An office building face recognition elevator control system

Abstract

The present invention discloses an office building face recognition elevator control system, comprising the following steps: by means of multimodal sensing devices installed at the entrance of the office building and in front of the elevator, the user's face, gait and behavioral action data are collected to generate the user's behavioral feature vector; the system performs instant identity authentication according to the vector, and dynamically allocates elevator authority based on the user's identity, historical behavior and real-time floor demand; the system also combines the user's historical usage data to intelligently predict the target floor, and allows the floor selection to be confirmed by gesture or voice; the elevator dispatching system utilizes a multivariable dynamic dispatching algorithm, comprehensively considers user behavior, elevator load, floor request, etc., and optimizes the elevator operation path in real time; with the support of a feedback mechanism, the system dynamically adjusts the dispatching strategy according to changes in user behavior, adapts to complex needs in the office building, and improves the elevator's operating efficiency and user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent building control, and in particular to an office building face recognition elevator control system. Background Art

[0002] In modern office building management, the efficiency and intelligence level of the elevator control system directly affect the user's travel experience and the management efficiency of the building. With the rapid development of artificial intelligence and Internet of Things technologies, traditional elevator control systems are gradually transforming towards intelligence and automation. Existing elevator control systems mainly rely on identity authentication methods such as physical buttons or card swiping. Although they have achieved elevator dispatch control to a certain extent, they still have problems such as low efficiency and poor user experience during peak hours and in complex environments.

[0003] Traditional elevator control systems usually rely on physical buttons, card swiping or password input for identity authentication. Although this method can meet basic needs in simple environments, in modern office buildings, especially during peak hours or in multi-user environments, the efficiency and security of identity authentication are particularly insufficient. Card swiping or button pressing requires users to interact directly with the device, which increases the complexity of operation and waiting time. This physical interaction method also has security risks such as information leakage or card loss, which can easily lead to unauthorized personnel entering the elevator or visiting specific floors, thereby posing security risks.

[0004] The existing elevator dispatching system lacks an intelligent dispatching mechanism. Most of them rely on a single floor request or elevator load to make simple dispatching decisions. This type of dispatching method performs poorly in an environment with complex demands and cannot be dynamically adjusted and optimized based on the user's identity, historical behavior or real-time demand. Elevator dispatching only focuses on the current floor request and elevator load, ignoring the priority differences between users. For example, VIP users and ordinary users have different needs for using elevators. The existing system finds it difficult to distinguish the priorities of such users, resulting in unreasonable resource allocation and the inability to achieve efficient resource utilization. This simple dispatching strategy is particularly inadequate during peak hours, with users waiting too long and the utilization rate of elevator resources also dropping significantly.

[0005] The elevator control system in the existing technology lacks the ability to integrate multimodal data. Usually, the system only relies on a single data source and cannot comprehensively judge the user's behavior status from multiple dimensions. Especially in modern office buildings, the diversity of user behavior and the complexity of needs make the scheduling strategy of a single data source very limited. The current elevator control system lacks the use of user historical behavior data and cannot make advance predictions based on the user's past usage habits or behavior patterns.

[0006] Another significant flaw of the existing system is the lack of an effective feedback mechanism and adaptive optimization capabilities. Most elevator systems do not analyze and utilize user feedback data and system operation results during operation. Even if bottlenecks or unreasonable situations occur in elevator scheduling, the system cannot automatically optimize the scheduling plan based on actual conditions. As a result, over the long term, the system's scheduling efficiency gradually decreases and cannot keep up with changes in user needs and environmental conditions. Especially during peak hours or emergencies, the system cannot optimize scheduling decisions by learning user behavior patterns or environmental changes, resulting in low scheduling efficiency and poor user experience.

[0007] Therefore, how to provide an office building face recognition elevator control system is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0008] One object of the present invention is to propose an office building face recognition elevator control system. The present invention makes full use of multimodal perception technology, intelligent scheduling algorithm and adaptive feedback optimization mechanism, and describes in detail the process of identity authentication, floor prediction and elevator scheduling optimization through multimodal data. It has the advantages of high identity authentication security, high scheduling efficiency, friendly user experience and strong system self-learning and optimization capabilities.

[0009] An office building face recognition elevator control system according to an embodiment of the present invention comprises the following steps:

[0010] S1. Install multimodal sensing cameras at the entrance of the office building and in front of the elevator to collect the user's facial, gait and behavioral action data. Perform the initial collection at the entrance and the secondary collection when approaching the elevator. Combine the multimodal data at the entrance and in front of the elevator to generate the user's behavioral feature vector;

[0011] S2. The system performs instant identity authentication based on the generated user behavior feature vector, dynamically allocates permissions based on user identity information, historical behavior, and real-time floor requirements, allows permission adjustment in special situations, and implements flexible permission management;

[0012] S3. The system intelligently predicts the user's target floor by analyzing the user's historical floor usage data and real-time behavior status, and allows the user to confirm and modify the floor through the behavior confirmation mechanism;

[0013] S4. The system adopts a situation-aware multivariate dynamic scheduling algorithm to monitor user behavior, elevator load, floor request density and user priority in real time. Through multivariate analysis, it dynamically optimizes the elevator scheduling strategy, prioritizes special needs, and realizes efficient allocation of resources.

