KR20230063949A - Apparatus and method for predicting congestion by subway car using deep learning - Google Patents

Abstract

A method for predicting congestion for each subway car using deep learning according to the present invention includes the steps of: receiving, by a data processing part, a message requesting congestion in a subway station and congestion for each subway car from a user device; selecting, by the data processing part, a plurality of operation times from the past of a reference point, the point in time when the message is received, and generating an input vector string including a plurality of input vectors corresponding to each of the selected operation times; performing, by a prediction part, calculations on the plurality of input vectors through a prediction network to derive the congestion for each subway car in the subway station at the next operation time of a certain number of times from the reference point.

Description

Apparatus and method for predicting congestion by subway car using deep learning}

The present invention relates to congestion prediction technology, and more particularly, to an apparatus and method for predicting congestion by subway compartment using deep learning.

The subway is composed of several connected rooms. In addition, the boarding and alighting tag is not performed when boarding the train, but the boarding and alighting tag is performed at the ticket gate when entering and exiting a subway station. This is a major feature that distinguishes it from buses, and due to these structural characteristics, the density of subways is distributed differently for each room. Also, in the same context, there is a limitation in the method of identifying or predicting the congestion level of each room.

Recently, attempts have been made to estimate the congestion of the subway. However, estimating the degree of congestion at that time has limitations in terms of service. Potential subway users in the future cannot identify congestion information in advance and select a boarding car, but can accept congestion information only when they arrive at the subway platform and when the train they are about to board arrives. As such, there is a lack of an appropriate methodology for predicting congestion at a future point in time.

Korean Patent Publication No. 2015-0018975 (published on February 25, 2015)

An object of the present invention is to provide an apparatus and method for predicting the degree of congestion for each subway compartment using deep learning.

In order to achieve the above object, a method for predicting congestion by subway compartment according to a preferred embodiment of the present invention includes the steps of receiving, by a data processing unit, a message requesting the congestion of a subway station and subway car from a user device; The data processing unit selects a plurality of dispatchers from the past of a reference point in time when the message is received, and generates an input vector sequence including a plurality of input vectors corresponding to each selected dispatcher; and deriving the degree of congestion of the subway car of the subway station at the time of the next dispatch of a predetermined number of times from the reference point by performing an operation on the plurality of input vectors through

The step of deriving the congestion degree of the subway station and the subway car includes the step of receiving a plurality of input vectors arranged in chronological order corresponding to a plurality of stages of the prediction network for each of the plurality of stages in the input layer of the prediction network; A plurality of hidden cells of the hidden layer of the prediction network perform an operation in which weights are applied to the state value of the previous stage and the input vector, which is the input value of the current stage, to calculate the state value of the current stage, and then calculate the state value of the next stage After the state value of the current stage is calculated for the last hidden cell among the plurality of hidden cells, a weight is applied to the calculated state value of the current stage, so that the subway of the subway station at the next dispatch of the predetermined number of times and calculating an output vector representing the degree of congestion of the cell.

In the method, before the step of receiving the message requesting the degree of congestion, the learning unit prepares learning data including a learning input vector column including a plurality of learning input vectors and an actual vector for learning that is a label corresponding to the learning input vector column. And, calculating an output vector by performing an operation on the plurality of input vectors for learning by a prediction network having unlearned weights, and loss, which is the difference between the output vector and the actual vector for learning, through a loss function Calculating , and performing optimization of modifying the weights of the prediction network so that the loss is minimized.

The step of generating the input vector sequence including the plurality of input vectors includes the step of selecting, by the data processing unit, a train arrangement having a predetermined multiple interval of the train dispatch interval of the subway station from a past time point based on the reference time point; and Collecting event information digitizing the dispatch time corresponding to the selected dispatch, the degree of congestion of the subway car corresponding to the dispatch time, meteorological information indicating weather conditions, and events contributing to whether or not people are crowded; and generating an input vector sequence by a processing unit by arranging a plurality of input vectors including the dispatch time, the degree of congestion of the subway car, the weather information, and the event information in order of dispatch time.

Here, the congestion degree of the subway car is characterized in that it is collected through the weight of the subway car.

In order to achieve the above object, an apparatus for predicting congestion by subway compartment according to a preferred embodiment of the present invention receives a message requesting the congestion of a subway station and a subway car from a user device, and receives the message A data processing unit that selects a plurality of dispatchers from the past of the reference time point, which is a starting point, and generates an input vector sequence including a plurality of input vectors corresponding to each selected dispatcher; and an operation for the plurality of input vectors through a prediction network and a predictor for deriving the degree of congestion of the subway car of the subway station at the time of the next dispatch of a predetermined number of times from the reference point in time.

