Building, elevator, elevator control device, and elevator control method

The AI-powered elevator control device accurately detects pet boarding and alighting by tracking companion movements, enhancing user convenience and safety in elevators with pets.

WO2026240637A1PCT designated stage Publication Date: 2026-11-19HYUNDAI ELEVATOR CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/KR2026/097042
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-15
Filing Date
2026-05-15
Publication Date
2026-11-19

AI Technical Summary

Technical Problem

Existing pet detection systems in elevators struggle with accurately identifying pet boarding and alighting, especially when obscured by passengers or luggage, and fail to consider the relationship between pets and their companions, leading to delayed operation of pet lamps and potential discomfort for other passengers.

Method used

An elevator control device that uses AI-based image analysis to establish a relationship between pets and their companions, determining boarding and alighting by tracking the companion's movement, and automatically controlling a pet lamp based on this relationship.

Benefits of technology

Enhances the accuracy of pet detection, ensures reliable operation of pet lamps, and reduces the likelihood of safety accidents by providing advance notice of pet boarding and alighting, improving user convenience and psychological stability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure KR2026097042_19112026_PF_FP_ABST
    Figure KR2026097042_19112026_PF_FP_ABST
Patent Text Reader

Abstract

An elevator control device according to one aspect of the present disclosure may be configured to: configure a paring on the basis of the relationship between a pet and a passenger (companion) by using images of a car and / or a platform; determine whether the pet is on board or not by using the pairing; and automatically control a pet lamp according to the result.
Need to check novelty before this filing date? Find Prior Art

Description

Building, elevator, elevator control device, and elevator control method

[0001] The present disclosure relates to a building, an elevator, an elevator control device, and an elevator control method. The present disclosure relates to a building, an elevator, an elevator control device, and an elevator control method that can be conveniently used for boarding and alighting a pet.

[0002] The number of households with pets (hereinafter referred to as "pets") is increasing. The types of pets, such as dogs, cats, and birds, are also increasing. Pets can ride in elevators. When a pet enters an elevator, other people may be startled or feel uncomfortable due to the pet. For this reason, there is a pet lamp that signals the presence of a pet. Patent document JP2001322775A can be cited as an example of technology related to the pet lamp. A companion of the pet (hereinafter referred to as "companion") can press the pet lamp to signal that the pet is currently in the elevator. However, the pet lamp has the inconvenience of requiring the companion to operate it. Consequently, the reality is that its application and use are delayed.

[0003] Regarding the aforementioned current technology, the inventor conceived of automatically detecting pets in an image (e.g., using artificial intelligence) and automatically operating the pet lamp. As a result, it was found that there are the following problems.

[0004] First, there is a problem in accurately detecting whether the pet has exited the elevator. This is because, when using video-based or weight sensors, if the pet is obscured by another passenger upon exiting, it is often mistakenly assumed that the pet is still on board.

[0005] Second, pets have small movements and unpredictable trajectories. Consequently, there is a significant problem with large detection blind spots in images. This is because small dogs or cats move between adult passengers' legs or behind objects, moving out of the camera's field of view. Therefore, simple object recognition-based systems frequently failed to detect pets.

[0006] Third, identification becomes more difficult in crowded elevator cars as the PET is obscured. In other words, since it is blocked by multiple passengers during congestion, there are frequent failures to recognize the PET.

[0007] Fourth, since the detection criterion is a single object (pet), it does not reflect the behavior or location of the companion. For example, if the companion is holding the pet or it is contained within a cage, the pet is not detected at all.

[0008] The means for solving the above technical problem are explained.

[0009] The above first to third tasks can be solved by providing technology that accurately recognizes whether a pet has boarded or disembarked from an elevator, resolves the problem of the pet not being detected when disembarking due to being obscured by other passengers or luggage, and enables detection of disembarking even when the pet is in a carrier, etc.

[0010] The above fourth task can be solved by providing a linkage judgment method that can estimate the relationship between a pet and a companion. Specifically, it can be solved by a relationship-based judgment method that considers the pet and the companion together, rather than a single object-centered approach.

[0011] The means for solving the above technical problem are explained in more detail.

[0012] An elevator control device according to one aspect of the present disclosure may be configured to establish a pairing based on the relationship between a pet and a passenger (companion) using an image of a car and / or a landing, determine the boarding / alighting of the pet using the pairing, and automatically control a pet ramp according to the result.

[0013] The above elevator control device may include a pet boarding judgment unit that pairs a pet and a passenger when the pet and the passenger have a predetermined relationship using a first image obtained from at least one of the car and the landing.

[0014] The elevator control device may include a pet disembarkation determination unit that determines that the paired pet has disembarked when a paired passenger disembarks, by using a second image obtained from at least one of the car and the landing.

[0015] The above control device may include a main control unit that controls the PET lamp to be turned on when the PET boarding judgment unit determines that the PET has boarded, and to turn off when the PET disembarking judgment unit determines that the PET has disembarked. Additionally, it may be configured to turn on the PET lamp when pairing is established.

[0016] The above pet disembarkation determination unit may be configured to determine that the pet has disembarked based on the disembarkation of the paired passenger, even if the disembarkation status of the pet is not directly confirmed on the image.

[0017] The riding conditions satisfying a predetermined relationship between the pet and the passenger in the above pet riding judgment unit may include at least one of: first, whether a leash connects the passenger and the pet in the first image; second, whether the pet and the passenger ride consecutively within a predetermined time in the first image; third, whether the passenger is holding the pet in the first image; fourth, whether the pet is boarded inside a carrier in the first image; and fifth, whether the passenger and the pet ride in a pre-registered state in the first image.

[0018] The disembarking condition in which the above pet disembarking judgment unit determines that the pet is disembarking may include at least one of whether the carrier carrying the first pet is disembarking in the second image and whether the companion holding the second pet is disembarking.