[0014] S5. After the user enters the elevator, the system monitors the user's behavior in real time and dynamically optimizes the elevator's running path based on requests from other floors and the current elevator load. During peak hours and special needs, the system instantly adjusts the dispatching path to achieve priority dispatching;

[0015] S6. The system automatically adjusts future scheduling strategies and floor selections based on user feedback data, continuously updates user behavior patterns through a feedback mechanism, and optimizes scheduling strategies based on floor usage frequency to adapt to the dynamic needs of office buildings.

[0016] Optionally, the S1 specifically includes:

[0017] S11. Install multimodal sensing cameras at the entrance of office buildings and in front of elevators to collect user facial, gait and behavioral data. The equipment has automatic light adjustment and dynamic background noise suppression functions, and can automatically adjust camera parameters according to environmental conditions to adapt to data collection in different environments.

[0018] S12: When the user enters the office building entrance, the system performs initial data collection and obtains facial image data , gait feature data and behavioral action sequence data , and generate preliminary behavioral feature vectors through light optimization and dynamic filtering ;

[0019] S13. When the user approaches the elevator, the system performs secondary data collection to collect facial images. , gait characteristics and behavioral action sequences , and generate secondary behavior feature vectors through motion analysis processing ;

[0020] S14, the initial eigenvector With the quadratic eigenvector Fusion to generate a complete user behavior feature vector ,Adopting a weight adaptive adjustment algorithm based on user behavior patterns, the weight allocation is dynamically adjusted according to user behavior frequency, historical behavior and current environmental conditions, to achieve weighted fusion of different behavior feature vectors.

[0021] Optionally, the S2 specifically includes:

[0022] S21, the system is based on the user's complete behavior feature vector For identity authentication, a multimodal feature fusion model is introduced , adjust the weights of facial, gait and behavioral action data:

[0023] ;

[0024] Among them, F, G, and A represent facial, gait, and behavioral action data, respectively. , , The weights are dynamically adjusted based on environmental conditions and user behavior;

[0025] S22. After completing identity authentication, the system , historical behavior data and real-time floor demand Dynamically allocate permissions. By introducing a behavior prediction model, the system can predict the user's access needs in advance and allocate permissions to them in advance.

[0026] S23, the system uses the context perception module to adjust user permissions in special situations, combined with environmental sensors And real-time feedback adjustment permissions:

[0027] ;

[0028] in, is a special situation parameter, Feedback signal for external sensor;

[0029] S24. The system further introduces a permission validity period management mechanism, which sets the permission validity period based on the user's behavior pattern. While allocating permissions, the system dynamically adjusts the permission validity period based on the user's behavior frequency and access requirements;

[0030] S25. When multiple users access the elevator at the same time, the system solves the permission conflict problem through the multi-user concurrent processing module. The system dynamically adjusts the elevator scheduling strategy according to the user's priority and real-time behavior characteristics, so that users with higher priority can obtain priority permissions when resources are limited.

[0031] Optionally, the S3 specifically includes:

[0032] S31, the system analyzes the user's historical floor usage data and real-time behavior status, combined with the user's behavior feature vector Perform floor prediction. Based on the correlation between the user's behavior characteristics and the floor, the system predicts the user's target floor.

[0033] S32, the system optimizes floor prediction by identifying the user's behavior pattern and combining the user's frequently visited floors and regular visit time periods;

[0034] S33. After the floor prediction is completed, the system detects the user's confirmation and modification operations through the behavior confirmation mechanism, allowing the user to confirm and modify the target floor through gestures or voice. The confirmed floor information will be used as the final floor selection;

[0035] S34. The system dynamically adjusts the floor prediction based on real-time environmental data:

[0036] ;

[0037] in, represents the revised floor prediction result, represents the initial floor predicted by the system based on historical data and behavioral characteristics, is the correction factor, is a correction function based on environmental data, Represents real-time environmental data.