In the prediction network, weights are applied to an input layer that receives a plurality of input vectors sorted in time order corresponding to a plurality of stages for each of the plurality of stages, and a state value of the previous stage and an input vector that is the input value of the current stage. After calculating the state value of the current stage by performing an operation to calculate the state value of the current stage, including a plurality of hidden cells that transfer the calculated state value to the next stage, wherein the last hidden cell among the plurality of hidden cells calculates the state value of the current stage Then, a hidden layer that calculates an output vector representing the degree of congestion of the subway car of the subway station at the time of the next dispatch of the predetermined number of times by applying a weight to the calculated state value of the current stage, and an output layer that outputs the output vector do.

The apparatus prepares learning data including a learning input vector column including a plurality of learning input vector columns and a learning actual vector that is a label corresponding to the learning input vector column, and a prediction network having unlearned weights is configured to generate the plurality of learning input vector columns. Calculating an output vector by performing an operation on an input vector for learning, calculating a loss, which is the difference between the output vector and the actual vector for learning, through a loss function, and modifying the weight of the prediction network so that the loss is minimized A learning unit that performs optimization is further included.

The data processing unit selects a train with an interval of a predetermined multiple of the dispatch interval of the subway station from a past time point based on the reference time point, and selects an dispatch time corresponding to the selected dispatch time, congestion of the subway car corresponding to the dispatch time, Weather information indicating weather conditions and event information digitizing events contributing to whether or not people are crowded are collected, and each input vector includes the dispatch time, congestion of the subway car, the weather information, and the event information. It is characterized in that an input vector sequence is generated by arranging in the order of allocation time.

Here, the congestion degree of the subway car is characterized in that it is collected through the weight of the subway car.

According to the present invention, it is possible to predict the degree of congestion for each train of a specific subway station and for each subway car. Accordingly, the user can use the subway avoiding congested locations or times.

1 is a diagram for explaining a system for predicting congestion by subway compartment using deep learning according to an embodiment of the present invention.
2 is a block diagram for explaining the configuration of a prediction server according to an embodiment of the present invention.
3 is a block diagram for explaining the detailed configuration of a prediction server according to an embodiment of the present invention.
4 is a diagram for explaining the configuration of a prediction network for predicting congestion according to an embodiment of the present invention.
5 is a diagram for explaining unit operations of a prediction network for predicting congestion according to an embodiment of the present invention.
6 is a flowchart illustrating a method of generating a prediction network for predicting the congestion of a subway car through learning according to an embodiment of the present invention.
7 will explain a method for predicting the degree of congestion for each subway compartment using deep learning according to an embodiment of the present invention.

Prior to the detailed description of the present invention, the terms or words used in this specification and claims described below should not be construed as being limited to a common or dictionary meaning, and the inventors should use their own invention in the best way. It should be interpreted as a meaning and concept corresponding to the technical idea of the present invention based on the principle that it can be properly defined as a concept of a term for explanation. Therefore, the embodiments described in this specification and the configurations shown in the drawings are only the most preferred embodiments of the present invention, and do not represent all of the technical ideas of the present invention, so various equivalents that can replace them at the time of the present application. It should be understood that there may be water and variations.

In the description and claims of embodiments of the present invention, a "network" is defined as one or more data links that enable the transfer of electronic data between computer systems and/or modules. When information is transmitted or provided to a computer system over a network or over a communication connection of any kind or kind, i.e., wired, wireless, or a combination of wired and wireless, the connection may be understood as a computer-readable medium. . Computer readable instructions include, for example, instructions and data that cause a general purpose or special purpose computer system to perform a particular function or group of functions. Computer executable instructions may be, for example, binary, intermediate format instructions, such as assembly language, or even source code. Moreover, the present invention relates to personal computers, laptop computers, handheld devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, mobile telephones, cellular phones, and mobile devices. In a network computing environment having various types of computer system configurations including communication terminals, smart phones, PDAs, pagers, etc., the above computer systems may be applied.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. At this time, it should be noted that the same components in the accompanying drawings are indicated by the same reference numerals as much as possible. In addition, detailed descriptions of well-known functions and configurations that may obscure the gist of the present invention will be omitted. For the same reason, some components in the accompanying drawings are exaggerated, omitted, or schematically illustrated, and the size of each component does not entirely reflect the actual size.

