Elevator control device

Abstract

An elevator control device 10 according to the present disclosure comprises a detection unit 11b, an assessment unit 11c, and a control unit 11. The detection unit 11b receives an "up-cabin weight," which is the weight of a passenger or passengers inside a cabin 20, and an "up-hall weight," which is the weight of a passenger or passengers in a hall. Based on the "up-cabin weight" and the "up-hall weight" received from the detection unit 11b, the assessment unit 11c performs an initial assessment as to whether the passenger(s) can enter the cabin 20 or not. The control unit 11a controls an elevator device 100 according to the assessment by the assessment unit 11c.

Description

[Technical field]

[0001] The present invention relates to an elevator control device. [State of the art]

[0002] PTL 1 discloses an elevator control device that photographs passengers in a hall and passengers in a car with a camera, predicts an area in the car occupied by the passengers when the passengers in the hall get into the car, compares the area with a total area in the car, and thereby predicts whether there are one or more passengers left behind in the hall, that is, one or more passengers who cannot get into the car. [Citation list][Patent literature]

[0003] [PTL 1] JP 2002-302348 A [Summary of the invention]

[0004] The elevator control device described above does not include any means for obtaining the weight of one or more passengers in the car and therefore has the problem that it cannot judge whether or not the passenger(s) can board the car taking into account the weight of the passenger(s) in the car.

[0005] The present disclosure is presented in view of the above-described problem, and it is an object of the present disclosure to judge, in an elevator control device, whether or not the passenger(s) can get into a car in consideration of a weight of the passenger(s) in a hall. [Means of solving the problem]

[0006] An elevator control device according to the present disclosure includes a detection unit that detects an on-car weight, which is a weight of a passenger or passengers inside a car, and an on-hall weight, which is a weight of a passenger or passengers in a hall, a judgment unit that makes a first judgment as to whether or not the passenger or passengers in the hall can get into the car based on the on-car weight and the on-hall weight detected by the detection unit, and a control unit that controls an elevator device according to the judgment of the judgment unit. [Advantageous effects of the invention]

[0007] According to the present disclosure, in the elevator control device, it can be judged whether or not the passenger(s) can get into a car by taking into account the weight of the passenger(s) in the hall. [Brief description of the drawings] [ Fig. 1] Fig. 1 is a figure showing the elevator apparatus including an elevator control device in a first embodiment. [ Fig. 2] Fig. 2 is a figure showing a configuration of the elevator apparatus including the elevator control apparatus in the first embodiment. [ Fig. 3] Fig. 3 is a flowchart showing the control when the control device in the first embodiment judges whether or not to hold a car in a hall. [ Fig. 4] Fig. 4 is a flowchart showing the control when the control device in the first embodiment receives the on-Hall weight. [ Fig. 5] Fig. 5 is a flowchart showing the control when the control device in the first embodiment issues a notification indicating a conditional stop. [ Fig. 6] Fig. 6 is a diagram showing information in a priority database of the control device of the first embodiment. [ Fig. 7] Fig. 7 is a flowchart showing control when the control device in the first embodiment updates a learned human weight model. [ Fig. 8] Fig. Figure 8 is a diagram of a three-layer neural network. [ Fig. 9] Fig. 9 is a flowchart showing the control when the control device of the second embodiment judges whether or not a car is to be stopped at a hall. [ Fig. 10] Fig. 10 is a flowchart showing the control when the control device in the third embodiment receives an on-Hall weight. [Description of the Embodiments]First Embodiment

[0008] An elevator device 100 including an elevator control device according to a first embodiment will be described in detail below based on the drawings. Note that like reference numerals in the corresponding drawings denote the same or corresponding components and steps.

[0009] Fig. 1 is a figure showing the entire elevator device 100 in a first embodiment. First, using Fig. 1, an overview of the entire elevator apparatus 100 and an operation of the present embodiment will be described.

[0010] The elevator device 100 includes a control device 10, which is an elevator control device, a car 20, a hall call device 31, a display device 32 that displays a message, an image pickup device 33 that photographs a hall, a rope 41, a drive device 42 that moves the car 20, and doors of unillustrated halls. Furthermore, the hall call device 31, the display device 32, and the image pickup device 33 are provided in the hall. The car 20 is provided with a car call device 21 and a load weighing device 22.

[0011] The hall call is used to perform registration to cause the car 20 to travel to a hall where an operation of calling the car 20 is performed using the hall call device 31 used in the hall. The call of the car causes the car 20 to travel to a hall on a floor designated by a passenger using the car call device 21 included in the car 20.

[0012] When the hall call and the car call are made, the control device 10 causes the car 20 to travel to the destination floors in the order in which the hall call and the car call are made. The destination floor is a floor to which the car 20 travels, as registered by the car call and the hall call.

[0013] It should be noted that when the car call is made to a hall on the way to the destination floor, car 20 will first stop in the hall on the floor specified by the car call. For example, even if the car call with the fifth floor as the destination floor is already made in the hall on the first floor, if the car call is made for the hall on the fourth floor, which is the hall on the way to the fifth floor, car 20 will first stop in the hall on the fourth floor and then travel to the hall on the fifth floor.

[0014] On the other hand, when the call is made by a passenger in a hall en route to the destination floor, the control device 10 obtains an on-car weight, which is a weight of the passengers inside the car 20, and an on-hall weight, which is a weight of the passengers in the hall. Subsequently, the control device 10 performs a first judgment to judge whether or not the passengers in the hall can board the car 20 based on the on-car weight and the on-hall weight. The control device 10 controls the elevator device 100 depending on a result of the first judgment.

