Intelligent shunting method and system for section station

Abstract

The invention relates to the technical field of intelligent shunting operation, in particular to an intelligent shunting method and system for a section station, and the method comprises the steps: obtaining first information which comprises vehicle attribute data, station yard state data and external constraint data corresponding to a section station to be subjected to shunting operation; performing feature extraction on the first information to obtain first feature information, second feature information and third feature information; sending the first feature information, the second feature information and the third feature information to a preset neural network to generate at least two candidate train marshalling sequence tables; the shunting efficiency of each candidate train marshalling sequence table is evaluated, and an evaluation result is obtained; screening the candidate train marshalling sequence table according to the evaluation result to obtain a screened train marshalling sequence table; and according to the screened train marshalling sequence table, shunting operation is conducted on the vehicles of the section station, and the scientificity and accuracy of transportation management are improved through a data driving mode.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent shunting operation, in particular to an intelligent shunting method and system for section station. BACKGROUND

[0002] In the railway transportation system, the section station as an important node undertakes a large number of train marshalling and unmarshalling and other shunting tasks. The traditional shunting operation mode of the section station mainly relies on experienced staff for manual operation, and highly depends on their experience judgment in the operation process. At the same time, the operation scene of the section station is extremely complex. There are not only mixed operation of different types of trains such as passenger trains and freight trains, but also significant differences in operation requirements and processes of various trains, which increases the complexity and difficulty of operation. It also often faces challenges such as equipment sudden failure, adverse weather influence, and temporary late arrival of trains. These unpredictable factors make the originally complex operation environment even more difficult to deal with. Especially during the busy period of shunting operation, due to the limitations of manual operation and the subjectivity of experience judgment, it is difficult to quickly and accurately make optimal decisions for complex operation situations. Therefore, the train waiting time for marshalling and unmarshalling is too long. The train is stranded in the station for a long time, not only wasting station resources, but also disrupting the operation plan of subsequent trains, having a chain reaction on the rhythm of the entire railway transportation, and seriously affecting the overall efficiency of railway transportation, restricting the improvement of railway transportation capacity and the improvement of service quality. SUMMARY

[0003] In order to improve the problems mentioned in the background, the present application provides an intelligent shunting method and system for section station. The technical solution adopted by the present application is as follows:

[0004] On the one hand, the present application provides an intelligent shunting method for section station, which comprises:

[0005] obtaining first information, the first information comprising vehicle attribute data, station state data and external constraint data corresponding to the section station to be completed shunting operation;

[0006] performing feature extraction on the first information to obtain first feature information, second feature information and third feature information, the first feature information comprising vehicle type, vehicle load and cargo nature, the second feature information comprising residual length of track, turnout conversion time and signal state change frequency, and the third feature information comprising arrival and departure time interval of train, priority of transportation task and weather condition;

[0007] sending the first feature information, the second feature information and the third feature information to a preset neural network to generate at least two candidate train marshalling sequence tables;

[0008] evaluate a shunting efficiency of each of the candidate train marshalling sequence tables to obtain evaluation results;

[0009] screen the candidate train marshalling sequence tables according to the evaluation results to obtain screened train marshalling sequence tables, the screened train marshalling sequence tables including a train marshalling sequence table with the highest shunting efficiency;

[0010] perform shunting operations on vehicles of the section station according to the screened train marshalling sequence tables.

[0011] In a second aspect, an embodiment of the present application provides an intelligent shunting system for a section station, the system comprising:

[0012] an acquisition module configured to acquire first information, the first information including vehicle attribute data, station yard state data and external constraint data corresponding to a section station to be subjected to a shunting operation;

[0013] a first processing module configured to perform feature extraction on the first information to obtain first feature information, second feature information and third feature information, the first feature information including vehicle type, vehicle load and cargo property, the second feature information including residual length of a track, turnout switching time and signal state change frequency, and the third feature information including arrival and departure time interval of a train, priority of a transportation task and weather condition;

[0014] a second processing module configured to send the first feature information, the second feature information and the third feature information to a preset neural network to generate at least two candidate train marshalling sequence tables;

[0015] an evaluation module configured to evaluate a shunting efficiency of each of the candidate train marshalling sequence tables to obtain evaluation results;

[0016] a third processing module configured to screen the candidate train marshalling sequence tables according to the evaluation results to obtain screened train marshalling sequence tables, the screened train marshalling sequence tables including a train marshalling sequence table with the highest shunting efficiency;

[0017] a fourth processing module configured to perform shunting operations on vehicles of the section station according to the screened train marshalling sequence tables.