First, a system for predicting congestion for each subway compartment using deep learning according to an embodiment of the present invention will be described. 1 is a diagram for explaining a system for predicting congestion by subway compartment using deep learning according to an embodiment of the present invention. Referring to FIG. 1, a system for predicting congestion by subway compartment using deep learning according to an embodiment of the present invention (hereinafter, abbreviated as 'prediction system') includes a prediction server 10, a user device 20 and information It includes a providing server 30.

The prediction server 10, the user device 20, and the information providing server 30 may communicate with each other through the network NW. For example, the network NW may include a wireless network such as wireless LAN (WLAN), Wi-Fi, Wibro, Wimax, High Speed Downlink Packet Access (HSDPA), Depending on the system implementation method, wired networks such as Ethernet, xDSL (ADSL, VDSL), HFC (Hybrid Fiber Coaxial Cable), FTTC (Fiber to The Curb), FTTH (Fiber To The Home) may be included. In addition, the network (NW) of the present invention may include, for example, a network consisting of a plurality of access networks and a core network connecting them. Also, the network NW according to the present invention may include the Internet. The Internet means a common open network, i.e., a public network, in which information is exchanged according to the TCP/IP protocol. Through this network (NW), the prediction server 10, the user device 20, and the information providing server 30 interwork with each other to form a prediction system according to the present invention.

The prediction server 10 is basically for predicting the congestion of the subway car according to the request of the user device 20 . To this end, the prediction server 10 collects past data related to the corresponding subway car from the information providing server 30, and uses the collected data to learn the prediction network (EN) to determine the location of the corresponding subway car. predict congestion. The prediction server 10 is an entity existing on a network and serves as a web server, a database server, and an application server.

Although the information providing server 30 is shown as one, it may be a plurality of servers providing corresponding data. For example, the information providing server 30 may be a server operated by the Subway Corporation, a server operated by the Korea Meteorological Administration, and a server operated by the Korea Centers for Disease Control and Prevention.

The user device 20 is a device used by a user, and the user device 20 accesses the prediction server 10, specifies the subway station and subway car to the prediction server 10, requests the degree of congestion, and the prediction server 10 It is possible to receive a predicted congestion degree of a subway car allocated to a corresponding subway station. The user device 20 may be a representative example of a smartphone, but is not limited thereto, and may be applied to various terminals such as all information communication devices, multimedia terminals, wired terminals, fixed terminals, and IP (Internet Protocol) terminals. For example, the terminal is a mobile terminal having various mobile communication specifications such as a mobile phone, a portable multimedia player (PMP), a mobile internet device (MID), a tablet PC, a phablet PC, and an information communication device, or a laptop computer, a personal computer, etc. may be a fixed terminal of

Next, the configuration of the prediction server 10 according to an embodiment of the present invention will be described in more detail. 2 is a block diagram for explaining the configuration of a prediction server according to an embodiment of the present invention. 3 is a block diagram for explaining the detailed configuration of a prediction server according to an embodiment of the present invention. 4 is a diagram for explaining the configuration of a prediction network for predicting congestion according to an embodiment of the present invention. 5 is a diagram for explaining unit operations of a prediction network for predicting congestion according to an embodiment of the present invention.

First, referring to FIG. 2 , the prediction server 10 includes a communication module 11, a storage module 12 and a control module 13.

The communication module 11 is for communicating with the user device 20 and the information providing server 30 through a network. The communication module 11 may include a modem that modulates a transmitted signal and demodulates a received signal in order to transmit and receive data through a network. The communication module 11 may transmit data received from the control module 13 through a network. Also, the communication module 11 may transmit data received through a network to the control module 13 .

The storage module 12 serves to store programs and data necessary for the operation of the prediction server 10, and may be divided into a program area and a data area. The program area may store a program for controlling the overall operation of the prediction server 10, an operating system (OS) for booting the prediction server 10, and applications for prediction. The data area is an area where data generated according to the operation of the prediction server 10 and data necessary for the operation of the prediction server 10 are stored. Various kinds of data stored in the storage module 120 can be deleted, changed, or added according to a user's manipulation.

The control module 13 may control the overall operation of the prediction server 10 and signal flow between internal blocks of the prediction server 10, and may perform a data processing function of processing data. The control module 13 may be a central processing unit (CPU), a graphic processing unit (GPU), a digital signal processor (DSP), or the like. Referring to FIG. 3 , the control module 13 includes a learning unit 100 , a data processing unit 200 and a prediction unit 300 .

