CN114559992B - Train safety protection method, model training method and device and electronic equipment - Google Patents
Train safety protection method, model training method and device and electronic equipment Download PDFInfo
- Publication number
- CN114559992B CN114559992B CN202210248260.5A CN202210248260A CN114559992B CN 114559992 B CN114559992 B CN 114559992B CN 202210248260 A CN202210248260 A CN 202210248260A CN 114559992 B CN114559992 B CN 114559992B
- Authority
- CN
- China
- Prior art keywords
- train
- data
- predicted
- target
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The application provides a train safety protection method, a model training device and electronic equipment, and belongs to the technical field of train safety protection. The train safety protection method comprises the following steps: acquiring first running data of a train within a preset time length; obtaining the predicted running data of the train in the next control period based on the first running data and a pre-trained running data prediction model; if the predicted running data triggers a preset safe speed protection curve of the train, acquiring a target predicted train speed and a target predicted stopping distance when the safe speed protection curve of the train is triggered for the first time in the predicted running data; determining a group of target time sequence data comprising a target predicted train speed and a target predicted stopping distance from a preset protection curve time sequence group; and outputting the running data of the train in the next control period according to the target time sequence data and the current state data of the train so as to enable the train to run according to the running data.
Description
Technical Field
The application relates to the technical field of train safety protection, in particular to a train safety protection method, a model training device and electronic equipment.
Background
In recent years, with the aging of train accurate positioning technology, it becomes possible to accurately control the safe stop of the train. In the prior art, the train is accurately stopped, a control strategy is obtained by analyzing historical running data of the train in each control period, and then the control strategy is selected in real time according to running data acquired when the train runs.
Disclosure of Invention
The application provides a train safety protection method, a model training device and electronic equipment, and aims to solve the problems that in the prior art, the requirement on the network environment of a train is high, and complex conditions in the train running process are difficult to deal with.
In a first aspect, the present application provides a train safety protection method, including: acquiring first running data of a train in a preset time length, wherein the first running data comprises first running speeds of the train at different moments and a first stopping distance representing the distance between the train and a safe stopping point; obtaining the predicted running data of the train in the next control period based on the first running data and a pre-trained running data prediction model, wherein the next control period is a preset time length; if the predicted running data triggers a preset safe speed protection curve of the train, acquiring a target predicted train speed and a target predicted stopping distance when the safe speed protection curve of the train is triggered for the first time in the predicted running data; determining a set of target time series data comprising the target predicted train speed and the target predicted stopping distance from a preset protection curve time series group, wherein the protection curve time series group comprises a plurality of sets of second train speeds and second stopping distances of trains at different moments; and outputting the operation data of the train in the next control period according to the target time sequence data and the current state data of the train so as to enable the train to operate according to the operation data.
In the embodiment of the application, the predicted running data of the train in the next control period is obtained through the first running data of the train in the preset time length and a pre-trained running data prediction model, the target predicted train speed and the target predicted stopping distance when the train safety speed protection curve is triggered for the first time in the predicted running data are obtained when the predicted running data trigger the preset safety speed protection curve of the train, a group of target time sequence data comprising the target predicted train speed and the target predicted stopping distance is further determined from the preset protection curve time sequence group, and the running data of the train in the next control period is output according to the target time sequence data and the current state data of the train, so that the train runs according to the running data. According to the scheme, the predicted running data of the train in the next control period is predicted through the pre-trained running data prediction model, and then when the predicted running data triggers the train safety speed protection curve, the running state of the train is controlled through the output running data of the train in the next control period, the safe running of the train is guaranteed, the operation can be completed without depending on a network, and when the complex situation in the running process of the train is faced, the running data of the train can be output through the pre-obtained predicted running data, so that the safety event which is possibly triggered can be responded to in advance, and the safe running of the train is further guaranteed.
With reference to the technical solution provided by the first aspect, in some possible implementations, the determining a set of target time series data including the target predicted train speed and the target predicted stopping distance from a preset protection curve time series group includes: acquiring at least one group of first time sequence data with a second parking distance equal to the target predicted parking distance in the protection curve time sequence group; determining at least one set of second time series data, from the first time series data, of which the error between a second train speed corresponding to a second parking distance equal to the target predicted parking distance and the target predicted train speed is within a preset threshold value; and determining target time-series data from the second time-series data.
