JP2026011564A - Group management device - Google Patents

Abstract

A group management device is provided that can predict the occurrence of passengers being left behind and appropriately allocate elevator cars.
[Solution] The group management device of an embodiment comprises a control unit that assigns one of a plurality of cars to respond to a hall call for each different floor, and an analysis unit that uses a learning model constructed by machine learning to calculate the probability of passengers being left behind at a target floor among the different floors where there is a hall call.When a first hall call is received, the analysis unit estimates the number of people waiting at the target floor, estimates the number of passengers in the first car assigned to respond to the first hall call, and calculates the probability from the number of people waiting and the number of passengers.If the probability is greater than or equal to a predetermined value, and when a second car among the plurality of cars is passing in the direction specified by the first hall call while the doors of the first car are open at the target floor, the control unit stops the second car that was scheduled to pass at the target floor.
[Selected Figure] Figure 2

Description

An embodiment of the present invention relates to a group management device.

In elevator systems, when responding to hall calls at crowded floors, there may be cases where not all passengers can board the car that responded, resulting in some passengers being left behind.

With the technology in Patent Document 1, when there is a possibility of long waiting times at a platform where a full passenger has passed, the car with the shortest response time that allows all passengers at the platform to board, or the car with the fewest passengers, is made to respond to that platform.

The technology in Patent Document 2, when there is a possibility that passengers will be left behind at a specific stop, evaluates the number of passengers waiting by multiplying the unresponsive time by the number of passengers waiting, and allocates elevators with short response times in proportion to the number of passengers waiting.

JP 2010-208708 A Japanese Patent Application Laid-Open No. 2023-081587

However, the technology in Patent Document 1 aims to shorten waiting times at platforms where passengers pass through due to a full bus. Therefore, it does not address the issue of passengers being left behind after a passenger has responded.

The technology in Patent Document 2 also acquires the number of people waiting at the platform and the number of people in the car, and controls the operation of the car based on these values. For example, an imaging device is installed at the platform, a gate is installed on the route to the platform, or equipment capable of communicating with users' terminals is installed, and the number of people waiting at the platform, etc. is acquired based on information obtained from these pieces of equipment. For this reason, the technology in Patent Document 2 requires the installation of additional equipment in the elevator system or at the location where the elevator is installed.

The problem that this embodiment aims to solve is to provide a group management device that can predict the occurrence of passengers being left behind and appropriately allocate cars.

An embodiment of the group management device controls the operation of multiple elevator cars between different floors, and includes: a control unit that assigns a car from the multiple cars to respond to a hall call for each of the different floors; and an analysis unit that uses a learning model constructed by machine learning to calculate the probability of passengers being left behind at a target floor from the different floors where the hall call was made. When a first hall call is made, the analysis unit estimates the number of passengers waiting at the target floor and the number of passengers in the first car assigned to respond to the first hall call, and calculates the probability from the number of passengers waiting and the number of passengers. If the probability is equal to or greater than a predetermined value, and there is a second car from the multiple cars passing in the direction specified by the first hall call while the doors of the first car are open at the target floor, the control unit stops the second car that was scheduled to pass at the target floor.

FIG. 1 is a diagram illustrating an example of the overall configuration of an elevator system according to an embodiment. FIG. 2 is a block diagram illustrating an example of a functional configuration of the elevator system according to the embodiment. FIG. 3 is a schematic diagram illustrating an example of the operation of the elevator system according to the embodiment. FIG. 4 is a schematic diagram illustrating another example of the operation of the elevator system according to the embodiment. FIG. 5 is a flowchart illustrating an example of a procedure for car allocation processing by the group management device according to the embodiment. FIG. 6 is a flowchart showing an example of a procedure for car allocation processing by a group control device according to a modified example of the embodiment.

(Example of elevator system configuration)
1 is a diagram illustrating an example of the overall configuration of an elevator system 10 according to an embodiment. As shown in FIG. 1, the elevator system 10 according to the embodiment includes a group control device 30, an elevator control device 40, a car 50, a hall 60, and a hoistway 70.

The elevator shaft 70 is provided across multiple floors inside a building or the like where the elevator system 10 is installed. Multiple cars 50 are configured to ascend and descend within the elevator shaft 70, allowing passengers to board and travel between different floors. Each floor of the building or the like is provided with a landing 60 where passengers can board and alight from the cars 50.

A rope 71, a counterweight 72, and a hoist 73 are provided within the elevator shaft 70.

