JP2021066575A - Elevator system - Google Patents

Abstract

To provide an elevator system that facilitates the measurement of the presence or absence of passengers and the number of passengers waiting at an elevator platform, and to improve the efficiency of elevator operation control.SOLUTION: The elevator system includes a sensor 10 that acquires information on passengers getting on and off the elevator. The sensor 10 is installed on the front side of an upper frame of a three-way frame 20 that is installed in the elevator platform 40, while facing the elevator platform 40. The elevator system further includes recognition means 300 that recognizes at least a part of the passenger's head, face, and pupil from data acquired by the sensor 10, and passenger identification means 310 that generates passenger identification information from the recognition result of the recognition means 300. Further, the passenger identification means 310 identifies at least one of the presence/absence of passengers and the number of passengers to generate the passenger identification information.SELECTED DRAWING: Figure 1A

Description

The present invention relates to an elevator system with a sensor.

In the current elevator operation, passengers press the up and down buttons installed at the landing to generate a car call on each floor, and the operation control device grasps the floor where the car call occurred and the position of the car of each elevator. Based on the condition of each elevator, such as the number of passengers in the car, which elevator is to be dispatched to which floor is determined. When a passenger gets on the arriving car and the passenger presses the desired floor number button, the destination floor is determined for the first time, and the operation control device determines the operation of the elevator based on a plurality of stop floors.

The elevator system cannot keep track of how many passengers are waiting for the car to arrive at the landing. Therefore, as mentioned above, there is a possibility that some passengers will not be able to board even if the car arrives at the landing. In addition, as described above, since the destination floor cannot be known unless the passengers board the car and press the floor number button, there is a high possibility that the optimum vehicle allocation has not been achieved.

There is a technology to take a picture of the landing by installing a camera in order to grasp the situation of the landing. As such a technique, there are those described in Patent Documents 1 and 2, for example.

Japanese Unexamined Patent Publication No. 2016-169096 Japanese Unexamined Patent Publication No. 2015-04467

In the technique described in Patent Document 1, as shown in FIG. 8, the camera is attached in a state of protruding from the three-sided frame in order to confirm the safety of getting on and off the wheelchair user, and the landing is photographed. In Patent Document 1, it is necessary to change the shooting direction up and down and left and right for each case so that a desired image can be shot, which complicates the adjustment work and captures a wide range of passengers from around the door for getting on and off. It was something I couldn't do. As described above, the technique described in Patent Document 1 cannot measure the number of passengers waiting at the landing and cannot identify each passenger in the entire landing.

In the technique described in Patent Document 2, for example, as disclosed in FIGS. 17 to 19, a camera is provided on the upper part of the three-sided frame, and the car floor surface and the in-situ floor are provided in an increasing direction from directly above. I try to shoot the gaps on the surface. Also in Patent Document 2, it is not possible to image a wider range of passengers than around the door for getting on and off, and it is not possible to measure the number of people waiting at the landing.

An object of the present invention is to provide an elevator system that facilitates measurement of the presence / absence and the number of passengers waiting at an elevator landing and improves the efficiency of elevator operation control.

In order to achieve the above object, the present invention is in an elevator system including a sensor for acquiring information of passengers getting on and off the elevator, the sensor is a front surface facing the elevator landing in an upper frame of a three-sided frame provided in the elevator landing. It is characterized by being prepared for.

Further, in addition to the above, the present invention includes a recognition means that recognizes at least a part of a passenger's head, face, and pupil from the data acquired by the sensor, and a passenger that generates passenger identification information from the recognition result of the recognition means. It is characterized by having an identification means.

According to the present invention, it is possible to provide an elevator system that facilitates measuring the presence / absence and the number of passengers waiting at an elevator landing and improves the efficiency of elevator operation control.

It is a front view of the elevator which concerns on Example 1. FIG. FIG. 1A is a sectional view taken along line I-I in FIG. 1A. FIG. 1A is a sectional view taken along line I-I in FIG. 1A. It is a front view of the elevator which concerns on Example 2. FIG. FIG. 2 is a sectional view taken along line II-II in FIG. 2A. It is a front view of the elevator which concerns on Example 3. FIG. It is a front view of the elevator which concerns on Example 4. FIG. It is a front view of the elevator which concerns on Example 5. It is a front view of the elevator which concerns on Example 6. FIG. 6 is a sectional view taken along line VI-VI in FIG. 6A. FIG. 6 is a sectional view taken along line VI-VI in FIG. 6A. It is an overall block diagram of the elevator system which concerns on Example 7. FIG. It is an overall block diagram of the elevator system which concerns on the modification of FIG. 7. It is a block diagram which shows the structure of a sensor controller. It is a figure which shows the processing flow of FIG. It is a process flow diagram of the number of people detection considering the overlap of people. It is a figure which shows the overlapping state of a person. It is a figure which shows the overlapping state of a person. It is a figure which shows the overlapping state of a person. It is a flow chart which shows the generation and storage processing of the passenger boarding / alighting information when a person cannot be identified at the time of getting off. It is a figure which shows boarding / alighting information. It is a flow chart which executes the vehicle allocation control based on a person's identification information. It is a figure which shows the arrangement example of a three-sided frame and a sensor in a building which operates a plurality of elevators. It is a flow diagram which eliminates the duplication of recognition processing by a plurality of sensors. It is a flow chart which carries out operation control based on the person existing in the area in charge. It is a flow chart which carries out operation control based on the person density of the area in charge.