[0019] An elevator control method according to another aspect of the present disclosure may include the step of acquiring at least one image in a time series, pairing the pet and the companion when the pet and the companion satisfy a predetermined relationship in a first image among the images, and providing a signal corresponding to the pairing in response to the movement of at least one of the paired companion and the pet in a second image among the images.

[0020] The above pairing may be performed at least one of the following: while the pet and the companion are boarding the car, and while the pet and the companion are inside the car. In the second image, if the companion disembarks from the car, it may be determined that the pet has also disembarked. In this case, even if the disembarking of the pet is not directly observed in the second image, it may be determined that the pet has disembarked in response to the disembarking of the companion.

[0021] In the above control method, the above predetermined relationship may include at least one of the following: a leash connecting the companion and the pet in the first image, the pet and the companion riding continuously within a predetermined time in the first image, the companion holding the pet in the first image, and the pet riding in a carrier in the first image.

[0022] The method may include the step of performing a first pairing with a companion and a pet that satisfy the above predetermined relationship, and storing the result of the first pairing as a learning model, and by repeating the above steps, the learning model may be configured to learn the pairing relationship between a pet and a family member of a household.

[0023] In the first image above, in response to at least one of the following: first, the pet and the companion satisfy a predetermined relationship, and second, the pet and the companion are registered in a learning model, a second pairing can be performed through the pairing of the pet and the companion in the first image.

[0024] The step of providing a signal corresponding to the pairing above may be provided in response to the paired pair disembarking. For example, it may be configured to provide a pet ramp-off signal in response to at least one of the following: first, when a companion in the pair disembarks; second, when the carrier carrying the pet disembarks; third, when a companion holding the pet disembarks; fourth, when both the pet and companion riding consecutively disembark; and fifth, when both the pet and companion connected by a leash disembark.

[0025] The present disclosure may have the following effects.

[0026] By determining whether the pet has disembarked using only the tracking of the companion, the turning on and off of the pet lamp can be controlled more reliably.

[0027] By utilizing AI-based pairing and tracking technology, the device and method can operate without malfunction even when the pet is small or when multiple passengers are on board.

[0028] Even when the PET is placed in a carrier and is not visually recognized, the accuracy of determining whether the PET has been unloaded can be maintained.

[0029] User convenience is enhanced because the pet lamp operates automatically.

[0030] By providing advance notice of the boarding and alighting status of paired pets and companions via a pet lamp and integrating it with door control based on the alighting status, it can reduce the possibility of safety accidents during the boarding and alighting process with pets, such as leash entrapment, and contribute to the realization of ESG values ​​by improving the convenience and psychological stability of various passengers.

[0031] In addition, the effects may be added to the detailed description of the present disclosure.

[0032] FIG. 1 is a block diagram schematically showing the configuration of an AI device according to an embodiment of the present disclosure.

[0033] FIG. 2 is a block diagram schematically showing the configuration of an AI server according to an embodiment of the present disclosure.

[0034] FIG. 3 is a drawing for explaining the configuration of an elevator control device according to an embodiment of the present disclosure.

[0035] FIG. 4 is a drawing illustrating an example of installation of a pet ramp according to an embodiment of the present disclosure, wherein (a) is an example of a landing panel and (b) is an example of a car panel.

[0036] FIG. 5 is a flowchart illustrating an elevator control method according to an embodiment of the present disclosure, wherein (a) describes the control relationship between an image and a pet ramp, (b) describes the relationship between pairing and disembarking determination, and (c) describes the relationship between pairing and pet ramp control.

[0037] FIG. 6 is a diagram illustrating a control device and a control method according to a second embodiment of the present disclosure, wherein (a) schematically illustrates an overview of the control method, (b) schematically illustrates a target for a learning model for determining a riding condition, and (c) schematically illustrates an edge device and server linkage configuration.

[0038] FIG. 7 is a drawing illustrating a PET unloading determination according to a third embodiment of the present disclosure, wherein (a) shows an overview of a control method and (b) shows a determination target of a learning model for determining unloading conditions.

[0039] Specific embodiments of the present disclosure are described in detail below with reference to the drawings. The spirit of the present disclosure is not limited to the embodiments described below. Those skilled in the art who understand the spirit of the present disclosure may easily propose other embodiments included within the scope of the same spirit by adding, changing, deleting, and adding components. However, such are also included within the scope of the spirit of the present disclosure. The present disclosure may include a plurality of embodiments. A first component of a first embodiment among the plurality of embodiments may be included by adding or changing a component of another embodiment. By combining a first component of a first embodiment among the plurality of embodiments and a second component of a second embodiment among the plurality of embodiments, other embodiments not specified in the plurality of embodiments may be provided. In the description of the drawings, identical or similar components are assigned identical or similar reference numbers regardless of drawing numbers and symbols, and redundant descriptions thereof may be omitted. The drawings are intended only to facilitate understanding of the embodiments of the present disclosure, and the technical spirit of the present disclosure is not limited by the attached drawings. It should be understood that all modifications, equivalents, and substitutions included within the spirit and scope of the present disclosure are included. In describing the embodiments of the present disclosure, detailed descriptions of related prior art may be omitted if it is determined that such detailed descriptions could obscure the essence of the embodiments disclosed herein. In describing the embodiments of the present disclosure, specific examples of related prior art may be included in the specification. In this case, the technical connection between the prior art and the present disclosure may be included within the technical spirit of the present disclosure. In the present disclosure, terms including ordinal numbers, such as first, second, etc., may be used to describe various components, but said components are not limited by said terms. Such terms may be used for the purpose of distinguishing one component from another.Where one component is described as being "connected" or "connected" to another component, it should be understood that it may be directly connected to or connected to the other component, or that there may be other components in between. Conversely, where one component is described as being "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. A singular expression may include a plural expression unless the context clearly indicates otherwise. In this disclosure, terms such as "comprising" or "having" are intended to indicate the presence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification. Therefore, the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof is not excluded.