[0015] Specifically, if no passenger can board the car, the car 20 is caused to bypass the hall, and if even one passenger can board the car, the car 20 is stopped in the hall where the call is made. Further, if the car 20 is caused to bypass (or bypass or overpass) the hall, and even if there is even one passenger who cannot board the car, even if the car 20 is stopped in the hall, a notice is displayed in the hall using the display device 32.

[0016] According to the present embodiment, a judgment as to whether or not the passenger(s) can board the car can be made taking into account the weight of the passenger(s) in the hall. By controlling whether or not the car 20 stops in the hall according to the judgment as to whether or not the passenger(s) can board the car, it is possible to avoid the situation where the car 20 cannot stop in the hall while the passenger(s) cannot board the car 20, and it is possible to improve the transportation efficiency of the passengers. Furthermore, when a passenger cannot board the car 20, the notice is given in the hall, whereby the passenger in the hall can make decisions such as stopping waiting for the car 20, and therefore the passenger in the hall can use the time effectively.

[0017] Next, using Fig. 2, a configuration of the control device 10 is described in detail. The control device 10 includes a processor 11 that performs the control, a storage unit 12, and an interface 13.

[0018] The processor 11 is a CPU (Central Processing Unit) and is connected to the storage unit 12 and the information exchange interface 13. The processor 11 includes a control unit 11a, a detection unit 11b, an evaluation unit 11c, and a model generation unit 11d.

[0019] The control unit 11a comprises software modules that control the detection unit 11b, the assessment unit 11c and the model generation unit 11d as well as the control of the entire elevator device 100 including the car 20, the hall call device 31, the display device 32, the image recording device 33, the drive device 42 and the doors of the halls not shown.

[0020] Specifically, the control unit 11a includes a software module that registers a destination floor of the car 20 when making a hall call and a car call. Further, the control unit 11a includes a software module that starts a process for judging whether the car 20 should stop in the hall or bypass the hall when making a call in the hall en route to the destination floor of the car 20. Furthermore, the control unit 11a includes a software module that controls the control of the drive device 42 to cause the car 20 to stop at the hall or bypass the hall, depending on the judgment of the judgment unit 11c described later. Additionally, the control unit 11a includes a software module that causes the display device 32 to display a message according to the judgment of the judgment unit 11c.

[0021] The acquisition unit 11b includes a software module that receives an on-car weight, i.e., the weight of the passenger(s) within the car 20, from the load weighing device 22. Furthermore, the acquisition unit 11b includes a software module that acquires an image captured by the image capture device 33, a software module that inputs the acquired image information into a learned model for passenger weight and outputs the weights of the passengers in the hall individually to acquire the passenger weight, and a software module that adds the acquired passenger weights to calculate the on-car weight of the entire hall. Furthermore, the acquisition unit 11b includes a software module that inputs the acquired image information into a learned model for attribute information and outputs the attribute information of the passengers in the hall individually to obtain the attribute information.

[0022] In the present disclosure, the on-hall weight means the weight of the passenger or passengers in the hall and includes the weight of a specific part (for example, the weight of a person) of the passengers in the hall and a sum of the weights of all the passengers in the hall.

[0023] Here, the learned model for human weight of the present embodiment is a learned model for outputting a weight of a passenger in the hall from the image information. Further, the learned model for attribute information is a learned model for outputting an attribute of the passenger in the hall from the image information. Note that the image information in the present embodiment is the image data itself, but it may be a set of features such as skeleton information extracted from the image data by a known algorithm. Furthermore, in the present disclosure, the feature is the classification of people such as a person carrying a stretcher, a pregnant woman, an elderly person, or a person carrying a heavy load.

[0024] The judging unit 11c includes a software module that judges whether or not the passenger in the hall can board the car 20 based on the on-car weight and the on-hall weight obtained from the detecting unit 11b with respect to the car 20 for which the call is made by the passenger in the hall on the way to the destination floor.

[0025] The model generation unit 11d includes a software module that generates a learned model for deriving the on-car weight from the image information using learning data including the image information of the person in the hall acquired by the image pickup device 33 through the acquisition unit 11b and the on-car weight acquired by the load weighing device 22 through the acquisition unit 11b when the person in the hall enters the car 20. In the present embodiment, the learning data generated based on a combination of the image data and the weight of the person in the image is stored in the storage unit 12 described later.In the present embodiment, the software module is a software module that updates the learning data by adding, using the updated learning data, a combination of image information of the passenger in the hall newly acquired by the image pickup device 33 and a difference between car weights acquired by the load weighing device 22 before and after the passenger in the hall gets into the car 20, which corresponds to the weight of the person in the image, to this learning data, and a software module that generates a learned human weight model that outputs a human weight when the image information is input. Here, the learning data is data in which the image information and the weight of the person in the image are linked.Further, the weight of the person in the image in the learning data of the present embodiment is a weight including luggage or the like held by the person in the image.

[0026] The storage unit 12 is a storage device configured by a non-volatile memory and a volatile memory. The storage unit 12 stores the learned model for human weight and the learned model for attribute information. Furthermore, the storage unit 12 stores a priority database 50 described later. Furthermore, the storage unit 12 temporarily stores information generated by the process of the processor 11, as well as information input via the interface 13 from the car 20, the hall call device 31, and the image pickup device 33.

[0027] The interface 13 includes connectors that connect electrical lines (not shown) and are connected to the car 20, the hall call device 31, the display device 32, the image pickup device 33, and the drive device 42. Note that the interface 13 also includes connectors required for other movements of the elevator device 100. Furthermore, the interface 13 can be connected to other configurations as a wireless communication device.