Hereinafter, examples of the present invention will be described with reference to the accompanying drawings. Similar components are designated by the same reference numerals, and the same description will not be repeated.

The various components of the present invention do not necessarily have to be independent of each other, and one component is composed of a plurality of members, a plurality of components are composed of one member, and a certain component is different. It is allowed that a part of one component overlaps with a part of another component.

1A is a front view of the elevator according to the first embodiment, and FIGS. 1B and 1C are cross-sectional views taken along the line II in FIG. 1A.

A three-sided frame 20 is provided at the entrance / exit of the elevator platform 40. The three-sided frame is attached to the upper part of the entrance and to the left and right. Since there is no frame at the bottom of the three-sided frame, the step due to the frame is eliminated and the ease of getting on and off the elevator car is improved. Since there are no steps, it is especially effective for wheelchairs and trolleys. Further, a door 25 for opening and closing the entrance / exit is attached to the three-sided frame 20.

The sensor 10 is provided in the upper riding scene 30 (the portion surrounded by the broken line) of the three-sided frame 20. The upper riding scene 30 is the front surface facing the elevator landing 40 in the upper frame of the three-sided frame 20. In the first embodiment, the sensor 10 is provided at the center position in the left-right direction of the upper frame in the upper riding scene 30 of the three-sided frame 20. When the sensor 10 is an image sensor 12 combined with a fisheye lens 11, as shown in FIG. 1B, when a wide-angle lens or a fisheye lens 11 having a wide angle of view (broken lines shown in the vertical direction in the figure) is selected for the fisheye lens 11, By incorporating the optical axis of the image sensor 12 perpendicular to the riding scene of the three-sided frame 20, it is possible to measure an arbitrary area or all of the space of the elevator landing 40.

Further, when it is desired to measure only the space from the installation height of the sensor 10 to the space between the floors of the elevator landing 40, it is preferable to use the lower side of the fisheye lens 11 from the horizontal plane. In this case, the image sensor 12 is combined only with the lower half of the fisheye lens 11, or the lower half of the measurement data captured by the image sensor 12 is used.

Further, as shown in FIG. 1C, even if a lens having a narrower angle of view than the above-mentioned lens is selected for the fisheye lens 11, the upper riding scene of the three-sided frame 20 is set so that the optical axis in the vertical direction is downward from the horizontal. By incorporating it in 30, it is possible to measure only the space between the installation height of the sensor 10 and the elevator landing 40. In the first embodiment, a sensor having a narrow vertical measurement range such as LiDAR (Light Detection and Ringing) can be similarly incorporated to measure a wide range of the elevator landing 40.

In this embodiment, since the measurement range can be secured for the landing space of the sensor 10, at least the upper body, especially the head, face, and pupils above the shoulders are affected by the overlap of passengers waiting at the landing. It becomes possible to measure. For example, an individual can be distinguished by the sensor 10 identifying the whorl shape, hairstyle, hat, etc. of the head with high accuracy. In addition, the sensor 10 can identify an individual with high accuracy. Further, the sensor 10 can distinguish an individual by discriminating the glow of the pupil with high accuracy.

By measuring the presence of the head, face, and pupil by the sensor 10, the number of people can be measured because it is equivalent to the presence of a person. If there is no overlap, the entire body can be measured, so it is possible to derive the occupied area. If the occupied area of each passenger can be derived, the filling rate in the car can be estimated in advance, so that the space margin in the car can be adjusted and the comfort level of the passengers can be improved. Further, by determining the direction in which the sensor 10 is incorporated in the upper riding scene 30 according to the landing space dimension, the measurement area of the sensor 10, and the dimension of the three-sided frame 20, the sensor 10 can be fixed at the time of manufacturing the three-sided frame 20. It is not necessary to adjust the top, bottom, left and right of the sensor 10 at the time of construction for each project in which the product is laid, and the workability is improved.

In some cases, the three-sided frame 20 is integrated with the wall surface of the building. In that case, the sensor 10 may be incorporated in the wall surface. In this embodiment, the specific structure of the door is omitted, but this embodiment can be applied to any structure such as a central opening and a single opening.

Next, Example 2 of the present invention will be described with reference to FIGS. 2A and 2B. FIG. 2A is a front view of the elevator according to the second embodiment, and FIG. 2B is a sectional view taken along line II-II in FIG. 2A. The same configurations as those in the first embodiment are designated by the same reference numerals, and detailed description thereof will be omitted.

In the second embodiment, the difference from the first embodiment is the installation position of the sensor 10. Depending on the shape of the landing, the elevator landing 40 space may be open to the elevator on either the left or right side. For example, in FIG. 2A, the wall 45 is located on the right side when the elevator landing 40 is viewed from the front, and the left side is open. In this case, the sensor 10 is shifted from the center position in the left-right direction of the upper riding scene 30 in the upper frame of the three-sided frame 20 and moved toward the wall 45 so that the sensor 10 can easily detect the periphery on the opposite side of the wall 45. Try to incorporate it.

Further, in FIG. 1C, the optical axis in the vertical direction is set to face downward from the horizontal as shown in the side sectional view, but in this embodiment, as shown in the upper sectional view of FIG. 2B, it is set. The optical axis in the left-right direction is set to be separated from the wall 45.