[0038] Optionally, the S4 specifically includes:

[0039] S41, the system processes user behavior status in real time , Elevator load , Floor Request Density and user permission priorities , and combined with historical behavior data Conduct trend forecasts and adjust scheduling strategies in advance to cope with possible peak periods and emergencies in the future;

[0040] S42, the system uses a multivariable dynamic scheduling algorithm based on real-time data obtained from multiple sources ,generates an intelligent scheduling strategy, which comprehensively analyzes user behavior, elevator load, floor request density, and user authority priority, so that the scheduling strategy dynamically adapts to changes in user demand and elevator load;

[0041] S43. Based on the analysis results, the system generates and adjusts the dispatching strategy so that the dispatching decision can adapt to the current user needs and environmental changes in real time and optimize the running order and stop floors of the elevators;

[0042] S44, when VIP users and emergency situations are detected, the system dynamically adjusts the priority to respond to special needs and dynamically adjusts the elevator operation plan. The priority processing mechanism increases the priority parameters to respond to high-priority users and emergency needs first;

[0043] S45. The system optimizes the path based on multi-dimensional real-time data, and optimizes the elevator's operating path by comprehensively considering the elevator load, floor request priority, request density and time sequence:

[0044] ;

[0045] in, Request weights for floors, The time order of the floor requests, is the elevator load condition, Request density for the floor, is the user's permission priority, is the balance coefficient.

[0046] Optionally, the S6 specifically includes:

[0047] S61. The system collects user behavior data in real time through multimodal sensing devices installed in elevators and office buildings. , Floor Request Density , Elevator load and environmental data , the data includes the user's behavioral characteristics inside and outside the elevator, the real-time elevator operation status and external environmental conditions, and the integrated data is transmitted to the central control system;

[0048] S62. Based on the collected feedback data, the system uses intelligent algorithms to analyze user behavior patterns, floor request change trends, and elevator load fluctuations, combining current feedback data and historical behavior data. , generate a user behavior trend prediction model, identify possible peak usage time periods and sudden changes in floor request density in advance, and dynamically adjust the scheduling plan based on the prediction results;

[0049] S63. Based on feedback data and prediction results, the system uses behavior pattern reconstruction technology to adjust user behavior patterns in real time , analyzes and optimizes user behavior characteristics through adaptive algorithms, and generates accurate user behavior models that combine historical behavior data and current feedback;

[0050] S64, the system is based on the reconstructed user behavior model and feedback data to generate real-time scheduling strategies , the intelligent decision-making engine comprehensively considers user behavior characteristics, elevator load and floor request priority, and dynamically adjusts the elevator operation order and stop floors;

[0051] S65. The system introduces a closed-loop feedback mechanism. After each elevator dispatch cycle, the system evaluates key indicators, identifies dispatch bottlenecks, and corrects dispatch strategies through adaptive optimization algorithms. The dispatch strategies are continuously optimized as user behavior and environmental factors change.

[0052] Optional modules include:

[0053] Multimodal perception module: collects user's facial, gait, and behavioral data at the entrance of the office building and in front of the elevator;

[0054] Identity verification module: Based on the behavioral feature vector generated by the user, identity verification is performed through a multimodal feature fusion model;

[0055] Permission management module: The system dynamically allocates access rights based on the user's identity information, historical behavior and real-time needs, flexibly manages the user's floor access rights, and adjusts permissions in special situations based on the environmental perception module to ensure flexible allocation of resources;

[0056] Floor prediction module: Based on the user's historical floor usage data and real-time behavior status, combined with the correlation between behavior characteristics and floors, predict the user's target floor;

[0057] Intelligent scheduling module: The system uses a multi-variable dynamic scheduling algorithm to monitor user behavior, elevator load, floor request density and user authority priority in real time, and dynamically optimizes the elevator scheduling strategy;

[0058] Path optimization module: The system calculates the elevator's operating path through an optimization algorithm based on user behavior patterns, elevator load, floor request density and priority, making scheduling decisions flexible to respond to changes in user demand;

[0059] Feedback learning module: During the elevator dispatching process, the system adjusts the future dispatching strategy in real time based on user feedback data. Through the adaptive optimization algorithm, the system can dynamically update the user's behavior pattern and dispatching strategy, realize closed-loop feedback and self-learning, and gradually improve the system's intelligence level and dispatching efficiency;

[0060] Behavior pattern reconstruction module: Based on user feedback data and trend prediction results, the system adjusts the user's behavior pattern in real time and generates a behavior feature model that combines historical data and current feedback;

[0061] Environmental perception module: monitors real-time environmental data in the office building, including crowd density, weather, and external environmental conditions;

[0062] Central control module: integrates and coordinates all functional modules, and uniformly manages data collection, analysis, decision-making and scheduling execution.

[0063] The beneficial effects of the present invention are:

[0064] The present invention introduces multimodal sensing technology, which enables the system to perform multi-dimensional identity authentication by collecting the user's facial, gait and behavioral action data. This not only improves the security of identity authentication, but also avoids the cumbersome operations of traditional card swiping or button interaction, and provides a more convenient contactless recognition experience. Through intelligent scheduling algorithms and multivariate analysis, combined with user behavioral characteristics, elevator load and floor request density, the scheduling strategy can be dynamically generated and adjusted. Compared with traditional systems that only rely on a single factor for scheduling, the scheduling algorithm of the present invention comprehensively considers data from multiple dimensions and can flexibly respond to different user needs. Especially in peak periods or emergency situations, it can give priority to responding to high-priority users, significantly reduce user waiting time, and improve elevator resource utilization efficiency.

