Data center elevator control method and related equipment

Abstract

The embodiment of the invention discloses a control method for a data center elevator and related equipment, and aims to solve the problem that invalid stopping is likely to occur in an existing elevator control mode. The method comprises the steps that when an elevator taking request sent by a target passenger for a data center elevator is received, an internal image of a lift car of the data center elevator is obtained; the available space and the remaining passenger number of the elevator in the data center are determined according to the internal images of the elevator car; acquiring emotion information of the target passenger; and generating a control instruction of the data center elevator according to the available space, the remaining passenger number and the emotion information, wherein the control instruction is used for indicating whether the data center elevator stops on the floor where the target passenger is located.

Description

Technical Field

[0001] This application relates to the field of elevator control technology, and in particular to a control method and related equipment for a data center elevator. Background Technology

[0002] Currently, existing elevator control methods still primarily rely on passenger-initiated operation, where passengers issue call commands via external up / down buttons or touch panels. Once the elevator receives a call command, it will proceed to the appropriate floor and stop according to the scheduling algorithm, provided that safety limits such as overload or exceeding the rated number of passengers are not triggered.

[0003] However, in actual operation, especially in scenarios with high passenger flow or where large items need to be transported, existing elevator control methods are prone to ineffective stopping: the elevator may not be overloaded or exceed its rated capacity, but the usable space inside the elevator car may be nearly saturated due to dense passenger traffic or the presence of large items. In this situation, the elevator will still respond to external calls and stop, but passengers waiting outside often cannot enter due to the crowded interior or choose to give up. This process does not achieve effective passenger transport, yet it completes a series of operations including deceleration, stopping, opening and closing, and then accelerating again. This ineffective stopping significantly reduces elevator operating efficiency and increases energy consumption and equipment wear.

[0004] Therefore, how to effectively reduce unnecessary parking, improve operational efficiency, and reduce energy consumption is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This application provides a control method for a data center elevator to solve the problem of invalid stops that easily occur in existing elevator control methods.

[0006] This application also provides a control device for a data center elevator, an electronic device, a computer-readable storage medium, and a computer program product.

[0007] The embodiments of this application adopt the following technical solutions: In a first aspect, embodiments of this application provide a control method for a data center elevator, including: When a passenger sends a request to ride the data center elevator, an image of the interior of the data center elevator car is obtained. The available space and remaining passenger capacity of the data center elevator are determined based on the images of the elevator car's interior. Obtain the emotional information of the target passenger; The control instructions for the data center elevator are generated based on the available space, the remaining number of passengers, and the emotional information. The control instructions are used to instruct the data center elevator whether to stop at the floor where the target passenger is located.

[0008] Secondly, embodiments of this application provide a control device for a data center elevator, including an image acquisition module, an information determination module, an emotion acquisition module, and an instruction generation module, wherein: The image acquisition module is used to acquire an image of the interior of the data center elevator when it receives a ride request from a target passenger for the data center elevator. The information determination module is used to determine the available space and remaining passenger capacity of the data center elevator based on the images inside the elevator car. An emotion acquisition module is used to acquire the emotional information of the target passenger; The instruction generation module is used to generate control instructions for the data center elevator based on the available space, the remaining number of passengers, and the emotional information. The control instructions are used to instruct whether the data center elevator should stop at the floor where the target passenger is located.

[0009] Thirdly, embodiments of this application provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the control method for a data center elevator as described above.

[0010] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the control method for a data center elevator as described above.

[0011] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the control method for a data center elevator as described above.

[0012] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects: The method provided in this application embodiment acquires real-time images of the interior of the data center elevator car to dynamically assess available space and remaining passenger capacity, and integrates emotional information of target passengers for multi-dimensional intelligent decision-making. This can effectively reduce the number of invalid stops of the data center elevator, improve the operating efficiency of the data center elevator, and reduce the energy consumption and equipment wear of the data center elevator. Attached Figure Description

[0013] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A schematic diagram of the framework of an elevator control decision system for a data center elevator provided in this application embodiment; Figure 2 A framework diagram of the multi-task intelligent recognition terminal in the elevator control decision system provided in the embodiments of this application; Figure 3 A schematic diagram illustrating the implementation process of a control method for a data center elevator provided in an embodiment of this application; Figure 4 A schematic diagram of the overall network structure of YOLOv5 provided in this application embodiment; Figure 5 A schematic diagram illustrating the implementation process of a method for determining the available space of a data center elevator, provided in an embodiment of this application; Figure 6 A schematic diagram of a CNN-based multi-task learning model structure is provided in an embodiment of this application; Figure 7 A schematic diagram illustrating an application process of the method provided in the embodiments of this application in practice; Figure 8 This application provides a schematic diagram of the specific structure of a control device for a data center elevator. Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0015] It should be understood that the training and prediction processes of the AI ​​models involved in the various embodiments of this specification all adhere to multiple legal and compliant principles, including legal data sources, compliant data content, compliant data governance, compliant training objectives and schemes, compliant training processes, compliant training environments and tools, and compliant ethical verification of training results, and comply with the requirements of Article 5 of the Patent Law. Among them: Data source legitimacy: All datasets used for AI model training were obtained through legal means, covering three categories: publicly authorized data, data authorized by partners, and self-collected compliant data. Publicly authorized data comes from compliant data sources following open-source licenses such as Apache 2.0, with complete copyright attribution and authorization scope clearly marked, and no unauthorized open-source code or data reuse. Data authorized by partners has been subject to formal data usage agreements, clearly defining the scope, duration, and confidentiality obligations, and possessing a complete authorization chain. For self-collected data involving personal information, strict informed consent procedures have been followed, and anonymization processes (including but not limited to field masking, feature anonymization, and differential privacy technology applications) have been implemented to remove personally identifiable information, fully complying with the requirements of relevant laws and regulations such as the "Interim Measures for the Administration of Generative Artificial Intelligence Services" and the "Personal Information Protection Law."

