Training system and training method

Abstract

To provide a training system associated with business systems, in which training is materialized that is close to the actual operation including conversation with the people concerned that is performed without via the business system.SOLUTION: Provided is a training system associated with business systems, comprising: a simulator for simulating a control mode that is the object of the business system; an agent having one or more simulations for each training state of training associated with the control mode; and a drive device for operating the simulator and the agent. The agent executes a simulation in response to at least one of a first request from a trainee and a second request related to training that are received from the drive device, and performs generation of a response to the first request and / or control of the training state pertaining to the control mode for the second request. The drive device replies the response generated by the simulation executed by the agent to the trainee and / or outputs the training state controlled by the simulation to the simulator.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a training system and a training method for a business system, and is particularly suitable for training related to train operation management operations. [Background technology]

[0002] Patent document 1 discloses a system that uses a device to easily create training scenarios from several parameter settings, and changes the scenario to suit the behavior of the trainee by creating rules in advance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2010-113105 Summary of the Invention [Problem to be solved by the invention]

[0004] In previous training for traffic management, such as the training simulation using the training scenario disclosed in Patent Document 1, a scenario prepared in advance was applied to the simulator of the controlled object and its control device, and training on operations on the traffic management system was conducted. On the other hand, when operating using a traffic management system, decisions are sometimes required based on information entered by the dispatcher without going through the traffic management system (such as conversations with the vehicle depot or driver), and it is not possible to provide training that includes this.

[0005] Developing an accurate simulator outside the scope of such a traffic control system is not realistic in terms of development time and man-hours compatibility. Furthermore, when simulating an area outside the scope of the traffic control system using natural language generation AI, etc., it is necessary to ensure that the simulated content is flawless and consistent with the control cycle of the control system.

[0006] Therefore, the present invention aims to provide technology that enables training related to business systems, including operation management systems, that is close to actual operations, including dialogue with relevant parties without going through the business system. [Means for solving the problem]

[0007] In order to solve the above problems, one representative training system of the present invention is a training system related to a business system, and includes a simulator that simulates the control mode targeted by the business system, an agent having one or more simulations for each training situation of the training related to the control mode, and a drive unit that operates the simulator and the agent, wherein the agent performs a simulation in response to at least one of a first request from the trainee and a second request related to the training received from the drive unit, thereby generating a response to the first request and controlling the training situation related to the control mode in response to the second request, and the drive unit at least performs an operation output to the simulator, the operation output being generated by the simulation performed by the agent. [Effects of the Invention]

[0008] According to the present invention, in training related to business systems including operation management systems, it is possible to realize training that is close to real-life operations, including unexpected situations and dialogue between the trainee and other parties. Problems, configurations, and effects other than those described above will become apparent from the following description of the preferred embodiments. [Brief explanation of the drawings]

[0009] [Figure 1] FIG. 1 is a block diagram showing a training system and its related configuration as an embodiment in which the present invention is applied to a train traffic control system. [Figure 2] FIG. 2 is a block diagram showing an example of functions and tables that make up the training drive system. [Figure 3] FIG. 3 is a diagram showing an example of a data flow of driving control by the training drive system. [Figure 4] FIG. 4 is a diagram showing an example of the table configuration of the control route planning timetable. [Figure 5] FIG. 5 is a diagram illustrating an example of the table configuration of the controlled line location status table. [Figure 6] FIG. 6 is a diagram showing an example of the table configuration of the serving line train presence status table. [Figure 7] FIG. 7 is a diagram illustrating an example of the table configuration of the accident situation table. [Figure 8] FIG. 8 is a diagram showing an example of the table configuration of the service route departure time prediction table. [Figure 9] FIG. 9 is a diagram showing an example of the table configuration of the accident prediction table. [Figure 10] FIG. 10 illustrates a first example of the configuration of the verification scenario table. [Figure 11] FIG. 11 illustrates a second example of the configuration of the verification scenario table. [Figure 12] FIG. 12 illustrates a third example of the configuration of the verification scenario table. [Figure 13] FIG. 13 is a diagram illustrating an example of the table configuration of the training scenario table. [Figure 14] FIG. 14 is a diagram illustrating an example of the configuration of train control information. [Figure 15] FIG. 15 is a diagram showing an example of the structure of the interactive dialogue. [Figure 16] FIG. 16 is a flowchart illustrating an example of the process flow of the event verification process. [Figure 17] FIG. 17 is a flowchart illustrating an example of the processing flow of the training process. [Figure 18] FIG. 18 is a flowchart illustrating an example of the processing flow of the control processing. [Figure 19]FIG. 19 is a diagram showing an example of the interactive screen. [Figure 20] FIG. 20 is a diagram showing an example of the prediction generation screen. [Figure 21] FIG. 21 is a diagram showing an example of the training status screen. [Figure 22] FIG. 22 is a diagram showing an example of the training control status screen. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Here, the embodiment is a case where the business system to which the training system according to the present invention is applied is a train traffic control system. However, the present invention is not limited to this embodiment, and the present invention can also be applied to business systems other than train traffic control systems. Furthermore, in the drawings, the same parts are denoted by the same reference numerals. [Example]

[0011] FIG. 1 is a block diagram showing a training system and its related configuration as an embodiment in which the present invention is applied to a train traffic control system. The operation simulator 30 simulates train operation on a predetermined line 1 (not shown). The traffic management system 20 performs route control of a route in which the traffic simulator 30 is involved (hereinafter referred to as a "controlled route") in accordance with input from the dispatcher (trainee) 10.