[0018] In a third aspect, an embodiment of the present application provides an intelligent shunting device for a section station, the device comprising a memory and a processor. The memory is configured to store a computer program; and the processor is configured to execute the computer program to implement steps of the intelligent shunting method for the section station.

[0019] In a fourth aspect, the embodiments of the present application provide a readable storage medium, and the readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the intelligent marshalling method for a section station.

[0020] The present application has the following beneficial effects:

[0021] The present application generates a candidate train marshalling sequence table by using a preset neural network, and evaluates and screens the marshalling efficiency, so that the train marshalling sequence table with the highest marshalling efficiency can be selected from multiple schemes, a data-driven mode is used to replace a traditional experience-driven mode, a neural network is used to process and analyze multiple data, the system adaptability and robustness are improved, and the system can adapt to a complex marshalling scene of a section station, and can generate a suitable marshalling scheme through data collection and analysis, regardless of the diversity of vehicle types, the change of station equipment states or the difference of external environmental conditions.

[0022] Other features and advantages of the present application will be described in the following description, and some will become apparent from the description, or will be learned through implementation of the embodiments of the present application. The purposes and other advantages of the present application can be achieved and obtained by the structure specifically pointed out in the written description and the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments, and the following drawings only show some embodiments of the present application, and therefore should not be considered as limiting the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0024] Figure 1 The figure is a flowchart of the intelligent marshalling method for a section station described in the embodiments of the present application.

[0025] Figure 2 The figure is a structure schematic diagram of the intelligent marshalling system for a section station described in the embodiments of the present application.

[0026] Figure 3 The figure is a structure schematic diagram of the intelligent marshalling equipment for a section station described in the embodiments of the present application.

[0027] In the figure, 800 is intelligent marshalling equipment for a section station, 801 is a processor, 802 is a memory, 803 is a multimedia component, 804 is an I / O interface, 805 is a communication component, 901 is an acquisition module, 902 is a first processing module, 903 is a second processing module, 904 is an evaluation module, 905 is a third processing module, and 906 is a fourth processing module. DETAILED DESCRIPTION

[0028] The technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings. Obviously, the described embodiments are part of, rather than all of, the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0029] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings of the embodiments of the present application. Obviously, the described embodiments are part of, rather than all of, the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed present application, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0030] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. Meanwhile, in the description of the present application, the terms “first”, “second”, and the like are merely used to distinguish description, and cannot be understood as indicating or implying relative importance.

[0031] Embodiment 1

[0032] The present embodiment provides an intelligent shunting method for a section station. It can be understood that a scene can be laid out in the present embodiment, for example: a scene of quickly disassembling, assembling, vehicle coupling and decoupling, and line switching operation for a train arriving at a section station.

[0033] Referring to Figure 1 , the present method includes steps S1, S2, S3, S4, S5, and S6, which specifically include:

[0034] Step S1: obtaining first information, the first information including vehicle attribute data, station yard state data, and external constraint data corresponding to a section station to be completed shunting operation;

[0035] In this step, the vehicle attribute data is collected in real time by electronic tags, sensors and other devices installed on the vehicle, including but not limited to the type of vehicle (such as passenger cars, trucks, special vehicles, etc.), vehicle number, load, volume, cargo name, cargo weight, cargo properties (such as flammable and explosive, fragile, etc.), vehicle length, vehicle length change, etc. These data will directly affect the rationality and safety of train formation; using various monitoring devices arranged in the station yard, such as track sensors, turnout position sensors, signal state monitors, etc., to obtain real-time station yard state data. Specifically, it includes the occupation of the track (vehicle information on each track, remaining length of the track, etc.), the position state of the turnout (opening direction, whether it is working normally, etc.), the display state of the signal (green light, yellow light, red light, etc.), and the position and state of the shunting locomotive (running speed, working state, etc.). Station yard state data is an important basis for developing reasonable marshalling plans and shunting operation plans; external constraint data is obtained from the railway transportation management system, such as train running time plan (including arrival time, departure time, passing stations and time, etc.), locomotive scheduling arrangement (number of available locomotives, locomotive running plan, etc.), special transportation task requirements (such as key material transportation, military train transportation, etc.), and environmental factors such as weather (adverse weather may affect the speed and safety of shunting operations).

[0036] In step S2, feature extraction is performed on the first information to obtain first feature information, second feature information and third feature information. The first feature information includes vehicle type, vehicle load and cargo properties, the second feature information includes track remaining length, turnout switching time and signal state change frequency, and the third feature information includes train arrival and departure time interval, transportation task priority and weather conditions.

