JP2010006613A - Traffic demand predicting device of elevator - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a traffic demand predicting device of an elevator capable of arranging elevator cars in flexible compliance with the change of the traffic situation. <P>SOLUTION: The elevator traffic demand predicting device to predict the number of persons coming waiting at an elevator hall of each floor is equipped with a personal authentication-attribute detection means for carrying out personal authentication and producing personal getting on/off data concerning getting on/off of a person based on detection of the personal attribute, a staying floor detection means for detecting staying time at each floor of the person from the personal getting on/off data, and a waiting person number predicting means for obtaining usage probability that the person uses the elevator based on the personal getting on/off data and the staying time and for predicting the number of persons coming waiting at each floor grouped into elevating and sinking of the elevator based on the usage probability of all elevator users. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to an apparatus for predicting traffic demand at each elevator hall.

  The conventional elevator traffic demand prediction device is a traffic demand prediction device for predicting the number of waiting passengers at each elevator landing. Each floor visitor number detection device detects the number of people staying on each floor at the present time. An elevator utilization rate learning device for learning an elevator utilization rate for the vehicle, and a waiting rate occurrence predicting device for predicting a wait occurrence rate for each floor direction from the elevator utilization rate and the current number of stayers on each floor, The number of passengers generated by the direction of the floor is predicted from the occurrence rate and the arrival interval of the car (see, for example, Patent Document 1).

JP-A-1-69480

  However, since the conventional elevator traffic demand forecasting device predicts the waiting time occurrence rate from the number of people staying on each floor and the elevator usage rate for each situation, it is difficult to follow the change in the traffic situation, and the traffic situation is not expected There is a problem that it is difficult to deal with. In addition, since only the number of people staying on each floor and the elevator usage rate are referred to, there is a problem that the individual traffic behavior characteristics are not reflected in the waiting time occurrence rate.

  An object of the present invention is to provide an elevator traffic demand prediction device that dispatches cars flexibly in response to changes in traffic conditions.

  An elevator traffic demand prediction device according to the present invention is an elevator traffic demand prediction device that predicts the number of waiting passengers at a landing on each floor of an elevator. From the personal authentication / attribute detection means for generating the getting-on / off data, the stay-floor detecting means for detecting the staying time of each person on the floor from the personal getting-on / off data, the personal getting-on / off data and the staying time, Waiting time occurrence number predicting means for obtaining the use probability of using the elevator and predicting the number of waiting times on each floor according to the rise and fall of the elevator from the use probability of all elevator users.

  The elevator traffic demand prediction device according to the present invention predicts the occurrence of waiting on the basis of the time series data of each individual and the stay floor of each individual. Since the car is dispatched in consideration of the individual behavior pattern, the waiting time of the individual who newly calls the car can be further shortened.

It is a general-view figure of the elevator in which the elevator traffic demand prediction apparatus concerning Embodiment 1 of this invention was installed. It is a functional block diagram of the elevator traffic demand prediction apparatus of Embodiment 1. It is a figure for demonstrating the data structure of time series data. It is a flowchart which shows the procedure of the update of the time series data which consist of boarding / alighting persons. It is a flowchart which shows the procedure which determines the standby floor of a cage | basket | car. It is a general-view figure of the elevator in which the elevator traffic demand prediction apparatus concerning Embodiment 2 of this invention was installed. It is a functional block diagram of the elevator traffic demand prediction apparatus of Embodiment 2. It is a general-view figure of the elevator in which the elevator traffic demand prediction apparatus concerning Embodiment 3 of this invention was installed. It is a functional block diagram of the elevator traffic demand prediction apparatus of Embodiment 3. It is a general-view figure of the elevator in which the elevator traffic demand prediction apparatus concerning Embodiment 4 of this invention was installed. It is a functional block diagram of the elevator traffic demand prediction apparatus of Embodiment 4.

Embodiment 1 FIG.
FIG. 1 is an overview diagram of an elevator in which an elevator traffic demand prediction apparatus according to Embodiment 1 of the present invention is installed. FIG. 2 is a functional block diagram of the elevator traffic demand prediction apparatus according to the first embodiment. FIG. 3 is a diagram for explaining the data structure of time-series data such as the number of passengers. FIG. 4 is a flowchart showing a procedure for updating time-series data including the number of passengers. FIG. 5 is a flowchart showing a procedure for determining the waiting floor of the car.