Each car 50 is connected to a counterweight 72 via a rope 71. The rope 71 is also stretched across a hoist 73, which drives the rope 71 to feed the car 50, allowing it to ascend and descend within the hoistway 70 while balancing with the counterweight 72.

Each elevator car 50 is equipped with a destination floor call device 51, an elevator door 52, and a load sensor 54.

The destination floor call device 51 is installed inside the elevator car 50 and is configured with buttons indicating destination floors, etc. When a user inside the elevator car 50 presses a specific destination floor button, the destination floor call is accepted by the destination floor call device 51, which then transmits the call to the group control device 30 via the elevator control device 40 (described below) and registers it. This allows the user to direct the elevator car 50 to the desired destination floor.

The car door 52 is located at the entrance of the car 50. The car door 52 is closed while the car 50 is moving within the elevator shaft 70, and opens when the car 50 arrives at the landing 60 on a specified floor, allowing passengers to board and disembark from the car 50.

The load sensor 54 is provided, for example, on the underside of the outside of the car 50, and detects the weight of the entire car 50. Since the weight of the car 50 itself is known, detecting the weight of the entire car 50 makes it possible to detect the total weight of passengers and other people inside the car 50. This makes it possible to estimate the number of passengers inside the car 50.

Each landing 60 on each floor is equipped with a landing call device 61 and a landing door 62.

The hall call device 61 is configured with up and down buttons, etc. When a user at the hall 60 presses either the up or down button, the hall call is accepted by the hall call device 61, and the call is sent to and registered in the group management device 30 via the elevator control device 40 (described below). This allows the user to call a car 50 heading to the desired destination floor at the hall 60.

The platform door 62 is located at the entrance to the car 50 within the platform 60. The platform door 62 remains closed until the car 50 arrives at the platform 60, and once the car 50 arrives, the platform door 62 opens in conjunction with the car door 52 of the car 50. This allows passengers to board and alight from the car 50.

The group control device 30 and elevator control device 40 are each configured as computers equipped with a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The group control device 30 and elevator control device 40 are connected by wire or wirelessly so that they can exchange information with each other.

An elevator control device 40 is provided for each individual car 50 and controls each corresponding car 50.

The group control device 30 controls the operation of the corresponding multiple cars 50 via the elevator control device 40. More specifically, the group control device 30 registers destination floor calls from the destination floor call device 51 and hall calls from the hall call device 61, and causes one of the multiple cars 50 to respond to these call registrations.

Figure 2 is a block diagram showing an example of the functional configuration of the elevator system 10 according to an embodiment.

As shown in Figure 2, the elevator car 50 is equipped with a destination floor call device 51, a car door 52, and a load sensor 54. The destination floor call device 51, the car door 52, and the load sensor 54 exchange various information with the elevator control device 40.

For example, when the destination floor call device 51 receives a destination floor call from a user, it transmits the destination floor call to the elevator control device 40. The car doors 52 open and close in response to control signals received from the elevator control device 40. The load sensor 54 detects the weight of the car 50 and transmits the weight to the elevator control device 40.

In addition, a hoist 73 is provided in the hoistway 70, which exchanges various information with the elevator control device 40. Specifically, the hoist 73 is driven in response to control signals received from the elevator control device 40, and raises and lowers the car 50 within the hoistway 70.

In addition, the platform 60 is equipped with a platform call device 61 and a platform door 62. These platform call devices 61 and platform doors 62 exchange various information with the group control device 30.

For example, when the platform call device 61 receives a platform call from a user, it transmits the platform call to the group control device 30. Furthermore, the platform door 62 opens and closes in response to control signals received from the group control device 30.

The elevator control device 40 has functional units, including a communication unit 41, a control unit 42, and a memory unit 45. These functional units are realized, for example, by the above-mentioned CPU provided in the elevator control device 40 expanding a control program stored in ROM or the like into RAM and executing it.

The communication unit 41 communicates with external devices of the elevator control device 40.

For example, the communication unit 41 exchanges various information with the destination floor call device 51, car door 52, and load sensor 54 installed in the car 50, as well as the hoist 73 installed in the elevator shaft 70, to receive status information about each of these components and to transmit control signals and the like that control each of these components. The communication unit 41 also receives destination floor calls from the destination floor call device 51 in the car 50.

For example, the communication unit 41 exchanges various information with the group control device 30, transferring status information for each of the above-mentioned units and destination floor calls from the destination floor call device 51. The communication unit 41 also receives instructions from the group control device 30 to each of the above-mentioned units.