[0028] The following are the other configurations of the elevator device 100 using Fig. 1 described.

[0029] The car call device 21 provided on the car 20 is a button device including a plurality of buttons in which names of floors are described, and is a button device that, when pressed by a passenger, outputs information about the presence of a car call and the pressed button to the control device 10.

[0030] The load weighing device 22 contained in the cabin 20 cyclically measures a weight of the entire interior of the cabin 20 and outputs the weight to the control device 10.

[0031] The hall call device 31 is located at each hall and includes a button device for making a hall call. The device includes a button for an ascending direction and a button for a descending direction. When the button is pressed by a passenger, the button device outputs the fact that a hall call is present, as well as information about the direction of the pressed button, to the control device 10.

[0032] The display device 32, which is a notification device, is provided in each hall and includes an LCD (liquid crystal display) that displays a message issued by the control device 10. Furthermore, the display device can be any device if it can indicate a message issued by the control device 10, and can be a projector or a voice. Furthermore, if a user can understand the meaning, a color of a lamp can indicate whether the car 20 is bypassing or stopping.

[0033] The image pickup device 33, which is a camera, is attached to each hall, photographs the hall, and outputs the photographed image to the control device 10.

[0034] The drive device 42 includes a traction machine for moving the cabin 20 according to an instruction issued by the control device 10. The traction machine moves the cabin 20 by winding the cable 41 connected to the cabin 20 and a counterweight (not shown).

[0035] The following describes an operation of the present embodiment using the Fig. 3 to Fig. 8 described. Fig. 3 is a flowchart showing the control of the control device 10 when judging whether to stop the car 20 at the hall or to cause the car 20 to bypass the hall.

[0036] In the present embodiment, the control unit 11a of the control device 10 receives the information output from the hall call device 31 or the car call device 21 via the interface 13 and registers the destination floor of the car 20 when a hall call or a car call is made by a passenger or a robot. The control unit 11a causes the car 20 to travel to the hall of the destination floor in the order in which the hall call and the car call are made. Specifically, an instruction is issued to the drive device 42, and the car 20 is caused to move to the destination floor.

[0037] When a call is made by a passenger in the hall on the way to the destination floor from the current position of the car 20, the control unit 11a starts a process of judging whether the car 20 should be made to stop in the hall where the call is made or the Fig. 3. Specifically, the process is started when the car call is made by the passenger through the button in the same direction as the destination floor of the car 20. More specifically, the control unit 11a starts the process when the passenger presses the up direction button of the button device included in the hall call device 31 and the control unit 11a receives the information output from the button device via the interface 13 when the destination floor of the car 20 is a higher floor than the hall. Note that when the car 20 is moving and the car 20 cannot be decelerated and stopped before reaching the hall, the car may be caused to bypass the hall without performing the judgment.

[0038] When the hall call is made by the passenger in the hall and the detection unit 11a starts the process, the detection unit 11b receives the in-car weight, i.e., the weight of the passenger or passengers inside the car 20, in step S1. Specifically, the weight of the entire interior of the car 20 output by the load weighing device 22 is received via the interface 13 and temporarily stored in the storage device 12. Subsequently, the detection unit 11b processes the process until step S2.

[0039] In step S2, the acquisition unit 11b obtains the on-hall weight, i.e., the weight of the passenger or passengers in the hall. More specifically, the image obtained by photographing the hall in which the hall call is made and output from the image pickup device 33 is received via the interface 13 and temporarily stored in the storage unit 12. Subsequently, the weight of each passenger or passengers in the hall and the total weight, i.e., the on-hall weights, are output from the image information of the captured image using the learned model for deriving the on-hall weight from the image information and temporarily stored in the storage unit 12. Subsequently, the acquisition unit 11b further processes the process in step S31. Details of step S2 will be explained later using Fig. 4 described in detail.

[0040] In step S31, the judging unit 11c judges whether or not all passengers in the hall where the hall call is made can board the car based on the on-car weight and the on-hall weight obtained from the detecting unit 11b. Specifically, the judging unit 11c judges whether or not the sum of the on-car weight and the sum of the weights of all passengers in the hall, temporarily stored in the storage unit 12, exceeds a rated load. If the sum does not exceed the rated load, the judging unit 11c judges that all passengers can board the car and processes the process to step S32. If the sum exceeds the rated load, the judging unit 11c judges that not all passengers can board the car and processes the process to step S33.

[0041] Note that, as described later, in the process after step S33, the control unit 11a causes the display device 32 to transmit the notification of whether or not the passenger(s) can board the car to the passenger(s) in the hall. Therefore, if the judgment unit 11c determines in step S31 that at least one or more passengers in the hall cannot board the car, the control unit 11a causes the elevator device 100 to transmit the notification of whether or not the passenger(s) can board the car to the passenger(s) in the hall.

[0042] In step S32, the control unit 11a issues an instruction to the drive device 42, causing the car 20 to stop in the hall where the hall call is being made. The control unit 11a then completes the process.

[0043] In step S33, the judging unit 11c judges, based on the on-car weight and the on-hall weight obtained from the detecting unit 11b, whether or not there is a passenger who can board the car among the passengers in the hall where the hall call is made. Specifically, the judging unit 11c calculates, for each of the passengers in the hall, a sum of the on-car weight temporarily stored in the storage unit 12 and the weight of each passenger in the hall, and judges whether or not there is a combination that is below the rated load. If the combination is below the rated load, the judging unit 11c judges that a passenger can board the car and processes the process to step S4.If there is no combination below the rated load, the judging unit 11c judges that there is no passenger who can get into the cabin and processes the process further to step S34.