As shown in FIG. 1, the elevator 1 includes a car 3 that moves up and down in the hoistway 2, a general call button 4 that is provided at the landing of each floor, a general destination floor registration device 5 that is provided in the car 3, The car 3 is composed of a car driving device 6 that moves up and down the car 3 and an elevator control device 7 that controls the car driving device 6.
A scale 8 for measuring the weight of the passenger is provided on the floor inside the car 3. In addition to the scale 8, the number of people may be counted using an image sensor, a photoelectric sensor, or the like.
The call button 4, the destination floor registration device 5, and the scale 8 are connected to the elevator control device 7 via the communication line 9.

The elevator traffic demand prediction device 11 is provided in the elevator control device 7.
As shown in FIG. 2, the elevator traffic demand prediction device 11 includes a car number measuring means 12, a passenger number calculating means 13, a time series data managing means 14, a staying person calculating means 15, a waiting time occurrence number predicting means 16, a prediction. It has function learning means 17 and car allocation floor determination means 18.

  The number-of-cars counting means 12 in the car measures the number of persons from the increase / decrease in the weight of the person on board, which is measured by the balance 8 installed in the car 3.

The boarding / alighting number calculating means 13 counts the data of increase / decrease in the number of persons in the car 3 obtained from the number-of-cars counting means 12 every predetermined period, and the number of passengers rising when the car 3 rises and stops at the floor f. n _{fL +} (t) and the number of up and down _vehicles n _{fUL +} (t) are calculated. Similarly, to calculate the falling number of passengers _{n fL-} (t) and down the getting-off number _{n fUL-} (t) when the car 3 is stopped to the floor f lowered. The number of passengers n _{fL +} (t), n _fL− (t) and the number of _passengers n _{fUL +} (t), n _fUL− (t) when the car is stopped use, for example, the method described in Japanese Patent Laid-Open No. 3-61273. Can be calculated.

As shown in FIG. 3, the time-series data management unit 14 calculates the number of passengers n _{fL +} (t), n _fL− (t ) And the number of _passengers n _{fUL +} (t) and n _fUL− (t) are updated in a first-in first-out manner. In FIG. 3, time-series data going back from time t of floor f to time (t−Z) is stored. In addition, it can be determined appropriately by examining the installation environment of the elevator 1 how long it goes back in the past.

The number-of-stays calculating unit 15 calculates the number of passengers _{every time the} number of passengers n _{fL +} (t), n _fL− (t) and the number of _passengers n _{fUL +} (t), n _fUL− (t) are calculated. The total number of passengers N _fL on the floor f is obtained from the equation (1), and the total number of passengers N _fUL on the floor f is _obtained from the equation (2). The counting is started after the total number of passengers and the total number of people getting off are initialized to zero during an unattended time such as midnight.

Next, the staying number calculating means 15 obtains the number of staying persons n _f (t) for each floor at the current time t according to the equation (3), and the time series data managing means 14 determines the number of staying persons as shown in FIG. Update series data.

It should be noted that where a lifting means such as a staircase or an escalator can be used in parallel with the elevator, a correction may be performed in consideration of the utilization ratio of the lifting means other than the elevator.
In addition, the number of people staying on the floor with the entrance to the outside of the building is a personal authentication device at the entrance such as IC card, RFID, biometric authentication by fingerprint or face, etc., and a person counting sensor such as an infrared sensor installed at the entrance gate You may correct | amend using the building entrance / exit number measured using.

The waiting time occurrence number predicting means 16 is the number of passengers n _{fL +} (i), n _fL− (i) (i = t) until the time T goes back in the past including the current time t stored in the time series data managing means 14. ~ T-T) and the number of visitors n _f (t) at the current time t, the predicted number P (hat) _{f +} (t) and P (hat) _f- (t) of the number of waiting customers generated for each floor direction are predicted. . The predicted number of waiting customer occurrences P _{f +} (t) waiting for an elevator that rises per predetermined time from the current time t to the future, and the predicted number of waiting customer occurrences P _f− (t) waiting for an elevator that descends are respectively Prediction is made using equations (4) and (5). For example, the calculation is performed using a function with the number of people n _f (t) staying on a certain floor f and the number of passengers n _{fL +} (t−i) from time t to time t−T as variables.

  Here, g is a waiting time occurrence prediction function, T is a maximum consideration time width, and (t−i) is a time of a predetermined cycle of t ≦ and ≦ t−T.

At this time, as an example of the waiting time occurrence number prediction function g, when the equation (6) that is a linear function is taken as an example, the waiting time occurrence number prediction number P _{f +} (t) waiting for the elevator rising from the equation (6) is obtained. Can do. In addition, as shown in Expression (7), it is possible to obtain the predicted number of waiting passengers P _f− (t) waiting for the elevator to descend. Note that b, d, and a _{0 to} a _T and c _{0 to} c _T are coefficients.
In addition, although the linear function was mentioned as an example as the waiting customer generation | occurrence | production number prediction function g, it is not restricted to this, It can predict similarly even if it uses other nonlinear functions.