According to this embodiment, since the sensor 10 is installed close to the wall side of the three-sided frame 20, it is possible to avoid the sensor 10 from detecting the space where passengers do not exist, and the measurement range of the sensor 10 Can be used effectively.

Next, Example 3 of the present invention will be described with reference to FIG. FIG. 3 is a front view of the elevator according to the third embodiment. The same configurations as those in the first embodiment are designated by the same reference numerals, and detailed description thereof will be omitted.

In the third embodiment, the difference from the first embodiment is that the sensor 10 is arranged on the indicator 50.

In FIG. 3, an indicator 50 is provided on the upper part of the three-sided frame 20 and the upper part on the right side. The indicator 50 includes a first indicator 50 (a) indicating the moving direction of the car 110 (see FIGS. 7, 8 and 14) and the position of the car 110, and a second indicator 50 indicating the arrival of the car 110. There are at least two types of indicators 50 (b). The indicator 50 (a) is often located above the three-sided frame 20, and even in this case, the indicator 50 (a) is in a position where the entire landing can be overlooked, so that it is suitable for incorporating the sensor 10 for measuring the passengers in the landing.

In the indicator 50 (a) of the present embodiment, the sensor 10 is incorporated between the vertical lamp indicating the moving direction of the car 110 and the floor number lamp indicating the position of the car 110. It is not limited to this position. The sensor 10 may be incorporated in any position as long as it is the indicator 50 (a).

The indicator 50 (b) is often located above or diagonally above the three-sided frame 20, and even in this case, since it is a position where the entire elevator landing 40 can be overlooked, it is suitable for incorporating a sensor 10 for measuring passengers at the landing. ing. In the indicator 50 (b) of the present embodiment, the sensor 10 is incorporated on the upper side of the lamp, but the present invention is not limited to this position.

According to this embodiment, by equipping the indicator 50 (a) or the indicator 50 (b) with the sensor 10, the passengers can be detected from a position higher than the height of the passengers, and the passengers waiting at the landing can be overlapped. It becomes possible to measure the head, face, and pupil accordingly.

Next, Example 4 of the present invention will be described with reference to FIG. FIG. 4 is a front view of the elevator according to the fourth embodiment. The same configurations as those in the first embodiment are designated by the same reference numerals, and detailed description thereof will be omitted.

In the fourth embodiment, the difference from the first embodiment is that the stereo camera 60 is incorporated into the upper riding scene 30 of the three-sided frame 20 as one of the sensors 10.

The stereo camera 60 is a kind of sensor that secures parallax by installing two image sensors 12a and 12b apart from each other and can measure a distance based on the parallax. As a measurement result, a distance image can also be acquired. In this embodiment, two image sensors 12a and 12b are embedded in the upper riding scene 30 of the three-sided frame 20. The distance between the image sensors 12a and 12b is a distance at which the parallax required to obtain the desired distance accuracy can be obtained. As shown in FIG. 1C, the image sensors 12a and 12b may be tilted horizontally downward.

In this way, by incorporating the stereo camera 60 capable of measuring the distance image into the upper riding scene 30 of the three-sided frame 20, passengers can be detected from a position higher than the height of the passengers, and the passengers waiting at the landing can be detected according to the degree of overlap of the passengers. It is possible to measure the distance to the head, face, and pupil.

Next, Example 5 of the present invention will be described with reference to FIG. FIG. 5 is a front view of the elevator according to the fifth embodiment. The same configurations as those in the first embodiment are designated by the same reference numerals, and detailed description thereof will be omitted.

In the fifth embodiment, the difference from the first embodiment is that the ToF sensor 70 is incorporated into the upper riding scene 30 of the three-sided frame 20 as one of the sensors 10.

The ToF (Time of Flight) sensor 70 is a sensor 10 capable of measuring the distance to an object by radiating light having a specific wavelength by modulating it and receiving and processing the light reflected by the object. It is a kind of. A separate light is required to irradiate light of a specific wavelength. A light 72 that emits a specific wavelength and a ToF sensor 70 are incorporated in the upper riding scene 30 of the three-sided frame 20. In this embodiment, the ToF sensor 70 is incorporated in the center of the upper riding scene 30, but the present invention is not limited to this position.

According to this embodiment, by incorporating the ToF sensor 70 capable of measuring the distance to an object into the upper riding scene 30 of the three-sided frame 20, a point cloud, which is a collection of distance data to passengers waiting at the landing, is obtained. be able to.

Next, Example 6 of the present invention will be described with reference to FIG. 6A is a front view of the elevator according to the sixth embodiment, and FIGS. 6B and 6C are cross-sectional views taken along the line VI-VI in FIG. 6A. The same configurations as those in the first embodiment are designated by the same reference numerals, and detailed description thereof will be omitted.

In the sixth embodiment, the difference from the first embodiment is that the millimeter wave sensor 80 is incorporated into the upper riding scene 30 of the three-sided frame 20 as one of the sensors 10.

In FIG. 6, the millimeter wave sensor 80 measures a distance to an object by modulating and transmitting a radio wave having a specific frequency and receiving and processing a reflected wave generated by the transmitted wave colliding with an object. It is a kind of sensor 10 that can be used.

