output a first dispatching strategy;

[0161] According to the first scheduling strategy, a preset second algorithm is applied to simulate an elevator running path in the digital twin model, dynamically adjust the elevator scheduling strategy, so as to reduce passenger waiting time and energy consumption, and obtain a second scheduling strategy.

[0162] According to the second scheduling strategy, the elevators in the building are managed.

[0163] It should be understood that, Fig. 2 The block diagram of the high-rise building elevator management system based on the space-time information platform shown is only schematic, and the number of each module shown does not limit the protection scope of the present application. The high-rise building elevator management system based on the space-time information platform provided in the embodiment can be used to execute each embodiment scheme of the corresponding high-rise building elevator management method based on the space-time information platform. For the specific implementation process, please refer to the description of each method embodiment, which is not repeated here.

[0164] It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the present application is not limited by the action sequence described, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily necessary for the present application.

[0165] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0166] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented by other ways. For example, the device embodiments described above are only schematic, and the division of the above units is only a logical function division, and there can be another division way in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the display or discussion of the coupling or direct coupling or communication connection between the units can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical or other forms.

[0167] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0168] In addition, each of the function units in each of the embodiments of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software function unit.

[0169] When the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable memory. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the above-mentioned method of each embodiment of the present application. The aforementioned memory includes: a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0170] Those of ordinary skill in the art can understand that all or part of the steps in the above-mentioned embodiments can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable memory, which can include a flash disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0171] The embodiments of the present application are described in detail above, and the principles and implementation manners of the present application are described by applying specific examples. The above description of the embodiments is only used to help understand the method of the present application and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present application, the specific implementation manner and application range can be changed, and the above description of the present application should not be understood as a limitation.

[0172] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements without departing from the spirit and scope of the present application, and can make various changes and modifications, including the combination of different functions and implementation steps, including the implementation of software and hardware, which are all within the protection scope of the present application.