The prediction function learning unit 17 learns parameters such as a coefficient of the waiting customer occurrence number prediction function g from the time series data obtained from the time series data management unit 14.
There are several possible learning methods. For example, when the waiting time occurrence number prediction function g is a linear function shown in Expression (6), parameters such as coefficients can be learned by the method of least squares.
Further, even when the waiting time occurrence number prediction function g is a non-linear function, it is possible to learn parameters such as coefficients by a method such as a neural network.

The car allocation floor determining means 18 waits for each floor direction during a predetermined time from the current time t obtained from the waiting time occurrence number prediction means 16 when the hall call and the in-car call are not assigned to the car 3. The waiting car dispatch floor is determined from the predicted number of passengers generated P _{f +} (t), P _f− (t). The car dispatch floor is a floor of F that minimizes the evaluation function V (F) of Expression (8). F is all floors. The function S (x, y) in the equation (8) is the moving time of the car 3 from the x floor to the y floor and the door opening time on the y floor. When x = y, S (x, y) = 0.

In Formula (8), in order to obtain F that minimizes the evaluation function V (F), for example, all floors from the lowermost floor to the uppermost floor are substituted for F to obtain a round robin.
The car dispatch floor determining means 18 determines F as the car dispatch floor that minimizes the evaluation function V (F) of the equation (8), but the predetermined time from the current time t of each floor shown in the equation (9). It is good also as a floor with the largest number of waiting-time customer generation predictions ahead.

In addition, when there are a plurality of cars 3, among the cars 3 to which calls are no longer assigned, first, the floor of F at which the evaluation function V (F) of the equation (8) is the minimum is selected for one car 3. It may be determined as a vehicle allocation floor, and the floor of F that has the second smallest evaluation function V (F) for the next car 3 may be determined as the vehicle allocation floor, and the vehicle allocation floor may be sequentially determined.
Alternatively, the vehicle allocation floor may be determined in order, such as the first floor with the maximum value of Equation (9) and the second largest floor.
In addition, the floor where the number of people staying is zero may be excluded from candidates for the allocation floor because there is almost no possibility of passengers getting on.

  The car driving device 6 moves the car 3 to the car dispatch floor determined by the car dispatch floor determining means 18. However, if a new hall call is assigned to the car 3 that is moving to the dispatch floor, it is moved to the landing call dispatch floor immediately after stopping at the nearest floor.

Next, a procedure for updating the time-series data including the number of passengers, the number of people getting off, and the number of people staying and the waiting number occurrence function prediction function will be described with reference to FIG.
In step 101, the boarding / alighting number calculating means 13 collects passenger weight data from the car number measuring means 12 every predetermined cycle, and proceeds to step 102.
In step 102, the boarding / alighting number calculating means 13 calculates the number of passengers boarding and the number of boarding / exiting persons from the weight data of passengers in a predetermined period, and proceeds to step 103. When the car 3 stops after ascending, the number of _uplifting passengers n _{fUL +} (t) and the number of rising passengers n _{fL +} (t) are calculated. Also, when the car is stopped after lowering, lowering alighting persons _{n fUL-} (t) and lowered number of passengers _{n fL-} (t) is calculated.
In step 103, the time series data management means 14 calculates the calculated number of _uplifting passengers n _{fUL +} (t) and the number of rising passengers n _{fL +} (t) or the number of descending passengers n _fUL− (t) and the number of descending passengers n _fL− ( t) is the latest data, and the time series data relating to the number of up and down passengers and the number of ascending and descending passengers and the number of descending and descending passengers and the number of descending passengers over a predetermined period in the past are updated and stored.
In step 104, the staying number calculation means 15 reads the data of the number of up and down passengers, the number of ascending passengers, the number of descending and descending passengers, and the number of descending passengers over a desired period stored in the time-series data management means 14, and formula (1) After calculating the number of visitors n _f (t) according to the equation (2), the calculated number of visitors n _f (t) is the latest data, and the time series data relating to the number of visitors over a predetermined period in the past is updated. The data is stored in the time series data management unit 14 and the process proceeds to Step 105.
In step 105, the prediction function learning means 17 reads the time series data of the rising number of passengers and the number of staying persons during the preset period (S + 1) from the time series data management means 14, and (S + 1) as in equation (10). The coefficients q and p _{0 to} p _T are obtained from the equations by the method of least squares, and are substituted into b = q, a ₀ = p ₀ ,..., A _T = p _T in equation (6). Similarly, the prediction function learning unit 17 reads the time series data of the number of passengers in the descent and the number of people in the period (S + 1) set in advance from the time series data management unit 14, obtains the coefficient by the least square method, ) D, c ₀ ,..., C _T, and the procedure for updating the waiting time occurrence number prediction function is completed.