[0044] In the present embodiment, the first judgment unit refers to step S31 and step S33, but the first judgment unit may be to judge whether the passenger or passengers in the hall can board the car 20 based on the on-car weight and the on-hall weight obtained from the detection unit 11b, and may be only step S31 or only step S32.

[0045] In step S4, the control unit 11a issues an instruction to the display device 32 to display a message indicating a conditional stop where not all passengers in the hall can board the car at the stop of the next car 20. Subsequently, the control unit 11a controls the process to step S32. Details of step S4 will be explained later using the Fig. 5 to 7.

[0046] In contrast, in step S34, the control unit 11a issues an instruction to the display device 32 to indicate that the car 20 is bypassing the hall. Subsequently, the control unit 11a controls the process to step S35. In step S35, the control unit 11a issues an instruction to the drive device 42, causing the car 20 to bypass the hall where the hall call is being made. Subsequently, the control unit 11a completes the process.

[0047] Accordingly, the control device 10 can make a judgment as to whether or not the passenger(s) can board the car, taking into account the weight of the passenger(s) in the hall. By controlling whether or not the car 20 stops in the hall according to the judgment as to whether the passenger(s) can board the car, it is possible to avoid the situation where the car 20 cannot stop in the hall while the passenger(s) cannot board the car 20, and to improve the efficiency of passenger transportation. Furthermore, when a passenger cannot board the car 20, the notice is displayed in the hall, whereby the passenger in the hall can determine not to wait for the car 20 or the like, and therefore the passenger in the hall can effectively use the time.

[0048] Next, a process of obtaining the on-Hall weight by the detecting unit 11b in step S2 using Fig. 4 is described in detail. Step S2 comprises four steps from step S21 to step S24.

[0049] In step S21, the acquisition unit 11b obtains image information. Specifically, the acquisition unit 11b receives the image obtained by photographing the hall where the hall call is made and output from the image pickup device 33 via the interface 13, and temporarily stores the image information in the storage unit 12. Subsequently, the acquisition unit 11b continues the process until step S22.

[0050] In step S22, the acquisition unit 11b inputs the image information acquired by the acquisition unit 11b in step S21 into the learned human weight model stored in the storage unit 12 and obtains the on-hall weight. Subsequently, the acquisition unit 11b continues the process in step S23. Specifically, the on-hall weight output by the acquisition unit 11b using the learned human weight model in step S22 is the human weight, which is the weight of each passenger in the hall.

[0051] In step S23, the acquisition unit 11b stores all human weights, ie, the on-hall weight obtained in step S22. Subsequently, the acquisition unit 11b processes the process until step S24.

[0052] In step S24, the detection unit 11b adds the passenger weights stored in the storage unit 12 in step S23, calculates the sum of the weights of all passengers in the hall, which is the on-hall weight, and stores it in the storage unit 12. Then, the detection unit 11b processes the process until step S31.

[0053] According to the above, it is possible to easily obtain the weights of the passenger(s) in the hall. Note that the method for obtaining the on-hall weight to solve the problem is not limited to this. For example, weight information can also be obtained from drinking water terminals owned by the passenger(s).

[0054] Next, step S4 is performed using Fig. 5 and Fig. 6 is described in detail. It comprises four steps, namely steps S41 to 44.

[0055] In step S41, the acquisition unit 11b obtains attribute information of the passenger or passengers in the hall. Specifically, the image information of the image obtained by photographing the hall where the hall call is made and stored in the storage unit 12 in step S21 is input into the learned attribute information model, and the attributes of all passengers in the hall to be issued are temporarily stored in the storage unit 12. Subsequently, the acquisition unit 11b processes the process until step S42.

[0056] In step S42, the judging unit 11c performs a second judgment to judge, based on the attribute information obtained from the acquiring unit 11b, whether or not there is a high-priority passenger who needs to board the car in the hall. Specifically, the judging unit 11c refers to the attributes of the passengers in the hall stored by the acquiring unit 11b in step S41 and the priority database 50 stored in the storage unit 12, and identifies priority information of the passengers in the hall. Fig. 6 is a diagram showing the information in the priority database 50. In the priority database 50, the attribute information 51 and the priority information 52 are stored by linking them together. Then, the judging unit 11c judges whether or not the priority information 52 of the highest priority or priority is present in the priority information of the passengers in the hall.

[0057] When the judging unit 11c judges that the passenger with high priority for boarding the car is present in the hall, that is, when the priority information 52 of the highest priority or priority is present in the priority information of the passengers in the hall, the judging unit 11c processes the process to step S43. Further, the judging unit 11c stores the information identifying the passenger whose priority information is the highest priority or priority in the storage unit 12.

[0058] On the other hand, when the judging unit 11c judges that there is no passenger with high priority to board the car in the hall, that is, when there is no priority information 52 of the highest priority or priority in the priority information of the passengers in the hall that is identified, the process is processed to step S44.

[0059] In step S43, the control unit 11a issues an instruction to the display device 32 to display a message indicating that priority for boarding the cabin is given to the passenger who has a high priority in step S42. Specifically, the message is created based on the information identifying the passenger whose priority is the highest, which is stored and output in the storage unit 12 in step S42. For example, if the identifying information is attribute information, the display device 32 is caused to display the message "please give priority to a pregnant woman when boarding the cabin" or the like. The identifying information may also be information other than attributes, for example, "please give priority to people wearing red clothing when boarding the cabin."Subsequently, the control unit 11a controls the process to step S32.