[0037] In this step, the first information needs to be preprocessed before feature extraction. The preprocessing includes checking the vehicle attribute data for abnormal values (such as load exceeding the reasonable range) in numerical data such as vehicle number, load, volume, etc., identifying and correcting or deleting abnormal values using statistical methods (such as 3σ principle). For text data such as cargo name, uniform coding is performed (such as using one-hot encoding) for subsequent processing; for discrete data such as track occupation and turnout position in station yard state data, standardization processing is performed to ensure data consistency. For time series data collected by devices such as track sensors and signal state monitors, denoising processing is performed (such as using moving average filtering) to improve data quality; for time-related data such as train running time plan in external constraint data, it is converted to a unified time format and the time interval is calculated. For text information such as special transportation task requirements, classification and coding are performed to enable it to be integrated into the subsequent feature extraction process.

[0038] It can be understood that the key features extracted from the vehicle attribute data include vehicle type, vehicle load and cargo property, wherein the vehicle type (passenger car, truck, special vehicle, etc.) as a classification feature, affects the priority and mode of marshalling; the vehicle load is used to evaluate the transportation capacity and balance of the train, which has an important influence on the marshalling sequence; the cargo property (flammable and explosive, fragile, etc.) determines the position of the vehicle in the marshalling to ensure the safety of transportation. The key features extracted from the station yard state data include the residual length of the track, the turnout conversion time and the signal state change frequency, wherein the residual length of the track reflects the capacity of the track to accommodate vehicles, which affects the parking and marshalling arrangement of the vehicles; the turnout conversion time is crucial for the time planning of the shunting operation and is a key factor for calculating the shunting efficiency; the signal state change frequency can reflect the busy degree of the station yard, which has an influence on the timing of the shunting operation. The key features extracted from the external constraint data include the arrival and departure time interval of the train, the priority of the transportation task and the weather condition, wherein the arrival and departure time interval of the train determines the available time window of the shunting operation, which affects the feasibility of the marshalling plan; the priority of the transportation task may require priority arrangement of vehicle marshalling and shunting operation; the weather condition (such as wind speed, snowfall, etc.) affects the speed and safety of the shunting operation and can be used as an environmental constraint feature.

[0039] Step S3, sending the first feature information, the second feature information and the third feature information to a preset neural network to generate at least two candidate train marshalling sequence tables;

[0040] In the step S3, it further includes a step S31 and a step S32, which specifically include:

[0041] Step S31, constructing a marshalling optimization feature vector according to the first feature information, the second feature information and the third feature information;

[0042] In this step, the collected vehicle attribute data, station yard state data and external constraint data are integrated and preprocessed to remove noise and outliers and unify the data format. Then, according to the characteristics and mutual relationships of these data, key features are extracted to construct a marshalling optimization feature vector. The feature vector will be used as the input data of the subsequent neural network model to generate reasonable candidate train marshalling sequence tables.

[0043] In the step S31, it includes a step S311, a step S312 and a step S313, which specifically include:

[0044] Step S311, fusing the first feature information, the second feature information and the third feature information to obtain a comprehensive feature set, wherein the comprehensive feature set includes at least two fused feature information;

[0045] In this step, different types of features are combined together using feature splicing to construct a comprehensive feature set including a variety of fused feature information.

[0046] In step S312, the feature information in the comprehensive feature set is screened by using a feature selection algorithm to obtain screened feature information.

[0047] In this step, the feature selection algorithm is used to screen the feature information in the comprehensive feature set, and features with high correlation or less impact on marshalling optimization are removed. Features that have a significant impact on marshalling decision are retained to reduce feature dimension and improve model training efficiency.

[0048] In step S312, it further includes steps S3121, S3122, S3123, S3124 and S3125, which specifically include:

[0049] In step S3121, the correlation between two feature information in the comprehensive feature set is calculated to obtain correlation information.

[0050] In this step, the way of calculating the correlation between features is not limited by the present application, including but not limited to using Pearson correlation coefficient. It is well known to those skilled in the art to calculate the correlation between feature information using Pearson correlation coefficient, so it will not be repeated here.

[0051] In step S3122, a correlation matrix is constructed according to the correlation information.

[0052] In step S3123, a first preset threshold is obtained.

[0053] In step S3124, the feature information in the comprehensive feature set is screened according to the first preset threshold and the correlation matrix to obtain a sub-feature set.

[0054] In this step, the correlation between all features in the feature set is calculated to obtain a correlation matrix. By analyzing the correlation matrix, the feature pairs with absolute correlation higher than the first preset threshold are found. For each pair of highly correlated features, the feature that has a more important impact on marshalling optimization is retained (which can be determined according to the experience of the skilled person), and the other feature is removed.

[0055] In step S3125, the sub-feature set is screened to obtain the screened feature information.