Next, a procedure for dispatching to the dispatch floor for waiting for the car 3 will be described with reference to FIG.
In step 201, the waiting time occurrence number predicting means 16 determines whether or not there is a call from the call button 4. Step 201 is repeated when there is a call. When there is no call, the process proceeds to step 202.
In step 202, the waiting time occurrence number predicting means 16 determines the number of rising passengers and the number of falling passengers for a predetermined period stored in the time-series data management means 14, that is, from the past (t-T) to the present (t). The time series data relating to the number of visitors is read, and the process proceeds to step 203.
In step 203, the waiting customer occurrence number prediction means 16 reads the waiting customer occurrence number prediction function stored in the prediction function learning means 17, and proceeds to step 204.
In step 204, the waiting customer occurrence number predicting means 16 calculates the waiting customer occurrence number predicted number according to the equations (6) and (7) based on the time series data read in step 202 and the waiting customer occurrence number prediction function. Predict and go to step 205.
In step 205, the car dispatch floor determining means 18 obtains an evaluation function V (F) according to the equation (8) from the predicted number of waiting customers and the predetermined required movement value. Then, the evaluation function V (F) is obtained over all the floors, the floor having the smallest value is determined as the dispatching floor for standby, and the car driving device 6 is instructed to dispatch the car 3 to the dispatching floor. Then, the procedure for determining the allocation floor for standby is completed.

Such an elevator traffic demand prediction device predicts the number of waiting passengers from the time series data on the number of passengers getting on and off the elevator and the number of people staying on each floor. Can do.
In addition, when an unexpected traffic situation occurs, the car can be dispatched in response to the circumstances.

  In addition, since the waiting car is dispatched based on the prediction of the number of waiting passengers generated for each floor direction, the waiting time of the passenger who newly calls the car can be shortened as compared with the prior art.

Embodiment 2. FIG.
FIG. 6 is an overview diagram of an elevator in which an elevator traffic demand prediction apparatus according to Embodiment 2 of the present invention is installed. FIG. 7 is a functional block diagram of the elevator traffic demand prediction apparatus according to the second embodiment.

As shown in FIG. 6, the elevator 20 in which the elevator traffic demand prediction device 21 according to the second embodiment is arranged is the same as the elevator 1 according to the first embodiment, with a special call button 22, a special destination floor registration device 23, a fire alarm. Since the machine 24 and the evacuation operation command device 25 are further provided and the others are the same, the same reference numerals are attached to the same parts and the description thereof is omitted.
The special call is a call for a specific person other than a general hall call button such as a wheelchair call using a wheelchair call button or a VIP call using a VIP call button. In the following description, unless otherwise specified, the special call is a general expression of calls for all specific persons such as wheelchair calls and VIP calls.
The special call button 22 is a button that is provided at the landing and can make a special call. The special call button 22 includes a wheelchair call button and a VIP call button.
The special destination floor registration device 23 is provided in the car 3 and can make a special call. The special destination floor registration device 23 has a function capable of distinguishing wheelchair calls and VIP calls.
The fire alarm 24 transmits fire detection information to the evacuation operation command device 25 based on a fire signal input from a fire detector (not shown). The evacuation operation command device 25 transmits an evacuation operation command to the elevator traffic demand prediction device 21 based on the fire detection information.

  As shown in FIG. 7, the elevator traffic demand prediction device 21 according to the second embodiment is different from a normal one when a function related to a special call is added to the elevator traffic demand prediction device 11 according to the first embodiment. Since the car delivery is different and the others are the same, the description of the same parts is omitted.

  The number-of-cars counting means 26 is used to increase or decrease the weight of the passenger when the car 3 stops due to a special call from the special call button 22 and when the car 3 stops due to a special call from the special destination floor registration device 23. Record the passenger weight increase / decrease data separately from the general call.

Next, the boarding / _alighting number calculating means 27 calculates the number of people boarding on the floor stopped by the special call button 22 as the special number of calling passengers n _{fAL +} (t), n _fAL− (t), and the special destination floor registration device 23. The number of people who got off at the floor stopped by the above is calculated as the special calling number of people n _{fAUL +} (t), n _fAUL− (t). As for the floor stopped without being related to the special call button 22 and the special destination floor registration device 23, the number of passengers n _{fL +} (t), n _fL− (t) The number of persons is calculated as n _{fUL +} (t) and n _fUL− (t). The method of measuring the number of passengers is obtained by the same method as the number of passengers calculating means 13 of the first embodiment, and the number of passengers is acquired for each type of special call. That is, the number of wheelchair climbing passengers, the number of wheelchair descending passengers, the number of wheelchair climbing passengers, the number of wheelchair descending passengers, the number of VIP climbing passengers, the number of VIP descending passengers, the number of VIP climbing passengers, and the number of VIP descending / falling passengers are acquired.
The time-series data management means 28 updates the time-series data of the normal boarding / alighting number and the special calling number of boarding / exiting persons with the data calculated by the boarding / alighting number calculation means 27 at predetermined intervals.