[0016] Data content compliance: The AI ​​model's dataset undergoes multiple screenings and cleaning processes to remove all content that may violate social morality or harm public interests. It contains no obscene, pornographic, violent, discriminatory, or information that endangers national or public safety, nor does it involve the illegal acquisition or use of genetic resources. For data in sensitive fields (such as healthcare and finance), an additional privacy-preserving computation module (including federated learning and secure multi-party computation technologies) ensures that the data is "usable but not visible," avoiding compliance risks during the original data transmission process and ensuring that the data application scenarios and uses comply with public order and good morals and industry regulatory requirements.

[0017] Data governance norms: A complete data traceability system is established during the AI ​​model training process to automatically record the source, collection time, annotation process, cleaning rules, and permission allocation of training data, generating traceable compliance reports to ensure that the data is verifiable throughout its entire lifecycle. The dataset annotation process for AI models is completed by a professional human R&D team, clearly defining the proportion of human creative contributions and avoiding reliance on AI-generated data that has not undergone substantial human modification, thus meeting the examination requirements for "human main contributions" in AI patent applications.

[0018] Training objectives and plans are compliant: The AI ​​model training objective focuses on improving the operational efficiency and user experience of data center elevators. Specifically, it is used to intelligently identify the emotional state of passengers and detect the number of people and available space in the elevator car in real time. The training data and model applications are strictly limited to the aforementioned specific business scenarios. The aim is to reduce unnecessary stops by optimizing scheduling decisions, thereby achieving energy conservation, emission reduction, and extending equipment life. The training scheme and final output results do not violate any mandatory provisions of laws and administrative regulations, do not harm the public interest or the legitimate rights and interests of others, and do not pose any potential risks of being used for illegal activities, privacy violations, or public safety disruptions. The model strictly adheres to the ethical principle of "intelligent for good."

[0019] Training process compliance: A closed-loop training framework is adopted to ensure compliance and controllability of the training process. The specific process is as follows: First, training samples are obtained through compliant data sources. After the aforementioned data cleaning and desensitization, they are input into the neural network model to generate preliminary training results. Second, an expert system is introduced to verify the preliminary results. Based on preset rules and human expert experience, the feasibility of the results is evaluated, and outputs that may pose ethical risks or compliance hazards are corrected (such as removing decision-making logic that violates public order and good morals, and adjusting model parameters that do not comply with safety regulations). Finally, the loss function weights are dynamically optimized based on expert system feedback to strengthen the model's learning of compliant results, avoid overfitting errors or non-compliant labels, and form a closed-loop control of "data input - model training - expert verification - parameter optimization - result feedback" to ensure that the entire training process complies with A5 ethical review requirements.

[0020] Training environment and tool compliance: AI model training is implemented using nationally licensed chips and a compliant training platform. All open-source frameworks and components used in the training process have obtained their corresponding licenses, and copyright statements and patent citation information are fully retained, with no instances of infringement or reuse. The training environment is built using virtual devices (containers / virtual machines) with fixed random seeds and initial parameter configurations to ensure the reproducibility of the training process. Furthermore, through access control and operation log recording, risks such as data leakage and parameter tampering during training are prevented, ensuring the security and compliance of the training process.

[0021] Training results ethical verification compliance: After the model is trained, it undergoes additional third-party ethical compliance assessment and algorithm filing review to verify that the model output does not violate social morality or harm public interests. For potentially sensitive scenarios (such as public services and intelligent decision-making), a special result verification mechanism is established to ensure that the model always complies with Article 5 of the Patent Law and relevant laws and regulations in practical applications.

[0022] In summary, the data and training process used in the AI ​​model of this specification strictly comply with the relevant provisions of Article 5 of the Patent Law and the Patent Examination Guidelines (2023 Edition), and there are no violations of laws, social ethics, public interests, or illegal use of genetic resources. It fully meets the compliance requirements for patent authorization.

[0023] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0024] This application addresses the operational scenario of data center elevators. These elevators, in addition to carrying passengers, may also carry critical data center infrastructure equipment, exacerbating the conflict between space occupancy, passenger congestion, and operational efficiency. Especially in scenarios with high passenger volume or the need to transport large items, existing elevator control methods are prone to ineffective stops: the elevator may not be overloaded or exceed its rated capacity, but the available space inside the elevator car may be nearly saturated due to dense passenger traffic or the placement of large critical data center infrastructure equipment. In this situation, the elevator will still respond to external calls and stop, but passengers waiting outside often cannot enter due to overcrowding or choose to give up. This process does not achieve effective personnel transport, yet it completes a series of operations including deceleration, stopping, opening, closing, and re-acceleration. This ineffective stop significantly reduces elevator operating efficiency and increases energy consumption and equipment wear.

[0025] To address the problems of ineffective stops that easily occur in existing elevator control methods, resulting in low elevator operating efficiency, high energy consumption, and severe wear, this application provides a control method for a data center elevator.

[0026] The subject executing this method can be the elevator control decision system of a data center elevator. Those skilled in the art will understand that the embodiments of this application use the elevator control decision system as an example to introduce the method, which is only an exemplary illustration and does not limit the scope of protection of the corresponding claims.

[0027] This application provides an elevator control decision system for a data center elevator. For example... Figure 1 As shown, the elevator control decision-making system mainly consists of four parts: an elevator interaction terminal, a multi-task intelligent recognition terminal, a target detection terminal, and an elevator control terminal. These parts interact and collaborate logically to jointly achieve intelligent judgment and control of elevator stopping behavior in the data center. (1) Elevator interface; The elevator interaction system is divided into an external interaction unit and an internal sensing unit. The external interaction unit is located on the elevator lobby floor and includes traditional up / down call buttons and a facial image capture device for waiting passengers. When a passenger presses the call button, the facial image capture device is simultaneously triggered to capture the passenger's facial image information.