[0056] In step S3125, it further includes steps S31251, S31252 and S31253, which specifically include:

[0057] Step S31251, divide the sub-feature set and the corresponding train efficiency index into a training set, a validation set and a test set, train a random forest model, and obtain a trained random forest model;

[0058] Step S31252, send the feature information in the sub-feature set to the trained random forest model, calculate the importance of each feature information, and obtain score information;

[0059] In this step, when constructing each decision tree of the random forest, for each node, a feature is selected for splitting to reduce the impurity of the split node. The importance of a feature is equal to the average of the impurity reduction amount caused by using it to split the node in all decision trees. The greater the impurity reduction amount, the greater the influence of the feature on the classification or regression result, and the higher the importance score. Therefore, the importance of each feature information can be measured by calculating the average of the impurity reduction amount caused by using the feature to split the node.

[0060] Step S31253, screen the sub-feature set according to the score information, and obtain screened feature information.

[0061] In this step, a second threshold is preset, features with score information lower than the second threshold are considered to have less influence on marshalling optimization, and are removed from the sub-feature set to obtain screened feature information. The screened feature information not only removes redundant features with high correlation, but also retains features that have important influence on marshalling optimization, and can be used for subsequent marshalling optimization model training and application, reducing the feature dimension and improving the model training efficiency.

[0062] Step S313, construct the marshalling optimization feature vector according to the screened feature information.

[0063] In this step, the screened key features are arranged in descending order of score information to construct the marshalling optimization feature vector. Each dimension of the feature vector corresponds to a key feature, and the value is the preprocessed and standardized feature value. The feature vector is normalized to map all feature values to the same range (such as [0, 1] or mean 0, standard deviation 1) to ensure the weight consistency of different features in subsequent model training. It should be noted that the screened key features are arranged in descending order of score information, and the features that have a greater impact on marshalling optimization are arranged in the front, so that important features can play a more prominent role in subsequent model training or analysis.

[0064] Step S32, send the marshalling optimization feature vector to the preset neural network model to generate at least two candidate train marshalling sequence tables.

[0065] In this step, pre-training is performed according to a large amount of historical train marshalling data and related vehicle, station yard and external constraint data. In the training process, the parameters of the network are adjusted, so that the model can learn the mapping relationship between different feature vectors and reasonable train marshalling sequences. The constructed marshalling optimization feature vector is input into the pre-trained neural network model, and the model generates a plurality of candidate train marshalling sequence tables according to the learned knowledge and patterns. Each candidate marshalling sequence table contains information such as the marshalling sequence of the vehicle, the arrangement mode of the vehicle on the track, and the possible shunting operation steps. The structure of the neural network model is not limited in the present application, including but not limited to deep neural network (DNN), recurrent neural network (RNN) or its variant long short-term memory network (LSTM).

[0066] Step S4, evaluating the shunting efficiency of each candidate train marshalling sequence table to obtain an evaluation result;

[0067] In step S4, steps S41, S42, S43 and S44 are further included, which specifically include:

[0068] Step S41, determining second information according to the candidate train marshalling sequence table, the second information including total time of vehicle coupling and uncoupling operation, total time of locomotive running and waiting time corresponding to the candidate train marshalling sequence table;

[0069] Step S42, calculating according to the second information to obtain total shunting time;

[0070] In this step, the specific calculation process of the total shunting time is as follows:

[0071] T=t1+t2+t3

[0072] In the above formula, T represents the total shunting time, t1, t2 and t3 represent the total time of vehicle coupling and uncoupling operation, the total time of locomotive running and the waiting time corresponding to the candidate train marshalling sequence table, respectively.

[0073] Step S43, determining third information according to the candidate train marshalling sequence table, the third information including total number of shunting hooks corresponding to coupling, pushing and putting in the marshalling process;

[0074] Step S44, evaluating the efficiency of the candidate train marshalling sequence table according to the third information and the total shunting time to obtain an evaluation result.

[0075] In this step, the specific calculation process of the evaluation result is as follows:

[0076] A=αT+βN+γT track

[0077] In the above formula, A represents the evaluation result, a, b, and g represent weight coefficients, T represents the total shunting time, N represents the total number of shunting hooks, T track represents the track occupancy time, and it should be noted that the weight coefficients a, b, and g are determined by using an adaptive method. The specific process is as follows: historical shunting operation data is used in combination with actual shunting efficiency, safety conditions, and resource utilization conditions and other indicators to train a model by using a machine learning algorithm (such as linear regression, neural network, etc.) to automatically learn the weight coefficients. In the training process, the model optimizes the objective function according to different combinations of weight coefficients, finds out the weight coefficients that make the objective function value optimal and most consistent with the actual situation, and determines the weight coefficients by using an adaptive method, which is beneficial to accurately evaluating the train marshalling sequence table corresponding to the section station.