[0012] The training drive system 100 simulates mutually connected lines and accidents on controlled lines as training situations that affect train operations, and controls part of the train operations. The training drive system 100 also comprises an interface unit 110 that controls input / output with the dispatcher 10 and trainee 40, a control unit 120 that controls input / output with the traffic management system 20 and traffic simulator 30, a communication unit 150 that controls input / output with the agent 50, a verification unit 140 that corrects the output results of the agent 50, and a cooperation management unit 130 that generates train control according to the training situation.

[0013] The agent 50 includes a control simulator 51 for each type of training situation. This control simulator 51 is a component capable of input and output in natural language using a generation AI or the like, and generates a dialogue response with the dispatcher 10 and control of the training situation in the training drive system 100. The control simulator 51 may also reflect the content of the dialogue with the dispatcher 10 in the control of the training situation. In this embodiment, the training situations simulated are the occurrence of an accident that disrupts train operation and train operation on a line (incoming line) 2 (not shown) that runs through to line 1. In the above configuration, the training system according to the embodiment of the present invention is made up of the training drive system 100, the traffic simulator 30, and the agent 50.

[0014] 2 is a block diagram showing an example of functions and tables constituting the training-driven system 100. The functions performed by the training-driven system 100 are comprised of an interface unit 110, a control unit 120, a cooperation management unit 130, a verification unit 140, and a communication unit 150, which are shown in (1) to (5) below, and each of these may be configured as either hardware or software. (1) Interface unit 110 The interface unit 110 has an interactive input function F111 and a screen display function F112. The dialogue input function F111 receives an input message 301 including a designation of a target control simulation 51 from the commander 10 and an input sentence in natural language, and stores it in the dialogue 208 for each control simulation 51. The screen display function F112 creates a screen that displays the interactive dialogue 208. It may also create a screen that displays a prediction dialogue 209 and some or all of the information contained in the various state tables 201-207 provided inside the training-driven system 100.

[0015] (2) Control unit 120 The control unit 120 includes a plan acquisition function F121, a track location acquisition function F122, and a control output function F123. The plan acquisition function F121 stores part or all of the timetable information 303 held by the traffic control system 20 in the control route plan timetable table 201 (FIG. 4). The train location acquisition function F122 stores train location information of the trains simulated by the operation simulator 30 in the control line train location status table 202 (FIG. 5). The control output function F123 stores the control information 203 generated by the cooperation management unit 130 as an operation output 305 to the operation simulator 30.

[0016] (3) Collaboration Management Department 130 The cooperation management unit 130 includes a training processing function F131 and a control processing function F132, and updates the state and control output according to a schedule. At that time, it checks for violations of operational rules in the state after application of each scheduled transition. The training processing function F131 updates the training status table 204 at every real time based on the training scenario table 205 (FIG. 13). This training status table 204 includes, for example, an access line train status table 204-A (FIG. 6) and an accident status table 204-B (FIG. 7). The control processing function F132 creates a control mode for the traffic simulator 30 that does not take into account the signal control performed by the traffic management system 20 as a target, based on the training situation table 204, and stores the control mode in the control information 203.

[0017] (4) Verification Unit 140 The verification unit 140 includes an event verification function F141. The event verification function F141 tests whether the prediction scenario table 207 generated by the agent 50 is feasible and corrects it so that it is feasible. The results of this correction are stored in the verification scenario table 206 (FIGS. 10 to 12). A training scenario table 205 (FIG. 13) may be created that includes only events that result in changes to the training status table 204.

[0018] (5) Liaison Department 150 The communication unit 150 has a dispatcher interaction request function F151 and a scenario prediction request function F152, and is responsible for predicting states and having interactions, and generates schedules for state transitions and control outputs in response to responses from the agents 50. The dispatcher dialogue request function F151 creates a dispatcher dialogue request 306 including the input statement stored in the dialogue dialog 208 and premise information including references to internal tables such as the training status table 204, and outputs it to the agent 50. In addition, a dispatcher dialogue response 307 corresponding to the dispatcher dialogue request 306 is received from the agent 50, and a response to the input statement from the target control simulation 51 is extracted and stored in the dialogue dialog 208. The scenario prediction request function F152 periodically or at any timing creates a scenario prediction request 308 including a statement requesting a prediction of changes in the internal situation of the traffic control system 20 and premise information including references to internal tables such as the training status table 204, and outputs the created scenario prediction request 308 to the agent 50. This scenario prediction request 308 may include training parameters input by the trainee 40. Furthermore, a scenario prediction response 309 corresponding to the scenario prediction request 308 is received from the agent 50, and a predicted scenario table 207 corresponding to the control simulation 51 is created. Here, the predicted scenario table 207 includes, for example, an incoming route departure time prediction table 207-A ( FIG. 8 ) and an accident prediction table 207-B ( FIG. 9 ). At this time, the recorded information of the scenario prediction request 308 and the scenario prediction response 309 may be stored in the prediction dialog 209.