Claims

1. A high-rise building elevator management method based on a space-time information platform, characterized by, The method comprises the following steps: acquiring elevator state data of elevators in a building; monitoring passenger flow in an elevator hall and in a car in real time by using machine vision technology to generate passenger state data; generating a digital twin model of an elevator system based on a space-time information platform according to the elevator state data; making intelligent decisions on elevator dispatching by using a preset first algorithm according to the passenger state data and the elevator state data, and outputting a first dispatching strategy; simulating an elevator running path in the digital twin model by applying a preset second algorithm according to the first dispatching strategy, dynamically adjusting the elevator dispatching strategy to reduce passenger waiting time and energy consumption, and obtaining a second dispatching strategy; managing the elevators in the building according to the second dispatching strategy; wherein the step of monitoring passenger flow in an elevator hall and in a car in real time by using machine vision technology to generate passenger state data comprises: acquiring first image data of the elevator hall and second image data of the car; analyzing the first image data in combination with a preset elevator waiting passenger behavior model to obtain elevator waiting passenger state data; analyzing the second image data in combination with a preset car passenger behavior model to obtain car passenger state data; labeling the elevator waiting passenger state data and the car passenger state data as the passenger state data; wherein the step of generating a digital twin model of an elevator system based on a space-time information platform according to the elevator state data comprises: establishing a three-dimensional scene coordinate system based on the space-time information platform, which comprises: establishing a global coordinate system with the center of the bottom floor of the building as the origin; establishing a local coordinate system with the center of the bottom of each elevator shaft as the origin; mapping and converting the global coordinate system and the local coordinate system; constructing a static model of the elevator system, which comprises: constructing three-dimensional models of elevator cars, guide rails, counterweight systems, and shafts in the elevator system in the three-dimensional scene coordinate system according to the three-dimensional data of the elevators and the three-dimensional data of the elevator shafts; calibrating the relative position relationship and motion constraint relationship between the elevator components; constructing a material library for the elevator components to assign corresponding material properties to each elevator component; constructing a dynamic model of the elevator system, which comprises: establishing a physical motion model of the elevator system, including a car motion model, a traction system model, and a counterweight system model; establishing an environmental model of the elevator system according to elevator internal environment data, including a temperature field model, an airflow field model, and a noise field model; establishing a component state model of the elevator system according to elevator accessory data, including the working state, wear state, and fault state of each component; fusing the static model and the dynamic model to generate a digital twin model of the elevator system, which comprises: establishing a corresponding relationship between the physical quantities in the dynamic model and the geometric features in the static model; updating the digital twin model in real time according to real-time collected elevator state data; establishing a bidirectional data interaction mechanism between the digital twin model and the physical elevator system; wherein the step of making intelligent decisions on elevator dispatching by using a preset first algorithm according to the passenger state data and the elevator state data, and outputting a first dispatching strategy comprises: extracting each floor waiting passenger state data from the waiting passenger state data in the passenger state data; generating elevator car state data according to the in-car passenger state data in the passenger state data and the elevator state data; extracting a first feature set from the each floor waiting passenger state data and the elevator car state data; intelligently deciding the elevator dispatching according to the first feature set and the first algorithm, and outputting a first dispatching strategy; wherein the second algorithm is an ant colony algorithm; and the step of applying the preset second algorithm to simulate the elevator running path in the digital twin model according to the first dispatching strategy, dynamically adjusting the elevator dispatching strategy to reduce passenger waiting time and energy consumption, and obtaining a second dispatching strategy, comprises: building a graph structure model of the elevator dispatching scene, including: mapping the floor nodes, elevator positions and target floors as the vertices of the graph; mapping the possible elevator running paths as the edges of the graph; and assigning the edges with initial pheromone concentration according to the first dispatching strategy; initializing the second algorithm parameters, including: setting the number of ants n to be adapted to the current dispatching task number; setting the maximum iteration number max_iter; setting the pheromone evaporation coefficient ρ1; and setting the state transition probability calculation parameters α1 and β1; building a fitness function, including: calculating the passenger average waiting time; calculating the total energy consumption of the elevator system; setting the waiting time weight w1 and the energy consumption weight w2 according to the actual dispatching scene; and defining the fitness as a weighted sum: f1=w1×waiting time+w2×energy consumption; The execution path search and optimization process includes: randomly assigning an initial node to each ant in the digital twin model; calculating the selection probability of the next node according to the state transition probability formula; calculating the state transition probability p according to the pheromone concentration τ and the heuristic information η ij ij ij = (τ ij ^α1 × η ij ^β1) / Σ(τ ij ^α1 × η ij ^β1); updating the global pheromone concentration: pheromone volatilization τ ij =(1-ρ1)×τ ij ; pheromone increase Δτ ij = Q / Lk, wherein Q is a constant and Lk is the fitness value of path k;​​ dynamically adjusting the optimization strategy: monitoring the real-time state changes in the digital twin model; dynamically adjusting the heuristic information according to the state changes; adaptively adjusting the α1, β1 parameter values; and updating the local optimal solution and the global optimal solution in each iteration; generating the second dispatching strategy: selecting the global optimal solution as the final dispatching scheme; converting the optimized path into specific dispatching instructions; and outputting detailed dispatching execution steps.

2. The high-rise building elevator management method based on the space-time information platform according to claim 1, characterized in that, The first algorithm is a fuzzy neural network algorithm; and the step of intelligently deciding the elevator dispatching according to the first feature set and the first algorithm, and outputting a first dispatching strategy, comprises: preprocessing the first feature set; building a fuzzy neural network structure to obtain the first algorithm, including: input layer configuration: setting the number of neurons to be consistent with the feature dimension, and defining the domain range of the input variables; fuzzy layer design: designing the membership function for each input variable, and determining the number of fuzzy rules; inference layer implementation: establishing a fuzzy rule base, and setting the rule weight; and output layer construction: defining the output variable set, and designing the defuzzification method; inputting the preprocessed first feature set into the first algorithm; executing the fuzzification process, including: calculating the membership degree of the input variable; and activating the relevant fuzzy rules; implementing fuzzy reasoning, including: executing rule reasoning; and aggregating the outputs of multiple rules to generate a comprehensive decision result; executing defuzzification, including: applying the barycentric method: calculating the crisp value of the output variable, and generating specific dispatching instructions; and outputting mapping conversion: mapping the crisp value to the dispatching action, and generating the execution timing; The first scheduling strategy is generated, including: outputting a scheduling priority; assigning a service order to each elevator; determining a target floor sequence; generating an execution plan; planning a running path; setting a time node; and formulating an emergency plan.