[0040]

[0041] <1st Embodiment>

[0042] [Composition and Function of Elevators]

[0043] An elevator may be installed in a building. Here, the building may be exemplified by reinforced concrete, steel frame structures, and modular buildings. The building may provide a platform and a hoistway for installing the elevator. The platform and the hoistway may form a component of the elevator. The platform may exemplify an access space on each floor where the car stops. Multiple platforms may be provided in the building. The hoistway may exemplify a passageway through which the car ascends and descends. Components such as a hoisting mechanism and a control module for raising and lowering the car may be placed at the top of the hoistway. A door module may be provided on the platform. The door module may open and close in conjunction with the car door.

[0044] The above hoisting machine can rotate the rope in both directions. The rope can be wound onto a sheave provided on the output side of the hoisting machine. The rope can extend in both directions with the hoisting machine as the geometric center. In other words, the rope can be supported by the sheave and have both ends extend downward by gravity. The car can be supported on the first side where the rope extends in one direction. A counterweight can be supported on the second side where the rope extends in the other direction. The counterweight can compensate for the car and the load mounted on the car.

[0045] The hoisting mechanism may be selectively driven according to the control of the control unit controlling the elevator. By moving the rope wound around the sheave, the car may move up and down in the elevator shaft. More parts may be required for the configuration and operation of the elevator. For example, a sheave around which the rope is wound may also be provided on the counterweight. Even if there is no description of each of the above parts in this disclosure, the applicant declares that the parts necessary for the configuration and operation of the elevator are included in this disclosure.

[0046] The above-mentioned elevator is installed in a high-rise building. When an emergency occurs on an upper floor of a high-rise building, response personnel can move rapidly using the above-mentioned elevator. Here, the emergency scene may refer to a place where an emergency situation exists. The above-mentioned response personnel may include firefighters, medical personnel, and disaster response personnel. The present disclosure is characterized in that the above-mentioned response personnel move rapidly to the above-mentioned emergency scene. The above-mentioned elevator may operate as a part of an emergency response device utilizing the above-mentioned elevator, as described below.

[0047]

[0048] [Artificial Intelligence]

[0049] In the following description, Artificial Intelligence refers to the field of researching artificial intelligence or the methodologies capable of creating it, while Machine Learning refers to the field of researching methodologies for defining and solving the various problems addressed in Artificial Intelligence. Machine Learning is also defined as an algorithm that improves performance on a task through continuous experience.

[0050] In the following description, an Artificial Neural Network (ANN) refers to a model used in machine learning that can be defined as a model capable of problem-solving, composed of artificial neurons (nodes) that form a network through the combination of synapses. An Artificial Neural Network can be defined by connection patterns between neurons in different layers, a learning process that updates model parameters, and an Activation Function that generates output values. The Artificial Neural Network may include an Input Layer, an Output Layer, and optionally one or more Hidden Layers. This can be referred to as a multilayer perceptron. Each layer includes one or more neurons, and the Artificial Neural Network may include synapses connecting the neurons. In an Artificial Neural Network, each neuron can output the function value of an Activation Function in response to input signals, weights, and biases input through the synapses.

[0051] The artificial neural network described above may include not only the Multilayer Perceptron (MLP) but also other neural networks. Specifically, it may include various modern neural network structures such as a Transformer Encoder / Decoder, which is a neural network module that extracts / generates contextual information based on self-attention; an Attention Mechanism, which is a mechanism that adds weights to important parts of the input; a Graph Neural Network (GNN), which is a neural network structure for processing graph data with a node / edge structure; a Diffusion Model, which is a model that generates data through a stochastic noise addition / removal process; a Sequence-to-Sequence Model (Seq2Seq), which is a structure that converts an input sequence into an output sequence; and a Multimodal Neural Network, which processes different forms of data simultaneously, such as text, images, and speech.

[0052] The above model parameters refer to parameters determined through learning, including the weights of synaptic connections and the biases of neurons. Hyperparameters refer to parameters that must be set before learning in a machine learning algorithm, including the learning rate, number of iterations, mini-batch size, and initialization function. The purpose of learning the above artificial neural network can be viewed as determining model parameters that minimize the loss function. The loss function can be used as an indicator to determine the optimal model parameters during the learning process of the artificial neural network. The above hyperparameters are not limited to conventional learning rate adjustment structures based on Stochastic Gradient Descent (SGD) and may include various techniques such as Adaptive Learning Rate (Adaptive LR), optimizers (Adam et al.), and modular learning rate separation structures.

[0053] The aforementioned machine learning can be classified into supervised learning, unsupervised learning, reinforcement learning, and self-supervised learning based on the learning method. Supervised learning refers to a method of training an artificial neural network with labels provided for the training data; a label can refer to the correct answer (or result value) that the neural network must infer when training data is input. Unsupervised learning refers to a method of training an artificial neural network without labels provided for the training data. Reinforcement learning refers to a learning method in which an agent defined within an environment is trained to select an action or sequence of actions that maximizes the cumulative reward in each state. Self-supervised learning is an artificial intelligence technique that learns by generating correct answers on its own from a large amount of unlabeled data. This method allows the AI ​​model to perform learning by utilizing the structure of the data itself to identify its characteristics.

[0054] Providing any learning model (learned model) of the present disclosure may utilize at least one of the learning methods. It is obvious that the learning model may be provided by applying one or at least two learning methods.

[0055]

[0056] [Artificial Intelligence (AI) Device]

[0057] FIG. 1 shows an AI device according to an embodiment of the present disclosure.