[0019] Next, an example of the data flow of the driving control based on each function described above will be shown. Figure 3 is a diagram showing an example of the data flow of the driving control by the training drive system 100. In response to the scenario prediction request function F152, a predicted scenario generated using the agent 50's route 2 control simulation 51-A is stored in the incoming route departure time prediction table 207-A, and a predicted scenario generated using the agent 50's accident situation control simulation 51-B is stored in the accident prediction table 207-B.

[0020] The event verification function F141 tests whether the generated prediction scenario table 207 (entry line departure time prediction table 207-A, accident prediction table 207-B) is feasible, and after any necessary corrections, stores it in the verification scenario table 206. At the same time, events that involve changes in the training status table 204 (entry line on-track status table 204-A, accident status table 204-B) are stored in the training scenario table 205.

[0021] The training processing function F131 of the collaboration management unit 130 updates the training status table 204 (incoming line status table 204-A, accident status table 204-B) based on the training scenario table 205, and also stores the necessary control mode from the training scenario table 205 in the control information 203. Furthermore, the control processing function F132 of the cooperation management unit 130 creates a control mode based on the training status table 204 and stores it in the control information 203.

[0022] 4 is a diagram showing an example of the table configuration of the control route plan timetable table 201. The control route plan timetable table 201 is a table that holds information on the planned timetable of the controlled route, and the held data items include a train ID 201a, a platform 201b, a running direction 201c, a stop time 201d, a departure time 201e, and a minimum stop time 201f.

[0023] The train ID 201a is an identifier assigned to each train. The platform 201b represents a combination of a station (A, B, C, D, ...) where the train stops and a platform (1, 2, ...). The direction of travel 201c indicates the direction of travel when the train departs from the platform. Here, the direction passing through stations A → B → C → D in this order is defined as A, and the direction passing through stations D → C → B → A in this order is defined as B. The stop time 201d indicates the time when the train stops at a specific platform. The departure time 201e indicates the time when the train departs from a specific platform. The minimum stop time 201f indicates the minimum time (sec) that a train will stop at a specified platform.

[0024] FIG. 5 is a diagram showing an example of the table configuration of the controlled line location status table 202. As shown in FIG. The control line location status table 202 is a table that holds information on trains currently on line 1, and the data items it holds include train ID 202a, current platform 202b, next platform 202c, running status 202d, and stop / running time 202e.

[0025] The train ID 202a is an identifier assigned to each train. The current platform 202b represents a pair of the station (A, B, C, D, ...) where the train is currently stopped and a platform (1, 2, ...), or a pair of the station and platform that the train between stations last passed through. The next platform 202c indicates the combination of the station and platform that the train will pass next on the timetable. The running state 202d indicates the running state of the train on the line, and mainly indicates whether the train is running between stations or stopped at a platform. The stop / running time 202e represents the stop time of a train stopped at a platform or the running time of a train running between stations. Fig. 5 shows an example in which a route having platforms 1 and 2 at stations A to D is set as the control target route 1.

[0026] FIG. 6 is a diagram showing an example of the table configuration of the serving line train location status table 204-A. The serving line train status table 204-A is a table that holds information on trains currently present on the serving line, and the data items it holds include train ID 204a, current platform 204b, next platform 204c, running status 204d, and stop / running time 204e.

[0027] The train ID 204a is an identifier given to each train. The current platform 204b indicates a combination of the station and platform where the train is currently stopped, or a combination of the station and platform that the train between stations last passed through. The next platform 204c indicates a combination of the station and platform that the train will pass next on the timetable. The running state 204d indicates the running state of the train on the line, and mainly indicates whether the train is running between stations or stopped at a platform. The stop / running time 204e represents the stop time of a train stopped at a platform or the running time of a train running between stations. FIG. 6 shows an example in which stations X to Z each have platforms 1 and 2, and the line that runs from station Z to station A is the service line.

[0028] FIG. 7 is a diagram showing an example of the table configuration of the accident situation table 204-B. The accident status table 204-B is a table that holds the status of the accident at the current time, and includes the following data items: accident ID 204f, location 204g, type 204h, current status 204i, occurrence time 204j, and resolution time 204k.

[0029] The accident ID 204f is an identification number assigned to each accident event. The location 204g indicates the location of the vehicle, track, etc. where the accident occurred. Type 204h indicates the pattern of accidents that occur, such as track failures and vehicle breakdowns. The current situation 204i indicates the current situation of the accident. The occurrence time 204j indicates the time when the accident occurred. · Resolution time 204k indicates the time when the accident was resolved.