The learning unit 100 is for generating a predictive network (EN) through learning using learning data. Different learning data is prepared for each subway station, subway car, and time point at which congestion is predicted, and the learning unit 100 generates a plurality of prediction networks (EN) using the different learning data. A prediction network (EN) according to an embodiment of the present invention is a deep neural network (DNN) and may be formed as a recurrent neural network. For example, the prediction network (EN) may include a recurrent neural network (RNN), long short-term memory models (LTSM), and a gated recurrent unit (GRU). In an embodiment of the present invention, the predictive network (EN) is described as being formed of a recurrent neural network (RNN), but is not limited thereto, and includes long short-term memory models (LTSM), gated recurrent units (GRUs), and the like. It can also be formed with other recurrent neural networks.

Referring to FIG. 4 , the data processing unit 200 is for generating a plurality of input vectors X1 to Xn. The data processing unit 200 selects a predetermined multiple interval (d×inter) of the allocation interval (inter) from among a plurality of allocations made from a predetermined past time point (P) based on the reference point (present), for each of a predetermined number of allocations. It collects the dispatch time of the selected train, the congestion level of the corresponding subway car of the assigned subway, weather information and event information of the dispatch time when the dispatch was made. For example, it is assumed that the prediction network (EN) is composed of five stages (S5, n = 5). Also, it is assumed that the multiple interval (d×inter) of the vehicle allocation interval (inter) is '2×inter'. Then, the data processing unit 200 selects 5 even-numbered vehicles out of 10 times from the past time point P to the reference time point B. In addition, the data processing unit 200 transmits the dispatch time of the selected train from the information providing server 30 through the communication module 11, the degree of congestion of the corresponding subway compartment of the assigned subway, weather information and event information of the dispatch time at which the dispatch was made. collect At this time, the data processing unit 200 may collect the dispatch time of the selected dispatch from the information providing server 30 . Here, the dispatch time includes the month, day, day and time of the dispatch. A load detection sensor for detecting weight is installed for each compartment of the subway, and the data processing unit 200 receives the weight detected in the corresponding subway car from the information providing server 30 and normalizes the received weight to determine the congestion level of the subway car. generate In addition, the data processing unit 200 may collect weather information of the dispatch time at which the selected dispatch was made from the information providing server 30 . Such meteorological information may include temperature, humidity, wind speed, weather (whether it rains or snows), and the like. Further, the data processing unit 200 may collect event information of the dispatch time at which the selected dispatch was made from the information providing server 30 . The event information includes information on various events that may affect the number of floating population. For example, the number of infected persons may be in a situation where an epidemic such as a pandemic situation is prevalent. As described above, after collecting the dispatch time, congestion, weather information, and event information, the data processing unit 200 includes the dispatch time, congestion of the subway car, weather information, and event information corresponding to the selected plurality of dispatches. A plurality of input vectors (X1 to Xn) are generated, and the plurality of generated input vectors (X1 to Xn) are sorted according to the order of allocation time, and an input vector sequence (Xt: X1 to Xn) including the plurality of input vectors (X1 to Xn) is arranged. ) is generated, and the generated input vector sequence (Xt: X1 to Xn) is provided to the prediction unit 300.

The prediction unit 300 is for predicting the degree of congestion of the corresponding subway car at the prediction time point F at which trains are dispatched after a predetermined number of times from the reference time point B using the prediction network EN. The prediction unit 300 inputs input vector strings (Xt: X1 to Xn) including a plurality of input vectors (X1 to Xn) to the prediction network (EN), and inputs the predicted time point (F), which is the operation result of the prediction network (EN). ) calculates the output vector (Y) representing the degree of congestion of the corresponding subway car. In other words, any one predictive network (EN) is an input vector sequence corresponding to an arrangement consisting of a predetermined multiple interval (d × inter) of an arrangement interval (inter) from a past time point (P) based on a reference time point (B). For (Xt: X1 to Xn), one or more operations to which weights are applied in the order of a plurality of stages (S1 to Sn) are performed to correspond to the prediction time point (F) at which the allocation after a predetermined number of times from the reference time point (B) is made. An output vector (Y) representing the degree of congestion in the subway car is calculated.

Each of the plurality of prediction networks (EN) is generated through learning for each subway station, subway car, and prediction point (F). The prediction network EN is composed of a plurality of stages S1 to Sn, and each stage includes an input layer (IL), a hidden layer (HL), and an output layer (OL). The length n of the stages S1 to Sn is preset.

The input layer IL includes a plurality of input buffers corresponding to each of the plurality of input vectors X1 to Xn, and inputs the plurality of input vectors X1 to Xn to the hidden layer HL. That is, a plurality of input vectors X1 to Xn arranged in chronological order are input to the input layer IL.