[0065] The system of the present invention can continuously monitor changes in user behavior and scheduling results through a closed-loop feedback mechanism and an adaptive optimization algorithm, and realize self-learning and optimization of scheduling strategies. The system can automatically adjust the scheduling strategy according to the user's behavior patterns and real-time needs to maintain long-term and efficient operation of the system, thereby solving the defect that the scheduling plan cannot be adjusted according to actual conditions in the existing system. Through self-learning and self-optimization mechanisms, the system can continuously improve scheduling efficiency and user experience in long-term operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0067] Figure 1 The present invention is a flowchart of an office building face recognition elevator control system. DETAILED DESCRIPTION

[0068] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0069] refer to Figure 1 , an office building face recognition elevator control system, comprising the following steps:

[0070] S1. Install multimodal sensing cameras at the entrance of the office building and in front of the elevator to collect the user's facial, gait and behavioral action data. Perform the initial collection at the entrance and the secondary collection when approaching the elevator. Combine the multimodal data at the entrance and in front of the elevator to generate the user's behavioral feature vector;

[0071] S2. The system performs instant identity authentication based on the generated user behavior feature vector, dynamically allocates permissions based on user identity information, historical behavior, and real-time floor requirements, allows permission adjustment in special situations, and implements flexible permission management;

[0072] S3. The system intelligently predicts the user's target floor by analyzing the user's historical floor usage data and real-time behavior status, and allows the user to confirm and modify the floor through the behavior confirmation mechanism;

[0073] S4. The system adopts a situation-aware multivariate dynamic scheduling algorithm to monitor user behavior, elevator load, floor request density and user priority in real time. Through multivariate analysis, it dynamically optimizes the elevator scheduling strategy, prioritizes special needs, and realizes efficient allocation of resources.

[0074] S5. After the user enters the elevator, the system monitors the user's behavior in real time and dynamically optimizes the elevator's running path based on requests from other floors and the current elevator load. During peak hours and special needs, the system instantly adjusts the dispatching path to achieve priority dispatching;

[0075] S6. The system automatically adjusts future scheduling strategies and floor selections based on user feedback data, continuously updates user behavior patterns through a feedback mechanism, and optimizes scheduling strategies based on floor usage frequency to adapt to the dynamic needs of office buildings.

[0076] In this implementation manner, the S1 specifically includes:

[0077] S11. Install multimodal sensing cameras at the entrance of office buildings and in front of elevators to collect user facial, gait and behavioral data. The equipment has automatic light adjustment and dynamic background noise suppression functions, and can automatically adjust camera parameters according to environmental conditions to adapt to data collection in different environments.

[0078] S12: When the user enters the office building entrance, the system performs initial data collection and obtains facial image data , gait feature data and behavioral action sequence data , and generate preliminary behavioral feature vectors through light optimization and dynamic filtering ;

[0079] S13. When the user approaches the elevator, the system performs secondary data collection to collect facial images. , gait characteristics and behavioral action sequences , and generate secondary behavior feature vectors through motion analysis processing ;

[0080] S14, the initial eigenvector With the quadratic eigenvector Fusion to generate a complete user behavior feature vector ,Adopting a weight adaptive adjustment algorithm based on user behavior patterns, the weight allocation is dynamically adjusted according to user behavior frequency, historical behavior and current environmental conditions, to achieve weighted fusion of different behavior feature vectors.

[0081] In this implementation, S2 specifically includes:

[0082] S21, the system is based on the user's complete behavior feature vector For identity authentication, a multimodal feature fusion model is introduced , adjust the weights of facial, gait and behavioral action data:

[0083] ;

[0084] Among them, F, G, and A represent facial, gait, and behavioral action data, respectively. , , The weights are dynamically adjusted based on environmental conditions and user behavior;

[0085] S22. After completing identity authentication, the system , historical behavior data and real-time floor demand Dynamically allocate permissions. By introducing a behavior prediction model, the system can predict the user's access needs in advance and allocate permissions to them in advance.

[0086] S23, the system uses the context perception module to adjust user permissions in special situations, combined with environmental sensors And real-time feedback adjustment permissions:

[0087] ;

[0088] in, is a special situation parameter, Feedback signal for external sensor;

[0089] S24. The system further introduces a permission validity period management mechanism, which sets the permission validity period based on the user's behavior pattern. While allocating permissions, the system dynamically adjusts the permission validity period based on the user's behavior frequency and access requirements;

[0090] S25. When multiple users access the elevator at the same time, the system solves the permission conflict problem through the multi-user concurrent processing module. The system dynamically adjusts the elevator scheduling strategy according to the user's priority and real-time behavior characteristics, so that users with higher priority can obtain priority permissions when resources are limited.

[0091] In this implementation, S3 specifically includes:

[0092] S31, the system analyzes the user's historical floor usage data and real-time behavior status, combined with the user's behavior feature vector Perform floor prediction. Based on the correlation between the user's behavior characteristics and the floor, the system predicts the user's target floor.