[0078] Step S5: screening the candidate train marshalling sequence table according to the evaluation result to obtain a screened train marshalling sequence table, wherein the screened train marshalling sequence table includes a train marshalling sequence table with the highest shunting efficiency;

[0079] Step S6: performing shunting operation on the vehicles of the section station according to the screened train marshalling sequence table.

[0080] The first characteristic information contains the properties of the goods, and the vehicle position can be reasonably arranged according to the danger degree, compatibility and the like of the goods during marshalling, so as to avoid safety accidents caused by improper placement of goods, and in combination with station yard state data (such as signal state change frequency and the like) and external constraint data (such as weather conditions and the like), various possible safety influencing factors can be fully considered when evaluating the marshalling sequence table. For example, under adverse weather conditions, the safety risk caused by weather factors can be reduced by adjusting the marshalling scheme and shunting operation plan, and the safety of personnel and equipment during shunting operation is ensured; the residual length information in the second characteristic information enables the track resources to be more reasonably utilized when generating the marshalling sequence table, so as to avoid idle or excessive occupation of the track, improve the utilization rate of the track, and fully play the role of the station yard infrastructure; the third characteristic information covers the arrival and departure time interval of the train, the priority of the transportation task and the like external constraint data, so that the technical scheme can flexibly cope with different transportation tasks and scheduling requirements. For the transportation task with high priority, the vehicle marshalling and shunting operation can be arranged in priority, so as to ensure the timely transportation of key materials or emergency tasks; at the same time, the shunting operation can be reasonably planned according to the train arrival and departure time interval, so as to avoid interference with the subsequent train operation, and ensure the overall coordination of the transportation system. Through the data-driven mode, the neural network is used to process and analyze various data, which can adapt to various complex and changeable station yard conditions and transportation demands. Whether the vehicle type is diverse, the station yard equipment state is changed or the external environmental conditions are different, the appropriate marshalling scheme can be generated through data collection and analysis, and the adaptability and robustness of the system are improved.

[0081] The step S6 further comprises a step S61, a step S62, a step S63 and a step S64, which specifically comprise:

[0082] The step S61 determines the shunting operation task of the vehicle according to the screened train marshalling sequence table, and generates a shunting operation sheet, wherein the shunting operation sheet comprises an operation sequence, an operation time and an operation task;

[0083] The step S62 acquires a space-time grid map corresponding to a section station and initializes the space-time grid map corresponding to the section station, to obtain an initialized space-time grid map;

[0084] In this step, the yard of the section station is digitally modeled, and the yard track is divided into a three-dimensional grid (X-axis: track position, Y-axis: time, Z-axis: locomotive ID) in time-space dimensions. Each grid unit has a unique identifier in time-space dimensions, representing the resource state of the yard at a specific time and space position. The initial resource state of the yard is marked in the time-space grid map, including the occupation of the track, the position of the turnout, the initial position of the shunting locomotive, etc. These resource information is mapped to the corresponding time-space grid unit, providing basic data for subsequent shunting path planning and conflict detection.

[0085] Step S63, mapping the shunting work order to the initialized time-space grid map to obtain a mapped time-space grid map;

[0086] In this step, each task in the generated shunting work order is mapped to the time-space resource grid map. According to the starting point, end point and expected work time of the task, the grid units occupied by the task in the time-space grid map are determined. Through this mapping, the resource demand and distribution of the shunting work in the time-space dimension can be intuitively displayed, and the spatial layout and time dimension information of the yard are presented in a visual manner. By mapping the shunting work order to the map, the specific location and time sequence of each shunting task (such as vehicle picking and dropping, line switching, marshalling, etc.) in the yard space can be clearly seen. The shunting personnel and management personnel can intuitively understand the process and progress of the entire shunting operation, facilitating monitoring and management, and timely discovering potential problems and conflict points. At the same time, since each grid unit represents the resource state of the yard at a specific time and space position, after mapping the shunting work order, the occupation of the track, turnout, shunting locomotive, etc. and the occupation time during the shunting process can be clearly displayed, which helps to determine whether there is resource waste or conflict, thereby providing a basis for optimizing resource allocation.

[0087] Step S64, performing shunting work on the vehicles of the section station according to the mapped time-space grid map.

[0088] The step S64 further includes steps S641, S642, S643, S644, S645, S646 and S647, which specifically include:

[0089] Step S641, performing conflict detection on the mapped time-space grid map to determine whether there is a conflict point, and obtaining a determination result;

[0090] In this step, since the section station is a complex shunting scene, in the shunting operation, the mapped space-time grid map needs to be detected for conflicts, to determine whether there is a conflict point (two trains may arrive at the same position at the same time) on the mapped space-time grid map.