[0028] The internal sensing unit, deployed inside the elevator car, includes real-time monitoring cameras, weight sensors (for detecting overload), and position sensors. The cameras continuously collect images of the car's interior to monitor passenger distribution, numbers, and the placement of critical infrastructure equipment in the data center; the weight sensors provide load status signals as the basis for triggering an overload alarm.

[0029] 2) Multi-task intelligent recognition terminal The multi-task intelligent recognition terminal receives facial images collected by the elevator interaction terminal and simultaneously performs two types of tasks, facial recognition and emotion recognition, through a neural network model trained based on multi-task learning, to obtain information such as the identity information and usual floor of the waiting passengers; and can infer the expected floor of the waiting passengers by combining the identity information and time information; at the same time, it outputs the emotional information such as the urgency of the waiting passengers to take the elevator, and uses the emotional information as one of the inputs of the elevator control terminal.

[0030] like Figure 2 The diagram shown is a framework diagram of a multi-task intelligent recognition terminal. As can be seen from the diagram, the overall processing flow of this terminal includes three stages: data acquisition, data preprocessing, and multi-task learning model inference.

[0031] During the data collection phase, when a passenger presses the elevator's up or down button from outside the elevator, the system simultaneously triggers and collects raw image data containing the passenger's face.

[0032] Secondly, in the data preprocessing stage, a series of data preprocessing operations are performed on the acquired raw facial images to improve image quality and standardize the input format. These preprocessing operations include: grayscale conversion to simplify color channels; face alignment to correct posture and ensure consistent facial feature positions; histogram equalization to enhance image contrast; and size standardization to scale the image to the fixed resolution required by the model.

[0033] Finally, in the inference stage of the multi-task learning model, the preprocessed face image is input into the neural network model trained based on multi-task learning, and two types of tasks, face recognition and emotion recognition, are executed simultaneously to obtain the identity information, usual floor, expected floor and emotion information of the waiting passengers.

[0034] 3) Target detection end The target detection end performs real-time detection based on images of the elevator car interior captured by the elevator's internal camera, identifies target objects such as passengers and equipment inside the elevator, counts the number of passengers and estimates equipment occupancy, thereby calculating the available space of the elevator, and sends the results, such as the available space, as input to the elevator control end.

[0035] 4) Elevator control terminal The elevator control unit receives multiple inputs, including emotional information from the multi-task intelligent recognition unit, detection results (remaining number of passengers / available space, etc.) from the target detection unit, and overload alarm information, and makes decisions and controls the stopping / braking strategies of the elevator during the current operation based on these inputs.

[0036] like Figure 1As shown, when a passenger initiates an elevator ride request in the elevator lobby, the elevator interaction terminal first receives the external up / down button signal and simultaneously captures the passenger's facial image; the facial image is sent to the multi-task intelligent recognition terminal to obtain passenger identity-related information and emotional information, and outputs the emotional information to the elevator control terminal.

[0037] Meanwhile, the internal camera of the elevator interface continuously collects images inside the car, and the target detection terminal performs real-time detection of passengers and data center critical infrastructure equipment inside the car, outputting the number of passengers in the car, the occupancy of data center critical infrastructure equipment, and the available space information obtained therefrom; the internal overload sensor collects whether the elevator is in an overload state and sends the overload state information to the elevator control terminal.

[0038] The elevator control unit integrates the above multi-source information to make a decision on whether to trigger the elevator to stop / brake for this elevator request, and outputs corresponding control commands to control the elevator operation. In this way, under the premise of meeting the constraints of safety and available space, and taking into account the urgency of the waiting passengers and the actual crowding / occupancy of the car, invalid stops are reduced and the elevator operation efficiency is improved.

[0039] Based on the above system architecture and its interaction process, in order to facilitate understanding of the technical solution of this application, the specific implementation of this application will be further described below with reference to the accompanying drawings.

[0040] Specifically, the implementation flow of the method provided in this application embodiment is as follows: Figure 3 As shown, it includes the following steps: Step 302: When a passenger sends a request to ride the data center elevator, obtain an image of the interior of the data center elevator car.

[0041] In this embodiment of the application, the target passenger refers to the passenger waiting for the elevator who sends an elevator request to the data center elevator at the current moment.

[0042] An elevator ride request refers to a request from a target passenger to trigger the data center elevator to proceed to the elevator lobby and provide passenger service. This request can be triggered by elevator buttons located on the outside of the elevator, which are the traditional up and down buttons.

[0043] In some embodiments, when a target passenger presses / touches the up or down button in the elevator lobby, the elevator interface samples and detects the button state and performs debouncing on the press signal. For example, a valid press is only determined if the button press continues for a preset stable time threshold. Once a valid press is determined, the elevator interface generates an elevator ride request, which includes at least: the target passenger's floor, the request direction (determined by the up / down button), and a request timestamp. The elevator ride request is then sent to the elevator control decision system of the data center elevator in the form of a signal or message, thereby enabling the target passenger to send an elevator ride request to the data center elevator.

[0044] When the elevator control decision system receives a passenger's request to use the elevator, it generates an image acquisition trigger command and sends this command to the elevator interface to trigger the interface to acquire images of the elevator's interior from the data center. These interior images refer to image data captured by cameras installed inside the elevator car. The image data can be a single still image or a short time-series video containing multiple frames.

[0045] Optionally, in order to ensure the correspondence between the images inside the elevator car and the elevator ride request in time, and to reduce latency, a timestamped image cache can be maintained on the camera side, such as a circular cache: when an elevator ride request arrives, the frame closest to the timestamp or multiple frames within a preset time window can be read from the cache based on the timestamp of the elevator ride request, thereby completing the acquisition of the images inside the elevator car of the data center.