Stay number calculation unit 29, based on the special call number of passengers _{n fAL + (t), n} fAL- (t) and the special call alighting persons _{n fAUL + (t), n} fAUL- (t), equation (11), formula ( In accordance with 12), the total number of special passengers N _fAL (t) and the total number of special passengers N _fAUL (t) are calculated. Then, based on the total special passengers _{N FAL} (t) and the total special alighting persons _{N Faul} (t), stored in accordance with the equation (13), obtains a special stay number _N fA (t), the time-series data management unit 28 To do. The normal number of visitors N _f (t) can be calculated from the equation (3) and stored in the time series data management means 28.

The waiting customer occurrence number prediction means 30 reads the special waiting user occurrence number prediction function stored in the prediction function learning means 31. Further, based on the special call rising passenger number n _{fAL +} (t) and the special staying person number N _fA (t) at the current time t stored in the time series data management means 28, a special waiting number occurrence function prediction function In accordance with the equation (14), the predicted number of special customers P (hat) _{fA +} (t) in a desired period from the present is predicted. Further, based on the special call descending passenger number n _fAL− (t) and the special staying person number N _fA (t), the special waiting time occurrence in a desired period from the present time according to the formula (15) as a special waiting time occurrence number prediction function The number prediction number P (hat) _fA- (t) is predicted. Note that the estimated number of waiting customers P (hat) _{f +} (t) and P (hat) _f− (t) of a person staying using an elevator by a general call is a waiting time occurrence prediction function. Prediction can be made by equations (4) and (5).

The car allocation floor determining means 32 is the estimated number of special waiting passengers obtained from the waiting occurrence number predicting means 30 when there is no evacuation driving command from the evacuation driving command device 25 and no call is assigned to the car 3. Based on the above, the allocation floor for waiting is determined.
It should be noted that the number of predicted waiting times for a specific special call that should be prioritized, such as the floor with the largest number of waiting calls for wheelchair calls or the floor with the largest number of waiting calls for VIP calls, is determined. The largest floor may be determined as the car standby floor.
On the other hand, when the car allocation floor determination unit 32 receives the evacuation operation command, the car allocation floor of the elevator that uses the floor with the highest predicted number of wheelchair call waiting customer occurrences obtained from the waiting occurrence number prediction unit 30 during evacuation. Determine as.

  The prediction function learning unit 31 learns the parameters of the special waiting customer occurrence number prediction function by the same method as the prediction function learning unit 17 of the first embodiment.

Such an elevator traffic demand forecasting device manages the number of passengers getting on and off, and the number of visitors related to general calls and special calls separately, and also predicts the number of waiting passengers for general calls and special calls. Traffic demand can be predicted by distinguishing passengers.
In addition, there is an effect that a floor with a high possibility that a passenger of a special call type that a building manager wants to prioritize is likely to be a car standby floor on the basis of the special call passenger traffic demand prediction.

  In addition, when a disaster occurs, the floor to which the car is dispatched is determined based on the number of waiting passengers related to the wheelchair call. can do.

  In addition, in the elevator used at the time of evacuation, there is an effect that the car can be preferentially dispatched to the floor on which a vulnerable person such as a wheelchair gets on. Although wheelchair users are described as examples of vulnerable people in disasters, elderly people, physically handicapped persons, etc. can be handled as necessary.

Embodiment 3 FIG.
FIG. 8 is a schematic diagram of an elevator provided with an elevator traffic demand prediction apparatus according to Embodiment 3 of the present invention. FIG. 9 is a functional block diagram of the elevator traffic demand prediction apparatus according to the third embodiment.
As shown in FIG. 8, the elevator 40 in which the elevator traffic demand prediction device 41 according to the third embodiment is arranged has an RFID reader 42 in the car 3 instead of the scale 8 of the elevator 1 according to the first embodiment. What is being done is different. Further, the elevator traffic demand prediction device 41 is managed in units of individuals, unlike the elevator traffic demand prediction device 11 of the first embodiment. The same parts as those in the first embodiment are denoted by the same reference numerals and the description thereof is omitted.