[0091] Step S642, when the judgment result is that there is a conflict point, the coordinate information of the conflict point is obtained;

[0092] Step S643, determining fourth information according to the coordinate information of the conflict point, the fourth information including the time of the vehicle arriving at the conflict point;

[0093] In this step, the specific calculation process of the fourth information is as follows:

[0094]

[0095] In the above formula, T p represents the time of the vehicle arriving at the conflict point, t p represents the current time, d represents the remaining track distance from the current position of the vehicle to the conflict point, v represents the current instantaneous speed, and a represents the current acceleration.

[0096] Step S644, establishing a preset objective function according to the fourth information;

[0097] In this step, the preset objective function is as follows:

[0098]

[0099] s.t.T p = T target ± ΔT

[0100] In the above formula, t0 and t f represent the start time and end time of integration respectively, a(t) represents the acceleration at time t, represents the rate of change of acceleration, σ represents a smoothing coefficient, which suppresses acceleration mutation, T p represents the time of the vehicle arriving at the conflict point, T target represents the target time, and ΔT represents the safety time margin. The above constraint condition requires that the time of the vehicle arriving at the conflict point must be within the range of the target time plus or minus the safety time margin, so as to ensure that the time of the vehicle arriving at the relevant position meets the safety and operation plan requirements, and to avoid conflicts.

[0101] Step S645, solving the acceleration curve according to the preset objective function;

[0102] In this step, solving the preset objective function is a technical solution known to those skilled in the art, and therefore will not be described here.

[0103] Step S646, control the acceleration of the vehicle according to the acceleration curve to obtain a control instruction;

[0104] In this step, the generated acceleration curve is converted into specific control instructions and sent to the corresponding shunting locomotive control system. The control instructions should include information such as the time point of acceleration adjustment and acceleration value to ensure that the locomotive can accurately perform the adjustment operation.

[0105] Step S647, control the acceleration of the vehicle corresponding to the control instruction for completing the shunting operation in the section station.

[0106] In this step, during the process of the locomotive executing the acceleration curve adjustment, the actual running state of the locomotive is continuously monitored, including parameters such as position, speed, acceleration, etc. By comparing with the preset acceleration curve, deviations that may occur are discovered and corrected in time. If it is found that the actual running does not match the expectation, such as the acceleration of the locomotive fails to reach the set value or the adjustment time deviates, the control instruction is adjusted in time to ensure that the locomotive runs according to the predetermined acceleration curve.

[0107] Embodiment 2:

[0108] As shown in Figure 2 , the embodiment provides an intelligent shunting system for a section station, which comprises an acquisition module 901, a first processing module 902, a second processing module 903, an evaluation module 904, a third processing module 905, and a fourth processing module 906, specifically comprising:

[0109] The acquisition module is used to acquire first information, wherein the first information comprises vehicle attribute data corresponding to a section station to be completed for shunting operation, station yard state data, and external constraint data.

[0110] The first processing module is used to extract features of the first information to obtain first feature information, second feature information, and third feature information, wherein the first feature information comprises vehicle type, vehicle load, and cargo nature, the second feature information comprises residual length of track, turnout conversion time, and signal state change frequency, and the third feature information comprises arrival and departure time interval of the train, priority of the transportation task, and weather condition.

[0111] The second processing module is used to send the first feature information, the second feature information, and the third feature information to a preset neural network to generate at least two candidate train marshalling sequence tables.

[0112] The evaluation module is used to evaluate the shunting efficiency of each of the candidate train marshalling sequence tables to obtain an evaluation result.

[0113] a third processing module, configured to filter the candidate train marshalling sequence table according to the evaluation result to obtain a filtered train marshalling sequence table, the filtered train marshalling sequence table including a train marshalling sequence table with the highest shunting efficiency;

[0114] a fourth processing module, configured to perform shunting work on vehicles at the section station according to the filtered train marshalling sequence table.

[0115] In one specific embodiment of the present disclosure, the second processing module further includes a first processing unit and a second processing unit, and specifically includes:

[0116] The first processing unit is configured to construct a marshalling optimization feature vector according to the first feature information, the second feature information, and the third feature information.

[0117] The second processing unit is configured to send the marshalling optimization feature vector to the preset neural network model to generate at least two candidate train marshalling sequence tables.

[0118] In one specific embodiment of the present disclosure, the first processing unit further includes a third processing unit, a fourth processing unit, and a fifth processing unit, and specifically includes:

[0119] The third processing unit is configured to fuse the first feature information, the second feature information, and the third feature information to obtain a comprehensive feature set, the comprehensive feature set including at least two fused feature information.