[0046] For example, suppose a camera inside a data center elevator captures images of the elevator car's interior at 25fps with a resolution of 1280×720, and maintains a 2-second circular buffer (approximately 50 frames) at the camera's side. When the elevator control decision system receives a passenger's request to board the elevator at 09:15:30 (including the timestamp 09:15:30, the requested floor being the 12th floor, and the direction being upward), it can issue an image acquisition trigger command within 50ms and read the closest frame of the elevator car's interior image from the buffer (e.g., the frame corresponding to 09:15:30, or the nearest frame with an error of no more than ±40ms); or it can read a sequence of 10 frames of the elevator car's interior image centered on that timestamp and within a time window of ±0.2s as the output of the elevator car's interior image. Thus, taking a single-frame JPEG (quality factor 80) as an example, assuming a single frame size of approximately 150KB, the end-to-end time from receiving the elevator request to obtaining the elevator car's interior image can be controlled to ≤200ms.

[0047] Step 304: Determine the available space and remaining passenger capacity of the data center elevator based on the images inside the elevator car.

[0048] After obtaining the image of the elevator car's interior, the acquired image can be input into a pre-trained object detection model for inference analysis to obtain labeled images of target objects in the elevator car's interior image. The labeled images include predicted bounding boxes used to identify the target objects. Then, based on the coordinate and size information of the predicted bounding boxes and the size information of the labeled images, the available space of the data center elevator is determined. Furthermore, the number of passengers in the data center elevator car is determined based on the labeled images. Finally, the remaining number of passengers is determined based on the number of passengers and the rated passenger capacity of the car.

[0049] This object detection model is a computer vision model based on deep convolutional neural networks, used to locate and identify objects in an input image. Its output typically includes the category of each object and a predicted bounding box for locating the object.

[0050] Annotated images refer to images on which prediction boxes are superimposed on images inside the car.

[0051] A bounding box is used to indicate the location of a target object in an annotated image. Target objects include at least passengers and equipment, such as network equipment and data center critical infrastructure equipment. Each identified target object in the annotated image can be identified by a rectangular bounding box. Each bounding box is uniquely determined by its coordinates in the image coordinate system (e.g., the x and y coordinates of the top-left corner) and its dimensions (width and height).

[0052] Optionally, considering the comprehensive requirements of elevator control scenarios for real-time performance, ease of deployment, and detection accuracy, this embodiment of the application may choose to use YOLOv5 as the object detection model when determining the available space and remaining passenger capacity of the data center elevator through the object detection model. This is because the YOLOv5 model has advantages such as a simple network structure, fast training convergence, high inference speed, and ease of deployment on embedded devices. Furthermore, it maintains good average accuracy (mAP) on general object detection datasets. Therefore, this choice ensures that the system can detect passengers and equipment inside the elevator car in real-time and accurately, meeting the timeliness and reliability requirements of elevator control decisions, and providing a stable and reliable foundation for subsequent calculations of available space and remaining passenger capacity.

[0053] like Figure 4As shown, the overall network structure of YOLOv5 can generally be divided into three parts: the backbone, the neck, and the head. The backbone uses CSPDarknet53 as a feature extractor to perform layer-by-layer convolution and downsampling on the input car interior image to obtain feature representations (also known as feature layers / feature maps) at different scales. In the backbone, three sets of feature layers at different scales can usually be selected as effective feature layers for input to subsequent detection branches to provide semantic and detailed information at different scales. For example, feature layers with spatial resolutions of 80×80, 40×40, and 20×20 can be selected.

[0054] The neck network is used to fuse multi-scale effective feature layers from the backbone network. YOLOv5 employs a PANet-based architecture for bidirectional feature transfer and fusion, including: In the top-down path, upsampling is used to transfer stronger overall semantic information from deep features to shallow features, thereby enhancing the semantic expressive power of shallow features. In the bottom-up path, downsampling is used to transmit richer details and localization information from shallow features back to the deep layers to enhance the detail representation capability of deep features. Meanwhile, the neck network introduces lateral connections to preserve key information from different levels, and concatenates lateral features with upsampling and downsampling features along the channel dimension using a concat method, thereby increasing the number of feature channels and improving the ability to describe image content. After this multi-scale fusion, three sets of enhanced effective feature layers are obtained, each targeting object detection tasks at different scales.

[0055] After feature extraction and fusion, the detection head is applied to the three enhanced effective feature layers. Each feature layer contains width, height, and number of channels, and can be considered as a set of multiple feature points, where each feature point corresponds to a set of channel feature vectors. The detection head predicts each feature point: on the one hand, it determines whether a target exists at that location; on the other hand, it regresses the bounding box parameters, adjusting the prior information through regression offsets, thereby outputting the final target detection result, including the target category, the coordinates of the predicted bounding box, and its size.

[0056] In one specific implementation, such as Figure 5 As shown, determining the available space for a data center elevator involves the following steps: (1) Before inputting the car interior image into the target detection model, the car interior image is preprocessed. The data preprocessing includes: image scaling, such as scaling and filling proportionally according to the input size of the target detection model, pixel normalization, standardization, channel arrangement conversion, etc., so that the car interior image meets the input format requirements of the target detection model.

[0057] (2) Input the preprocessed interior image of the car into a target detection model, such as YOLOv5, for inference to obtain detection results of multiple target objects. Each detection result includes at least: the target object category, such as passenger or equipment, and the prediction box corresponding to the target object.

[0058] The prediction bounding box contains coordinate information and size information. The coordinate information is used to represent the position of the prediction bounding box in the labeled image, which can be represented by the coordinates of the center point of the prediction bounding box or the coordinates of the upper left / lower right corner. The size information is used to represent the size of the prediction bounding box, which usually includes the width and height of the prediction bounding box.