Based on instructions from the group management device 30, the control unit 42 generates control signals to control the destination floor call device 51, car door 52, and load sensor 54 installed in the car 50, as well as the hoist 73 installed in the elevator shaft 70, and controls each of these components via the communication unit 41.

The memory unit 45 stores control programs and control parameters that realize the functions of the elevator control device 40.

The group management device 30 has functional units, including a communication unit 31, a control unit 32, an analysis unit 33, a call management unit 34, and a memory unit 35. These functional units are realized, for example, by the above-mentioned CPU of the group management device 30 expanding into RAM and executing a control program stored in ROM or the like.

The communication unit 31 communicates with devices external to the group management device 30.

For example, the communication unit 31 communicates with the elevator control device 40 and obtains, via the elevator control device 40, status information for each part of the car 50, destination floor calls from the destination floor call device 51, and detection results from the load sensor 54. The communication unit 31 also transmits commands to the elevator control device 40 to control the car 50.

For example, the communication unit 31 communicates with the platform call device 61 and platform door 62 installed at the platform 60, and acquires their status information and platform calls from the platform call device 61. The communication unit 31 also transmits control signals that control each part of the platform 60.

The control unit 32 generates control signals to control the hall call device 61 and hall doors 62 installed at the hall 60, and controls these components via the communication unit 31. The control unit 32 also generates instructions to cause the car 50 to respond to destination floor calls obtained from the destination floor call device 51 or hall calls obtained from the hall call device 61, and transmits the generated instructions to the elevator control device 40 via the communication unit 31.

When a platform call is received from a platform 60 at a specified floor, the analysis unit 33 uses a learning model 37 constructed by machine learning to calculate the probability that a passenger will be left behind at that platform 60 when a specified car 50 is made to respond to the platform call. An occurrence of a passenger being left behind at a platform 60 means, for example, that a passenger is unable to fit into the car 50 that responded to the platform call.

The learning model 37 is obtained by learning data such as the operating status of multiple cars 50 in the elevator system 1 through machine learning, and is stored, for example, in the memory unit 35 of the group control device 30.

Data that is the subject of machine learning includes, for example, data showing the operation status of multiple elevator cars 50 by time period, day of the week, season, or weather. The operation status of multiple elevator cars 50 includes the location of multiple elevator cars 50 at any given time, the allocation status to platform calls or destination floor calls, and whether or not there are any passengers left behind at each platform 60 and the number of passengers left behind.

Whether or not passengers are left behind at an individual platform 60 can be estimated based on, for example, whether or not a call is made for a car 50 traveling in the same direction within a specified time after a platform call at that platform 60 is answered. Furthermore, if there are passengers waiting to board at a platform 60, the presence or absence of passengers left behind, as well as the length of time it takes for a specific car 50 to respond, can also be used to estimate. The number of passengers left behind when a passenger is left behind can be estimated based on, for example, the number of passengers who board a car 50 traveling in the same direction after the departure.

Furthermore, the number of passengers waiting to board, as well as the flow of people, may vary depending on the time of day, day of the week, season, or weather. Furthermore, passengers may be dressed lightly or heavily depending on the season, and depending on the weather, more passengers may carry rain gear, and the number of passengers that can fit in elevator 50 may also vary depending on these conditions.

The above-described machine learning is performed by providing the above data to, for example, a computer (not shown) equipped with machine learning capabilities. The machine learning computer constructs a learning model 37 that can calculate the probability of leftovers based on the provided data. The learning model 37 constructed by the machine learning computer is stored in the memory unit 35 or the like so that it can be used by the analysis unit 33 of the group management device 30.

However, the group management device 30 may also have a machine learning function. In this case, the above data may be stored in the group management device 30, and the group management device 30 itself may perform machine learning to construct the above learning model 37.

When a platform call occurs at a specific platform 60, the analysis unit 33 applies the operating status of multiple elevators 50 at that time, the time of day when the platform call occurred, the day of the week, season, weather, etc. to the learning model 37 to estimate the number of people waiting at the platform 60 where the platform call occurred.

The analysis unit 33 also estimates the number of passengers in the car 50 assigned to respond to the platform call. The number of passengers can be estimated, for example, from the detection results of a load sensor 54 installed in the car 50. Alternatively, the analysis unit 33 may estimate the number of passengers in the car 50 by applying various conditions at the time to the learning model 37, just as when estimating the number of people waiting at the platform 60.

Furthermore, the analysis unit 33 calculates the probability that passengers will be left behind at the platform 60 based on the estimated number of people waiting at the platform 60 where the platform call was made and the number of passengers in the car 50 assigned to the platform call. Note that the greater the number of people waiting at the platform 60 and the greater the number of passengers in the car 50, the higher the probability that passengers will be left behind at the platform 60.