[0058] The AI ​​device (100) may be implemented as a stationary device or a mobile device, such as an elevator, TV, CCTV, terminal, mobile phone, smartphone, desktop computer, laptop, navigation, wearable device, robot, and vehicle. The AI ​​device (100) may include a communication unit (110), an input unit (120), a learning processor (130), a sensing unit (140), an output unit (150), a memory (170), and a processor (180). The AI ​​device (100) may be used by a person receiving information. The AI ​​device (100) may be used by a person providing information.

[0059] The communication unit (110) can transmit and receive data with external devices, such as other AI devices or AI servers (200), using wired or wireless communication technology. For example, the communication unit (110) can transmit and receive sensor information, user input, learning models, and control signals with external devices. Communication technologies used by the communication unit (110) may include GSM (Global System for Mobile communication), CDMA (Code Division Multi Access), LTE (Long Term Evolution), 5G, WLAN (Wireless LAN), Wi-Fi (Wireless-Fidelity), Bluetooth, RFID (Radio Frequency Identification), Infrared Data Association (IrDA), ZigBee, and NFC (Near Field Communication).

[0060] The above input unit (120) can acquire various types of data. At this time, the input unit (120) may include a camera for inputting video signals, a microphone for receiving audio signals, a user input unit for receiving information from a user, etc. Here, the camera or microphone may be treated as a sensor, and the signal acquired from the camera or microphone may be referred to as sensing data or sensor information. The above input unit (120) can acquire training data for model learning and input data to be used when acquiring output using the learning model. The input unit (120) may also acquire unprocessed input data. In this case, the processor (180) or the learning processor (130) may extract input feature points as a preprocessing step for the input data.

[0061] The above learning processor (130) can train a model composed of an artificial neural network using training data. Here, the trained artificial neural network may be referred to as a learning model. The learning model can be used to infer result values ​​for new input data other than the training data. The inferred value can be used as a basis for judgment to perform a certain action. At this time, the learning processor (130) can perform AI processing together with the learning processor (240) of the AI ​​server (200). At this time, the learning processor (130) may include memory integrated or implemented in the AI ​​device (100). At this time, the learning processor (130) may be implemented using memory (170), external memory directly coupled to the AI ​​device (100), or memory maintained by an external device.

[0062] The sensing unit (140) can acquire at least one of internal information of the AI ​​device (100), surrounding environment information of the AI ​​device (100), and user information using various sensors. Sensors included in the sensing unit (140) include a proximity sensor, an illuminance sensor, an accelerometer, a magnetic sensor, a gyroscope, an inertial sensor, an RGB sensor, an IR sensor, a fingerprint recognition sensor, an ultrasonic sensor, a light sensor, a microphone, a lidar, a radar, etc.

[0063] The output unit (150) can generate outputs related to sight, hearing, or touch. At this time, the output unit (150) may include a display unit that outputs visual information, a speaker that outputs auditory information, and a haptic module that outputs tactile information.

[0064] The memory (170) can store data that supports various functions of the AI ​​device (100). For example, the memory (170) can store input data, training data, training models, and training history obtained from the input unit (120).

[0065] The processor (180) may determine at least one executable action of the AI ​​device (100) based on information determined or generated using a data analysis algorithm or a machine learning algorithm. The processor (180) may perform the determined action by controlling the components of the AI ​​device (100). To this end, the processor (180) may request, search, receive, or utilize data from the learning processor (130) or memory (170). Additionally, the processor (180) may control the components of the AI ​​device (100) to execute a predicted action or an action deemed desirable among the at least one executable action. At this time, if the processor (180) requires the connection of an external device to perform the determined action, it may generate a control signal to control the external device and transmit the generated control signal to the external device. The processor (180) may obtain intent information regarding user input and determine the user's requirements based on the obtained intent information. At this time, the processor (180) can obtain intent information corresponding to the user input by using at least one of a Speech To Text (STT) engine for converting voice input into a string or a Natural Language Processing (NLP) engine for obtaining intent information of natural language. At this time, at least one of the STT engine or the NLP engine may be composed of an artificial neural network in which at least a part is trained according to a machine learning algorithm.

[0066] And, at least one of the STT engine or NLP engine may be learned by the learning processor (130), learned by the learning processor (240) of the AI ​​server (200), or learned by the distributed processing thereof.

[0067] The processor (180) may collect history information, including the operation details of the AI ​​device (100) or user feedback regarding the operation, and store it in memory (170) or a learning processor (130), or transmit it to an external device such as an AI server (200). The collected history information may be used to update a learning model. The processor (180) may control at least some of the components of the AI ​​device (100) to run an application program stored in memory (170). Furthermore, the processor (180) may combine two or more of the components included in the AI ​​device (100) to operate the application program.

[0068] FIG. 2 shows an AI server according to an embodiment of the present disclosure.

[0069] Referring to FIG. 2, the AI ​​server (200) may refer to a device that trains an artificial neural network using a machine learning algorithm or uses a trained artificial neural network. Here, the AI ​​server (200) may be composed of multiple servers to perform distributed processing, or it may be defined as a communication network. At this time, the AI ​​server (200) may include a part of the AI ​​device (100) as a component. The AI ​​device (100) may include a part of the AI ​​server (200) as a component. The AI ​​server (200) and the AI ​​device (100) may perform at least a part of the AI ​​processing together.

[0070] The AI ​​server (200) may include a communication unit (210), memory (230), a learning processor (240), and a processor (260). The communication unit (210) may transmit and receive data with an external device, such as an AI device (100). The memory (230) may include a model storage unit (231). The model storage unit (231) may store a model (or artificial neural network, 231a) that is being learned or has been learned through the learning processor (240). The learning processor (240) may train the artificial neural network (231a) using learning data. The learning model may be used while mounted on the AI ​​server (200) of the artificial neural network, or it may be used while mounted on an external device, such as an AI device (100). The learning model may be implemented in hardware, software, or a combination of hardware and software. If part or all of the learning model is implemented in software, one or more instructions constituting the learning model may be stored in the memory (230). The processor (260) can use a learning model to infer a result value for new input data and generate a response or control command based on the inferred result value.