An antenna 82 that realizes the functions of transmitting and receiving radio waves is embedded in the upper riding scene 30 of the three-sided frame 20. In this embodiment, the transmitting antenna and the receiving antenna are mixedly mounted on the planar antenna. Since the antenna pattern looks like a special pattern, a blindfold plate 84 or the like through which millimeter waves pass may be provided in front of the antenna 82. When the directivity of transmission and reception of the antenna 82 is wide, as shown in FIG. 6B, the blind plate 84 and the antenna 82 can be coaxial with each other in the horizontal direction, so that the upper riding scene 30 of the three-sided frame 20 can be easily set. It can be incorporated. When the directivity of transmission and reception of the antenna 32 is narrow, the opening of the blindfold plate 84 is the same, and the position of the millimeter wave sensor 80 is installed above the blindfold plate 84 and downward in the horizontal plane, so that the elevator landing 40 The measurement range can be all the spaces of.

According to this embodiment, the distance to the object can be measured, and by incorporating the millimeter wave sensor 80 into the upper riding scene 30 of the three-sided frame 20, a point cloud, which is a collection of distance data to the passengers waiting at the landing, is created. Obtainable.

Next, Example 7 of the present invention will be described with reference to FIG. 7. FIG. 7 is an overall configuration diagram of the elevator system according to the seventh embodiment. The same configurations as those of Examples 1 to 6 are designated by the same reference numerals, and detailed description thereof will be omitted. In the seventh embodiment, the configuration for realizing the first to sixth embodiments will be described.

In FIG. 7, the elevator 100 includes a car 110, a counterweight 150, a rope 160, a main engine 120 that moves the car 110 up and down, an elevator controller 130 that controls the main engine 120, and a destination on the car 110. It is composed of a car controller 140 that processes floor buttons and controls the car door. The car 110 and the counterweight 150 are connected by a rope 160.

The group management controller 200 communicates with the plurality of elevator controllers 130 via the communication path 180 in order to control vehicle allocation and the like so as to efficiently operate the plurality of elevators 100. The sensor 10 of the upper riding scene 30 of the three-sided frame 20 embedded in the three-sided frame 20 on each floor of the plurality of elevators 100 is connected to the sensor controller 170. The sensor controller 170 is connected to the group management controller 200 via a communication path 190.

With such a configuration, the elevator system can measure and identify the number of passengers in the elevator landing 40. The group management controller 200 can dispatch elevators according to the number of people measured on each floor.

In this embodiment, the sensor controller 170 is provided in each elevator 100, but the present invention is not limited to this. If the processing capacity of the sensor controller 170 is high, all the sensors 10 may be connected to one sensor controller 170 for processing. Further, when the processing capacity of the group management controller 200 is high, all the sensors 10 may be connected to the group management controller 200 for processing.
[Modification example]
Next, a modified example will be described with reference to FIG. FIG. 8 is an overall configuration diagram of the elevator system according to the modified example of FIG. 7. The same configurations as those in FIG. 7 are designated by the same reference numerals, and detailed description thereof will be omitted. In the modified example, the processing of the sensor controller 170 and the group management controller 200 in FIG. 7 is configured to be executed in the cloud 240. The cloud 240 can communicate via the communication network 230. There is a communication relay facility 220 such as a router for connecting to the communication network 230. When the cloud 240 is a global cloud, the communication network 230 is a wide area communication network. When the cloud 240 is on-premises, the communication network 230 becomes a LAN.

The car controller 140 connects to the communication relay facility 220 via the communication path 180, and can communicate with the cloud 240. The sensor 10 incorporated in the upper riding scene 30 of the three-sided frame 20 installed for each elevator 100 at the elevator landing 40 on each floor is connected to the communication relay facility 220 via the communication path 190 to form a cloud. Communication with 240 becomes possible.

According to the modified example, by configuring as described above, the mechanism and equipment of the elevator 100 that must physically exist in the hoistway where the elevator 100 is installed and the new sensor 10 are left on the building side. Since group management processing and sensor processing that do not need to be installed near the elevator 100 can be centralized in the cloud 240, only the minimum mechanism and equipment of the elevator 100 need to be laid, and construction and maintenance work can be performed. It can be simplified.
[Sensor controller configuration]
Next, the configuration of the sensor controller will be described with reference to FIG. FIG. 9 is a block diagram showing the configuration of the sensor controller 170. The data measured by the sensor 10 is transmitted to the sensor controller 170. The sensor controller 170 includes a recognition means 300, a passenger identification means 310, a storage means 320, a learning means 330, a call guessing means 340, and a coordinate conversion means 350.

The data acquired by the sensor 10 is transmitted to the recognition means 300, and the recognition means 300 recognizes whether or not at least a part of a person, a passenger's head, a face, and a pupil exists, and if the recognition means can be recognized, the sensor 10 is used. The reference position information is also generated as a recognition result.

The recognition result of the recognition means 300 is transmitted to the passenger identification means 310, and the passenger identification means 310 identifies the passenger from the recognition result. The passenger identification means 310 identifies at least one of the presence or absence of passengers and the number of passengers, generates passenger identification information, and outputs the information to the storage means 320 and the call guessing means 340.

Depending on the measurement accuracy of the sensor 10, the passenger identification means 310 may only determine the presence or absence of passengers, or may count the number of passengers. Even if it can be counted as the number of passengers, there are cases where the number of passengers can be obtained as an absolute value, and there are cases where it is output as a range value such as N or more.