If the probability of passengers being left behind is equal to or greater than a predetermined value, the control unit 32 described above will, for example, arrange for other cars 50 to head to the platform 60 in addition to the car 50 assigned to respond to the platform call.

The call management unit 34 stores the destination floor call obtained from the destination floor call device 51 in the memory unit 35 and registers the destination floor call. The call management unit 34 also stores the platform call obtained from the platform call device 61 in the memory unit 35 and registers the platform call.

The memory unit 35 stores control programs and control parameters that realize the functions of the group control device 30. The memory unit 35 also stores the call registration database 36 and the learning model 37 described above. The call registration database 36 stores call registrations such as destination floor call registrations and platform call registrations registered by the call management unit 34.

(Example of elevator system operation)
Next, an operation example of the elevator system 10 according to the embodiment will be described with reference to Fig. 3 and Fig. 4. Fig. 3 is a schematic diagram showing an example of the operation of the elevator system 10 according to the embodiment. Fig. 4 is a schematic diagram showing another example of the operation of the elevator system 10 according to the embodiment.

In the example shown in Figure 3, it is assumed that a hall call for travel to a lower floor is generated at hall 60 on a specific floor. Furthermore, when the hall call is generated, of the multiple cars 50, cars 50 No. 1 to No. 3 are located on floors above the floor where the hall call was generated, and it is possible to have one of these cars 50 respond to the hall call. It is assumed that the group control device 30 has decided to have, for example, car 50 No. 1 respond to the hall call from among cars 50 No. 1 to No. 3.

The No. 1 car 50 descends from an upper floor and opens its doors when it arrives at the platform 60 where the platform call was made at time t1. This allows passengers who have been waiting at that platform 60 for a car 50 heading to a lower floor to board the No. 1 car 50. The No. 1 car 50 then closes its doors at time t3 and departs from the platform 60 on that floor for the lower floor.

The elevators 50 of elevators 2 and 3 that were not assigned to the above-mentioned hall calls pass through the hall 60 where the hall call occurred at times t2 and t4, respectively. Time t2 is between time t1, when the doors of elevator 50 of elevator 1 open, and time t3, when the doors of elevator 50 of elevator 1 close. Time t4 is after time t3, when the doors of elevator 50 of elevator 1 close.

Here, if there are a large number of people waiting at platform 60, or if there are a large number of people already aboard car No. 1 50, even if car No. 1 50 responds to the platform call, not all passengers at platform 60 may be able to board car No. 1 50, resulting in passengers being left behind at platform 60.

Even in this case, the next platform call cannot be accepted while the doors of the No. 1 car 50 are open. Therefore, the No. 2 car 50 passes through the platform 60 without stopping. Passengers left behind at the platform 60 can finally make the next platform call after seeing the No. 2 car 50 pass by and the No. 1 car 50 depart. As a result, the No. 3 car 50 following the No. 1 and No. 2 cars will stop at the platform 60, for example, to accommodate the left-behind passengers.

To avoid such a situation, in the group management device 30 of this embodiment, the analysis unit 33 calculates the probability that passengers will be left behind at the platform 60 on the floor where the platform call was made. That is, for example, when a platform call is made, the analysis unit 33 estimates the number of people waiting at that platform 60. Furthermore, when the control unit 32 determines, for example, the allocation of a car 50 to respond to the platform call, the analysis unit 33 estimates the number of passengers in the allocated car 50.

This allows the analysis unit 33 to complete calculation of the probability of passengers being left behind around the time when the first car 50 arrives at the platform 60. In this case, it is more preferable that the calculation of the probability of passengers being left behind be completed before the first car 50 arrives at the platform 60.

If the probability of passengers being left behind calculated by the analysis unit 33 is equal to or greater than a predetermined value, the control unit 32 begins to consider other cars 50 that can respond to the platform 60, following the assigned No. 1 car 50.

More specifically, the control unit 32 checks whether there are any other cars 50 that can respond to the platform 60, such as cars 50 that are passing through the platform 60 in the direction specified by the platform call while the doors of car 50 No. 1 are open, that is, between times t1 and t3. In the example of Figure 3, this would be car 50 No. 2, which is passing through the platform 60 from an upper floor to a lower floor, from which the platform call was made.