The RFID reader 42 is installed inside the car 3 and reads personal identification information (hereinafter referred to as personal ID) of the passenger from a portable information storage medium 43 possessed by the passenger. The RFID reader 42 is connected to the elevator traffic demand prediction device 41 via the communication line 9.
The elevator control device 45 controls the car driving device 6 to control the running / stopping of the car 3, and after the car 3 stops, the door is opened during passenger boarding / exiting and the door is closed after boarding / exiting is completed.

Next, the elevator traffic demand prediction apparatus 41 will be described with reference to FIG.
The personal authentication / attribute detection means 46 receives the personal ID from the RFID reader 42 and the car running / stop and door opening / closing signals from the elevator control device 45. Then, the personal authentication / attribute detection means 46 reads the personal IDs of all the passengers from the RFID reader 42 when the door is opened and boarding / exiting is completed and the door is closed, and is included in the boarding registration data. When no personal ID is read, it is determined that an individual related to the newly read personal ID has boarded and registered in the boarding registration data. Conversely, when the read personal ID is not in the boarding registration data, It is determined that the person related to the personal ID disappeared from the boarding registration data has got off.
The personal authentication / attribute detection means 46 includes the determined time, the personal ID of the person who gets on the board together with the boarding, the floor of the boarding person, the personal ID of the person who gets off together with the getting off and the floor of the boarding board. The personal getting-on / off data is transmitted to the stay floor detecting means 47 and the personal time-series data managing means 48.

The stay floor detecting means 47 obtains the time staying on the floor for each individual based on the personal getting-on / off data from the time of getting off and the time of getting on, and transmits it to the personal time-series data managing means 48.
The personal time series data management means 48 acquires the personal getting-on / off data from the personal authentication / attribute detecting means 46 and the staying time from the stay floor detecting means 47, and updates and stores them as the getting-on / off time series data of each individual.

The waiting customer occurrence number predicting means 49 reads the waiting function occurrence prediction function (16) h _j or (17) v _j from the prediction function learning means 50. Then, the use probabilities R (hat) _{jf +} (t) and R (hat) _jf− (t) are _obtained according to the equations (16) and (17) based on the personal getting-on / off data and the stay time. R (hat) _{jf +} (t) is a use probability that passenger j gets on the elevator that rises in a predetermined period from the current time t toward the future, and R (hat) _jf− (t). Is a use probability that the passenger j gets on the elevator that descends in the predetermined period from the current time t to the future from the stay floor f. u _{jf +} (t) is an elevator utilization rate in which passenger j gets on the rising elevator in the time zone including the past time t from the stay floor f, and u _jf− (t) is a time zone including the past time t. It is the elevator utilization rate that passenger j got on from the staying floor f to the descending elevator. w _jf (t) is the staying time at the floor f where the passenger j is currently staying.

Typical examples of the personal waiting time occurrence prediction functions h _j and v _j are shown in Expression (18) and Expression (19). Here, α _j , β _j , γ _j and η _j are coefficients for passenger j, and k and l are coefficients common to all passengers.
It should be noted that the individual waiting customers occurrence prediction function _h j, as _{v j} past the elevator utilization _{u jf + (t), u} jf- (t) only, or may be used only stay time _w jf (t).

  The waiting customer occurrence number predicting means 49 performs the traffic demand prediction for each floor direction by adding the use probabilities of all the target individuals.

The prediction function learning means 50 learns parameters such as coefficients of the personal waiting time occurrence prediction functions h _j and v _j from the personal getting-on / off data, the stay time, and the elevator usage rate. As this learning method, for example, when the personal waiting time occurrence prediction functions h _j and v _j are linear functions, α _j , β _j , γ _j and η _j can be learned by the least square method. Further, even when the waiting time occurrence prediction functions h _j and v _j are in other function formats, parameters such as coefficients can be learned by a technique such as a neural network.

  Such an elevator traffic demand prediction device predicts elevator traffic demand reflecting each person's behavioral characteristics in order to predict the occurrence of waiting on the basis of each person's getting-on / off time series data and each person's stay floor. be able to.

  In addition, since the car is dispatched in consideration of the individual behavior pattern, it is possible to further shorten the waiting time of the person who newly calls the car.

  In addition, although the example which detects personal ID superimposed on the radio wave which RFID transmits is given, it is personal using a portable wireless terminal such as a mobile phone or a wireless LAN or biometric authentication based on biological information such as a fingerprint, an iris, and a face. The ID may be detected.