[0093] S32, the system optimizes floor prediction by identifying the user's behavior pattern and combining the user's frequently visited floors and regular visit time periods;

[0094] S33. After the floor prediction is completed, the system detects the user's confirmation and modification operations through the behavior confirmation mechanism, allowing the user to confirm and modify the target floor through gestures or voice. The confirmed floor information will be used as the final floor selection;

[0095] S34. The system dynamically adjusts the floor prediction based on real-time environmental data:

[0096] ;

[0097] in, represents the revised floor prediction result, represents the initial floor predicted by the system based on historical data and behavioral characteristics, is the correction factor, is a correction function based on environmental data, Represents real-time environmental data.

[0098] In this implementation, S4 specifically includes:

[0099] S41, the system processes user behavior status in real time , Elevator load , Floor Request Density and user permission priorities , and combined with historical behavior data Conduct trend forecasts and adjust scheduling strategies in advance to cope with possible peak periods and emergencies in the future;

[0100] S42, the system uses a multivariable dynamic scheduling algorithm based on real-time data obtained from multiple sources ,generates an intelligent scheduling strategy, which comprehensively analyzes user behavior, elevator load, floor request density, and user authority priority, so that the scheduling strategy dynamically adapts to changes in user demand and elevator load;

[0101] S43. Based on the analysis results, the system generates and adjusts the dispatching strategy so that the dispatching decision can adapt to the current user needs and environmental changes in real time and optimize the running order and stop floors of the elevators;

[0102] S44, when VIP users and emergency situations are detected, the system dynamically adjusts the priority to respond to special needs and dynamically adjusts the elevator operation plan. The priority processing mechanism increases the priority parameters to respond to high-priority users and emergency needs first;

[0103] S45. The system optimizes the path based on multi-dimensional real-time data, and optimizes the elevator's operating path by comprehensively considering the elevator load, floor request priority, request density and time sequence:

[0104] ;

[0105] in, Request weights for floors, The time order of the floor requests, is the elevator load condition, Request density for the floor, is the user's permission priority, is the balance coefficient.

[0106] In this implementation manner, S6 specifically includes:

[0107] S61. The system collects user behavior data in real time through multimodal sensing devices installed in elevators and office buildings. , Floor Request Density , Elevator load and environmental data , the data includes the user's behavioral characteristics inside and outside the elevator, the real-time elevator operation status and external environmental conditions, and the integrated data is transmitted to the central control system;

[0108] S62. Based on the collected feedback data, the system uses intelligent algorithms to analyze user behavior patterns, floor request change trends, and elevator load fluctuations, combining current feedback data and historical behavior data. , generate a user behavior trend prediction model, identify possible peak usage time periods and sudden changes in floor request density in advance, and dynamically adjust the scheduling plan based on the prediction results;

[0109] S63. Based on feedback data and prediction results, the system uses behavior pattern reconstruction technology to adjust user behavior patterns in real time , analyzes and optimizes user behavior characteristics through adaptive algorithms, and generates accurate user behavior models that combine historical behavior data and current feedback;

[0110] S64, the system is based on the reconstructed user behavior model and feedback data to generate real-time scheduling strategies , the intelligent decision-making engine comprehensively considers user behavior characteristics, elevator load and floor request priority, and dynamically adjusts the elevator operation order and stop floors;

[0111] S65. The system introduces a closed-loop feedback mechanism. After each elevator dispatch cycle, the system evaluates key indicators, identifies dispatch bottlenecks, and corrects dispatch strategies through adaptive optimization algorithms. The dispatch strategies are continuously optimized as user behavior and environmental factors change.

[0112] In this embodiment, the following modules are included:

[0113] Multimodal perception module: collects user's facial, gait, and behavioral data at the entrance of the office building and in front of the elevator;

[0114] Identity verification module: Based on the behavioral feature vector generated by the user, identity verification is performed through a multimodal feature fusion model;

[0115] Permission management module: The system dynamically allocates access rights based on the user's identity information, historical behavior and real-time needs, flexibly manages the user's floor access rights, and adjusts permissions in special situations based on the environmental perception module to ensure flexible allocation of resources;

[0116] Floor prediction module: Based on the user's historical floor usage data and real-time behavior status, combined with the correlation between behavior characteristics and floors, predict the user's target floor;

[0117] Intelligent scheduling module: The system uses a multi-variable dynamic scheduling algorithm to monitor user behavior, elevator load, floor request density and user authority priority in real time, and dynamically optimizes the elevator scheduling strategy;

[0118] Path optimization module: The system calculates the elevator's operating path through an optimization algorithm based on user behavior patterns, elevator load, floor request density and priority, making scheduling decisions flexible to respond to changes in user demand;

[0119] Feedback learning module: During the elevator dispatching process, the system adjusts the future dispatching strategy in real time based on user feedback data. Through the adaptive optimization algorithm, the system can dynamically update the user's behavior pattern and dispatching strategy, realize closed-loop feedback and self-learning, and gradually improve the system's intelligence level and dispatching efficiency;

[0120] Behavior pattern reconstruction module: Based on user feedback data and trend prediction results, the system adjusts the user's behavior pattern in real time and generates a behavior feature model that combines historical data and current feedback;

[0121] Environmental perception module: monitors real-time environmental data in the office building, including crowd density, weather, and external environmental conditions;

[0122] Central control module: integrates and coordinates all functional modules, and uniformly manages data collection, analysis, decision-making and scheduling execution.