In addition, the control unit 32 calculates the amount of time that a passenger will not be answered if a passenger is actually left behind. Here, the amount of time that is not answered is, for example, the time from when the first car 50 that responded to the above-mentioned platform call departs from that platform 60 until when a car 50 that can respond to a further platform call from the left-behind passenger arrives at that platform 60.

In the example of Figure 3, as described above, after the departure of car No. 1 50, the car 50 that can respond to a new platform call is car No. 3 50. Therefore, the unanswered time for the left-behind passengers is from time t3, which is the departure time of car No. 1 50, to time t4, when car No. 3 50 arrives at the platform 60.

If the unresponsive time indicated by times t3 to t4 exceeds a predetermined time, the control unit 32 determines to stop the second car 50, which is scheduled to pass through the same platform 60, at the platform 60 while the doors of the first car 50 are open, and to allow the remaining passengers to respond. An example of this is shown in Figure 4.

As shown in Figure 4, after the arrival of the car 50 of the first train at time t1, the car 50 of the second train arrives at the platform 60 at time t2 and opens its doors. Then, at time t3, the car 50 of the first train closes its doors and departs from the platform 60, and then, at time t3' after time t3, the car 50 of the second train closes its doors and departs from the platform 60. If there is not a high probability of further passengers being left behind, the car 50 of the third train will pass through the platform 60 at time t4 as scheduled.

As described above, when a platform call occurs at a specific floor and the probability of passengers being left behind at that platform 60 is equal to or greater than a specific value, and there is another car 50 passing through that platform 60 while the doors of the car 50 assigned to that platform call are opening, and further, if the unanswered time after the departure of the originally assigned car 50 exceeds a specific time, the group control device 30 of the embodiment assigns the car 50 that was scheduled to pass while the doors of the assigned car 50 were opening to respond to the left-behind passengers.

(Example of elevator system processing)
Next, an example of processing by the group control device 30 of the embodiment will be described with reference to Fig. 5. Fig. 5 is a flow diagram showing an example of the procedure of processing for allocating cars 50 by the group control device 30 according to the embodiment.

As shown in FIG. 5, the group control device 30 monitors whether a hall call is made at any of the halls 60 (step S101). The group control device 30 continues monitoring as long as no hall call is made (step S101: No).

When a platform call is made at any of the platforms 60 (step S101: Yes), the analysis unit 33 estimates the number of people waiting at the platform 60 on the floor where the platform call was registered, using, for example, the learning model 37 stored in the memory unit 35 (step S102).

Furthermore, once the control unit 32 assigns a car 50 to respond to the hall call, the analysis unit 33 estimates the number of passengers in the car 50 responding to the hall call (step S103). As described above, the number of passengers may be estimated, for example, using the learning model 37, or may be estimated based on the detection results of a load sensor 54 installed in the car 50 responding to the hall call.

The analysis unit 33 calculates the probability that passengers will be left behind at the platform 60 based on the number of people waiting at the platform 60 and the number of passengers in the car 50 estimated in steps S102 and S103 (step S104). The control unit 32 determines whether the calculated probability of passengers being left behind is equal to or greater than a predetermined value (step S105).

If the probability of passengers being left behind is equal to or greater than a predetermined value (Step S105: Yes), the control unit 32 checks the operation status of the multiple cars 50 (Step S106). Based on the operation status of the multiple cars 50, the control unit 32 determines whether there is another car 50 scheduled to pass through the platform 60 in the same direction as the car 50 whose door is open while the car 50 that responded to the platform call is opening (Step S107).

If there is a car 50 scheduled to pass through that platform 60 (step S107: Yes), the control unit 32 determines whether the time without a response from the remaining passengers will be a predetermined time or longer after the car 50 that responded to the platform call closes its doors (step S108).

If the unresponsive time is equal to or longer than the predetermined time (step S108: Yes), the control unit 32 calculates the number of passengers that can fit in the car 50 identified in the processing of step S107 (step S109). More specifically, the control unit 32 estimates the current number of passengers from the detection results of, for example, the load sensor 54 installed in the car 50, and can further calculate the number of passengers that can fit from the current number of passengers.

The control unit 32 stops the number of cars 50 that can accommodate all of the remaining passengers at the stop 60 where the stop call occurred, based on the number of passengers that can be accommodated in the cars 50 identified in the processing of step S107 (step S110). Note that if the number of cars 50 identified in the processing of step S107 is less than the number that can accommodate all of the remaining passengers, it may decide to stop all of these cars 50.