In addition, the passenger identification means 310 generates guess information such as age and gender, or identification information that cannot identify an individual but can be distinguished for each passenger, and also generates passenger information corresponding to the identification information.

The storage means 320 stores at least the identification information, the time information, and the boarding / alighting floor number. The learning means 330 has means such as machine learning and deep learning, and learns the identification information, the time information, and the boarding / alighting floor number stored in the storage means 320, and as a result of the learning. Generate a trained network.

The learning means 330 stores the learned network in the storage means 320.

The call guessing means 340 guesses the car call from the identification information generated by the passenger identification means 310 and the network stored in the storage means 320.

The coordinate conversion means 350 uses the position information of the recognition target based on the sensor 10 output from the recognition means 300 in a three-dimensional space whose height direction is the built-in position of the sensor 10 with the plane of the elevator landing 40 as a reference plane. Convert to the spatial coordinate system again.

In this embodiment, the sensor controller 170 is provided with the recognition means 300, the passenger identification means 310, the storage means 320, the learning means 330, the call guessing means 340, and the coordinate conversion means 350, but these means are provided in the group management controller 200. Alternatively, it may be prepared for the cloud 240.

In this embodiment, the passenger identification means 310 is provided to identify at least one of the presence or absence of passengers and the number of passengers and generate passenger identification information. By controlling the operation of the elevator based on the passenger identification information, the group management controller 200 can operate according to the presence or absence of passengers and the number of passengers, and can improve the efficiency of elevator operation.
[Processing flow by measuring the number of people]
FIG. 10 is a diagram showing a processing flow of FIG.

In FIG. 10, the sensor 10 measures the elevator landing 40 and generates measurement data (step S10).

The recognition means 300 processes the measurement data measured by the sensor 10 and extracts the data that can be recognized as a person (step S20). For example, when the sensor 10 is an image sensor, the recognition means 300 has deep learning, and outputs area information that can be recognized as a person and an accuracy value that is a person.

The passenger identification means 310 determines the number of passengers from the data extracted by the recognition means 300 from the accuracy value of a person or the like (step S30).

The call guessing means 340 dispatches an elevator 100 having a vacant car according to the number of people waiting on the floor of the elevator landing 40 from the number of passengers (identification information) identified by the passenger identification means 310. If there is not enough space in one elevator 100, operation control is executed so that a plurality of elevators 100 are dispatched (step S50). In the conventional elevator 100, passengers can know that a call has occurred at that time by pressing the up and down buttons on the landing floor, but it is not possible to know how many passengers there are, so an appropriate number of elevators can be used. It was not possible to dispatch 100 to the floor. In addition, it was not possible to know the passengers who stopped using the elevator 100 after pressing the button. In this way, when the use of the elevator 100 is stopped and there are no waiting passengers, the vehicle is immediately redistributed, and a predetermined door is opened for unnecessary stoppage or boarding of passengers who are not at the elevator landing 40. It is possible to avoid wasting maintenance time.
[Passenger identification means 310 for updating the number of passengers]
Next, the number update process of the passenger identification means 310 will be described with reference to FIGS. 11 and 12.

FIG. 11 is a processing flow diagram for detecting the number of people in consideration of the overlap of people, and FIGS. 12A to 12C are diagrams showing the overlapping state of people.

Depending on the accuracy of the sensor 10 and the processing method of the means of the recognition means 300, the recognition result cannot be constantly obtained. In addition, the degree of overlap between people changes due to the delicate movements of people, and the recognition result also changes. Passengers waiting for the arrival of the car 110 at the elevator platform 40 are not standing upright and waiting, but constantly turning around, shifting their weight from side to side, and over the shoulders of others to check the indicator 50. There is an opportunity to secure a sufficient measurement area for the sensor 10 because it looks into the sensor 10. As in the present invention, by installing the sensor 10 in the upper riding scene 30 of the three-sided frame 20 (the front surface of the upper frame of the three-sided frame 20 facing the elevator landing 40), it is possible to measure the crowns of a plurality of passengers. However, depending on the degree of overlap, it is possible to recognize the face part, the pupil, and the like.

At the elevator landing 40, since a plurality of passengers waiting for the arrival of the car 110 are waiting at an arbitrary place, overlap may occur from the viewpoint of the sensor 10. If one passenger gets behind another passenger, it is impossible to measure with the general sensor 10 as described above. Depending on the degree of overlap due to the movement of passengers, it often happens that it cannot be recognized or not.

The passenger identification means 310 acquires the recognition result from the recognition means 300 (step S100). The passenger identification means 310 determines whether or not the newly acquired recognition result is the same from the result of being recognized as a memorized person (step S110). If it is the same as the stored recognition result, the number of people is maintained as it is (step S130). If it is different from the stored recognition result, the result recognized as a new person is stored (step S120). Since new passengers have been added, the number of passengers is added and updated (step S130).

The overlap and the recognition result at the time of waiting for passengers at the landing will be described with reference to FIG. In FIG. 12A, person 1 and person 2 are recognized as humans with high accuracy because the upper body can be completely measured by human recognition. The person 3 who has a large overlap with the person 1 is likely not recognized as a person.