[0120] The fourth processing unit is configured to filter the feature information in the comprehensive feature set by using a feature selection algorithm to obtain filtered feature information.

[0121] The fifth processing unit is configured to construct the marshalling optimization feature vector according to the filtered feature information.

[0122] It should be noted that, as for the system in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments of the method, and will not be described in detail here.

[0123] Embodiment 3:

[0124] Corresponding to the above method embodiments, the present embodiment also provides an intelligent shunting device for a section station. The intelligent shunting device for a section station described below can be mutually corresponding to the intelligent shunting method for a section station described above.

[0125] Figure 3 is a block diagram of an intelligent shunting device 800 for a section station according to an exemplary embodiment. As shown in FIG. 8, the intelligent shunting device 800 includes a first processing module 810, a second processing module 820, a third processing module 830, and a fourth processing module 840. Figure 3As shown, the intelligent shunting device 800 for block station can include a processor 801, a memory 802. The intelligent shunting device 800 for block station can further include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0126] The processor 801 is configured to control overall operations of the intelligent shunting device 800 for block station to complete all or part of the steps of the above intelligent shunting method for block station. The memory 802 is configured to store various types of data to support operations of the intelligent shunting device 800 for block station, which can include, for example, instructions for any application or method operating on the intelligent shunting device 800 for block station, and application-related data, such as contact data, sent and received messages, pictures, audio, video, and the like. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 803 can include a screen and an audio component. The screen can be a touch screen, for example, and the audio component is configured to output and / or input audio signals. For example, the audio component can include a microphone configured to receive external audio signals. The received audio signals can be further stored in the memory 802 or transmitted through the communication component 805. The audio component further includes at least one speaker configured to output audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, which can be a keyboard, a mouse, a button, and the like. The buttons can be virtual buttons or physical buttons. The communication component 805 is configured to enable wired or wireless communication between the intelligent shunting device 800 for block station and other devices. The wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G or 4G, or one or more of them or a combination of one or more of them, so the corresponding communication component 805 can include a Wi-Fi module, a Bluetooth module, an NFC module.

[0127] In an example embodiment, the intelligent shunting device 800 for the section station can be implemented by one or more Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor or other electronic components for executing the intelligent shunting method for the section station described above.

[0128] In another example embodiment, a computer readable storage medium including program instructions that, when executed by a processor, implement the steps of the intelligent shunting method for the section station described above is also provided. For example, the computer readable storage medium can be the memory 802 described above including program instructions executable by the processor 801 of the intelligent shunting device 800 for the section station to complete the intelligent shunting method for the section station described above.

[0129] Embodiment 4:

[0130] Corresponding to the method embodiments above, in this embodiment, a readable storage medium is also provided, which can be referred to in conjunction with the intelligent shunting method for the section station described above.

[0131] A readable storage medium, on which a computer program is stored, the computer program being executed by a processor to implement the steps of the intelligent shunting method for the section station of the method embodiments described above.

[0132] The readable storage medium can be specifically a U disk, a mobile hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and various readable storage media that can store program codes.

[0133] The above only describes the preferred embodiments of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0134] The above merely describes the present application and its embodiments, which are not restrictive, and the drawings only show one of the embodiments of the present application, and the actual structure is not limited thereto. In general, if a person skilled in the art is inspired by the above, without departing from the purpose of the present application, without creative design, similar structure and embodiments of the technical solution are not creative, and should belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A smart shunting method for section stations, characterized in that, include: Obtain first information, which includes vehicle attribute data, station status data, and external constraint data corresponding to the section station where the shunting operation is to be completed; Feature extraction is performed on the first information to obtain first feature information, second feature information and third feature information. The first feature information includes vehicle type, vehicle load and cargo nature; the second feature information includes remaining track length, turnout switching time and signal status change frequency; the third feature information includes train arrival and departure time interval, transportation task priority and weather conditions. The first feature information, the second feature information, and the third feature information are sent to a preset neural network to generate at least two candidate train formation order tables; The shunting efficiency of each of the candidate train formation sequence tables is evaluated to obtain the evaluation results; Based on the evaluation results, the candidate train formation sequence list is screened to obtain the screened train formation sequence list, which includes the train formation sequence list with the highest shunting efficiency. Shunting operations are carried out on the vehicles at the section stations according to the selected train formation sequence table.

2. The intelligent shunting method for section stations according to claim 1, characterized in that, The first feature information, the second feature information, and the third feature information are sent to a preset neural network to generate at least two candidate train formation order tables, including: Construct a grouping optimization feature vector based on the first feature information, the second feature information, and the third feature information; The optimized train formation feature vector is sent to the preset neural network model to generate at least two candidate train formation order tables.