If the probability of passengers being left behind is less than a predetermined value (step S105: No), if there are no cars 50 scheduled to pass through the platform 60 while the doors of a car 50 that responded to a platform call are open (step S107: No), or if the unanswered time after the doors of a car 50 that responded to a platform call are closed is less than a predetermined time (step S108: No), no processing is performed to stop any more cars 50 in addition to the initially assigned cars 50.

This completes the elevator car 50 allocation process performed by the group management device 30 in this embodiment.

(Overview)
A group control device is known that controls the operation of multiple elevator cars. The group control device controls the operation of multiple elevator cars, for example, to operate at equal intervals in order to equalize waiting times at each floor. In this case, if passengers are left behind when a hall call is answered, the left-behind passengers must wait for the car that answered the hall call to depart before making a new hall call, which results in a long unanswered time for these passengers.

For this reason, the group control device prevents passengers from being left behind by, for example, allowing cars with more than a certain number of passengers to pass through without responding to calls from the platform, or by assigning cars with fewer than a certain number of passengers.

However, it is difficult to completely eliminate backlogs with the above measures, and it is also difficult to shorten the unresponsive time when backlogs do occur.

According to the group control device 30 of this embodiment, a learning model 37 constructed by machine learning is used to calculate the probability that passengers will be left behind at the target floor where the hall call occurred among different floors. In this way, by using a learning model 37 based on machine learning, it is possible to predict the occurrence of passengers being left behind and appropriately allocate elevator cars 50.

According to the group control device 30 of this embodiment, if the probability of passengers being left behind is equal to or greater than a predetermined value, and if one of the multiple cars 50 assigned to respond to a hall call is at the target floor where the call was made and the car 50 is opening its doors, and there is another car 50 passing in the direction specified by the hall call, the other car 50 that was scheduled to pass through is stopped at the target floor.

In this way, by stopping a car 50 at a platform 60 that passes by without being able to make an additional platform call while the doors of the originally assigned car 50 are open, it is possible to prevent the unanswered time for left-behind passengers from becoming too long.

According to the group control device 30 of this embodiment, if the probability of passengers being left behind is equal to or greater than a predetermined value and if the non-response time after the doors of the assigned car 50 closes is equal to or greater than a predetermined time, another car 50 is stopped at the target floor. In this way, whether or not to stop another car 50 is determined based on the non-response time being equal to or greater than a predetermined time, so it is possible to avoid making additional responses inadvertently and prevent additional responses from being wasted.

According to the group control device 30 of this embodiment, the probability of passengers being left behind is calculated by applying the learning model 37 to at least one of the estimation of the number of people waiting at the target floor where the hall call was made and the estimation of the number of passengers in the elevator car 50 at the time of the response. This makes it possible to calculate the probability of passengers being left behind without, for example, installing additional equipment in the elevator system 10.

According to the group management device 30 of this embodiment, the number of people waiting at a target floor is estimated using the learning model 37. This makes it possible to obtain an estimate of the number of people waiting at each landing 60 without installing additional equipment such as an imaging device at each landing 60.

According to the group control device 30 of this embodiment, the number of passengers in the car 50 at the time of response is estimated from the detection results of the load sensor 54 installed in the car 50. Elevator systems 10 are typically already equipped with a load sensor 54 for detecting whether the car 50 is overloaded. By using the detection results of the load sensor 54, which is an existing facility, a more accurate number of passengers in the car 50 can be obtained. Because there is no need to estimate the number of passengers in the car 50, the burden of building a learning model 37 using machine learning can also be reduced.

In the above-described embodiment, when the additional car 50 responds to the left-behind passengers, an announcement encouraging them to board the additional car 50 may be made using a speaker (not shown) installed at the platform 60 or the car 50, or a display board or the like installed at the platform 60.

Furthermore, in the above-described embodiment, the learning model 37 may be updateable as appropriate. Specifically, for example, if an additional elevator 50 is sent to respond to left-behind passengers but no passengers board, the learning model 37 can be updated to further reflect the situation at the time, such as the number of passengers in the originally assigned elevator 50 after the response or the amount of available space in the elevator 50, so as to enable more accurate predictions.

(Modification)
Next, a processing example of a group management device according to a modified example of the embodiment will be described with reference to Fig. 6. The group management device according to the modified example has a different condition for determining whether to stop an additional car 50 from that of the group management device 30 according to the above-described embodiment.

Figure 6 is a flow diagram showing an example of the procedure for allocating cars 50 by a group management device according to a modified embodiment.

As shown in Figure 6, in the modified group control device, steps S201 to S204 are processed in the same manner as steps S101 to S104 described above. That is, when there is a hall call at a given floor (step S201: Yes), the modified group control device also calculates the probability that a passenger will be left behind at that hall 60 (steps S202 to S204).