[0060] It should be noted that, in the present disclosure, the notification that a particular passenger is given priority for boarding the cabin means that there is a passenger among the other passengers who cannot board the cabin, and is therefore comprehensively included in the notification of whether or not the passengers can board the cabin. Furthermore, in the present embodiment, when a plurality of passengers are determined to be high priority, the control unit 11a causes the display device 32 to display a notification of the priority of all the passengers determined to be high priority.

[0061] In step S44, the control unit 11a issues an instruction to the display device 32 to display a message indicating the number of people who can board the car. Subsequently, the control unit 11a controls the process to step S32.

[0062] It should be noted that in the present disclosure, the notification indicating the number of persons who can board the cabin means that there is a passenger among the passengers who cannot board the cabin, and is therefore included in the notification of whether the passenger(s) can board the cabin or not.

[0063] Thus, the passenger(s) waiting in the hall where the hall call is made can know more accurately whether or not they can board the car when car 20 stops, and use their time more effectively. Furthermore, if there is a passenger with the attribute who should board the car first, the passenger can easily board car 20 first.

[0064] Next, using Fig. 7 and Fig. 8 describes the updating of the learned model for human weight. Fig. 7 is a flowchart showing the control of the control device 10 when the learned human weight model is updated.

[0065] In the present embodiment, the control unit 11a starts a process of updating the learned human weight model when the car 20 arrives at the hall which is the hall where the hall call is made but is not at the destination floor of the hall call.

[0066] When the control unit 11a starts the process, the detection unit 11b obtains an image of the hall through a process similar to step S21 in step S71 and processes the process until step S72.

[0067] In step S21, the detection unit 11b judges whether the number of passengers in the hall in the image stored in the storage unit 12 in step S71 is one or not. If the number of passengers is one, the detection unit 11b advances the process to step S73, and if it is not one, the detection unit 11b terminates the process.

[0068] In step S73, the on-cab weight is obtained by a process similar to step S1, and the process proceeds to step S74.

[0069] In step S74, the control unit 11a opens the hall gate and the car door 20 and processes the process to step S75. In step S75, the control unit 11a closes the hall door and the car door 20 through a normal process when a door closing command is issued by the user's operation of the car call device 21 or a certain amount of time passes. Then, the process proceeds to step S76.

[0070] In step S76, the detection unit 11b obtains the on-cab weight through a process similar to step S1 and step S73 and further processes the process to step S77.

[0071] In step S77, the detection unit 11b calculates a difference between the on-car weights obtained from the load weighing device 22 before and after the passenger(s) board(s) the car 20 in the hall. Specifically, the difference between the on-car weights stored in the storage unit 12 in step S73 and step S76 is calculated and stored in the storage unit 12. Subsequently, the detection unit 11b processes the process until step S78. The difference between the on-car weights obtained from the load weighing device 22 before and after the passenger(s) board(s) the car 20 in the hall is the weight of the passenger(s) in the hall.

[0072] In step S78, the model generation unit 11d updates learning data generated based on the combination of the image information stored in the storage unit 12 and the weight of the person in the image. Specifically, the model generation unit 11d adds the combination of the image information acquired by the acquisition unit 11b in step S71 and the difference between the weights in the car 20 detected by the load weighing device 22 before and after the passenger(s) board the car 20 in the hall, calculated by the acquisition unit 11b in step S77, to the learning data as the image information and the weight of the person in the image. Subsequently, the model generation unit 11d processes the process until step S79.

[0073] In step S79, the model generation unit 11d generates a learned model for deriving the weight in the cabin from the image information using the learning data including the image data of the passenger(s) in the cabin and the difference between the weights in the cabin captured by the image pickup device 22 before and after the passenger(s) boarded the cabin 20. Then, the process proceeds to step S710. Specifically, the model generation unit 11d generates the learned model for the human weight using the learning data updated in step S78.

[0074] A generation processing process of the learned model by the model generation unit 11d will be described in detail. A learning algorithm used by the model generation unit 11d in the present embodiment is the well-known supervised learning. As an example, a case where a neural network is applied will be described.

[0075] The model generation unit 11d learns the on-hall weight of a passenger through so-called supervised learning based on a neural network model, for example. Supervised learning refers herein to a method in which the learning device is given a set of input and result data to thereby learn features in the learning data and derive results from the inputs.

[0076] The neural network is configured with an input layer consisting of a large number of neurons, an intermediate layer consisting of a large number of neurons (hidden layers), and an output layer consisting of a large number of neurons. The number of intermediate layers can be one, two, or more.

[0077] In the case of a three-layer neural network, as in Fig. For example, as shown in Figure 8, when a plurality of inputs are input to the input layers (X1 to X3), their values ​​are multiplied by the weights W1 (w11 to w16) to be input to the intermediate layers (Y1 to Y2), and their results are further multiplied by the weights W2 (w21 to w26) to be output from the output layers (Z1 to Z3). The outputs vary depending on the values ​​of the weights W1 and W2.

[0078] In the present embodiment, the neural network learns the weight of a passenger in the cabin, which is the weight in the cabin, by so-called supervised learning according to the learning data generated based on the combination of the image data and the weight of the person in the image, including the combination of the image data of the passenger or passengers in the cabin acquired by the acquisition unit 11b and the difference between the weights in the cabin acquired by the load weighing device 22 before and after the passenger or passengers board the cabin 20.

[0079] In other words, the neural network learns by adjusting the weights W1 and W2 so that the result output by the output layer after inputting the image data, i.e., the image information, to the input layer is close to the weight of the person in the image. The model generation unit 11d generates and outputs the learned human weight model by performing the learning as above.

[0080] In step S710, the model generation unit 11d stores the learned human weight model generated in step S79 in the storage unit 12 to update it. At this time, the previously learned human weight model is deleted. Then, the model generation unit 11d completes the process.