[0059] In this embodiment of the application, each prediction box can be uniquely defined by its geometric parameters, and the calculation formula is as follows:

[0060]

[0061]

[0062]

[0063] in, and These represent the x and y coordinates of the center point of the prediction box, respectively. and These represent the width and height of the prediction box, respectively, in pixels. This indicates the coordinates of the top-left corner of the grid where the center point of the prediction box is located; This indicates the offset of the center point of the prediction box relative to the coordinates of the top-left corner of the grid. This indicates the scaling ratio of the predicted bounding box width and height relative to the preset anchor point bounding box width and height. This indicates the width and height of the corresponding preset anchor point bounding box. The preset anchor point bounding box is a set of baseline rectangles with different widths and heights, pre-defined before the object detection model is trained. Its dimensions are determined based on the statistical distribution of device and passenger targets in the training dataset. The Sigmoid activation function maps the input to the interval (0, 1).

[0064] The output of the object detection model is a labeled image containing the coordinates and dimensions of all predicted bounding boxes.

[0065] (3) Calculate the available space based on the coordinate information (including passenger coordinate information and equipment coordinate information), size information and size information of the target object given by the prediction box and the labeled image.

[0066] First, determine the total pixel area of ​​the labeled image itself based on the size information of the labeled image.

[0067] Secondly, the pixel area occupied by each target object in the labeled image is determined based on the size information of the predicted bounding box of each target object in the labeled image, and the pixel areas occupied by all target objects are summed to obtain the total pixel area occupied by all target objects in the labeled image.

[0068] Then, the unoccupied pixel area in the labeled image is calculated based on the total pixel area of ​​the labeled image itself and the total pixel area occupied by all target objects in the labeled image.

[0069] Finally, based on the unoccupied pixel area in the labeled image and a preset spatial conversion factor, the unoccupied pixel area is converted into the actual physical area, thereby obtaining the usable space of the car. This spatial conversion factor can be obtained through camera calibration and is used to determine the mapping relationship between the image pixel coordinate system and the physical coordinate system of the car floor.

[0070] Step 306: Obtain the target passenger's emotional information.

[0071] In this embodiment of the application, when obtaining the emotional information of the target passenger, a face image containing the target passenger can be obtained first; then the face image is input into the trained multi-task learning model to output the face recognition result and emotional information of the target passenger, and the face recognition result is used to associate the emotional information of the target passenger.

[0072] Here, the facial image refers to image data acquired by an external facial image acquisition device of the elevator, which at least contains the facial region of the target passenger. In some embodiments, multiple frames of images can be continuously acquired within a preset time window after the button is triggered, for example, 0.5 to 2 seconds, and the frame with the highest clarity of the facial region can be selected as the facial image; or multiple frames of images can be used as the facial image to enhance robustness.

[0073] Optionally, to improve the accuracy and stability of emotion recognition, the face image can be preprocessed before being input into the multi-task learning model. Preprocessing may include: grayscale processing, face alignment, equalization, denoising, and standardization. Standardization can be understood as normalizing / normalizing the mean and variance of pixel values, making the input distribution more stable, thereby improving the consistency of model inference.

[0074] Next, the preprocessed face images are input into the trained multi-task learning model to obtain the target passenger's emotional information and face recognition results. The multi-task learning model sets a primary task and an auxiliary task for the face images. The primary task is emotion recognition, used to output emotional information; the auxiliary task is face recognition, used to output the target passenger's face recognition results. The core idea is to allow the multi-task learning model to simultaneously learn these two related but different tasks, utilizing the correlation between tasks to share the learned feature representations, thereby improving the generalization ability and robustness of the primary task (emotion recognition). .in, The multi-task learning model includes a shared layer and a task-related layer connected to the shared layer. The shared layer extracts common features shared by the primary and auxiliary tasks. The task-related layer includes a first task-related network and a second task-related network, which are independent of each other. The first task-related network extracts features needed for emotion recognition from the common features. The second task-related network extracts features needed for face recognition from the common features. The face recognition results output by the second task-related network are used to associate the emotional information of the target passenger.

[0075] In one alternative implementation, considering that convolutional neural networks (CNNs) can extract multi-level features such as image edges, textures, and shapes, and that sharing convolutional kernel parameters can reduce the number of parameters and computational complexity, multi-task learning can be integrated into CNNs to construct a CNN-based multi-task learning model structure, such as... Figure 6 As shown, the shared layer consists of multiple convolutional and pooling layers, whose parameters are jointly optimized by two tasks during training. This layer is responsible for extracting low-level common visual features (such as edges, textures, and shapes) from the preprocessed face image input and gradually combining them into high-level semantic features. The shared representations of all tasks are learned in this layer, and the parameter sharing mechanism of the convolutional kernels effectively reduces model complexity and the risk of overfitting.

[0076] The task-related layer comprises two independent task-related networks. The first task-related network receives general features from the shared layer, further extracts expression-related features through its proprietary CNN layer, and then passes through a fully connected layer and a Softmax classifier to output the current passenger's emotional information. This could be a probability distribution vector representing the confidence level of whether the emotional information belongs to the category of anxiety or calmness. The second task-related network, based on the general features from the shared layer, extracts identity features through another proprietary CNN layer, and then processes these features through a fully connected layer to output a face recognition result. This result could be a feature vector or a direct identity ID, used to associate with pre-stored passenger identity files in the system.

[0077] In this embodiment, before inputting the preprocessed face image into the trained multi-task learning model, the multi-task learning model needs to be trained. Optionally, to train the trained multi-task learning model, a multi-task learning objective function with adaptive weights can be used. This multi-task learning objective function is the weighted sum of the loss function of the main task (emotion recognition task) and the loss function of the auxiliary task (face recognition task), as follows:

[0078] This represents the multi-task learning objective function that incorporates adaptive weights; , These represent the weighted hyperparameters for the primary and secondary tasks, respectively. , and These represent the training parameters for the shared layer, the main task layer, and the auxiliary layer, respectively. and Let represent the loss functions for the main task and the auxiliary task, respectively. This represents the sentiment information output by the multi-task learning model. It represents the actual emotional information of the target customer. This represents the face recognition result output by the multi-task learning model. This indicates the actual face recognition result.