  In addition, although the example of detecting the stay floor of each individual based on personal boarding / exiting data has been given, the detection of the stay floor on the floor where there is an entrance to the outside of the building can be detected when entering or leaving the building, such as an IC card You may correct | amend using each person's entrance / exit information obtained from personal authentication apparatuses, such as authentication by a card | curd, authentication by portable radio | wireless terminals, such as RFID, biometric authentication by biometric information, such as a fingerprint, an iris, and a face. That is, an individual who is determined not to be in the building from the entrance / exit information is corrected as having no stay floor.

  In addition, an RFID reader may be provided in the vicinity of an escalator that is a lifting means other than an elevator or a stair landing, so that the stay floor of each individual may be accurately detected even when the lifting means other than the elevator is used.

  In addition, the floor information of each individual stay at each time point can be used for various building management systems. For example, in the security system, it is possible to detect unauthorized access to an area of a certain floor or detect an illegal behavior pattern deviating from a general behavior pattern from stay floor information in a building.

  In addition, in a management system for commercial facilities such as department stores, it is possible to analyze customer purchase behavior by detecting the movement route between floors of each individual and create guidance adapted to each individual from the information on the floors in the building. .

Further, in the presence management system, it is possible to detect the current stay floor of a certain individual from the stay floor information in the building, and to automatically reflect each individual's destination in the scheduler.
In addition, in the housing complex management system, it is possible to detect residents who have not moved from the residence floor for a certain period of time based on the stay floor time series data of residents living alone, etc., and check the health status of the residents. it can.

Embodiment 4 FIG.
FIG. 10 is a schematic diagram of an elevator provided with an elevator traffic demand prediction apparatus according to Embodiment 4 of the present invention. FIG. 11 is a functional block diagram of the elevator traffic demand prediction apparatus according to the fourth embodiment.
The elevator 60 provided with the elevator traffic demand prediction device 61 according to the fourth embodiment is different from the elevator 40 according to the third embodiment in that a fire alarm 24 is further provided. The fire alarm 24 and the evacuation operation command device 25 are the same as those described in connection with the second embodiment, and a description thereof will be omitted.
As shown in FIG. 10, the elevator traffic demand prediction device 61 of the fourth embodiment is different in that an evacuation operation command is input to the car allocation floor determining means 51 of the elevator traffic demand prediction device 41 of the third embodiment. Yes. The elevator control device 62 includes an elevator traffic demand prediction device 61 and an evacuation operation command device 25.

Next, the elevator traffic demand prediction apparatus 61 will be described with reference to FIG.
The special passenger means a passenger registered as an elderly person, a wheelchair user, a VIP, or the like. A special attribute for identifying a special passenger is assigned to the personal ID. Further, passengers such as wheelchair users and elderly persons are given the disaster vulnerable person attribute as disaster vulnerable persons.
Next, about each component of the elevator traffic demand prediction apparatus 61, the part regarding a special passenger and a disaster weak person is demonstrated. Since general users are the same as those in the third embodiment, description thereof is omitted.

The personal authentication / attribute detection means 46 generates special passenger boarding / exiting data related to boarding / exiting special passengers based on the detection of special attributes from the personal ID. Further, the personal authentication / attribute detection means 46 generates disaster vulnerable person getting-on / off data related to getting on and off the disaster vulnerable person based on the detection of the disaster vulnerable person attribute from the personal ID.
The stay floor detection means 47 obtains the stay time of the special passenger on the stay floor from the special passenger boarding / exiting data and the stay time of the vulnerable person on the stay floor from the disaster weak person boarding / exiting data.
The personal time-series data management means 48 updates each time-series data using the special passenger boarding / alighting data, the disaster vulnerable person boarding / exiting data, the staying time of the special passenger and the disaster vulnerable person on the stay floor.
The special waiting customer occurrence number predicting means 64 calculates the use probability based on the special passenger boarding / alighting data and the staying time among the personal boarding / alighting data and the staying time obtained from the personal time-series data management means 48. ). And the use probability regarding a special passenger is calculated over the whole special passenger, and the special passenger waiting number of generation according to each floor direction is estimated.
Also, the special waiting time occurrence number predicting means 64 calculates the use probability based on the disaster vulnerable person getting on / off data and the staying time among the personal getting on / off data and the staying time obtained from the personal time-series data managing means 48. Calculate according to (19). And the use probability regarding a disaster weak person is calculated over the whole disaster weak person, and the disaster weak person waiting number of occurrence generation according to each floor direction is estimated.

When no evacuation operation command is received, the car dispatch floor determining unit 63 determines a floor with a large number of special passenger waiting occurrences obtained from the special passenger occurrence number predicting unit 64 as a preferential car dispatch floor.
On the other hand, when the evacuation operation command is received, the car dispatching level determining unit 63 determines the floor having a large number of disaster vulnerable person waiting generations obtained from the special waiting number occurrence number predicting unit 64 as the preferential dispatching level of the car. .
When the stay floor detection means 47 receives the evacuation operation command, the stay floor detection means 47 determines that the individual staying on the floor predetermined as the evacuation floor has completed the evacuation.