[0030] FIG. 8 is a diagram showing an example of the table configuration of the service route departure time prediction table 207-A. The incoming line departure time prediction table 207-A is a table for storing information on trains currently on the incoming line for training purposes, and the stored data items include train ID 207a, platform 207b, direction of travel 207c, stop time 207d, departure time 207e, and minimum stop time 207f.

[0031] The train ID 207a is an identifier assigned to each train. Platform 207b indicates the combination of station and platform where the train stops. The direction of travel 207c indicates the direction of travel when the train departs from the platform. Here, the direction passing through stations X → Y → Z → A in this order is designated as A, and the direction passing through stations A → Z → Y → X in this order is designated as B. The stop time 207d indicates the time when the train stops at the platform. The departure time 207e indicates the time when the train departs from the platform. -Minimum stop time 207f indicates the minimum time (sec) that a train will stop at a platform.

[0032] FIG. 9 is a diagram showing an example of the table configuration of the accident prediction table 207-B. The accident prediction table 207-B is a table that holds information on accidents that occur during training, and includes the following data items: accident ID 207g, location 207h, type 207i, occurrence time 207j, and resolution time 207k.

[0033] The accident ID 207g is an identification number assigned to each accident event. Location 207h indicates the location of the vehicle, track, etc. where the accident occurred. The type 207i indicates the pattern of an accident that occurs, such as a track failure or vehicle breakdown. The occurrence time 207j indicates the time when the accident occurred. · Resolution time 207k indicates the time when the accident was resolved.

[0034] 10 to 12 are diagrams showing an example of the table configuration of the verification scenario table 206, divided into first to third sections. The verification scenario table 206 is a table that holds event information that accompanies predicted situation changes, and includes the following data items: event target 206a, type 206b, location 206c, scheduled time 206d, processing status 206e, and processing time 206f.

[0035] The event target 206a is an identifier of the target on which the event occurs, and corresponds to a train ID or an accident ID. The type 206b is a pattern of an event that occurs, and indicates the stopping or starting of a train, the occurrence and resolution of an accident, and the like. The location 206c indicates the location where the event occurs. The scheduled time 206d indicates the time at which an event is expected to occur before the verification process 141. The processing status 206e indicates whether the event has been processed or is pending in the verification process 141. The processing time 206f indicates the time when the event was successfully applied in the verification process 141.

[0036] FIG. 13 is a diagram showing an example of the configuration of the training scenario table 205. The training scenario table 205 is a table in the verification scenario table 206 that stores events including changes in the training situation 204, and stores data items such as event target 205a, type 205b, location 205c, and time 205d.

[0037] The event target 205a is an identifier of the target on which the event occurs, and corresponds to a train ID or an accident ID. The type 205b is a pattern of an event that occurs, and indicates the stopping or starting of a train, the occurrence and resolution of an accident, and the like. The location 205c indicates the location where the event occurs. The time 205d indicates the time when the event was successfully applied in the verification process 141.

[0038] FIG. 14 is a diagram showing an example of the configuration of the train control information 203. The train control information 203 stores the details of the control that the training drive system 100 performs on the operation simulator 30, and the details include a target 203a, a time 203b, and a control 203c.

[0039] The target 203a is an identifier representing the target to be controlled, and corresponds to a train ID or a track block. The time 203b indicates the time at which control is performed. The control 203c indicates the content of the control.

[0040] FIG. 15 is a diagram showing an example of the structure of the interactive dialog 208. As shown in FIG. The dialogue 208 includes a dialogue subject 208a, a time 208b, a type 208c, and a sentence 208d.

[0041] Based on the configuration of the training-driven system shown in FIG. 15 and FIG. 2, an example of interactive training in which the dispatcher 10 confirms the position of the next train scheduled to enter the station will be described below. When the dispatcher 10 specifies the Line 2 control simulation 51-A as the dialogue target and inputs "Where is the next train entering Line 1 now?" as the input message 301, the dialogue input function F111 of the interface unit 110 stores in the dialogue dialog 208, as shown in FIG. 15, "Line 2 control simulation" as the dialogue target 208a, the input acceptance time (12:00:00) as the time 208b, "input" as the type 208c, and the input message 301 as the sentence 208d.

[0042] In response to this input message, the dispatcher interaction request function F151 of the communication unit 150 creates, for example, the following dispatcher interaction request 306. "You are the dispatcher for Line 2 and must answer inquiries by referring to the train location information and operation forecast for Line 2. For the train location information, you refer to (access to training status 204), and for the operation forecast, you refer to (access to training scenario table 205). The content of the inquiry is as follows: 'Where is the next train entering Line 1?'"

[0043] The agent 50 passes this dispatcher dialogue request 306 to the line 2 control simulator 51-A, and returns a dispatcher dialogue response 307. The following response sentence is an example of an assumed dispatcher dialogue response 307. Answer: "Train Axx6 left station platform Y1 60 seconds ago and is scheduled to arrive at station platform A1 at 12:04:00." Here, the agent 50 may perform input and output with the control simulation 51 in natural language using a generation AI or the like.