The hidden layer HL includes a plurality of hidden cells HC corresponding to a multiple of the number n of stages. That is, although the hidden layer HL is illustrated as one layer, it may be formed of two or more layers. The plurality of hidden cells HC calculates the output vector Y by performing one or more operations to which weights are applied to the plurality of input vectors X1 to Xn in a circular manner.

The output layer OL outputs the output vector Y calculated by the hidden cell HCn of the last stage Sn. The output vector (Y) represents the degree of congestion of the corresponding subway car at the predicted time point (F) at which trains are dispatched after a predetermined number of times from the reference time point (B).

Meanwhile, referring to FIG. 5 , the hidden cell HC is composed of one or more operations to which weights (Weight, Wx, Wh, and Wy) are applied. Here, the operation means an operation to which an activation function is applied. Activation functions may include sigmoid, hyperbolic tangent (tanh), exponential linear unit (ELU), rectified linear unit (ReLU), leaky ReLU, Maxout, Minout, and Softmax. In addition, in one hidden cell (HC), the weight is the input weight Wx applied to the input vector (Xt), the state weight Wh applied to the state value Ht-1 of the previous stage, and the output applied to the output value Yt Include the weight Wy. For example, the following Equation 1 may be exemplified as an operation to which the weight applied to the hidden cell HC is applied.

Here, b is a threshold or bias. In particular, the tanh function and the ReLU function can be changed to other activation functions. Referring to Equation 1 and FIG. 5, each of the plurality of hidden cells HCt has a state value Ht- of the previous stage St-1 calculated by the hidden cell HCt-1 of the previous stage St-1. 1) and the state value (Ht) and output value (Yt) of the current stage (St) by performing an operation in which weights (W: Wh, Wx, Wy) are applied to the input value (Xt) of the own stage (St) can be calculated.

The input layer IL of the prediction network EN converts each of the plurality of input vectors X1 to Xn arranged in time order into a corresponding hidden cell of the hidden layer HL, that is, the first to nth hidden cells HC1 to HCn). Then, each of the first to n-th hidden cells HC1 to HCn of the hidden layer HL has a state value of the previous stage St-1 calculated by the hidden cell HCt-1 of the previous stage St-1 ( The state value (Ht) of the current stage (St) is calculated by performing an operation in which weights (state weight and input weight) are applied to the input value (Xt) of the stage (St) of Ht-1) and itself (HCt). After that, the calculated state value (Ht) is transferred to the hidden cell (HCt+1) of the next stage (St+1). For example, the first hidden cell HC1 applies the state weight Wh to the initial state value H0 and applies the input weight Wx to the first input vector X1 to obtain the state of the first stage S1. Calculate the value H1. In the case of the first hidden pixel HC1, since there is no previous stage, the initial state value H0 is used. Subsequently, the second hidden cell HC2 applies the state weight Wh to the state value H1 of the first stage S1, which is the previous stage, and applies the input weight Wx to the second input vector X2. Then, the state value H2 of the second stage S2 is calculated. The third hidden cell HC3 applies the state weight Wh to the state value H2 of the previous stage, the second stage S2, and applies the input weight Wx to the third input vector X3. This is an equation for calculating the state value H3 of the third stage S3. In this way, the plurality of hidden cells (HCt) perform an operation in which the state and input weights (Wh, Wx) are applied to the state value (Ht-1) of the previous stage and the input vector (Xt), which is the input value of the current stage. Then, the state value (Ht) of the current stage is calculated, and the calculated state value (Ht) is transferred to the next stage (St+1).

During the execution of such an operation, the state weight (Wh) is applied to the state value (Hn-1) of the n-1th stage (Sn-1), which is the previous stage, of the n-th hidden cell (HCn), which is the last hidden cell. The state value Hn of the n-th stage Sn, which is the current stage, is calculated by performing an operation to which the input weight Wx is applied to the input vector Xn, which is the input value of the stage Sn. Then, the n-th hidden cell HCn calculates the output vector Y by performing an operation of applying the output weight Wy to the state value Hn of the n-th stage Sn, which is the current stage.