The control unit assigns a car 50 to respond to the hall call, and after the assigned car 50 responds, the doors close (step S205). The control unit determines whether the probability of passengers being left behind calculated in the processing of step S204 is equal to or greater than a predetermined value (step S206).

If the probability of passengers being left behind is equal to or greater than a predetermined value (step S206: Yes), the control unit determines whether the number of passengers in the elevator 50 after the response was equal to or greater than a predetermined number when departing from the platform 60 (step S207). The number of passengers in the elevator 50 after the response can be estimated, for example, from the detection results of the load sensor 54 installed in the elevator 50.

Furthermore, if the number of passengers in the elevator car 50 after the response is equal to or greater than the predetermined number (step S207: Yes), the possibility that passengers have been left behind increases even further.

The control unit further determines whether there is another platform call from that platform 60 within a predetermined time after the responded car 50 departs from that platform 60, heading in the same direction as the responded car 50 (step S208). If there is an additional platform call within the predetermined time (step S208: Yes), it is highly likely that the call is from a passenger who was left behind.

When assigning elevator cars 50 to respond to additional hall calls, the control unit prioritizes shortening the response time (step S209).

Normally, when the control unit assigns cars 50, it selects a specific car 50 by equally considering several conditions, such as the load of each car 50, the time it takes for each car 50 to respond, and the number of hall calls at other floors. In other words, the weighting of these conditions is equalized, and cars 50 that meet these conditions to an equal extent are assigned. This can improve the overall operating efficiency of the elevator system.

However, when all of steps S206 to S208 are met and an additional car 50 is assigned, the car 50 is selected by, for example, placing a higher weight on the time required for each car 50 to respond than on the other conditions. This makes it possible to minimize the unresponsive time in situations where a high probability of passengers being left behind is predicted.

In order to allocate an additional car 50 with priority given to shortening the response time, the control unit may change the allocation of an already determined car 50, for example, to have it respond to a hall call for another floor. Furthermore, the control unit may not allocate the additionally allocated car 50 to other calls until the response to the additional hall call in the processing of step S208 has finished.

If the probability of passengers being left behind is less than a predetermined value (step S206: No), if the number of passengers in the initially assigned car 50 after the response is complete is less than a predetermined number (step S207: No), or if there are no additional calls from the platform within a predetermined time period (step S208: No), no process is performed to stop any additional cars 50 in addition to the initially assigned car 50.

This completes the elevator car 50 allocation process performed by the modified group management device.

According to the modified group management device, if the probability of a passenger being left behind is equal to or greater than a predetermined value, and if an additional platform call from the same direction as the assigned car 50 is received at the target floor within a predetermined time after the assigned car 50 departs from the target floor, a predetermined car 50 is assigned from among the multiple cars 50, with shortening the response time to the additional platform call being given priority.

This allows hall calls resulting from a passenger being left behind to be distinguished from normal hall calls. Furthermore, in such cases, an additional car 50 is allocated with priority given to shortening the response time, further reducing the unanswered time when a passenger is left behind.

In accordance with the modified group management device, if there is an additional hall call within a specified time period and the number of passengers on the elevator 50 departing from the target floor is equal to or greater than a specified number, an additional elevator is allocated with a priority on shortening the response time. This makes it possible to more reliably determine when passengers are left behind and prevents inadvertent allocations that prioritize response time.

In accordance with the modified group management device, additional cars 50 are allocated with the priority given to shortening response times, and the allocation of other cars 50 that have already been determined is changed. This allows for more flexible allocation of additional cars when passengers are left behind, further reducing the unresponsive time when passengers are left behind.

In the modified group management device, after an additional car 50 is assigned, no additional assignments are made to that car 50. This further reduces the unresponsive time when a passenger is left behind.

The group management device of the modified example can use the modified example method in addition to the method of the above-described embodiment. In this case, for example, the modified example method may be applied when there is no corresponding car 50 in the processing of step S107 in Figure 5 of the above-described embodiment. Also, even if the non-response time is less than the predetermined time in the processing of step S108 in Figure 5 of the above-described embodiment, the modified example method may be applied again to assign an additional car 50.

Furthermore, in the above-described modified example, when the additional car 50 responds to the left-behind passengers, a speaker (not shown) installed at the platform 60 or the car 50, or a display board or the like installed at the platform 60 may be used to inform the passengers of the arrival time of the additional car 50 and encourage them to board the additional car 50.