[0044] Upon receiving this dispatcher dialogue response 307, the dispatcher dialogue request function F151 stores in the dialogue dialog 208, as shown in Figure 15, "Route 2 control simulation" as the dialogue target 208a, the time when the response was received (12:00:20) as the time 208b, "Response" as the type 208c, and the previous response sentence as the sentence 208d.

[0045] 16 is a flowchart showing an example of the process flow of the event verification process F141 (FIG. 2). Each processing step in this flowchart is executed by the verification unit 140, but the description of this execution entity will be omitted below. The event verification process F141 is a process for determining an event schedule to be actually applied to training, based on the incoming route departure time prediction table 207-A and the accident prediction table 207-B, which are event predictions extracted by the scenario prediction request function F152.

[0046] In step S141, each table for the current time shown in Figures 4 to 7 (control route planning timetable table 201, control route track status table 202, incoming route track status table 204-A, and accident status table 204-B) is duplicated, and each table for verification (control route planning timetable table 201', control route track status table 202', incoming route track status table 204'-A, and accident status table 204'-B) is created.

[0047] In step S142, information including the scheduled times of events such as train departures and stops, and accident occurrences and resolutions, for a certain period of time prior to the current time is extracted from the verification control route plan timetable 201', the access route departure time prediction table 207-A, and the accident prediction table 207-B, and this information is stored in chronological order in the verification scenario table 206. At this time, "unprocessed" is stored in the event processing status 206e of the verification scenario table 206. To perform a simulation from the current time until a certain time has elapsed, with the current time set as the initial value of the internal verification time, the verification process shown in steps S143 to S145 is executed for each event in the verification scenario table 206 at each time.

[0048] After the verification process shown in steps S143 to S145 is completed, the internal time is advanced in step S146, and the stopping / running time of trains whose running status in the control line track status table 202' and the incoming line track status table 204'-A is "stopped at track number" or "running between stations" is added. Then, after the verification is completed up to a certain time, step S147 is executed.

[0049] In step S143, for events whose processing status 206e in the verification scenario table 206 is "unprocessed," it is checked whether the internal time has passed the scheduled time 206d of the event. For events whose time has passed (Yes), in step S144, it is checked whether the verification tables, the control route plan timetable table 201', the control route track status table 202', the incoming route track status table 204'-A, and the accident status table 204'-B, conflict with the obstruction conditions.

[0050] Here, an obstacle condition is a condition that determines whether an event cannot be executed. For example, examples of obstacle conditions for a train departure event include the train's current platform being the event's departure platform (condition 1), there being no trains in the section to which it is approaching (condition 2), there being no entry prohibition in the section to which it is approaching (condition 3), and the stop time exceeding the minimum stop time (condition 4). If any of the conditions is not met, it is determined that a conflict has occurred.

[0051] Examples of obstacle conditions for a train stop event include the following: the next platform of the train is the platform of the event (condition 1); there is no train in the section of the destination (condition 2); the section of the destination is not prohibited from entry (condition 3); and the running time has exceeded the minimum on-track time (condition 4). If any of the conditions is not met, it is determined that a conflict has occurred. However, the following examples do not set conditions for the occurrence of track faults or for the elimination of such faults.

[0052] In step S144, if the obstruction conditions are not met (Yes), the event processing in step S145 updates the verification tables, the control line location status table 202', the incoming line location status table 204'-A, and the accident status table 204'-B, according to the event content, and stores "Processed" in the processing status 206e of the verification scenario table 206 and the internal time in the processing time 206f.

[0053] Next, the verification process of steps S143 to S145 will be explained using an example in which the verification tables (control route planning timetable table 201', control route track status table 202', incoming route track status table 204-A', and accident status table 204'-B with no records) which are copies of the tables shown in Figures 4 to 7 are used, and the processing status 206e of the verification scenario table 206 shown in Figures 10 to 12 is all "unprocessed" and the processing time 206f is all blank, and the internal time is set to 12:00:00 as the initial value and advanced by one second in step S146, and steps S143 to S145 are repeated.

[0054] If the internal time is 12:01:00, in step S143, this internal time is compared with the scheduled event time, and five pieces of data with a scheduled time 206d of 12:01:00 (the top five pieces of data in Figure 10) are extracted from the verification scenario table 206 in Figure 10. The determination in step S144 will be explained for a departure event in which train Axx2 departs from platform C1 at station C1 and heads for platform D1 at station D1.

[0055] The above condition 1 is not violated because the current platform of train Axx2 is the same as that of station C1, as determined by referring to the verification control line location status table 202'. The above condition 2 is not violated because the verification control line location status table 202' is referenced and the current platform 202b of the other train is not platform D1. The condition 3 is not violated because the verification accident situation table 204'-B is empty. Condition 4 is not violated because the internal time progression in step S146 has made the stop / running time 202e of train Axx2 60 seconds, which is longer than the minimum stop time 201f of 30 seconds in the verification control route plan timetable 201'.