As such, in the prediction network (EN) according to the embodiment of the present invention, the input layer (IL) receives a plurality of input vectors (X1 to Xn) arranged in time order corresponding to each of the plurality of stages (S1 to Sn). It is input for each stage (S1 to Sn) of Then, the plurality of hidden cells HC1 to HCn-1 excluding the last hidden cell HCn of the hidden layer HL are the state values Ht-1 of the previous stage St-1 and the input of the current stage St After calculating the state value (Ht) of the current stage by performing an operation to which the state and input weights (Wx, Wh) are applied to the input vector (Xt), which is a value, the calculated state value is used as the next stage (St+1) forward to Then, the last hidden cell (HCn) has a state and an input weight (Wx, After calculating the state value (Hn) of the current stage (Sn) by performing an operation to which Wh) is applied, the output vector (Y) is obtained by applying the output weight (Wy) to the calculated state value (Hn) of the current stage. yield

The output vector (Y) represents the degree of congestion of the corresponding subway car at the predicted time point (F) at which trains are dispatched after a predetermined number of times from the reference point (B). As such, if the prediction network (EN) calculates the output vector (Y) , The prediction unit 300 derives the congestion level of the corresponding subway car at the prediction time point F according to the output vector Y, and transmits the derived congestion level to the user device 20 through the communication unit 11.

Next, a method for generating a prediction network for predicting congestion in a subway car through deep learning according to an embodiment of the present invention will be described. 6 is a flowchart illustrating a method of generating a prediction network for predicting the congestion of a subway car through learning according to an embodiment of the present invention.

Referring to FIG. 6 , the learning unit 100 initializes the prediction network EN in step S110. This means initializing the weights of the prediction network (EN), that is, input weight Wx, state weight Wh, and output weight Wy.

Then, the learning unit 100 prepares learning data in step S120. The training data includes a learning input vector column including a plurality of learning input vectors and a ground truth for learning that is a label for the learning input vector column. The learning unit 100 may collect learning data using past data from the information providing server 30 through the communication module 11 . A plurality of input vectors for learning are prepared according to the size (number of stages) of the input layer IL of the initialized prediction network EN. For example, if the size (the number of stages) of the input layer IL of the prediction network EN is 5, 5 training input vectors are required according to time order. Accordingly, if the multiple interval (d×inter) of the vehicle allocation interval (inter) is ‘3×inter’, the third of 15 vehicle allocations from the past time point (P) to the reference time point (B) based on the reference time point (B). Select 5 for the dispatch interval. Accordingly, the learning unit 100 collects the dispatch time of the selected train, the congestion level of the corresponding subway car of the dispatched subway, weather information and event information of the dispatch time at which the dispatch was made, and normalizes them to generate an input vector for learning. . In addition, the output vector for learning may be collected according to the number of times the prediction time point (F) is from the reference time point (B). For example, when the predicted time point (F) is the third train dispatch after the reference time point (B), the learning unit 100 is transferred from the information providing server 30 through the communication module 110 after the reference time point (B) of the corresponding subway station. The weight of the subway car in the third dispatch is collected, and a ground truth for learning representing the degree of congestion is generated by normalizing the collected weight.

Subsequently, the learning unit 100 inputs a plurality of input vectors for learning to the prediction network EN including weights that have not been learned, that is, initialized in step S130. Then, the prediction network (EN) calculates an output vector by performing an operation on a plurality of input vectors for learning in step S140. That is, the prediction network (EN) is a learning input vector sequence (Xt: For X1 to Xn), by performing one or more calculations to which weights are applied in the order of a plurality of stages (S1 to Sn), the corresponding subway car at the predicted time point (F) at which the train is assigned a predetermined number of times from the reference point (B). An output vector representing the degree of congestion is calculated.

Then, the learning unit 100 calculates a loss, which is the difference between the output vectors corresponding to the plurality of input vectors for learning and the ground truth for learning, which is a label corresponding to the plurality of input vectors for learning, through the loss function in step S150. and optimizes the weight of the prediction network (EN) so that the loss is minimized.

The learning unit 100 may generate the predictive network (EN) by repeating steps S120 to S150 described above using different learning data until the loss is less than a predetermined target value.

According to the procedure described above, a plurality of prediction networks (EN) are generated for each subway station, subway car, and prediction point (F). Then, a method for predicting the degree of congestion in a subway car using a plurality of predictive networks (EN) will be described. 7 will explain a method for predicting the degree of congestion for each subway compartment using deep learning according to an embodiment of the present invention.

Referring to FIG. 7 , the data processing unit 200 receives a message requesting the congestion level of subway stations and subway cars from the user device 20 through the communication module 11 in step S210. That is, in the message requesting the degree of congestion, the requested subway station and subway car are specified and transmitted.