[0081] According to the above, the learned human weight model is updated, and the detection unit 11b can obtain the on-hall weight more accurately.

[0082] Note that in the present embodiment, the learned model for human weight is stored in advance in the storage unit 12. This is the result of storing the learned model, which was created through so-called supervised learning according to the image data generated by another learning device based on the combination of the image information and the weight of the person in the image, in the storage device 12. The learned model for attribute information is also created in a similar manner and stored in the storage unit 12.

[0083] Second Embodiment If the judging unit 11c judges in step S33 of the first embodiment that there is no passenger who can board the car, the process is advanced to step S34, and the process in which the car 20 bypasses the hall is performed. In the present embodiment, the car 20 is caused to stop in the hall if there is a passenger with high priority for boarding the car in the hall, even if it is judged after the first judgment that there is no passenger who can board the car. Hereinafter, a difference from the first embodiment will be explained mainly using Fig. 9 described.

[0084] First, a configuration of the present embodiment will be described. In the present embodiment, an elevator device 100 includes a display device similar to a display device 32 in a hall within a car 20.

[0085] A control unit 11a of the present embodiment includes a software module that causes the car 20 to stop in a hall regardless of a result of the first judgment when a judgment unit 11c judges that there is a high-priority passenger who needs to board the car in the hall by a second judgment.

[0086] Next, an operation of the present embodiment will be described. In the present embodiment, when the judging unit 11c judges that there is no passenger entering the cabin in step S33 in Fig. 3, the process is processed to step S51.

[0087] In step S51, an acquisition unit 11b obtains attribute information of the passenger or passengers in the hall, similarly to step S41 in the first embodiment. Then, the process proceeds to step S51. In step S52, the judging unit 11c performs the second judgment to judge whether or not a high-priority passenger needs to board the car, similarly to step S42. Then, when the judging unit 11c judges that a high-priority passenger needs to board the car in the hall, it processes the process further to step S53. On the other hand, when it is judged that there is no high-priority passenger to board the car in the hall, that is, when there is no priority information 52 with the highest priority or priority in the priority information of the passenger or passengers in the hall that is determined, the process proceeds to step S55.

[0088] Note that in step S42, it is judged that a high-priority passenger is to be boarded if the passenger with the highest priority or priority of the priority information 52 is present, but in the judgment in step S52, an evaluation standard thereof may be changed. For example, in the present embodiment, it is judged that there is a high-priority passenger for boarding the cabin if the passenger with the highest priority of the priority information 52 is present.

[0089] In step S53, the control unit 11a issues an instruction that causes the display device 32 and the display device within the car 20 to display a message that the passenger determined to have high priority in step S52 can board the car with priority. The process then proceeds to step S54. In step S54, the control unit 11a stops the car 20 in the hall where the hall call is being made, similarly to step S32, and completes the process.

[0090] In step S55, the control unit 11a issues an instruction to the display device 32 to display a notification that the car 20 is bypassing the hall, similar to step S34, and advances the process to step S56. In step S56, the control unit 11a causes the car 20 to bypass the hall where the hall call is being made, similar to step S35, and completes the process.

[0091] According to the above, even in the case where the car 20 bypasses the hall in the first embodiment, the passenger with high priority for boarding can get into the car 20. This is particularly useful when there is a passenger who has difficulty waiting in the hall for a long time (for example, when there is a passenger with the attribute of a stretcher or a pregnant woman who has the highest priority of the priority information 52 in Fig. 6), and can reduce the waiting time of such a passenger. Furthermore, in the present embodiment, the display device is provided within the cabin 20 so that the passenger(s) in the cabin 20 can know in advance that they need to get off en route.

[0092] Third Embodiment In step S2 of the first embodiment, the detection unit 11b performs the calculation of the on-hall weight using the learned model for the human weight. In the first embodiment, no distinction is made between the weight of people and the weight of baggage. In the present embodiment, the on-hall weight can be calculated more accurately using a learned model for the baggage weight in addition to the learned model for the person weight. Below, a difference from the first embodiment will be explained mainly by using Fig. 10 described.

[0093] First, a configuration of the present embodiment will be described. A learning unit 12 of the present embodiment stores the learned model for human weight and the learned model for luggage weight.

[0094] In the first embodiment, the human weight learning model is generated according to the learning data created based on the combination of the image data, i.e., the image information, and the weight (the weight including the luggage or the like held by the person in the image) of the person in the image. In the present embodiment, the information used to generate a learned human weight model and derive an on-height weight using this model is a set of features of the human body in an image extracted from the image data by a known algorithm.The learned model for human weight in the present embodiment is a learned model created according to learning data based on a combination of the feature amount of the human physique in the image, which is the image data, and the weight (weight of the person himself without luggage or the like, which can be inferred from the physique) of the person in the image.

[0095] A learned model for the weight of luggage is a learned model generated according to data based on a combination of image information and a weight of the luggage held by the person in the image. In the present embodiment, the image information used to generate the learned model for the luggage weight and to derive the on-hand weight using this model is information about the holding manner of the luggage of the person in the image, which is extracted from the image data by a known algorithm. Since people change the way they hold their luggage depending on the weight of the luggage, the weight of the luggage can be derived from the way it is held. For example, if a person holds the luggage with both hands and bends down, it is obvious that the person is carrying very heavy luggage.Furthermore, a person holding the luggage in one hand can be recognized as carrying light luggage. Note that in the present disclosure, the on-hall weight includes the weight of the luggage of the passenger or passengers in the hall.