[0056] Therefore, it is determined that the event does not violate any of conditions 1 to 4 and therefore does not conflict with the obstruction conditions. This departure event is processed in step S145, and in the verification control line location status table 202', "Running between stations" is entered in the running state 202d of the train with train ID Axx2, "00" is entered in the stop / running time 202e, "Processed" is entered in the processing status 206e of the corresponding record in the verification scenario table 206, and the internal time 12:01:00 is entered in the processing time 206f.

[0057] If the verification process of steps S143 to S145 proceeds without any problems until the internal time reaches 12:06:00, in step S143, the internal time is compared with the scheduled event time, and six pieces of data with a scheduled time 206d of 12:06:00 (the six pieces of data at the bottom of FIG. 10) are extracted from the verification scenario table 206 of FIG. 10. Of these, the event of an accident occurring due to a track fault between stations B1 and C1, which corresponds to event "Accident 1," passes the judgment of step S144 because no fault conditions have been set for it, and by the process of step S145, information on the accident with accident ID 207g of "1" in the accident prediction table 207-B of FIG. 9 is stored in the verification accident status table 204'-B, and "occurring" is entered in the current status 204i.

[0058] Thereafter, if the verification process of steps S143 to S145 proceeds without any problems until the internal time reaches 12:08:00, a departure event is extracted in step S143 from the verification scenario table 206 in FIG. 11, in which train Axx6, scheduled to depart from platform B1 at 12:08:00 and head toward platform C1. Regarding the determination of this departure event in step S144, the result is that condition 3 is not satisfied because there is a track fault between stations B1 and C1 in the verification accident situation table 204'-B, and the condition conflicts with the fault condition (No). Therefore, this departure event does not proceed to step S145 at 12:08:00. From then on, until step S144 is passed, step S143 extracts a departure event in which train Axx6, scheduled at 12:08:00, departs from platform B1 at station B1 and heads toward platform C1 at station C1.

[0059] If the internal time subsequently progresses to 12:30:00, in step S143, the conditions of step S144 are not set for the accident recovery event between stations B1 and C1 for "Accident 1" with a scheduled time of 12:30:00 extracted from the verification scenario table 206 in Figure 12, and therefore in step S145 the current status 204i of the accident with accident ID 204f of "1" in the verification accident status table 204'-B is set to "resolved."

[0060] Next, when the internal time has progressed to 12:30:01, the departure event of train Axx6, whose scheduled time is 12:08:00 and which departs from platform B1 at station C1, extracted in step S143, is as follows: The above condition 1 is not violated because the current platform of train Axx6 is the same as that of station B1, as determined by referring to the verification control line location status table 202'. The condition 2 is not violated because the current platform 202b of the other train is not platform C1 at station 202' when the verification control line location status table 202' is referenced. The condition 3 is not violated because there is no accident in the verification accident status table 204'-B whose current status 204i is "occurring." Condition 4 is not violated because the stop / running time 202e of train Axx2 is set to 60 seconds in step S146, which is longer than the minimum stop time 201f of 30 seconds in the verification control route plan timetable 201'.

[0061] Therefore, it is determined that the event does not violate any of conditions 1 to 4 and therefore does not conflict with the obstruction conditions. This departure event is processed in step S145, and in the verification control line location status table 202', "Running between stations" is entered in the running state 202d' of the train ID Axx6, "00" is entered in the stop / running time 202e, "Processed" is entered in the processing status 206e of the corresponding record in the verification scenario table 206, and the internal time 12:30:01 is entered in the processing time 206f. Through the above process, the event is put on hold until the conditions for execution are met, and the scenario is corrected to an executable one.

[0062] In step S147, the train movement and accident situation changes on the service line from the verification scenario table 206 are stored in the training scenario table 205. In addition, an event corresponding to the transfer from service line 2 to service line 1 is stored in the control information 203. By the processing of steps S143 to S147, an executable training scenario table 205 can be created from the prediction scenario table 207 generated by the control simulator 51.

[0063] 17 is a flowchart showing an example of the processing flow of the training process F132. Each processing step in this flowchart is executed by the cooperation manager 130, but the description of this executing entity will be omitted below. In step S131, it is monitored whether the current time of the event read from the training scenario table 205 has passed the event time in the training scenario table 205. At this time, it may be simultaneously monitored whether the obstacle conditions for the event have been cleared.

[0064] If the event time has passed (Yes), the event is applied to the current training situation in step S132. That is, if the event type 205b is "departure", the running state 204d of the corresponding train ID 204a in the serving line location status table 204-A (FIG. 6) is switched to "running between stations", and the running time is stored in the stop / running time 204e. Also, if the event type 205b is "stopped at platform number," the running status 204d of the corresponding train ID 204 in the serving line track status table 204-A is switched to "stopped at platform number," and the current platform number 204b, next platform number 204c, and stop / running time 204e are updated based on the training scenario table 205.