In FIG. 12B, when the person 3 is displaced, the pupil is exposed to the sensor, and the pupil is recognized with high accuracy, so that the person can be recognized as a new person different from the person 1 and the person 2.

In FIG. 12C, when the person 3 is further displaced, the face is exposed to the sensor, and the face is recognized with high accuracy, so that the person 3 is re-recognized as a person. In this way, by using recognition processes having different regions such as faces and eyes according to the degree of overlap, it is possible to determine the number of people even when people overlap.
[Passenger identification means 310 boarding / alighting processing flow]
Next, with reference to FIG. 13, a process when a person cannot be identified at the time of getting off will be described. FIG. 13 is a flow chart showing generation and storage processing of passenger boarding / alighting information when a person cannot be identified at the time of getting off.

Since the sensor 10 is installed in the upper riding scene 30 of the three-sided frame 20 or the indicator 50, the passenger faces backward with respect to the sensor 10 when getting off, so that the front of the face and the pupil can be within the measurement range. However, it is difficult to identify a person. Therefore, it is assumed that the same person has a high probability of boarding again from the disembarkation floor. Since the head is within the measurement range, as described above, if the person can be identified based on the whorl shape, hairstyle, hat, etc., the person can be identified even when getting off.

The passenger identification means 310 determines that the person has boarded the car 110 based on the person's identification information, the recognition position, and the like (step S200).

The passenger identification means 310 confirms whether the boarding information (boarding time, boarding floor) is stored in the storage means 320 from the floor corresponding to the person's identification information (step S210).

If it is not stored, the passenger identification means 310 registers the boarding information (boarding time, boarding floor) from the floor with the person's identification information in the storage means 320 (step S220). At this point, there is no disembarkation information (disembarkation floor) corresponding to boarding. When the person gets on the train from the disembarkation floor (step S200), it is confirmed whether the boarding information is stored from the floor corresponding to the person's identification information (step S210). If it is stored and the disembarkation floor information is not stored, the floor is stored as the disembarkation floor information (step S230). When the person can be identified when getting off by the head (step S240), the passenger identification means 310 confirms whether the boarding information is stored from the floor corresponding to the person's identification information (step S210). .. After that, it is the same as the above flow. By repeating the above flow, the passenger identification means 310 identifies the passenger boarding / alighting information 360 from the recognition result of the passenger's head and stores it in the storage means 320.

FIG. 14 is a diagram showing boarding / alighting information 360. The storage means 320 stores the passenger identifier, the boarding time associated with the identifier, the boarding floor on which the passenger boarded, and the disembarking floor on which the passenger disembarked, as boarding / alighting information 360.

For example, the passenger A to which the identifier A is assigned gets on the train from the first floor to the third floor where the workplace is located at 07:30, which is the commuting time. At 12:00 at lunch, Passenger A boarded from the 3rd floor to the 10th floor where the dining room is located. Passenger B to which the identifier B is assigned gets on the train from the 1st floor to the 4th floor where the workplace is located at 08:00, which is the commuting time. At 14:00 when the meeting is held, Passenger B boarded from the 4th floor to the 8th floor where the meeting room is located. Passenger D, who is given the identifier D, gets on the train from the 1st floor to the 4th floor where the workplace is located at 07:30, which is the commuting time. At 12:00 at lunch, Passenger D boarded from the 4th floor to the 10th floor where the dining room is located.

From the boarding / alighting information 360, it can be understood that passenger A, passenger D, and passenger E go to work at 07:30 and board the elevator from the first floor to head for the workplace. Similarly, from the boarding / alighting information 360, it can be understood that passengers B and E are heading from the 4th floor to the 8th floor at 14:00 in accordance with the holding of the conference.

In addition, there is a possibility that the boarding / alighting information 360 may remain, such as the boarding floor could be confirmed but the getting-off floor could not be confirmed. Such boarding / alighting information 360 is arranged / deleted by selecting, for example, a time zone in which passengers are less used, such as midnight. When the accumulated boarding / alighting information 360 is learned by a learning means 330 such as machine learning or deep learning, a learned network can be constructed.
[Vehicle allocation control flow based on boarding / alighting information 360]
FIG. 15 is a flow chart for executing vehicle allocation control based on human identification information.

In FIG. 15, the sensor 10 measures the elevator landing 40 and generates measurement data (step S10).

The recognition means 300 processes the measurement data detected by the sensor 10 and extracts the data that can be recognized as a person (step S20).

The passenger identification means 310 generates the above-mentioned identification information from the recognition data extracted by the recognition means 300 (step S40).

The call guessing means 340 estimates the passenger's disembarkation floor from the identification information (recognition result) and the boarding / alighting information 360 from the above-mentioned learned network (step S300). For example, when the time is 07:30, as shown in FIG. 14, passenger A, passenger D, and passenger E board from the first floor and get off at the third and fourth floors, respectively. The call guessing means 340 estimates the boarding floor from the identification information, the time, the boarding floor, and the boarding / alighting information 360.

Further, the call guessing means 340 registers the call in the storage means 320 with the estimated drop-off floor as the destination floor (step S310).

The group management controller 200 redistributes the vehicle based on the newly registered call (step S50).