[0071]

[0072] [Elevator Control Device]

[0073] FIG. 3 is a drawing illustrating a control device for an elevator of the present disclosure. FIG. 4 is a drawing illustrating a pet ramp, wherein (a) is a panel of the landing and (b) is a drawing of the car panel. Refer to FIG. 3 and 4.

[0074] The elevator control device (1) can analyze an image that includes a pet and a companion. Here, the image may include each entity getting on or off between the landing and the car, and each entity being inside the car. The image may include time-series changes. The elevator control device (1) can determine at least one of the pet getting on or off the elevator car in the image. The elevator control device can pair the pet and the companion in the image. Here, pairing may mean that the pet and the companion form a pair. The elevator control device can determine that the pet is getting off using the pairing. It may determine that the pet is getting off as a reaction to the companion forming the pair getting off. The elevator control device (1) may include the AI ​​device (100).

[0075] In the present disclosure, “pairing” may mean matching a pet and a passenger as a single pair by determining that a companion relationship exists between the pet and the passenger detected in an image. The pairing may be performed based on at least one of the positional relationship within the image, the travel time relationship, the leash connection relationship, the holding posture, the carrier boarding status, pre-registration information, and the judgment result of a learning model. The pairing may be maintained during a single boarding event and released after a disembarkation determination, and may be stored as a registration relationship through iterative learning according to an embodiment.

[0076] The elevator control device (1) may include a judgment unit (11) that determines whether a pet is boarding or alighting. The judgment unit (11) may include at least one learning model. Here, the learning model can perform judgments necessary for the pet's boarding and alighting. The control device (1) may include a main control unit (21) that additionally controls parts of the elevator according to the judgment of the judgment unit (11). The main control unit (21) can control the on / off of the pet ramp (2) in response to the pet's boarding and alighting. The main control unit (21) can control the opening and closing of the elevator door (3) in response to the pet's boarding and alighting. For example, the closing of the door can be delayed until the pet's alighting is completed. Here, the door may include a car door and a landing door.

[0077] The above judgment unit (11) is described. The above judgment unit may include a pet riding judgment unit (11a) that determines whether the pet is riding through the image. The pet riding judgment unit can determine whether the pet is riding. The pet riding judgment unit can pair the pet and the companion if the pet and the companion have a predetermined relationship. When paired, the companion and the pet can form a pair. The pet riding judgment unit can determine that the pet is riding by using the pet and the companion riding together and by pairing the pet and the companion. For example, the companion can ride while holding the pet. An image of the companion holding the pet can be input into the pet riding judgment unit (11a). In the image, the pet's image may be obscured by the companion. However, the pet may be briefly exposed to the image for a predetermined period of time. When the companion rides, the pet's exposure time to the image occurs in conjunction with changes in position / direction / posture. This can be used to pair the pet and the companion. A predetermined learning model can determine a companion and a pet held by the companion. To this end, the learning model can be trained in advance.

[0078] When the above pet boarding judgment unit (11a) determines that a pet has boarded, the main control unit (21) can turn on the pet lamp (2). When the above pet boarding judgment unit (11a) determines that the pairing is established, the main control unit (21) can turn on the pet lamp (2). The above pet lamp (2) may include both the pet lamp (P2) of the platform panel (P1) and the pet lamp (C2) of the car panel (C1). A user viewing the above pet lamp can know that the pet is riding in the corresponding car. The user can know in advance that the pet has boarded and respond accordingly. For example, they can use the next car.

[0079] The above judgment unit may include a pet disembarkation judgment unit (11b) that determines the disembarkation of the pet through the above image. The pet boarding judgment unit (11a) can determine that the pet's boarding is accompanied by the companion's boarding by pairing the pet with the companion. A predetermined learning model can track the companion and determine the companion's movement as the pet's movement. The pet disembarkation judgment unit (11b) can determine that the companion is disembarking from the car in the image. The pet disembarkation judgment unit can determine that the pet has disembarked when the companion disembarks. The pet disembarkation judgment unit can determine that the pet has disembarked when the companion disembarks, even if the disembarkation of the pet is not determined. Determining the companion's disembarkation as the pet's disembarkation is because the pet can be considered to be moving together with the companion.

[0080] To explain with an example, the pet may not be observed in an image of the companion disembarking. The car's camera may be positioned on the opposite side of the car's door. Accordingly, the image acquired by the camera may be an image of the companion's back. The image of the companion disembarking does not include an image of the pet being held by the companion. In this case, the disembarking of the paired companion can be determined as the disembarking of the paired pet.

[0081] When the above pet disembarkation judgment unit (11b) determines that the pet has disembarked, the main control unit (21) can turn off the pet lamp (2). The above pet lamp (2) may include both the pet lamp (P2) of the platform panel (P1) and the pet lamp (C2) of the car panel (C1). A user viewing the above pet lamp can know that the pet has disembarked from the corresponding car. The user can know in advance that the pet has disembarked and respond accordingly. For example, the user can board the current car.

[0082] The above judgment unit (11) may include a predetermined judgment algorithm. Preferably, the judgment unit may include a learning model. A server (4) that learns the learning model may be connected to the control device (1). The server may provide the learning model by learning a number of images. The server may acquire images from the control device (1). The server may include the AI ​​server (200). The elevator control device (1) may be placed adjacent to the current elevator. For example, the elevator control device (1) may be installed as part of the building control system of the building where the elevator is operated. The server (4) may be spaced apart from the building where the elevator is operated and may be, for example, a large-capacity server of the elevator service provider and the elevator manufacturer.