[0079] Step 308: Generate control instructions for the data center elevator based on available space, remaining passenger capacity, and emotional information.

[0080] Among them, the control command is used to control whether the data center elevator stops at the floor where the target passenger is located.

[0081] In some embodiments, the elevator control terminal can first determine whether the data center elevator still has room for passenger capacity based on the remaining number of passengers, that is, whether the remaining number of passengers is greater than or equal to 1.

[0082] When the number of passengers remaining is less than 1, it means that the data center elevator can no longer carry new passengers under the constraint of its rated passenger capacity. At this time, a control command can be directly generated to control the data center elevator to not stop at the floor where the target passenger is located.

[0083] When the remaining number of passengers is greater than or equal to 1, the system further determines whether the preset stopping conditions are met based on the available space. The preset stopping condition is that the available space is greater than or equal to a preset available space threshold. The preset available space threshold can be pre-configured based on factors such as the size of the car, the minimum standing area required for a single person to enter, and data center equipment handling scenarios.

[0084] If the available space meets the preset stopping conditions, a control command is generated to control the data center elevator to stop at the floor where the target passenger is located.

[0085] In some implementations, to meet service demands in high-urgency situations and avoid a decline in passenger experience due to insufficient available space (i.e., available space being less than a preset available space threshold) leading to non-stop service, a further mechanism can be developed. When available space does not meet the preset stopping conditions, the emotional intensity of the target passenger's desire to ride the data center elevator can be determined based on emotional information. This emotional intensity is then compared to a preset emotional intensity threshold. When the emotional intensity is greater than the preset threshold, a control command is generated to control the data center elevator to stop at the target passenger's floor. When the emotional intensity is less than or equal to the preset threshold, a control command is generated to control the data center elevator to not stop at the target passenger's floor. This balances operational efficiency with the target passenger's urgent need for elevator access when available space is insufficient.

[0086] Among them, emotional intensity is a numerical indicator derived from emotional information, used to quantify the urgency / anxiety of the target passenger in wanting to take the data center elevator. When the emotional information is a classification result (such as anxious / calm), anxiety can be mapped to a higher emotional intensity, and calmness can be mapped to a lower emotional intensity; when the emotional information is a classification probability or regression score, the probability / score can be directly used as the emotional intensity (with a value of, for example, 0 to 1).

[0087] To further enhance intelligent service capabilities, some embodiments can also perform identity recognition and automatic floor pre-selection based on the facial images of target passengers. Specifically, after acquiring a facial image containing the target passenger, the facial image can be input into the preceding multi-task learning model for facial recognition to output the target passenger's identity information, such as user ID / name / employee number, etc.

[0088] The reason for identifying passenger information first is that in data center scenarios, passengers' elevator destinations often exhibit clear personal patterns, such as fixed workstation floors, fixed server room / maintenance floors, or fixed meeting floors, and these patterns show a stable distribution across different time periods. Therefore, by using passenger information, their historical elevator usage behavior can be traced, allowing the system to infer their desired floor without increasing the operational burden on the target passenger. Based on this, the elevator control decision system can retrieve historical elevator usage behavior data associated with the identified information from a pre-established elevator usage behavior database. This historical elevator usage behavior data includes information on the high-frequency floors reached by the target passenger at different time periods. For example, the day can be divided into multiple intervals (morning peak, noon, evening peak, or hourly granularity), and the frequency of visits to the target passenger's historical floors can be counted within each time period to obtain the high-frequency floors reached during that time period.

[0089] Subsequently, the target floor that the target passenger expects to reach is determined based on historical elevator behavior data. For example, a high-frequency arrival floor that matches the current time period is selected as the target floor, or the final target floor is determined by combining the weight of the most recent arrival floor and the frequency of access among multiple candidate high-frequency arrival floors.

[0090] Furthermore, to avoid false triggering due to premature floor selection before the target passenger enters the elevator car, in this embodiment, automatic floor selection can be performed only after the target passenger is detected entering the data center elevator car. For example, the entry of the target passenger can be confirmed by target detection results from an image inside the car, or by detection signals from door zone sensors, infrared beams, or load changes. After entry is detected, a floor pre-selection command is generated based on the target floor and sent to the elevator control terminal, so that the elevator automatically completes the floor selection operation for the target floor.

[0091] This entry detection-triggered pre-selection method reduces the number of manual floor selection steps for target passengers in the elevator car, significantly reducing interaction costs and improving elevator efficiency, especially when carrying equipment or in crowded situations. It also enhances the reliability and safety of automatic floor selection, reducing invalid commands. Furthermore, the strategy based on high-frequency arrival floors at different times helps improve the accuracy of target floor prediction, thereby reducing error correction operations and additional stops caused by incorrect floor selection, further improving the scheduling efficiency and operational experience of data center elevators.

[0092] The method provided in this application embodiment acquires real-time images of the interior of the data center elevator car to dynamically assess available space and remaining passenger capacity, and integrates the emotional information of target passengers for multi-dimensional intelligent decision-making. This can effectively reduce the number of invalid stops of the data center elevator, improve the operating efficiency of the data center elevator, and reduce the energy consumption and equipment wear of the data center elevator.

[0093] The following describes how the methods provided in the embodiments of this application are applied in practice, taking into account real-world scenarios.

[0094] Please see Figure 7 This is a schematic diagram illustrating an application process of the method provided in this application embodiment. The process specifically includes the following steps: (1) First, determine whether an alarm message (i.e., an overload signal) has been received from the elevator weight sensor. If yes, for safety reasons, the elevator control decision system will directly output the decision result without stopping, and the process will end. If no, proceed to the next step.