In this way, by estimating the destination floor based on the passenger boarding / alighting information 360 in the past, it is possible to efficiently control the operation of the elevator 100.
[Measurement of the number of people in multiple elevators]
Next, a state in which a plurality of elevators are installed will be described. FIG. 16 is a diagram showing an arrangement example of the three-sided frame 20 and the sensor 10 in a building that operates a plurality of elevators. FIG. 16 is a bird's-eye view from the ceiling.

In FIG. 16, the six elevators 101, 102, 103, 104, 105, and 106 are in a general arrangement in which three elevators are arranged facing each other. Each of the plurality of elevators 101, 102, 103, 104, 105, 106 is provided with a sensor 10.

The coordinate system is the X-axis in the horizontal left-right direction, the Z-axis in the depth direction, and the Y-axis in the height direction from the passage 46 side toward the elevator landing 40. Since the back of the elevator landing 40 is a dead end due to the wall 45, as shown in FIG. 2, the axis of the sensor 10 is oriented in the opening direction of the elevator landing 40 space, and the installation position of the sensor 10 is above the three-sided frame 20. It is shifting toward the back of the riding scene 30. It can be seen that by combining the measurement ranges of the six (plural) sensors 10, almost the entire area of the elevator landing 40 space can be covered. By converting the position of the recognition object obtained by processing such as human recognition, head recognition, face recognition, and pupil recognition by the sensor 10 into a coordinate system by the coordinate conversion means 350, the position is mapped to the elevator landing 40 space. Can be done.

In FIG. 16, 401, 402, 403, 404, 405, and 406 are defined as regions corresponding to the elevators 101, 102, 103, 104, 105, and 106. This is the virtual area that each elevator is in charge of.

According to this embodiment, the axis of the sensor 10 is directed toward the opening of the elevator landing 40 space, and the installation position of the sensor 10 is shifted toward the back of the upper riding scene 30 of the three-sided frame 20, so that the elevator landing 40 space Can be detected comprehensively.

Next, the recognition process by the plurality of sensors 10 will be described with reference to FIG. FIG. 17 is a flow diagram for eliminating duplication of recognition processing by the plurality of sensors 10.

As shown in FIG. 16, when a plurality of elevators 100 are installed, a plurality of sensors 10 exist at an arbitrary landing, the measurement areas overlap, and the same person is recognized in duplicate. Need to be eliminated.

The sensor 10 measures the elevator landing 40 and generates measurement data (step S10).

The recognition means 300 processes the measurement data measured by the sensor 10 and extracts the data that can be recognized as a person (step S20).

The passenger identification means 310 generates the above-mentioned identification information from the recognition data extracted by the recognition means 300 (step S40).

The coordinate conversion means 350 converts the identification information into coordinates based on the installation position information of the sensor 10 and stores the identification information in the storage means 320 as the coordinate information of the elevator landing 40 (step S42). It is assumed that the installation position information of the sensor 10 is stored in the storage means 320 in advance.

The recognition means 300 determines whether or not the sensor 10 for which the measurement has not been completed remains (step S44).

If it remains, the process returns to step S10 to continue the flow. If not, the passenger identification means 310 eliminates duplication of the identification information stored based on the coordinate information (step S46). After deduplication, the passenger identification means 310 determines as a unique identification result with no duplication (step S48). The confirmed result is stored in the storage means 320. Even when a plurality of sensors 10 are present, it is possible to obtain a single identification result by eliminating duplication in this way.
[Individual number measurement method in a plurality of elevators 100]
Next, a method of measuring the number of individuals in the plurality of elevators 100 will be described with reference to FIG. FIG. 18 is a flow chart for performing operation control based on a person existing in the area in charge. Overlapping portions in regions 401, 402, 403, 404, 405, 406 are eliminated by the method shown in FIG.

In FIG. 18, the group management controller 200 selects an arbitrary elevator 100 from a plurality of elevators and determines the area in charge of the elevator 100 (step S100).

The group management controller 200 acquires the number of people in each region portion of the plurality of elevators via the sensor controller 170 (step S110).

The group management controller 200 confirms the vacancy status of the car 110 of the elevator 100 in each area portion of the acquired number of people, and confirms whether or not each of the above-mentioned number of people can be accommodated (step S120).

The group management controller 200 selects the elevator 100 as a vehicle allocation candidate when there is a space that can be accommodated in the elevator in each area portion (step S130).

If there is a vacancy but the above-mentioned all cannot be accommodated, the group management controller 200 selects another candidate with the number of persons that cannot be accommodated as the shortage (step S140).

When there is no vacancy, the group management controller 200 selects another candidate so that all the above-mentioned elevators 100 can be entrusted to another elevator 100 (step S145).

As described above, it is possible to determine a vehicle allocation candidate for operation control based on the number of passengers waiting in the area of the elevator landing 40 in charge of any elevator 100. In the vehicle allocation process, the vehicle allocation candidate generated separately from the vehicle allocation candidate is comprehensively determined to determine the final vehicle allocation. According to the flow of this embodiment, it is possible to efficiently execute the vehicle allocation process for the number of people waiting at the elevator landing 40.
[Other examples of individual number measurement methods in a plurality of elevators 100]
Next, another example of the method of measuring the number of individuals of the plurality of elevators 100 will be described with reference to FIG. FIG. 19 is a flow chart for executing operation control based on the density of people in the area in charge.

The group management controller 200 determines the area in charge of each elevator 100 (step S100).

The group management controller 200 acquires the number of people in each region portion of the plurality of elevators via the sensor controller 170 (step S110).