3. The high-rise building elevator management method based on the space-time information platform according to claim 2, characterized in that, The first algorithm is a fuzzy neural network algorithm; before the step of outputting the first scheduling strategy according to the passenger state data and the elevator state data and using the preset first algorithm to make an intelligent decision on elevator scheduling, the step of optimizing the first algorithm by using a third algorithm is further included, specifically: Initializing a parameter optimization space, including: determining fuzzy neural network parameters to be optimized, including membership function parameters of fuzzy rules, connection weight parameters of neural networks, and threshold value parameters of neurons; setting value ranges and constraint conditions of each parameter; and establishing a parameter coding scheme to discretize a continuous parameter space; Constructing an optimization objective function, including: designing evaluation indexes, including scheduling decision accuracy, system response time, and calculation resource consumption; defining an objective function: f2 = q1*accuracy + q2*(1 / response time) + q3*(1 / resource consumption); and setting weights q1, q2, and q3 of each index; Configuring the third algorithm, including: initializing an ant colony, including adapting the number of ants m to the parameter dimension and randomly assigning initial parameter combinations; setting algorithm control parameters, including an information pheromone evaporation coefficient ρ2, a local search probability p_local, and a global search probability p_global; and defining state transition rules, including selecting parameter values based on a roulette strategy and considering constraint relationships between parameters; Performing a parameter optimization process, including: evaluating a current parameter combination, including testing fuzzy neural network performance on a verification data set and calculating an objective function value; updating a local optimal solution, including recording an optimal parameter combination found by each ant and updating local information pheromone concentration; updating a global optimal solution, including comparing and updating a global optimal parameter combination and updating global information pheromone concentration; and applying an adaptive mechanism, including dynamically adjusting a search step and adjusting control parameters according to optimization progress; Implementing convergence control, including: setting a termination condition, including reaching a maximum number of iterations, the objective function value converging, and parameter changes being less than a threshold value; and applying an early stopping mechanism, including monitoring optimization effects and avoiding overfitting; Updating the fuzzy neural network, including: updating the network using the optimized parameters, including reconstructing membership functions, updating connection weights, and adjusting threshold parameters; and verifying optimization effects, including evaluating performance on a test data set and comparing index changes before and after optimization.

4. The high-rise building elevator management method based on the space-time information platform according to claim 3, characterized in that, The method for constructing the waiting passenger behavior model includes: Collecting basic behavior data, including: acquiring image sequence data, i.e., recording motion trajectories of waiting passengers, capturing posture change information, collecting facial expression data, and collecting gesture action data; collecting operation behavior data, i.e., recording key operation sequences, acquiring card swiping information, and collecting voice instruction data; and collecting environmental parameters, i.e., recording waiting area crowding, monitoring environmental noise levels, and collecting lighting condition data; Behavior features are extracted from the basic behavior data, including: processing spatiotemporal features, i.e. analyzing motion trajectory features, extracting position distribution features, and calculating velocity and acceleration features; identifying interaction features, i.e. extracting human-machine interaction modes, analyzing crowd interaction features, and identifying group behavior modes; and analyzing emotional features, i.e. identifying facial expression changes, analyzing body language features, and evaluating emotional states; A behavior classification system is established according to the extracted behavior features, including: defining basic behavior types, i.e. normal waiting-for-elevator behavior, emergency travel behavior, and special demand behavior; constructing complex behavior modes, i.e. identifying behavior sequence correlations, extracting behavior combination features, and establishing behavior transition rules; and designing abnormal behavior identification, i.e. defining abnormal behavior features, establishing early warning triggering mechanisms, and formulating response strategies; A probabilistic state transition model is constructed according to the behavior features and the behavior classification system, including: establishing a state space, i.e. defining a behavior state set, designing a state feature vector, and determining state transition conditions; calculating transition probabilities, i.e. based on historical data statistics, applying Bayesian inference, and updating a probability matrix; and realizing state prediction, i.e. predicting next state probabilities, evaluating prediction reliability, and dynamically adjusting a prediction model; Model self-adaptation optimization is realized to obtain a waiting-for-elevator passenger behavior model, including: designing a feedback mechanism, i.e. collecting prediction error data, analyzing error distribution features, and adjusting model parameters; performing online learning, i.e. updating a behavior feature library, optimizing classification rules, and improving prediction accuracy; and realizing scene adaptation, i.e. identifying scene changes, adjusting model parameters, and updating prediction strategies.