[0065] On the other hand, if the event type 205b is "accident occurrence," information on the accident with a matching accident ID 207g from the accident prediction table 207-B (Fig. 9) is stored in the accident situation table 204-B (Fig. 7). At that time, if the event type is accident recovery, the corresponding accident is deleted from the accident situation table 204-B (Fig. 7). Additionally, in step S132, the processed event is deleted from the training scenario table 205 (Fig. 13).

[0066] 18 is a flowchart showing an example of the processing flow of the control process F131. Each processing step in this flowchart is executed by the cooperation management unit 130, but the description of this execution entity will be omitted below. In step S133, the control line train location status table 202 is read, and the following processing is carried out for each train in the train location stored in the control line train location status table 202.

[0067] In step S134, it is checked whether the running state of the train in question is stopped at platform No. If it is stopped at platform No. (Yes), in step S135, the planned departure time of the train in question (201e in FIG. 4) is read from the control route plan timetable 201, and this planned departure time is compared with the time obtained by adding the minimum stopping time to the stopping time, and the later time is set as the controlled departure time (203b in FIG. 14) and stored in the control information 203.

[0068] In step S136, it is checked whether the affected location 204g of the accident whose current status 204i in the accident status table 204-B is "occurring" is interfering with the running of trains on the line. If there is an interfering event (Yes), in step S137, accident response control such as delaying departure times or stopping trains between stations is stored in the control information 203.

[0069] FIG. 19 is a diagram showing an example of the interactive screen. The dialogue screen includes a display area E100 for the dialogue history for each control simulation 51. The dialogue history display area E100 includes an input message display M001 and an output message display M002.

[0070] The input message display M001 displays a sentence 208d in which the type 208c of the interactive dialog 208 in FIG. 15 is "input." The output message display M002 displays a sentence 208d in which the type 208c of the interactive dialogue 208 in FIG. 15 is "answer." Moreover, the display in the display area E100 is in the order of time 208b of the interactive dialog 208 in FIG.

[0071] FIG. 20 is a diagram showing an example of the prediction generation screen. The prediction generation screen includes a display area E101 when a training scenario prediction for each control simulation 51 is requested. When a scenario prediction request is made, the display area E101 displays input / output messages with the agent, including message M003 and message M004.

[0072] In message M003, a message indicating a request for input of prediction in the prediction dialog 209 is displayed. Message M004 displays a message indicating a response to the request input in message M003. Here, the display in the display area E101 is in the order of time 208b of the interactive dialogue 208.

[0073] FIG. 21 is a diagram showing an example of the training status screen. The screen display includes a display area for the training status for each control simulation 51. The display area for the training status for line 2, which is a service line, includes a display E201 of the order of trains entering the line from line 2, a display E202 of the order of trains leaving the line to line 2, and a display E203 of predicted train operations on line 2. In addition, the display area for the accident training status includes a display E204 of predicted accident details.

[0074] The train entry time display E201 displays train entry / departure information extracted from the training scenario table 205. The train departure time display E202 displays train entry / departure information extracted from the training scenario table 205.

[0075] The operation prediction display E203 displays the contents of the arrival route departure time prediction table 207-A in the verification unit 140 for each time from the current time until a certain time later. In addition, the accident details display E204 displays the contents of the accident prediction table 207-B.

[0076] FIG. 22 is a diagram showing an example of the training control status screen. The training control status screen includes a train operation status display E205 for the controlled line and a train control information display E206. The train operation status display E205 displays the train location in the controlled line location status table 202. The train control information display E206 displays the control output of each train and the control output of each track in the control information 203.

[0077] The event verification process F141 (FIG. 2) allows the changes in the training situation generated by the agent 50 to be corrected to an applicable order. Furthermore, the events corrected by the event verification process F141 through the training process F131 are applied as a training scenario table 205, and the control of the operation simulator 30 based on the training scenario table 205 is created in the control process F132. This prevents the breakdown of control output by the agent 50, and enables the agent 50 and the traffic management system 20 to perform training control with independent control periods.

[0078] According to the above-described embodiment, the present invention includes at least the following aspects. <Aspect 1> A training system related to a business system includes a simulator that simulates the control mode targeted by the business system, an agent having one or more simulations for each training situation of the training related to the control mode, and a drive device that operates the simulator and the agent, wherein the agent performs a simulation in response to at least one of a first request from a trainee and a second request related to the training received from the drive device, thereby generating a response to the first request and controlling the training situation related to the control mode in response to the second request, and the drive device at least performs an operation output to the simulator, the operation output being generated by the simulation performed by the agent and the training situation controlled by the simulation performed by the agent.

[0079] <Aspect 2> In the training system according to the first aspect, the drive device generates the first request in an interactive format between the trainee and the simulation.

[0080] <Aspect 3> In the training system according to the first or second aspect, the driving device generates the first request in a dialogue format in natural language using a generation AI.