The group management controller 200 calculates the person density based on the area in each area portion and the acquired number of people (step S150).

Density calculation is repeated for all elevators 100 (step S160).

When the calculation of the density in each area in charge of all the elevators 100 is completed, the group management controller 200 rearranges the order of allocating the elevators 100 according to the magnitude of the density (step S170).

The group management controller 200 is determined as a vehicle allocation candidate in order (step S180).

In the vehicle allocation process, the vehicle allocation candidate generated separately from the vehicle allocation candidate is comprehensively determined to determine the final vehicle allocation. According to the flow of this embodiment, the passengers waiting for the car are not always uniformly dispersed. Focusing on the densities of the areas 401, 402, 403, 404, 405, and 406 in charge of each elevator 100, the density is high. By preferentially allocating the elevator 100 corresponding to the area, the convenience of passengers can be improved.

As described above, according to each embodiment, it is possible to provide an elevator system that facilitates measurement of the presence / absence and the number of passengers waiting at the elevator landing and improves the efficiency of elevator operation control.

The present invention is not limited to the above-described examples, and includes various modifications. The above-described examples have been described in detail in order to explain the present invention in an easy-to-understand manner, and are not necessarily limited to those having all the described configurations.

10 ... sensor, 11 ... fish-eye lens, 12 ... image sensor, 20 ... three-sided frame, 30 ... upper riding scene, 32 ... antenna, 40 ... elevator landing, 45 ... wall, 46 ... passage, 50 ... indicator, 60 ... stereo camera, 70 ... ToF sensor, 72 ... Light, 80 ... Millimeter wave sensor, 82 ... Antenna, 84 ... Blind plate, 100, 101, 102, 103, 104, 105, 106 ... Elevator, 110 ... Car, 120 ... Main engine, 130 Elevator controller, 140 ... basket controller, 150 ... counterweight, 160 ... rope, 170 ... sensor controller, 180, 190 ... communication path, 200 ... group management controller, 230 ... communication network, 240 ... cloud, 300 ... recognition means, 310 ... Passenger identification means, 320 ... Storage means, 330 ... Learning means, 340 ... Call guessing means, 350 ... Coordinate conversion means, 360 ... Boarding / alighting information

Claims (15)

In an elevator system equipped with a sensor that acquires information on passengers getting on and off the elevator
The elevator system is characterized in that the sensor is provided on the front surface of the upper frame of the three-sided frame provided in the elevator landing, which faces the elevator landing. In claim 1,
It is characterized by including a recognition means that recognizes at least a part of a passenger's head, face, and pupil from the data acquired by the sensor, and a passenger identification means that generates passenger identification information from the recognition result of the recognition means. Elevator system to do. In claim 2,
The passenger identification means is an elevator system characterized in that the presence or absence of passengers and at least one of the number of passengers are identified to generate passenger identification information. In claim 3,
An elevator system including a group management controller that controls the operation of the elevator based on the identification information. In claim 3,
An elevator system characterized in that it is provided with a call guessing means for allocating an elevator according to the number of people waiting on the floor of the elevator landing from the number of people identified by the passenger identification means. In claim 5,
The calling guessing means is an elevator system characterized in that a plurality of elevators are dispatched when one elevator is insufficient. In claim 2,
The passenger identification means is an elevator system characterized in that passenger boarding / alighting information is identified from a recognition result of a passenger's head and stored in a storage means. In claim 7,
The elevator system is characterized in that the boarding / alighting information includes a passenger identifier, a boarding time associated with the identifier, a boarding floor on which the passenger boarded, and a boarding floor on which the passenger disembarked. In claim 8.
An elevator system characterized by having a learning means for machine learning and deep learning of the boarding / alighting information. In claim 7,
An elevator system including a call guessing means for guessing a passenger's getting-off floor from the identification information and the boarding / alighting information. In claim 2,
A plurality of the elevators are provided, and the sensors are provided in each of the plurality of elevators.
A coordinate conversion means for converting the identification information by the installation position information of the sensor and storing the identification information as the coordinate information of the elevator landing in the storage means is provided.
The passenger identification means is an elevator system characterized by eliminating duplication of identification information stored based on the coordinate information of the coordinate conversion means. 11.
It is equipped with a group management controller that controls the operation of the plurality of elevators.
The group management controller acquires the number of people in each area portion of the plurality of elevators, and if there is a vacancy that can be accommodated in the elevator in each area portion, the elevator with the vacancy is selected as a vehicle allocation candidate. Elevator system featuring. 11.
It is equipped with a group management controller that controls the operation of the plurality of elevators.
The group management controller acquires the number of people in each area of the plurality of elevators, calculates the person density based on the area in each area and the acquired number of people, and responds to the calculated magnitude of the person density. An elevator system characterized in that the order of the elevators to be dispatched is rearranged. In claim 1,
The elevator system is characterized in that the sensor is provided at a center position in the left-right direction of the upper frame in the three-sided frame. In an elevator system including a sensor that acquires information on passengers getting on and off the elevator, a first indicator that indicates the position of the car of the elevator, and a second indicator that indicates the arrival of the car.
The first indicator and the second indicator are provided above the three-sided frame provided in the elevator landing.
The elevator system, wherein the sensor is provided on the first indicator or the second indicator.

Images