[0123] Embodiment 1:

[0124] In order to verify the feasibility of the present invention in implementation, the present invention is applied to the face recognition elevator control system of the "XX Building" office building in Beijing. The building has a total of 20 floors. During the morning and evening peak hours every day, there are a large number of employees in the building, and the elevator load pressure is extremely high. The traditional elevator dispatching system is difficult to effectively cope with the peak demand. The system collects the user's facial, gait and behavioral action data in real time through multimodal sensing devices installed at the building entrance and in front of the elevator, providing intelligent support for elevator dispatching.

[0125] In the past, the office building's elevator system relied on users to swipe cards or press buttons to select floors for operation. During the morning rush hour, users had to queue up to wait for the elevator, and the average waiting time was long. In particular, the elevator was unable to intelligently assign priorities based on user needs, resulting in unreasonable resource allocation. Elevators often stopped at multiple floors, extending the operating time of each run, seriously affecting elevator efficiency and user experience.

[0126] After applying the system of the present invention, users no longer need to carry access cards when entering the building, nor do they need to manually select floors. Through multimodal data collection such as facial recognition and gait analysis, the system can quickly verify the user's identity and predict their target floor. After the user is identified by the camera equipment in front of the elevator, the system dynamically allocates permissions based on the user's historical behavior and real-time floor requirements, and the floor selection can be automatically completed without additional operations.

[0127] The system uses an intelligent scheduling algorithm and dynamically adjusts the elevator scheduling strategy through real-time monitoring and analysis of multi-dimensional data, such as user behavior characteristics, elevator load, floor request density and user permission priority. For example, during the morning rush hour, the system recognizes that multiple VIP users have entered the building. The system prioritizes elevators to serve these users and flexibly schedules the elevator order for other users based on the elevator load. This priority-based scheduling method greatly improves the operating efficiency of the elevator and reduces user waiting time.

[0128] The dispatching effect of the system was verified through data comparison. During the one-month test, we observed the use of elevators by users during peak hours in the morning and evening, and especially compared the performance before and after the introduction of the system. The test time was from September 1, 2023 to September 30, 2023, and the test location was "XX Building". In the traditional system, the average waiting time for users during peak hours in the morning is 5 minutes, and the elevator running time is 10 minutes, and each operation requires frequent stops at multiple floors. After the introduction of this system, the average waiting time of users was reduced to 2 minutes, a 60% reduction compared to the traditional system; the elevator running time was also reduced from the original 10 minutes to 6 minutes, and the overall operating efficiency was improved by 40%. These improvements have significantly improved the travel experience of users in the building, especially during peak hours, when elevator dispatch has become smoother.

[0129] In order to further demonstrate the system's emergency response capabilities, a fire alarm simulation exercise was conducted on September 15. After the system recognized the fire alarm signal through the environmental perception module, it quickly adjusted the elevator dispatching strategy, prioritized the dispatching of elevators to safe floors, restricted the elevator access rights of ordinary users, and ensured safe evacuation. During the exercise, the system completed the adjustment of elevator dispatching within 30 seconds, and all personnel were safely evacuated within 3 minutes. Traditional elevator systems cannot achieve such a fast response in emergency situations.

[0130] The system also has an adaptive feedback mechanism that can continuously optimize the dispatch strategy by collecting user behavior data, real-time elevator operation status and dispatch results. In the first few days of the test, the system gradually learned and adjusted the dispatch rules. By the third week, the average waiting time for users was further shortened to 1.5 minutes, and the accuracy of floor prediction reached 98%. Through the system's self-learning and optimization capabilities, the dispatch strategy can become more intelligent over time and gradually adapt to different time periods and user habits.

[0131] Table 1 Test data comparison table

[0132]

[0133] It can be seen from the data that the system of the present invention has significantly improved the elevator dispatching efficiency and user experience in practical applications. Through intelligent scheduling algorithms and multimodal data collection, the system shortens the waiting time from the traditional 5 minutes to 1.5 minutes, reduces the elevator running time to about 5 minutes, and increases the floor prediction accuracy from 70% to 98%. In emergency situations, the system can give priority to emergency needs within 30 seconds and complete user evacuation within 3 minutes. Overall, the system of the present invention performs well in improving elevator resource utilization, optimizing scheduling paths, and handling emergencies, which greatly solves the efficiency and safety problems in traditional systems.