Then, the data processing unit 200 moves from the past time point (P) to a predetermined multiple interval (d × inter) of the dispatch interval (inter) of the corresponding subway station based on the reference time point (B), which is the time point at which the message is received in step S220. The selected arrangement is selected, and an input vector sequence (Xt: X1 to Xn) including a plurality of input vectors (X1 to Xn) corresponding to each selected arrangement is generated. Step S220 will be described in more detail as follows. In other words, the data processing unit 200 selects a train dispatch made at a predetermined multiple interval (d × inter) of the train dispatch interval (inter) of the corresponding subway station from the past time point (P) based on the reference time point (B). Then, the data processing unit 200 receives the dispatch time corresponding to the previously selected dispatch from the information providing server 30 through the communication module 11, the congestion level of the subway car corresponding to the dispatch time, and the meteorological conditions of the dispatch time. It collects event information that digitizes events that contribute to the clustering of people corresponding to the indicated weather information and the dispatch time. Subsequently, the data processing unit 200 arranges a plurality of input vectors (X1 to Xn), each of which includes dispatch time, subway car congestion, weather information, and event information, in order of dispatch time, and input vector sequence (Xt: X1). to Xn).

Next, the prediction unit 300 inputs an input vector sequence (Xt: X1 to Xn) including a plurality of input vectors (X1 to Xn) to the prediction network (EN) in step S230. That is, a plurality of input vectors (X1 to Xn) arranged in order of dispatch time are input to the prediction network (EN).

Then, in step S240, the prediction network (EN) applies weights in the order of the plurality of stages (S1 to Sn) to the input vector sequence (Xt: X1 to Xn), that is, to the plurality of input vectors (X1 to Xn) One or more operations are performed to calculate an output vector (Y) representing the degree of congestion of the corresponding subway car of the corresponding subway station at the predicted time point (F), which is the time point at which trains are dispatched after a predetermined number of times from the reference time point (B).

This step S240 will be described in more detail with reference to FIGS. 4 and 5. The input layer IL of the prediction network EN transmits a plurality of input vectors X1 to Xn arranged in time order corresponding to the plurality of stages S1 to Sn of the prediction network for each of the plurality of stages S1 to Sn. receive input Then, each of the plurality of hidden cells HC1 to HCn of the hidden layer HL of the prediction network EN has a state value Ht-1 of the previous stage St-1 and an input value of the current stage St After calculating the state value (Ht) of the current stage (St) by performing an operation in which weights (state weight and input weight) are applied to the input vector (Xt), the calculated state value (Ht) is transmitted to the next stage do. At this time, the last hidden cell (HCn) among the plurality of hidden cells (HC1 to HCn) applies a weight (output weight) to the calculated state value (Hn) of the current stage to obtain the An output vector representing the degree of congestion is calculated.

Then, the prediction unit 300 transmits the degree of congestion according to the output vector Y calculated by the prediction network EN to the user device 20 through the communication module 11 in step S250. Then, the user device 20 can display the corresponding congestion level so that the user can view it.

On the other hand, the method according to the embodiment of the present invention described above may be implemented in the form of a program readable by various computer means and recorded on a computer-readable recording medium. Here, the recording medium may include program commands, data files, data structures, etc. alone or in combination. Program instructions recorded on the recording medium may be those specially designed and configured for the present invention, or those known and usable to those skilled in computer software. For example, recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magneto-optical media such as floptical disks ( It includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media and ROM, RAM, flash memory, etc. Examples of program commands may include high-level language wires that can be executed by a computer using an interpreter, as well as machine language wires such as those produced by a compiler. These hardware devices may be configured to act as one or more software modules to perform the operations of the present invention, and vice versa.

The present invention has been described above using several preferred examples, but these examples are illustrative and not limiting. As such, those skilled in the art to which the present invention belongs will understand that various changes and modifications can be made according to the doctrine of equivalents without departing from the spirit of the present invention and the scope of rights set forth in the appended claims.

10: prediction server 11: communication module
12: storage module 13: control module
20: user device 30: information provision server
100: learning unit 200: data processing unit
300: prediction unit

Claims (10)