[0096] In the present embodiment, the acquisition unit 11b includes a software module that acquires a feature set of the human body in the image and the information about the holding manner of the person in the image from the image data acquired by the image pickup device 33. Further, the acquisition unit 11b includes a software module that obtains human weights by inputting the acquired sets of human body features into the learned human weight model and outputs the weights of the passengers in the hall one by one. The acquisition unit 11b also includes a software module that obtains the baggage weight by inputting the acquired information about the holding manner of people into the learned baggage weight model and outputs the weights of the passengers' baggage in the hall one by one.The registration unit 11b also includes a software module that calculates the total weight of the person and their luggage for each passenger. Furthermore, the registration unit 11b includes a software model that calculates the total weight of all passengers in the hall.

[0097] Next, an operation of the present embodiment will be described. In the present embodiment, the process of step S1 in Fig. 3 to the Fig. 10 shown step S81.

[0098] In step S81, the acquisition unit 11b obtains the information about the image. Specifically, the acquisition unit 11b receives the image obtained by photographing the hall where the hall call is made and output from the device 33 via the interface 13, and temporarily stores the image data in the storage unit 12. Subsequently, the acquisition unit 11b extracts the quantities of the physique characteristics of the passenger(s) in the hall and the information about the manner of holding the baggage in the hall from the image data and stores them in the storage unit 12 in association with each of the passengers. Subsequently, the acquisition unit 11b further processes the process until step S82.

[0099] In step S82, the acquisition unit 11b inputs the feature quantities of the body structure of the passengers, which are the image information acquired by the acquisition unit 11b in step S81, into the learned human weight model stored in the storage unit 12 and obtains the weights of the passengers themselves one by one. Then, the acquisition unit 11b processes the process until step S83.

[0100] In step S83, the detection unit 11b stores all the weights of the passengers themselves obtained in step S82. Then, the detection unit 11b processes the process until step S84.

[0101] In step S84, the detection unit 11b inputs the information about the way passengers hold their luggage, which is the image information acquired by the detection unit 11b in step S81, into the learned luggage weight model stored in the storage unit 12 and obtains the weights of the luggage one by one. The detection unit 11b then processes the process until step S85.

[0102] In step S85, the detection unit 11b adds the weight of the passenger himself obtained in step S82 and the weight of the baggage obtained in step S84 for each of the passengers, calculates the on-hall weight of each of the passengers, and stores all these weights in the storage unit 12. Then, the detection unit 11b processes the process until step S86.

[0103] In step S86, the detection unit 11b adds the on-hall weights of the respective passengers stored in step S85, calculates the total hall weight, and stores the total hall weight in the storage unit 12. Subsequently, the detection unit 11b processes the process until step S31.

[0104] According to the above, the control device 10 can calculate the on-hall weight more accurately.

[0105] The present embodiments have been described so far, but the present invention is not limited to the embodiments. Modified examples are shown below.

[0106] In the embodiments, the detection unit 11b obtains the in-hall weight through the machine learning method using the learned model. However, for solving the problem, the method for obtaining the in-hall weight is not limited to this. For example, the weights of the passenger(s) may be stored in advance in the terminals carried by the passenger(s) in the hall, and the weights of the passenger(s) in the hall may be obtained through wireless communication.

[0107] In the embodiments, the detection unit 11b outputs the weight and attribute of each passenger using the learned model, but the weights or attributes of a plurality of people may also be output. For example, a learned model created based on the data created by combining the image data in which a plurality of people are photographed or the feature amounts of the physiques and the sum of the human weights of the plurality of people in the image may be stored in the storage unit 12 as the learned human weight model. Then, the total weight of the hall can be calculated directly from the image data of the hall photographed by the image pickup device 33.

[0108] Furthermore, a part of the software modules included in the acquisition unit 11b and the learned models stored in the storage unit 12 may be placed on a cloud server. For example, the acquisition unit 11b may obtain the image information in step S21, then output the image information to the cloud server, and obtain the human weight from the cloud server after the cloud server performs the process of step S22. In other words, a device for deriving the on-hall weight may be provided separately from the elevator control device. Furthermore, in the embodiments, the learned model is stored in advance in the storage unit 12, but the learning data may be collected from the beginning by controlling the update of the learned model for the human weight performed using Fig. 7 is used to generate the learned model.

[0109] Note that in the embodiments, the case where supervised learning is applied to the learning algorithm used by the model generation unit 11d is described, but the learning algorithm is not limited to this. For the learning algorithm, in addition to supervised learning, forced learning, unsupervised learning, semi-supervised learning, or the like may be applied. Furthermore, as the learning algorithm used in the model generation unit 11d, deep learning that learns the extraction of the feature set itself may also be used, and machine learning may be performed by other known methods such as genetic programming, functional logic programming, and a support vector machine.

[0110] Furthermore, the model generation unit 11d may perform human weight learning according to learning data generated for a plurality of elevator devices 100. Note that the model generation unit 11d may obtain learning data from a plurality of elevator devices 100 serving in the same area, or may learn the human weight using learning data collected from a plurality of elevator devices 100 independently serving in different areas. Furthermore, it is also possible to add or remove the elevator device 100 from which learning data is being collected on the way to or from the destination. Furthermore, a learning device that has learned the human weight for a specific elevator device 100 may be applied to the various elevator devices 100, and the human weight may be relearned and updated for the various elevator devices 100.

[0111] In the embodiments, the judging unit 11c judges whether the passenger(s) in the hall can board the car based on the on-car weight and the on-hall weight. Additionally, a camera may be mounted to photograph the interior of the car, and a free space and a floor-projected area of ​​the passenger(s) in the hall may be compared to judge whether the passenger(s) in the hall can board the car.