[0134] The present invention not only solves the problems of long waiting time and complex floor selection in the traditional elevator dispatching system, but also greatly improves the resource utilization of the elevator. Through multimodal sensing technology, intelligent dispatching algorithm and adaptive feedback mechanism, elevator dispatching becomes more flexible and efficient, providing users with a better user experience. Test data shows that while improving dispatching efficiency, this system ensures the stability and adaptability of the elevator system in different usage scenarios, achieving an important breakthrough in intelligent building management.

[0135] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A facial recognition elevator control method for office buildings, characterized in that: The steps include: S1. Install multimodal sensing cameras at the entrance of the office building and in front of the elevator to collect the user's facial, gait and behavioral action data. Perform the initial collection at the entrance and the secondary collection when approaching the elevator. Combine the multimodal data at the entrance and in front of the elevator to generate the user's behavioral feature vector; S2. The system performs instant identity authentication based on the generated user behavior feature vector, dynamically allocates permissions based on user identity information, historical behavior, and real-time floor requirements, allows permission adjustment in special situations, and implements flexible permission management; S3. The system intelligently predicts the user's target floor by analyzing the user's historical floor usage data and real-time behavior status, and allows the user to confirm and modify the floor through the behavior confirmation mechanism; S4. The system adopts a situation-aware multivariate dynamic scheduling algorithm to monitor user behavior, elevator load, floor request density and user priority in real time. Through multivariate analysis, it dynamically optimizes the elevator scheduling strategy, prioritizes special needs, and realizes efficient allocation of resources. S5. After the user enters the elevator, the system monitors the user's behavior in real time and dynamically optimizes the elevator's running path based on requests from other floors and the current elevator load. During peak hours and special needs, the system instantly adjusts the dispatching path to achieve priority dispatching; S6. The system automatically adjusts future scheduling strategies and floor selection based on user feedback data, continuously updates user behavior patterns through the feedback mechanism, and optimizes scheduling strategies based on floor usage frequency to adapt to the dynamic needs of office buildings; The S1 specifically includes: S11. Install multimodal sensing cameras at the entrance of office buildings and in front of elevators to collect user facial, gait and behavioral data. The equipment has automatic light adjustment and dynamic background noise suppression functions, and can automatically adjust camera parameters according to environmental conditions to adapt to data collection in different environments. S12: When the user enters the office building entrance, the system performs initial data collection and obtains facial image data , gait feature data and behavioral action sequence data , and generate preliminary behavioral feature vectors through light optimization and dynamic filtering ; S13. When the user approaches the elevator, the system performs secondary data collection to collect facial images. , gait characteristics and behavioral action sequences , and generate secondary behavior feature vectors through motion analysis processing ; S14, the initial eigenvector With the quadratic eigenvector Fusion to generate a complete user behavior feature vector ,Adopt a weight adaptive adjustment algorithm based on user behavior patterns,dynamically adjust the weight allocation according to user behavior frequency,historical behavior and current environmental conditions,and achieve weighted fusion of different behavior feature vectors; The S2 specifically includes: S21, the system based on the user's complete behavior feature vector For identity authentication, a multimodal feature fusion model is introduced , adjust the weights of facial, gait and behavioral action data: ； Among them, F, G, and A represent facial, gait, and behavioral action data, respectively. , , The weights are dynamically adjusted based on environmental conditions and user behavior; S22. After completing identity authentication, the system , historical behavior data and real-time floor demand Dynamically allocate permissions. By introducing a behavior prediction model, the system can predict the user's access needs in advance and allocate permissions to them in advance. S23, the system uses the context perception module to adjust user permissions in special situations, combined with environmental sensors And real-time feedback adjustment permissions: ； in, is the adjusted user permission value, and is the weight coefficient, is a special situation parameter, is the feedback signal of the external sensor, For external environment data; S24. The system further introduces a permission validity period management mechanism, which sets the permission validity period based on the user's behavior pattern. While allocating permissions, the system dynamically adjusts the permission validity period based on the user's behavior frequency and access requirements; S25. When multiple users access the elevator at the same time, the system solves the permission conflict problem through the multi-user concurrent processing module. The system dynamically adjusts the elevator scheduling strategy according to the user's priority and real-time behavior characteristics, so that users with higher priority can obtain permissions first when resources are limited; The S3 specifically includes: S31, the system analyzes the user's historical floor usage data and real-time behavior status, combined with the user's behavior feature vector Perform floor prediction. Based on the correlation between the user's behavior characteristics and the floor, the system predicts the user's target floor. S32, the system optimizes floor prediction by identifying the user's behavior pattern and combining the user's frequently visited floors and regular visit time periods; S33. After the floor prediction is completed, the system detects the user's confirmation and modification operations through the behavior confirmation mechanism, allowing the user to confirm and modify the target floor through gestures or voice. The confirmed floor information will be used as the final floor selection; S34. The system dynamically adjusts the floor prediction based on real-time environmental data: ； in, represents the revised floor prediction result, represents the initial floor predicted by the system based on historical data and behavioral characteristics, is the correction factor, is a correction function based on environmental data, Represents real-time environmental data; The S4 specifically includes: S41, the system processes user behavior status in real time , Elevator load , Floor Request Density and user permission priorities , and combined with historical behavior data Conduct trend forecasts and adjust scheduling strategies in advance to cope with possible peak periods and emergencies in the future; S42, the system uses a multivariable dynamic scheduling algorithm based on real-time data obtained from multiple sources ,generates an intelligent scheduling strategy, which comprehensively analyzes user behavior, elevator load, floor request density, and user authority priority, so that the scheduling strategy dynamically adapts to changes in user demand and elevator load; S43. Based on the analysis results, the system generates and adjusts the dispatching strategy so that the dispatching decision can adapt to the current user needs and environmental changes in real time and optimize the running order and stop floors of the elevators; S44, when VIP users and emergency situations are detected, the system dynamically adjusts the priority to respond to special needs and dynamically adjusts the elevator operation plan. The priority processing mechanism increases the priority parameters to respond to high-priority users and emergency needs first; S45. The system optimizes the path based on multi-dimensional real-time data, and optimizes the elevator's operating path by comprehensively considering the elevator load, floor request priority, request density and time sequence: ； in, Request weights for floors, The time order of the floor requests, is the elevator load condition, Request density for the floor, is the user's permission priority, is the balance coefficient; The S6 specifically includes: S61. The system collects user behavior data in real time through multimodal sensing devices installed in elevators and office buildings. , Floor Request Density , Elevator load and environmental data , the data includes the user's behavioral characteristics inside and outside the elevator, the real-time elevator operation status and external environmental conditions, and the integrated data is transmitted to the central control system; S62. Based on the collected feedback data, the system uses intelligent algorithms to analyze user behavior patterns, floor request change trends, and elevator load fluctuations, combining current feedback data and historical behavior data. , generate a user behavior trend prediction model, identify possible peak usage time periods and sudden floor request density changes in advance, and dynamically adjust the scheduling plan based on the prediction results; S63. Based on feedback data and prediction results, the system uses behavior pattern reconstruction technology to adjust user behavior patterns in real time , analyzes and optimizes user behavior characteristics through adaptive algorithms, and generates accurate user behavior models that combine historical behavior data and current feedback; S64, the system is based on the reconstructed user behavior model and feedback data to generate real-time scheduling strategies , the intelligent decision-making engine comprehensively considers user behavior characteristics, elevator load and floor request priority, and dynamically adjusts the elevator operation order and stop floors; S65. The system introduces a closed-loop feedback mechanism. After each elevator dispatch cycle, the system evaluates key indicators, identifies dispatch bottlenecks, and corrects dispatch strategies through adaptive optimization algorithms. The dispatch strategies are continuously optimized as user behavior and environmental factors change.