[0083]

[0084] [Elevator Control Method]

[0085] FIG. 5 is a flowchart illustrating a control method for an elevator according to the present disclosure, wherein (a) explains the relationship between an image and a pet ramp, (b) explains the relationship between a pairing and disembarking, and (c) explains the relationship between a pairing and a pet ramp. Refer to FIG. 5.

[0086] [[Pairing setup (refer to Fig. 5(a))]]

[0087] An image is acquired (S1). The image may include a pet and a companion. In the image, the companion may be continuously included over time. In the image, the pet may be obscured, disappear, or become unobservable over time. For this reason, it may be more difficult for the learning model to recognize the pet than to recognize the companion.

[0088] A predetermined learning model can analyze the relationship between the pet and the companion. The predetermined learning model can pair the pet and the companion (S2). For example, if the companion is holding the pet, the pet and the companion can be paired.

[0089] Pairing between a pet and a companion may imply that the pet has a specific relationship with the companion. For example, it may imply that the pet is owned by the companion. For instance, if the pet moves away from the companion despite the pairing relationship, it may indicate that the companion is mismanaging the pet. In response to this, an alarm signal may be provided to the companion. For instance, if the image is acquired from a car or a landing, it may be determined that the pet disembarks along with the companion when the companion disembarks from the car. Based on this determination, the pet disembarkation judgment unit may transmit a control signal to the main control unit. For instance, if the image is acquired from a park, if the pet moves away from the companion and approaches another person, it may be determined that the pet is engaging in an aggressive situation (against others). The control system may provide an emergency signal to the user.

[0090] [[Car boarding and alighting and pairing control (refer to FIG. 5(b))]]

[0091] Images of a pet and a passenger are obtained through images of a car or a platform (S1). The images may include a pet and a passenger who can be identified as a companion of the pet. It can be determined that the pet and the passenger have a certain relationship. For example, the passenger may be holding the pet. Whether the pet and the passenger have a certain relationship, that is, a relationship between a pet and a companion, can be determined by a learning model. For example, a learning model can be provided by learning images of a companion holding a pet. The learning model can determine whether the passenger in the image has a relationship between the pet and a companion. In this case, the pet and the passenger can be paired (S20).

[0092] The above pairing can be performed while the passenger and pet are boarding the car (while moving) and while the passenger and pet are inside the car (while stationary). Accordingly, when the companion and pet are boarding together, the pairing can be reliably established. The established pairing can be prevented from changing while riding. The established pairing can be nullified after disembarking. Accordingly, the control method can be operated reliably.

[0093] After the above pairing, it can be determined that a relationship between the pet and the passenger is established. When the passenger disembarks, it can be determined that the pet has also disembarked (S30). Even if the disembarking of the pet is not visible in the disembarking image, it can be determined that the pet is disembarking in response to the companion disembarking. This is because, except in accident situations, the paired companion and the pet disembark together.

[0094] [[Pairing and PET lamp control (refer to FIG. 5(c))]]

[0095] The acquisition of images of the pet and the passenger through an image of the car or platform (S1), and the pairing of the pet and the passenger (S20) may be as described above. Along with the pairing, the pet lamp may be turned on (S21). By turning on the pet lamp, other users can know that the pet has boarded.

[0096] After the above pairing, it can be determined that a relationship between the pet and the passenger is established. When the passenger disembarks, it can be determined that the pet has also disembarked (S30). At the same time, the pet lamp that was in the ON state can be turned OFF. By turning off the pet lamp, other users can know the disembarking status of the pet.

[0097]

[0098] <2nd Example>

[0099] The second embodiment of the present disclosure includes a specific technique for determining whether a pet is riding. This embodiment describes only the parts that differ from or are added to the first embodiment, and the description of the first embodiment is to be applied.

[0100] FIG. 6 is a drawing illustrating an elevator control device and method according to the present disclosure, wherein (a) is a method for controlling an elevator, (b) is a part of an elevator control device, and (c) is a drawing showing the overall configuration of an elevator control device in a simplified manner.

[0101] Referring to FIG. 6(a), a passenger and a pet board in the image (S110). Whether boarding has occurred can be determined through image analysis. Image analysis can be performed by a learning model. It is determined whether the passenger and the pet are registered (S210), and if they are not registered, a first pairing can be performed (S310). The first pairing can be performed by the image analysis. It is determined whether the passenger and the pet are registered (S210), and if they are registered, a second pairing can be performed (S410).

[0102] The first pairing mentioned above may mean that the current passenger and the pet establish a companion-pet relationship. The second pairing mentioned above may mean that the current passenger and the pet disembark together as companion-pets. In other words, the second pairing may mean disembarking together in addition to the first pairing.

[0103] It is explained that the first pairing is established by performing the first pairing above, and that the second pairing is established by performing the second pairing above. FIG. 6(b) shows a target that is determined by at least one learning model analyzing the image. The learning model may be recorded in the pet ride judgment unit (11a) or transmit information when necessary. An example is provided for recording. The learning model is described in detail.

[0104] The 12th learning model (12) can determine whether the leash connects the passenger and the pet in the image above. If the leash connects the two entities, the first pairing between the companion and the pet is established (S310), and this can be stored in the database. The 13th learning model (13) can determine whether the pet and the passenger ride consecutively within a predetermined time in the image above. If the pet and the passenger ride together within a predetermined time, the first pairing between the companion and the pet is established (S310), and this can be stored in the database. This is because the passenger can freely follow the pet when it is not crowded. The 14th learning model (14) can determine whether the passenger is holding the pet in the image above. If the passenger is holding the pet, the first pairing between the companion and the pet is established (S310), and this can be stored in the database. The 15th learning model (15) can determine whether a pet is riding in the carrier in the above image. If a pet is riding in the carrier, the first pairing between the companion and the pet is established (S310), and can be stored in the database.