[0112] In the embodiment, the classification of a person, such as a stretcher, a pregnant woman, an elderly person, or a person carrying large luggage, is represented as the attributive information used by the judging unit 11c for the second judgment. Of course, the classification of a person constituting the attributive information is not limited to this. For example, a tired person or the like may be used.

[0113] Furthermore, the means for obtaining the attribute information may be a combination of a variety of means. For example, the attribute information of the fatigue level may be treated independently of the other attribute information. A feature set of an expression may be extracted from the image of a passenger's face in the hall using a known algorithm. The fatigue level may be estimated using a learned model that estimates the fatigue level from the feature set of the expression. This allows the attribute information that the passenger is tired to be obtained.

[0114] In the embodiment, the control unit 11a causes the display device 32 to display the priority messages of all passengers classified as high, but may limit the passengers whose priority messages are displayed by the display device 32 according to the priority level and the number of passengers who can board the cabin. For example, the attributive information of people carrying large luggage and the attributive information of tired people may be treated separately, the priority order may be determined based on the other attributive information, and then the priority order may be determined in order of luggage weight and fatigue level.

[0115] In the embodiments, the elevator device 100 includes only one car 20, but the elevator device 100 may include a plurality of cars 20, and the control device 10 may include a group control function. For example, if there is a passenger who cannot board a particular car 20 among the passengers in the hall, the control unit 11a may cause the car 20 to bypass the hall and cause another car 20 to stop in the hall.

[0116] In the present embodiment, the hall call device 31 includes the up-and-down button and the down-and-down button, but it may also be the hall call device 31 that allows the passenger to make a hall destination call indicating a destination floor in the hall. In this case, the control device 10 can detect in advance which hall the passenger will disembark from in the car 20.Therefore, in the embodiment, the control unit 11a decides whether or not the car 20 in the hall where the hall call is made causes the hall to be bypassed, or issues the priority notification, but in addition, the judging unit 11c may judge whether or not the passenger or passengers in the hall can board the car based on an on-car weight after the passenger or passengers in the car 20 get off at the destination floor and an on-hall weight in the hall that is the destination floor of the hall call. [List of reference symbols]

[0117] 10 Control device, 11 Processor, 11a Control unit, 11b Acquisition unit, 11c Judgment unit, 11d Model generation unit, 12 Storage unit, 13 Interface, 20 Car, 21 Car call device, 22 Load weighing device, 31 Hall call device, 32 Display device, 33 Image pickup device, 41 Rope, 42 Drive device, 50 Priority database, 51 Attribute information, 52 Priority information, 100 Elevator device QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] JP 2002-302348 A

[0003]

Claims

[1] Elevator control device comprising: a detection unit that detects an on-cabin weight, which is a weight of a passenger or passengers inside a cabin, and an on-hall weight, which is a weight of a passenger or passengers in a hall; a judging unit that makes a first judgment as to whether the passenger or passengers in the hall can board the cabin or not, based on the on-cabin weight and the on-hall weight received from the detection unit; and a control unit that controls an elevator device according to the judgment of the judgment unit. [2] The elevator control device according to claim 1, wherein the detection unit obtains image information of the passenger or passengers in the hall photographed by an image pickup device, outputs the on-hall weight from the captured image information of the passenger or passengers in the hall by using a learned model for deriving the on-hall weight from the image information, and thereby obtains the on-hall weight. [3] Elevator control device according to claim 2, wherein the image information includes information about a manner of holding luggage of the passenger or passengers in the hall, and the learned model is used to derive a weight of the baggage from the information about the manner of holding the baggage of the passenger or passengers in the hall, which is included in the image information. [4] Elevator control device according to one of claims 1 to 3, wherein the detection unit receives the on-cabin weight obtained from a load weighing device of the cabin, the elevator control device further comprises a model generation unit that uses learning data including image information of the passenger or passengers in the hall and a difference between the car weights detected by the load weighing device before and after the passenger or passengers in the hall board the car, and that generates a learned model for deriving the car weight from the image information. [5] The elevator control device according to any one of claims 1 to 4, wherein, when the judging unit judges that at least one or more of the passengers in the car cannot get into the car based on the first judgment, the control unit causes the elevator device to notify the passenger or passengers in the hall whether or not the passenger or passengers in the hall can get into the car. [6] Elevator control device according to claim 5, wherein the detection unit receives attribute information of the passenger or passengers in the hall, the judging unit performs a second judgment to judge whether there is a passenger in the hall who has a high priority to enter the cabin based on the attribute information received from the detecting unit, and when the judging unit judges that at least one or more of the passengers in the hall cannot get into the car based on the first judgment, and judges that there is a passenger in the hall with the high priority to get into the car based on the second judgment, the control unit causes the elevator device to notify that the priority to get into the car is given to the passenger with the high priority to get into the car. [7] The elevator control device according to any one of claims 1 to 4, wherein, when the judging unit judges that the passenger or passengers in the hall can get into the car based on the first judgment, the control unit causes the car to stop in the hall, and when the judging unit judges that the passenger or passengers in the hall cannot get into the car, the control unit causes the car to bypass the hall. [8] Elevator control device according to claim 7, wherein the detection unit receives attribute information of the passenger or passengers in the hall, the assessment unit makes a second assessment as to whether there is a high-priority passenger in the hall who is to board the cabin, based on the attribute information received from the detection unit, and if the judgment unit judges that there is a high-priority passenger in the hall to board the car based on the second judgment, the control unit causes the car to stop in the hall regardless of the first judgment.

Images