2. An office building face recognition elevator control system, executing the office building face recognition elevator control method according to claim 1, characterized in that: Includes the following modules: Multimodal perception module: collects user's facial, gait, and behavioral data at the entrance of the office building and in front of the elevator; Identity verification module: Based on the behavioral feature vector generated by the user, identity verification is performed through a multimodal feature fusion model; Permission management module: The system dynamically allocates access rights based on the user's identity information, historical behavior and real-time needs, flexibly manages the user's floor access rights, and adjusts permissions in special situations based on the environmental perception module to ensure flexible allocation of resources; Floor prediction module: Based on the user's historical floor usage data and real-time behavior status, combined with the correlation between behavior characteristics and floors, predict the user's target floor; Intelligent scheduling module: The system uses a multi-variable dynamic scheduling algorithm to monitor user behavior, elevator load, floor request density and user authority priority in real time, and dynamically optimizes the elevator scheduling strategy; Path optimization module: The system calculates the elevator's operating path through an optimization algorithm based on user behavior patterns, elevator load, floor request density and priority, making scheduling decisions flexible to respond to changes in user demand; Feedback learning module: During the elevator dispatching process, the system adjusts the future dispatching strategy in real time based on user feedback data. Through the adaptive optimization algorithm, the system can dynamically update the user's behavior pattern and dispatching strategy, realize closed-loop feedback and self-learning, and gradually improve the system's intelligence level and dispatching efficiency; Behavior pattern reconstruction module: Based on user feedback data and trend prediction results, the system adjusts the user's behavior pattern in real time and generates a behavior feature model that combines historical data and current feedback; Environmental perception module: monitors real-time environmental data in the office building, including crowd density, weather, and external environmental conditions; Central control module: integrates and coordinates all functional modules, and uniformly manages data collection, analysis, decision-making and scheduling execution.

Images