[0105] The first pairing mentioned above may refer to a pairing between a first pet and a first companion. The first pet may also be paired with a second companion. Similarly, the first companion may also be paired with a second pet. At least two such pairings for a pet may be included. At least two pairings for a companion may be included. The pairings may form a set of at least two first pairings. The collection of at least two first pairings may be called a class. All pairings for a given family may be stored in the class. Here, all pairings may provide a learning model. In other words, they may include a learning model for the class. For example, multiple other passengers (other family members) may be identified as companions for a given pet. That is, a second pairing may be established even if another person rides with the same pet. In the step of performing the first pairing (S310) above, the passenger and pet for whom the first pairing is established may provide the 16th learning model (S311). In other words, when it is determined that the first pairing is established, the passenger and pet at that time may be trained as a new 16th learning model. The 16th learning model may include a learning model for the class. The pair of passenger and pet may be registered in the 16th learning model. There may be at least three entities to be trained. At this time, the passenger may use facial and body information, and the pet may use body information and color information. Of course, other information may also be used.

[0106] The above 16th learning model (16) can determine whether a registered passenger and a pet are boarding in the image (S210).

[0107] When the first pairing is possible by the above 12 to 15 learning models, and when the first pairing has already been established by the above 16 learning model, a second pairing can be performed for the current passenger and the pet (S411). Here, when the first pairing is possible, it can be said that the above 12 to 15 learning models are in a state where they can determine that the first pairing is possible for the current passenger and the pet. That is, it may satisfy a predetermined relationship. If the above 12 to 15 learning models determine that the first pairing is impossible, the second pairing may not be performed. If the second pairing is not performed, the method of the present disclosure may not be applied. In this case, it may be a normal operation of an elevator in a state without pairing.

[0108] The second pairing described above includes the possibility of the first pairing being established, and the possibility of determining that the pet has also disembarked when the passenger disembarks. The fact that two entities in an image can be paired in the first way by the learning models 12 to 15, and that two entities that have formed the first pairing are recognized by the image by the learning model 16, can be referred to as a boarding condition. Here, the boarding condition can be described as a condition for determining that the paired companion and the pet have boarded, or a condition in which the companion boards the pet and forms a pair. The boarding condition may be a condition that satisfies a predetermined relationship between the pet and the companion (as the owner of the pet).

[0109] Subsequently, in a pair (a pair of companion and pet) where the second pairing is established, it can be determined that the pet has also disembarked when the companion (i.e., passenger) disembarks (S411). The system can track the movement of the pair until disembarking. The movement of the pair may include both the pet and the companion, or at least one of them. In many cases, the movement of the companion can be tracked. This is because the companion is easier to observe than the pet, and the companion manages the pet.

[0110] Refer to FIG. 6(c). It may be difficult for all of the above 12 to 16 learning models to be learned and stored in the artificial intelligence device (100) as an edge device. This is because it is necessary to ensure hardware resource constraints and operational speed of the edge device. Accordingly, the storage of the above 12 to 16 learning models can be performed at the edge device (100). The learning process of the above 12 to 16 learning models can be performed by the artificial intelligence server (200) as a main server. Data for learning can be transmitted from the edge (100) to the server (200). The learning models can be transmitted from the server (200) to the edge (100).

[0111]

[0112] <Third Embodiment>

[0113] The third embodiment of the present disclosure includes a specific technique for determining the unloading of a PET. This embodiment describes only the parts that differ from or are added to the first and second embodiments, and the descriptions of the first and second embodiments shall apply.

[0114] FIG. 7 is a drawing illustrating an elevator control device and method of the present disclosure, wherein (a) is a method for controlling an elevator and (b) is a part of an elevator control device.

[0115] Referring to FIG. 7(a), the image may include a pair of a second paired companion and a pet (a pair of companion and pet). The pair may be tracked on the image. Tracking may include movement, direction, speed, and attitude. During the movement of the pair, the pair may begin disembarking (S510). Here, the initiation of disembarking may be determined by movement toward the door and arrival at the disembarking floor. Subsequently, the door may be opened and the pet lamp may flash (S511). The flashing of the pet lamp may indicate the movement status of the pet. The flashing of the pet lamp may notify the user that there may be a time extension for boarding and disembarking due to the pet.

[0116] Subsequently, when the disembarking condition is satisfied, the above pet lamp can be turned off (S520). The above disembarking condition may correspond to the above boarding condition. The above disembarking condition may mean a condition in which it is determined that the pet has disembarked. In other words, it may mean a condition in which a pair satisfying the above boarding condition for the car can be determined to have disembarked from the car.

[0117] FIG. 7(b) shows a target that is determined by at least one learning model analyzing the image. The learning model may be recorded in the pet disembarkation judgment unit (11b) or transmit information when necessary. An example is provided for recording. The learning model will be described in detail.

[0118] The 22nd learning model (22) can determine whether a companion in the second paired pair disembarks. If a companion in the pair disembarks, the entire pair disembarks (S520), and the pet lamp can be turned off. The 23rd learning model (23) can determine whether the carrier carrying the pet disembarks. If the carrier disembarks, the pet disembarks (S520), and the pet lamp can be turned off. The 24th learning model (24) can determine whether a companion holding the pet disembarks. If the companion disembarks, the pet disembarks (S520), and the pet lamp can be turned off. The 25th learning model (25) can determine whether the pet and companion riding consecutively disembark. When both the pet and the companion disembark, the pet lamp can be turned off as the pet has disembarked (S520). The 26th learning model (26) can determine whether both the pet and the companion connected by the leash have disembarked. When both the pet and the companion disembark, the pet lamp can be turned off as the pet has disembarked (S520). Turning off the pet lamp can be performed by providing an off signal for the pet lamp.

[0119] The above 23 to 26 learning models may be applied when the above 1 pairing is not established and the above 2 pairing is established. For example, they may be preferably applied when riding in a car with a pet for the first time, or when the amount of learning of the learning model is insufficient.