While several embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments may be embodied in a variety of other forms, and various omissions, substitutions, and modifications may be made without departing from the spirit of the invention. These embodiments and their variations are within the scope and spirit of the invention, and are also included in the scope of the invention and its equivalents as set forth in the claims.

10...Elevator system, 30...Group control device, 31...Communication unit, 32...Control unit, 33...Analysis unit, 34...Call management unit, 35...Memory unit, 36...Call registration DB, 37...Learning model, 40...Elevator control device, 50...Cab, 51...Destination floor call device, 54...Load sensor, 60...Landing, 61...Landing call device.

The group control device of an embodiment is a group control device that controls the operation of a plurality of elevator cars between different floors, and includes a control unit that assigns a car among the plurality of cars to respond to a hall call for each of the different floors, and a learning model that is constructed by machine learning based on data showing the operation status of the plurality of cars , and estimates at least one of the number of people waiting at a target floor among the different floors where the hall call occurred and the number of people in a car that is assigned to respond to the hall call, and calculates the probability that passengers will be left behind at the target floor. and an analysis unit that calculates the probability of a first floor call being received, wherein when a first floor call is received, the analysis unit estimates the number of people waiting at the target floor, estimates the number of people riding in a first car assigned to respond to the first floor call, and calculates the probability from the number of people waiting and the number of people riding in the first car, and when the probability is equal to or greater than a predetermined value and there is a second car among the multiple cars passing in the direction specified by the first floor call while the doors of the first car are open at the target floor, the control unit stops the second car that was scheduled to pass at the target floor.

Claims (10)

A group control device that controls the operation of multiple elevator cars between different floors,
a control unit that assigns cars among the plurality of cars to respond to hall calls for the different floors;
an analysis unit that calculates a probability that a passenger will be left behind at a target floor where the hall call was made among the different floors, using a learning model constructed by machine learning;
The analysis unit
When a first hall call is received, the number of people waiting at the target floor is estimated, the number of people riding in a first car assigned to respond to the first hall call is estimated, and the probability is calculated from the number of people waiting and the number of people riding;
The control unit
When the probability is equal to or greater than a predetermined value, and when there is a second car among the plurality of cars passing in the direction specified by the first hall call while the door of the first car is open at the target floor, the second car that was scheduled to pass is stopped at the target floor.
Group management device. The control unit
If the probability is equal to or greater than a predetermined value and if a non-response time after the door of the first car is closed is equal to or greater than a predetermined time, the second car is stopped at the target floor.
The group management device according to claim 1 . The control unit
When the probability is equal to or greater than a predetermined value, and when a second hall call in the same direction as the first hall call is also at the target floor within a predetermined time after the first hall call departs from the target floor, assign a third hall call from among the plurality of halls, with shortening the response time to the second hall call as a priority condition.
The group management device according to claim 1 . The control unit
When the second hall call is received within the predetermined time and the number of passengers in the first car that departed from the target floor is equal to or greater than a predetermined number, the third car is assigned with a priority given to shortening the response time.
The group management device according to claim 3 . The control unit
changing the allocation of other cars that have already been determined in order to allocate the third car with the shortening of the response time as a priority condition;
The group management device according to claim 3 . The control unit
after allocating the third car, no additional allocations are made to the third car;
The group management device according to claim 3 . The analysis unit
calculating the probability by applying the learning model to at least one of the estimation of the number of people waiting at the target floor and the estimation of the number of people riding in the first car at the time of the response;
The group management device according to claim 1 . The analysis unit
Estimating the number of people waiting on the target floor using the learning model;
The group management device according to claim 7. The first car includes:
a load sensor capable of detecting the number of passengers in the first car is provided;
The analysis unit
The number of passengers in the first car at the time of the response is estimated from the detection result of the load sensor.
The group management device according to claim 7. A group control device that controls the operation of multiple elevator cars between different floors,
a control unit that assigns cars among the plurality of cars to respond to hall calls for the different floors;
an analysis unit that calculates a probability that a passenger will be left behind at a target floor where the hall call was made among the different floors, using a learning model constructed by machine learning;
The analysis unit
When a first hall call is received, the number of people waiting at the target floor is estimated, the number of people riding in a first car assigned to respond to the first hall call is estimated, and the probability is calculated from the number of people waiting and the number of people riding;
The control unit
When the probability is equal to or greater than a predetermined value, and when a second hall call in the same direction as the first hall call is also at the target floor within a predetermined time after the first hall call departs from the target floor, assign a third hall call from among the plurality of halls, with shortening the response time to the second hall call as a priority condition.
Group management device.