5. The high-rise building elevator management method based on the space-time information platform according to claim 4, characterized in that, The method for constructing the in-car passenger behavior model comprises: Multi-modal data in the car are collected, including: acquiring visual data, i.e. collecting passenger body posture data, recording passenger position distribution data, and identifying face orientation information; collecting audio data, i.e. recording voice command information, collecting environmental sound features, and identifying abnormal sound events; and recording sensor data, i.e. collecting weight change data, monitoring vibration information, and acquiring temperature and humidity parameters; Passenger features are extracted from the collected multi-modal data in the car, including: analyzing spatial features, i.e. calculating passenger density distribution, extracting station mode features, and identifying movement path features; identifying interaction behavior, i.e. analyzing human-machine interaction modes, extracting interpersonal interaction features, and identifying group behavior modes; and evaluating emotional states, i.e. analyzing facial expression changes, identifying body language features, and evaluating tension level indicators; A behavior classification model is established according to the passenger features, including: defining basic behavior categories, i.e. normal elevator riding behavior, emergency condition behavior, and special demand behavior; constructing complex behavior modes, i.e. identifying behavior sequence rules, extracting behavior combination features, and establishing behavior transition rules; and designing an abnormal detection mechanism, i.e. defining abnormal behavior features, establishing early warning thresholds, and formulating emergency response strategies; In combination with passenger characteristics and classification models, an in-car passenger behavior model is obtained, including: constructing a state transition network, i.e., defining a state space, calculating transition probabilities, and predicting the next state; applying time series prediction, i.e., analyzing historical behavior sequences, predicting future behavior trends, and evaluating prediction reliability; and performing multi-modal fusion, i.e., integrating multi-source data information, optimizing prediction results, and improving prediction accuracy; A safety monitoring mechanism is designed for the in-car passenger behavior model, including: implementing real-time monitoring, i.e., monitoring congestion levels, detecting abnormal behavior, and assessing safety risks; establishing an early warning system, i.e., setting multi-level warning thresholds, defining warning trigger conditions, and developing warning response strategies; and executing emergency handling, i.e., identifying emergency conditions, initiating emergency plans, and recording event logs.

6. A high-rise building elevator management system based on a space-time information platform, for executing the high-rise building elevator management method based on a space-time information platform according to any one of claims 1 to 5, characterized in that, It includes: a server, a communication module, and an elevator; the server is configured to: obtain elevator state data of elevators in a building; use machine vision technology to monitor passenger flow in elevator lobbies and cars in real time, generating passenger state data; based on a space-time information platform, generate a digital twin model of the elevator system according to the elevator state data; based on the passenger state data and the elevator state data, use a pre-set first algorithm to make intelligent decisions on elevator dispatching, outputting a first dispatching strategy; based on the first dispatching strategy, apply a pre-set second algorithm to simulate elevator operating paths in the digital twin model, dynamically adjust the elevator dispatching strategy to reduce passenger waiting time and energy consumption, and obtain a second dispatching strategy; manage the elevators in the building according to the second dispatching strategy.

Images