[0120] The above 25th and 26th learning models may be applied even when the first pairing and the second pairing are established. This is because the separation of the pet and the companion may cause problems for the safety of the pet and the user.

[0121] Until the above disembarking conditions are met, the PET lamp may be turned on and / or the time for opening the door (including a delay in the closing time) may be extended. Upon the fulfillment of the above disembarking conditions, the door may be closed and / or a notification of disembarking may be sent to a companion of the same class.

[0122] According to the present disclosure, automatic operation of the pet lamp can be performed in an elevator.

Claims

1. A pet boarding judgment unit that pairs the pet and the passenger when the pet and the passenger have a predetermined relationship in at least one first image among the car and the platform; A pet disembarkation determination unit that determines that the paired pet has disembarked when the paired passenger disembarks in at least one second image of the car and the platform; and A main control unit is included that turns on the PET lamp when the PET boards at the PET boarding judgment unit and turns off the PET lamp when the PET disembarks at the PET disembarking judgment unit. Elevator control device.

2. In Paragraph 1, The above PET disembarkation determination unit is an elevator control device that determines that the PET has disembarked even if it is not determined whether the PET has disembarked.

3. In Paragraph 1, The above main control unit is a control device for an elevator that turns on the PET lamp when the pairing is established in the PET boarding judgment unit.

4. In Paragraph 1, The riding condition satisfying a predetermined relationship between the pet and the passenger in the above pet riding judgment unit is, Determining whether a leash connects the passenger and the pet in the first image, determining whether the pet and the passenger ride consecutively within a predetermined time in the first image, determining whether the passenger is holding the pet in the first image, determining whether the pet is riding inside a carrier in the first image, and determining whether the passenger and the pet ride in the first image while in a pre-registered state. including at least one of Elevator control device.

5. In Paragraph 1, The disembarking condition in which the above PET disembarking judgment unit determines that the PET is disembarking is, In the second image above, determining whether the carrier carrying the pet is disembarking, and determining whether the companion holding the pet is disembarking. including at least one of Elevator control device.

6. Acquiring at least one image in a time series; and In the first image among the above images, if the pet and the companion satisfy a predetermined relationship, pairing the pet and the companion; and A method for controlling an elevator, comprising providing a signal corresponding to the pairing in response to the movement of at least one of the paired companion and the pet in the second image among the above images.

7. In Paragraph 6, A method of controlling an elevator, wherein the above pairing is performed while the pet and the companion are boarding the car, and while the pet and the companion are inside the car.

8. In Paragraph 7, A method for controlling an elevator in which, in the second image above, it is determined that the pet has also disembarked when the companion disembarks from the car.

9. A method for controlling an elevator according to claim 8, wherein, even when the disembarking of the pet is not visible in the second image, the pet is determined to be disembarking in response to the companion disembarking.

10. In Paragraph 6, A method for controlling an elevator, performing at least one of turning on a pet lamp when the above pairing is established, and providing a signal to turn off the pet lamp when the above companion moves.

11. In Paragraph 6, The above predetermined relationship is, In the first image above, a leash connecting the companion and the pet; in the first image above, the pet and the companion riding continuously within a predetermined time; in the first image above, the companion holding the pet; and in the first image above, the pet riding inside a carrier. A method for controlling an elevator, comprising at least one of the following.

12. In Paragraph 11, For the companion and the pet satisfying the predetermined relationship above, a first step of performing a first pairing, and a second step of storing the result of the first pairing as a learning model; A method for controlling an elevator, wherein the above first step and the above second step are repeated, and the learning model learns all pets and all family members of a household.

13. In Paragraph 12, In response to satisfying at least one of the cases in the first image above, where the pet and the companion satisfy a predetermined relationship, and where the pet and the companion are registered in the learning model, Performing a second pairing as the pairing of the pet and the companion in the first image above, Elevator control method.

14. In Paragraph 6, Providing a signal corresponding to the above pairing is provided for the paired pair to disembark from the elevator, and Providing an off signal for the pet lamp by having the entire pair disembark when a companion in the pair disembarks. When the carrier carrying the above pet disembarks, the above pet has disembarked, thereby providing an off signal for the above pet lamp. When a companion holding the pet disembarks, the pet is deemed to have disembarked, thereby providing an off signal for the pet lamp. When both the pet and the companion who rode consecutively have disembarked, providing an off signal for the pet lamp as the pet has disembarked, and When both the pet connected by a leash and the companion disembark, the pet is deemed to have disembarked, thereby providing an off signal for the pet lamp. A method for controlling an elevator, comprising at least one of the following.

15. An elevator to which any one of the devices or methods of paragraphs 1 through 14 is applied.

16. A building to which any one of the devices or methods of paragraphs 1 through 14 is applied.

17. A pairing determination unit that determines a companion relationship between a pet and a companion in a first image obtained from at least one of a car and a platform, and matches the pet and the companion as a pair; A companion disembarkation determination unit that determines whether the companion included in the pair disembarks in a second image obtained from at least one of a car and a platform; A pet disembarkation determination unit that determines that the pet included in the pair has disembarked when the disembarkation of the companion is determined, even if the pet is not detected in the second image above or the disembarkation of the pet is not directly confirmed; and It includes a main control unit that outputs a guidance signal indicating the pet boarding status based on the judgment result of the pairing judgment unit, the companion disembarkation judgment unit, or the pet disembarkation judgment unit. The above guidance signal includes at least one of an on / off signal of a pet lamp, a display output signal, a voice guidance signal, a door control signal, and a control server transmission signal. Elevator control device.

18. In Paragraph 17, The above guidance signal includes a door control signal, and The main control unit outputs a door control signal that delays the closing of the car door or landing door until the pet disembarkation determination unit determines that the pet has disembarked. Elevator control device.