[0095] (2) Determine whether the real-time number of passengers in the elevator car has reached the elevator's rated passenger capacity. If yes, it means the elevator is full, and the elevator control decision system will output a decision not to stop. If no, proceed to the next step of more refined capacity judgment.

[0096] (3) Determine whether the real-time passenger count has reached the preset capacity threshold, for example, the rated number of passengers minus 2. If it has not reached the threshold, it indicates that the elevator car is relatively spacious, and the elevator control decision system tends to output the decision to stop. If it has reached or exceeded the threshold, it means that the elevator is starting to become crowded, and further judgment needs to be made in conjunction with spatial information, and the process proceeds to the next step.

[0097] (4) Based on the calculated available space in the car, determine whether it meets the preset space requirements (i.e., whether it is greater than or equal to the preset available space threshold). If it does, it indicates that there is still enough physical space to accommodate new passengers or equipment, and the elevator control decision system outputs the decision result to stop. If it does not meet the requirements, it indicates that the space is already cramped. At this time, passenger emotions will be introduced for the final flexible judgment dimension, and the process will proceed to the last step.

[0098] (5) When the elevator is saturated in both capacity and space, the elevator control decision system queries the emotional information of waiting passengers. This information is provided by the multi-task intelligent recognition terminal and is usually classified as urgent (e.g., anxious) or slow (e.g., calm).

[0099] If the emotional information is mild, it indicates that the passenger's needs are not urgent. To avoid unnecessary stops, the elevator control decision system outputs the final decision not to stop.

[0100] If the emotional information indicates urgency, it means that the passenger has a high degree of urgency. The elevator control decision system will prioritize responding to this need and output the final decision to stop.

[0101] The method provided in this application embodiment acquires real-time images of the interior of the data center elevator car to dynamically assess available space and remaining passenger capacity, and integrates the emotional information of target passengers for multi-dimensional intelligent decision-making. This can effectively reduce the number of invalid stops of the data center elevator, improve the operating efficiency of the data center elevator, and reduce the energy consumption and equipment wear of the data center elevator.

[0102] To address the problem of invalid stops that easily occur in existing elevator control methods, this application provides a control device for a data center elevator, the specific structure of which is shown in the schematic diagram below. Figure 8 As shown, the system includes an image acquisition module 801, an information determination module 802, an emotion acquisition module 803, and an instruction generation module 804. The functions of each module are as follows: The image acquisition module 801 is used to acquire images of the interior of the data center elevator car when it receives a ride request from a target passenger for the data center elevator. Information determination module 802 is used to determine the available space and remaining passenger capacity of the data center elevator based on the images inside the car. The emotion acquisition module 803 is used to acquire the emotional information of the target passenger; The instruction generation module 804 is used to generate control instructions for the data center elevator based on available space, remaining passenger capacity, and emotional information. The control instructions are used to instruct the data center elevator whether to stop at the floor where the target passenger is located.

[0103] Optionally, the information determination module 802 is used for: The image inside the car is input into the target detection model to obtain an labeled image of the target object in the image inside the car. The labeled image includes a prediction box used to identify the target object. Based on the coordinate and size information of the predicted bounding box and the size information of the labeled image, the available space of the data center elevator is determined; In addition, the number of passengers in the elevator car of the data center was determined based on the labeled images; The remaining number of passengers is determined based on the number of passengers and the car's rated passenger capacity.

[0104] Optional, the emotion acquisition module 803 is used for Obtain facial images containing the target passenger; The face image is input into the trained multi-task learning model, which outputs the face recognition result and emotional information of the target passenger. The multi-task learning model sets a primary task and an auxiliary task for face images. The primary task is emotion recognition, which is used to output emotional information; the auxiliary task is face recognition, which is used to output the face recognition results of the target passenger. The multi-task learning model includes a shared layer, which is used to extract common features shared by the main task and the auxiliary task. The task-related layer is connected to the shared layer. The task-related layer includes a first task-related network and a second task-related network that are independent of each other. The first task-related network is used to extract the features required for emotion recognition from general features; The second task-related network is used to extract the features required for face recognition from general features; the face recognition results output by the second task-related network are used to associate emotional information of the target passenger.

[0105] Optionally, the instruction generation module 804 is used for: When the number of passengers remaining is greater than or equal to 1, determine whether the available space meets the preset docking conditions based on the available space. When the available space does not meet the preset stopping conditions, the emotional intensity of the target passenger's expectation to take the data center elevator is determined based on emotional information. When the emotional intensity is greater than the preset emotional intensity threshold, a control command is generated to control the data center elevator to stop at the floor where the target passenger is located. The preset docking condition is that the available space is greater than or equal to the preset available space threshold.

[0106] Optionally, the instruction generation module 804 is also used for: When the available space does not meet the preset stopping conditions, the emotional intensity of the target passenger's expectation to take the data center elevator is determined based on emotional information. When the emotional intensity is less than or equal to a preset emotional intensity threshold, a control command is generated to prevent the data center elevator from stopping at the floor where the target passenger is located.

[0107] Optionally, the control device for the data center elevator is also used for: Identify the target passenger's identity information based on their facial image; Obtain historical elevator behavior data associated with identity information, including the high-frequency arrival floor information of the target passenger at different time periods; Based on historical elevator behavior data, determine the target floor that the target passenger expects to reach; When a target passenger is detected entering the elevator car in the data center, a floor pre-selection instruction is generated based on the target floor, so that the elevator can automatically complete the floor selection operation for the target floor.

[0108] The device provided in this application can dynamically assess available space and remaining passenger capacity by acquiring real-time images of the interior of the data center elevator car, and make multi-dimensional intelligent decisions by integrating the emotional information of target passengers. This can effectively reduce the number of invalid stops of the data center elevator, improve the operating efficiency of the data center elevator, and reduce the energy consumption and equipment wear of the data center elevator.