In the method for predicting the degree of congestion by subway compartment,
receiving, by a data processing unit, a message requesting a congestion level of a subway station and a subway car of a subway from a user device;
selecting, by the data processing unit, a plurality of dispatchers from the past of a reference point in time when the message is received, and generating an input vector sequence including a plurality of input vectors corresponding to each selected dispatcher;
deriving a degree of congestion of the subway car of the subway station when a predetermined number of next trains are allocated from the reference point by a predictor unit performing an operation on the plurality of input vectors through a prediction network;
characterized in that it includes
A method for predicting congestion. According to claim 1,
The step of deriving the congestion degree of the subway station and the subway car
receiving, for each of the plurality of stages, a plurality of input vectors arranged in chronological order corresponding to a plurality of stages of the prediction network in an input layer of the prediction network; and
The plurality of hidden cells of the hidden layer of the prediction network calculates the state value of the current stage by performing an operation in which weights are applied to the state value of the previous stage and the input vector, which is the input value of the current stage, and then calculates the state value of the current stage forward to the next stage,
The last hidden cell among the plurality of hidden cells calculates the state value of the current stage, and then applies a weight to the calculated state value of the current stage to indicate the degree of congestion of the subway car of the subway station at the time of the next dispatch of the predetermined number of times. calculating an output vector;
characterized in that it includes
A method for predicting congestion. According to claim 1,
Before the step of receiving the message requesting the degree of congestion,
preparing, by a learning unit, learning data including a learning input vector sequence including a plurality of learning input vectors and an actual measurement vector for learning that is a label corresponding to the learning input vector sequence;
calculating an output vector by performing an operation on the plurality of input vectors for learning by a prediction network having unlearned weights;
calculating a loss, which is a difference between the output vector and the actual vector for training, through a loss function, and performing optimization of modifying weights of the prediction network so that the loss is minimized;
characterized in that it further comprises
A method for predicting congestion. According to claim 1,
The step of generating an input vector sequence including the plurality of input vectors
Selecting, by the data processing unit, an interval of a predetermined multiple of an interval of an interval of the subway station from a past point in time based on the reference point in time;
Collecting, by the data processing unit, dispatch time corresponding to the selected dispatch time, congestion level of the subway car corresponding to the dispatch time, meteorological information indicating weather conditions, and event information digitizing events contributing to whether or not people are crowded; and
generating an input vector sequence by the data processing unit by arranging a plurality of input vectors including the dispatch time, the degree of congestion of the subway car, the weather information, and the event information in order of dispatch time;
characterized in that it includes
A method for predicting congestion. According to claim 4,
The congestion level of the subway car is
Characterized in that it is collected through the weight of the subway car
A method for predicting congestion. In the apparatus for predicting the degree of congestion by subway compartment,
Receive a message requesting the congestion level of subway stations and subway cars of the subway from the user device;
a data processing unit that selects a plurality of dispatchers from the past of a reference point in time when the message is received, and generates an input vector sequence including a plurality of input vectors corresponding to each selected dispatcher; and
a prediction unit for deriving a degree of congestion of the subway car of the subway station when a predetermined number of next trains are allocated from the reference point by performing an operation on the plurality of input vectors through a prediction network;
characterized in that it includes
A device for predicting congestion. According to claim 6,
The prediction network
an input layer that receives a plurality of input vectors arranged in time order corresponding to a plurality of stages for each of the plurality of stages;
Includes a plurality of hidden cells that calculate the state value of the current stage by performing an operation in which weights are applied to the state value of the previous stage and the input vector, which is the input value of the current stage, and then transfer the calculated state value to the next stage but
The last hidden cell among the plurality of hidden cells calculates the state value of the current stage, and then applies a weight to the calculated state value of the current stage to indicate the degree of congestion of the subway car of the subway station at the time of the next dispatch of the predetermined number of times. a hidden layer that calculates an output vector; and
an output layer outputting the output vector;
characterized in that it includes
A device for predicting congestion. According to claim 6,
Preparing learning data including a learning input vector column including a plurality of learning input vectors and an actual measurement vector for learning that is a label corresponding to the learning input vector column;
A prediction network having unlearned weights calculates an output vector by performing an operation on the plurality of input vectors for learning,
Calculating a loss, which is the difference between the output vector and the actual vector for learning, through a loss function, and performing optimization to modify the weights of the prediction network so that the loss is minimized
learning department;
characterized in that it further comprises
A device for predicting congestion. According to claim 6,
The data processing unit
Based on the reference point in time, selecting an interval consisting of a predetermined multiple of the interval of the interval of the subway station from a past point in time;
Collecting dispatch time corresponding to the selected dispatch time, congestion level of the subway car corresponding to the dispatch time, meteorological information indicating weather conditions, and event information digitizing events contributing to whether or not people are crowded,
Characterized in that an input vector sequence is generated by arranging a plurality of input vectors each including the dispatch time, the degree of congestion of the subway car, the weather information, and the event information in order of dispatch time.
A device for predicting congestion. According to claim 9,
The congestion level of the subway car is
Characterized in that it is collected through the weight of the subway car
A device for predicting congestion.

Images