3. The intelligent shunting method for section stations according to claim 2, characterized in that, Constructing a grouping optimization feature vector based on the first feature information, the second feature information, and the third feature information includes: The first feature information, the second feature information, and the third feature information are fused to obtain a comprehensive feature set, which includes at least two fused feature information. The feature information in the comprehensive feature set is filtered using a feature selection algorithm to obtain the filtered feature information; The optimized grouping feature vector is constructed based on the filtered feature information.

4. The intelligent shunting method for section stations according to claim 3, characterized in that, The feature information in the comprehensive feature set is filtered using a feature selection algorithm to obtain the filtered feature information, including: Calculate the correlation between two feature information in the comprehensive feature set to obtain correlation information; Construct a correlation matrix based on the correlation information; Obtain the preset first threshold; The feature information in the comprehensive feature set is filtered according to the preset first threshold and the correlation matrix to obtain a sub-feature set; The sub-feature set is filtered to obtain the filtered feature information.

5. The intelligent shunting method for section stations according to claim 4, characterized in that, The filtered feature information is obtained by filtering the sub-feature set, including: The sub-feature set and the corresponding shunting efficiency index are divided into training set, validation set and test set to train the random forest model, and the trained random forest model is obtained. The feature information in the sub-feature set is sent to the trained random forest model to calculate the importance of each feature information and obtain the score information. The sub-feature set is filtered based on the score information to obtain the filtered feature information.

6. The intelligent shunting method for section stations according to claim 1, characterized in that, Shunting operations are performed on the vehicles at the section stations according to the selected train formation sequence list, including: Based on the filtered train formation sequence table, the shunting operation tasks of the vehicles are determined, and a shunting operation order is generated. The shunting operation order includes the operation sequence, operation time, and operation tasks. Obtain the spatiotemporal grid map corresponding to the segment station and initialize the spatiotemporal grid map corresponding to the segment station to obtain the initialized spatiotemporal grid map. The shunting operation order is mapped onto the initialized spatiotemporal raster map to obtain the mapped spatiotemporal raster map; The shunting operation of vehicles at the section station is carried out based on the mapped spatiotemporal grid map.

7. The intelligent shunting method for section stations according to claim 6, characterized in that, Based on the mapped spatiotemporal grid map, shunting operations are performed on vehicles at the section station, including: The mapped spatiotemporal raster map is subjected to conflict detection to determine whether there are conflict points and obtain the judgment result. When the judgment result indicates that a conflict point exists, the coordinate information of the conflict point is obtained; The fourth information is determined based on the coordinate information of the conflict point, and the fourth information includes the time when the vehicle arrived at the conflict point. A preset objective function is established based on the fourth information; The acceleration curve is solved according to the preset objective function; The vehicle's acceleration is controlled based on the acceleration curve to obtain control commands; The acceleration of the vehicles that have completed shunting operations within the section station is controlled according to the control commands.

8. An intelligent shunting system for section stations, characterized in that, include: The acquisition module is used to acquire first information, which includes vehicle attribute data, station status data, and external constraint data corresponding to the shunting operation section station to be completed. The first processing module is used to extract features from the first information to obtain first feature information, second feature information and third feature information. The first feature information includes vehicle type, vehicle load and cargo nature; the second feature information includes remaining track length, turnout switching time and signal status change frequency; and the third feature information includes train arrival and departure time interval, transportation task priority and weather conditions. The second processing module is used to send the first feature information, the second feature information and the third feature information to a preset neural network to generate at least two candidate train formation order tables; The evaluation module is used to evaluate the shunting efficiency of each of the candidate train formation sequence tables and obtain the evaluation results. The third processing module is used to filter the candidate train formation sequence table according to the evaluation results to obtain the filtered train formation sequence table, which includes the train formation sequence table with the highest shunting efficiency. The fourth processing module is used to perform shunting operations on the vehicles at the section station according to the filtered train formation sequence table.

9. The intelligent shunting system for section stations according to claim 8, characterized in that, The second processing module includes: The first processing unit is configured to construct a grouping optimization feature vector based on the first feature information, the second feature information, and the third feature information; The second processing unit is used to send the grouping optimization feature vector to the preset neural network model to generate at least two candidate train grouping order tables.

10. The intelligent shunting system for section stations according to claim 9, characterized in that, The first processing unit includes: The third processing unit is used to fuse the first feature information, the second feature information and the third feature information to obtain a comprehensive feature set, wherein the comprehensive feature set includes at least two fused feature information. The fourth processing unit is used to filter the feature information in the comprehensive feature set using a feature selection algorithm to obtain the filtered feature information. The fifth processing unit is used to construct the grouping optimization feature vector based on the filtered feature information.

Images