  In addition, in an elevator with a plurality of cars, when all the cars are not wheelchair use number machines, the wheelchair use number machine is preferentially placed on the floor having a large number of the waiting times for disaster vulnerable persons obtained from the special waiting number occurrence number prediction means 64. Wait.

Such an elevator traffic demand prediction device predicts the number of waiting customers according to personal attributes such as wheelchairs, elderly people, VIPs, etc., and therefore can provide a driving service adapted to the individual.
In addition, in elevators used during evacuation, cars can be dispatched according to the floor of stay of disaster vulnerable persons such as wheelchair users and senior citizens who need priority rescue, so rescue vulnerable persons with priority. Can evacuate efficiently.

  Moreover, since the stay floor detection means 47 can check the person who got off at the evacuation floor as an evacuation completion person, an evacuation delay person can be detected at an early stage.

  1, 20, 40, 60 Elevator, 2 hoistway, 3 car, 4 call button, 5 destination floor registration device, 6 car drive device, 7, 45, 62 elevator control device, 8 scale, 9 communication line, 11, 21 , 41, 61 Elevator traffic demand forecasting device, 12, 26 Number of passengers measuring means, 13, 27 Number of passengers getting on and off means, 14, 28 Time series data management means, 15, 29 Number of staying person calculating means, 16, 30, 49 Waiting Customer generation number prediction means, 17, 31, 50 Prediction function learning means, 18, 32, 51, 63 Car dispatch floor determination means, 22 Special call button, 23 Special destination floor registration device, 24 Fire alarm, 25 Evacuation operation command Device, 42 RFID reader, 43 portable information storage medium, 46 personal authentication / attribute detection means, 47 stay floor detection means, 48 personal time series data management means 64 special waiting passengers occurrence number prediction means.

Claims (5)

In the elevator traffic demand forecasting device that predicts the number of waiting passengers at the landing on each floor of the elevator,
Personal authentication / attribute detection means for performing personal authentication and generating personal boarding / exiting data related to personal boarding / exiting based on detection of personal attributes;
A stay floor detecting means for detecting a stay time of each person on the floor from the personal getting-on / off data;
Waiting occurrence that determines the use probability that the individual uses the elevator from the personal getting-on / off data and the stay time, and predicts the number of waiting occurrences on each floor according to the rise and fall of the elevator from the use probability of all elevator users Number prediction means,
An elevator traffic demand prediction apparatus characterized by comprising: The personal attribute includes a special passenger attribute that identifies an individual registered as a special passenger,
The personal authentication / attribute detection means generates special passenger boarding / exiting data related to boarding / exiting the special passenger based on the detection of the special passenger attribute,
The stay floor detection means calculates a stay time on the stay floor of the special passenger from the special passenger boarding / alighting data,
For each floor, obtain the use probability that the special passenger will use the elevator from the special passenger boarding / alighting data and the stay time, and the number of waiting occurrences in each floor according to the rise and fall of the elevator from the use probability of all special passengers A means for predicting the number of special waiting customers that predicts
The elevator traffic demand prediction apparatus according to claim 1, wherein   The car allocation floor determination means for determining the allocation floor of the car that waits for the floor with the smallest evaluation function of the estimated number of waiting passengers and the condition factor required for allocation of vehicles. The elevator traffic demand prediction apparatus described in 2. The special attribute includes a disaster vulnerable person attribute that identifies a disaster vulnerable person,
The personal authentication / attribute detection means generates disaster weak person getting on / off data related to getting on and off the disaster weak person based on detection of the disaster weak person attribute,
The stay floor detection means detects the stay floor of the vulnerable person from the disaster vulnerable person getting on and off data,
The above-mentioned special passenger occurrence prediction means predicts the number of waiting passengers related to disaster vulnerable persons by rising and falling elevators from the disaster vulnerable person stay floor,
The car dispatch floor determining means preferentially determines a floor with a large number of waiting passengers related to a predicted disaster vulnerable person as a car dispatch floor when an evacuation operation command accompanying the occurrence of a disaster is input. The elevator traffic demand prediction apparatus according to claim 3.   3. The stay floor detection means, when an evacuation operation command accompanying the occurrence of a disaster is input, judges that a person who has got off the floor previously determined as an evacuation floor is an evacuation completion person. The elevator traffic demand prediction apparatus as described in any one of thru | or 4.

Images