[0109] Figure 9 To illustrate the hardware structure of an electronic device according to various embodiments of this application, the electronic device may include a processor 901 and a memory 902 storing computer program instructions. Specifically, the processor 901 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of this application.

[0110] Memory 902 may include mass storage for data or instructions. For example, and not limitingly, memory 902 may include a hard disk drive (HDD), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 902 may include removable or non-removable (or fixed) media. Where appropriate, memory 902 may be internal or external to an electronic device. In a particular embodiment, memory 902 may be a non-volatile solid-state memory.

[0111] In one embodiment, memory 902 may be read-only memory (ROM). In one embodiment, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0112] The processor 901 reads and executes computer program instructions stored in the memory 902 to implement any of the data center elevator control methods in the above embodiments.

[0113] In one example, the electronic device may also include a communication interface 903 and a bus 910. Wherein, as... Figure 9 As shown, the processor 901, memory 902, and communication interface 903 are connected through bus 910 and complete communication with each other.

[0114] The communication interface 903 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0115] Bus 910 includes hardware, software, or both, that couples components of an electronic device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 910 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0116] Furthermore, in conjunction with the data center elevator control method in the above embodiments, this application embodiment can provide a computer-readable storage medium for implementation. This computer-readable storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the data center elevator control methods in the above embodiments.

[0117] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0118] The above description is merely a specific implementation example of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0119] Secondly, those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0120] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0121] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0122] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0123] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0124] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0125] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0126] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0127] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of this application should be included within the scope of the claims of this application.

Claims

1. A control method for a data center elevator, characterized in that, include: When a passenger sends a request to ride the data center elevator, an image of the interior of the data center elevator car is obtained. The available space and remaining passenger capacity of the data center elevator are determined based on the images of the elevator car's interior. Obtain the emotional information of the target passenger; The control instructions for the data center elevator are generated based on the available space, the remaining number of passengers, and the emotional information. The control instructions are used to instruct the data center elevator whether to stop at the floor where the target passenger is located.

2. The method as described in claim 1, characterized in that, The available space and remaining passenger capacity of the data center elevator are determined based on the interior images of the elevator car, including: The image inside the car is input into the target detection model to obtain an labeled image of the target object in the image inside the car. The labeled image includes a prediction box for identifying the target object. Based on the coordinate and size information of the predicted bounding box and the size information of the labeled image, the available space of the data center elevator is determined; And, based on the labeled image, determine the number of passengers inside the elevator car of the data center; The remaining number of passengers is determined based on the number of passengers and the rated passenger capacity of the car.

3. The method as described in claim 1 or 2, characterized in that, Obtaining the emotional information of the target passenger includes: Obtain a facial image containing the target passenger; The face image is input into the trained multi-task learning model, which outputs the face recognition result of the target passenger and the emotional information. The multi-task learning model sets a primary task and an auxiliary task for the face image. The primary task is an emotion recognition task, which is used to output the emotion information. The auxiliary task is a face recognition task, which is used to output the face recognition result of the target passenger. The multi-task learning model includes: a shared layer, used to extract common features shared by the main task and the auxiliary task; A task-related layer connected to the shared layer, the task-related layer comprising a first task-related network and a second task-related network that are independent of each other; The first task-related network is used to extract features required for emotion recognition from the general features; The second task-related network is used to extract features required for face recognition from the general features; the face recognition result output by the second task-related network is used to associate the emotional information of the target passenger.

4. The method according to any one of claims 1 to 3, characterized in that, Based on the available space, the remaining passenger capacity, and the emotional information, control instructions for the data center elevator are generated, including: When the remaining number of passengers is greater than or equal to 1, determine whether the available space meets the preset docking conditions based on the available space. When the available space does not meet the preset stopping conditions, the emotional intensity of the target passenger's desire to ride the data center elevator is determined based on the emotional information. When the emotional intensity is greater than a preset emotional intensity threshold, a control command is generated to control the data center elevator to stop at the floor where the target passenger is located. The preset docking condition is that the available space is greater than or equal to a preset available space threshold.

5. The method as described in claim 4, characterized in that, The method further includes: When the available space does not meet the preset stopping conditions, the emotional intensity of the target passenger's desire to ride the data center elevator is determined based on the emotional information. When the emotional intensity is less than or equal to the preset emotional intensity threshold, a control command is generated to control the data center elevator to not stop at the floor where the target passenger is located.

6. The method as described in claim 1 or 3, characterized in that, The method further includes: Identify the target passenger's identity information based on the target passenger's facial image; Obtain historical elevator behavior data associated with the identity information, the historical elevator behavior data including the high-frequency arrival floor information of the target passenger in different time periods; Based on the historical elevator behavior data, the target floor that the target passenger expects to reach is determined; When the target passenger is detected entering the elevator car of the data center, a floor pre-selection instruction is generated according to the target floor, so that the elevator can automatically complete the floor selection operation of the target floor.

7. A control device for a data center elevator, characterized in that, It includes an image acquisition module, an information determination module, an emotion acquisition module, and an instruction generation module, wherein: The image acquisition module is used to acquire an image of the interior of the data center elevator when it receives a ride request from a target passenger for the data center elevator. The information determination module is used to determine the available space and remaining passenger capacity of the data center elevator based on the images inside the elevator car. An emotion acquisition module is used to acquire the emotional information of the target passenger; The instruction generation module is used to generate control instructions for the data center elevator based on the available space, the remaining number of passengers, and the emotional information. The control instructions are used to instruct whether the data center elevator should stop at the floor where the target passenger is located.

8. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the control method for a data center elevator as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the control method for a data center elevator as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the control method for a data center elevator as described in any one of claims 1 to 6.

Images