[0081] <Aspect 4> A training system according to any one of the above aspects 1 to 3, wherein the driving device comprises an interface unit that processes input related to a first request from the trainee and screen display for the trainee; a communication unit that makes at least one of a first request and a second request to the agent and at least one of creating a response to the first request through dialogue between the trainee and the simulation based on the response, and creating a predictive scenario for the training situation from the current time point to a certain time point after the second request; a verification unit that tries out the predictive scenario to verify its feasibility and, if there is a problem, makes corrections to create a feasible verification scenario, and also creates a training scenario based on the verification scenario; a coordination management unit that checks the training situation based on the training scenario and creates control information for the simulator; and a control unit that inputs control information to the simulator and operates the simulator.

[0082] <Aspect 5> In the training system described in aspect 4 above, the screens displayed to the trainee for each simulation include a dialogue screen showing the history of the dialogue, a prediction generation screen showing the requests and responses when creating a prediction scenario, and a training status screen showing the training status.

[0083] <Aspect 6> In the training system according to the fourth or fifth aspect, the business system is a train traffic control system, and the trainee is a dispatcher.

[0084] <Aspect 7> A training method for a business system using software, using a simulator that simulates the control aspects targeted by the business system and an agent having one or more simulations for each training situation of training related to the control aspects, the training method comprising the steps of making at least one of a first request from the trainee and a second request related to the training to the agent, a step of creating a reply through dialogue between the trainee and the simulation based on the response from the agent to the first request, a step of creating a prediction scenario for the training situation from the current time point to a certain time after the elapse of time based on the response from the agent to the second request, a step of trying out the prediction scenario to verify its feasibility and, if there is any problem, making corrections to create a feasible verification scenario, a step of creating a training scenario based on the verification scenario, a step of checking the training situation based on the training scenario and creating control information for the simulator, and a step of inputting the control information into the simulator and operating the simulator.

[0085] Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications are possible within the scope of the gist of the present invention. [Explanation of symbols]

[0086] 10 dispatcher (trainee), 20 operation control system, 30 operation simulator, 40 Trainer, 50 Agent, 51 Control Simulation, 100 Training Driven System, 110 interface section, F111 dialogue input function, F112 dialogue display function, F113 status display function, 120 control unit, F121 plan acquisition function, F122 on-track acquisition function, F123 control output function, 130 Coordination Management Department, F131 Training Processing Function, F132 Control Processing Function, 140 Verification section, F141 Event verification function, 150 Contact section, F151 Interaction request function, F152 Event request function, 201 Control route planning timetable, 202 Control route on-track status table, 203 control information, 204 training status table, 205 training scenario table, 206 Verification scenario table, 207 Prediction scenario table, 208 Interactive dialogue, 209 Predictive dialogue

Claims

1. A training system for a business system, a simulator for simulating a control mode targeted by the business system, an agent having one or more simulations for each training situation of training relating to the control mode, and a drive device for operating the simulator and the agent, the agent executes the simulation in response to at least one of a first request from the trainee received from the drive device and a second request related to the training, thereby performing at least one of generating a response to the first request and controlling a training situation related to the control manner in response to the second request; The driving device at least one of sending the response generated by the simulation performed by the agent to the trainee and outputting the training situation controlled by the simulation performed by the agent as an operation output to the simulator. A training system characterized by:

2. 10. The training system of claim 1, The drive device generates the first request in an interactive manner between the trainee and the simulation. A training system characterized by:

3. 3. The training system of claim 2, The driving device generates the first request as the dialogue format in natural language using a generation AI. A training system characterized by:

4. 4. A training system according to claim 1, further comprising: The drive device is an interface unit that processes input related to the first request from the trainee and screen display for the trainee; a communication unit that makes at least one of the first request and the second request to the agent, and that performs at least one of creating a response to the first request through a dialogue between the trainee and the simulation based on the response, and creating a prediction scenario for the training situation from the current time until a certain time has elapsed, in response to the second request; a verification unit that tests the prediction scenario to verify feasibility, corrects any impediments, and creates a feasible verification scenario, and also creates a training scenario based on the verification scenario; a cooperation management unit that checks the training situation based on the training scenario and creates control information for the simulator; a control unit that inputs the control information to the simulator and operates the simulator; A training system comprising:

5. 5. The training system of claim 4, The screens to be displayed to the trainee include, for each simulation, a dialogue screen showing a history of the dialogue, a prediction generation screen showing requests and responses when creating the prediction scenario, and a training status screen showing the training status. A training system characterized by:

6. 5. The training system of claim 4, The business system is a train operation management system, and the trainee is a dispatcher. A training system characterized by:

7. A training method for a business system using software, comprising: A simulator that simulates a control mode targeted by the business system and an agent having one or more simulations for each training situation of training related to the control mode are used, making at least one of a first request from a trainee and a second request regarding the training to the agent; generating a dialogue response between the trainee and the simulation based on a response from the agent to the first request; creating a prediction scenario for the training situation from the present time to a certain time point after the elapse of the predetermined time based on a response from the agent to the second request; a step of testing the prediction scenario to verify its feasibility, and correcting any impediments to create a feasible verification scenario; creating a training scenario based on the validation scenario; confirming the training situation based on the training scenario and creating control information for the simulator; inputting the control information into the simulator to operate the simulator; A training method having the following features.

Images