In the embodiment of the application, at least one group of first time series data with a second stopping distance equal to a target predicted stopping distance exists in a protection curve time series group, at least one group of second time series data with an error between a second train speed corresponding to the second stopping distance equal to the target predicted stopping distance and the target predicted train speed within a preset threshold value is determined from the first time series data, and finally, target time series data is determined from the second time series data.
In combination with the technical solution provided by the first aspect, in some possible implementations, the method further includes: acquiring a training data set, wherein the training data set comprises historical driving data of N different control periods, and each historical driving data comprises the driving speed of a train at different time and the stopping distance of the train at different time; and training the driving data prediction model by using the training data set, and updating parameters of the driving data prediction model by using a back propagation method and a gradient descent method during training until an error value of the predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain the trained driving data prediction model.
In the embodiment of the application, the driving data prediction model is trained by using the training data set comprising the historical driving data of N different control periods, and the parameters of the driving data prediction model are updated by a back propagation method and a gradient descent method, so that the predicted driving data of the train in the next control period, which is predicted by the finally obtained trained driving data prediction model, is more accurate.
With reference to the technical solution provided by the first aspect, in some possible implementations, the method further includes: acquiring actual running data of the train in the same period as the predicted running data; determining a prediction error value based on the actual driving data and the predicted driving data; and carrying out error compensation correction on the output value of the pre-trained running data prediction model based on the prediction error value.
In the embodiment of the application, the accuracy of the output value of the pre-trained running data prediction model is further improved by acquiring the actual running data of the train in the same period as the predicted running data, determining the prediction error value based on the actual running data and the predicted running data, and performing error compensation correction on the output value of the pre-trained running data prediction model based on the prediction error value.
In a second aspect, the present application provides a driving data prediction model training method, including: acquiring a training data set, wherein the training data set comprises historical driving data of N different control periods, and each historical driving data comprises the driving speed of a train at different time and the stopping distance of the train at different time; and training the driving data prediction model by using the training data set, and updating parameters of the driving data prediction model by using a back propagation method and a gradient descent method during training until an error value of the predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain the trained driving data prediction model.
In a third aspect, the present application provides a train safety device, comprising: the train safety stopping system comprises an acquisition module and a processing module, wherein the acquisition module is used for acquiring first running data of a train in a preset time length, and the first running data comprises first running speeds of the train at different moments and a first stopping distance representing the distance between the train and a safety stopping point; the processing module is used for obtaining the predicted running data of the train in the next control period based on the first running data and a pre-trained running data prediction model, wherein the next control period is a preset time length; the processing module is further used for acquiring a target predicted train speed and a target predicted stopping distance when the train safety speed protection curve is triggered for the first time in the predicted running data if the predicted running data triggers the train safety speed protection curve; the processing module is further configured to determine a set of target time series data including the target predicted train speed and the target predicted stopping distance from a preset protection curve time series group, where the protection curve time series group includes a plurality of sets of second train speeds and second stopping distances of trains at different times; the processing module is further used for outputting the running data of the train in the next control period according to the target time sequence data and the current state data of the train, so that the train runs according to the running data.
With reference to the technical solution provided by the third aspect, in some possible embodiments, the processing module is specifically configured to obtain at least one set of first time series data in the protection curve time series group, where a second parking distance is equal to the target predicted parking distance; determining at least one set of second time series data, from the first time series data, of which the error between a second train speed corresponding to a second parking distance equal to the target predicted parking distance and the target predicted train speed is within a preset threshold value; and determining target time-series data from the second time-series data.
With reference to the technical solution provided by the third aspect, in some possible embodiments, the train safety protection device further includes a training module, where the training module is configured to obtain a training data set, where the training data set includes historical driving data of N different control cycles, and each of the historical driving data includes a driving speed of a train at different time and a stopping distance of the train at different time; the training module is further used for training the driving data prediction model by using the training data set, and during training, updating parameters of the driving data prediction model by using a back propagation method and a gradient descent method until an error value between the predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain the trained driving data prediction model.
In a fourth aspect, an embodiment of the present application provides an electronic device, including: a memory and a processor, the memory and the processor being connected; the memory is used for storing programs; the processor is configured to invoke a program stored in the memory to perform the method according to the embodiment of the first aspect and/or any possible implementation manner in combination with the embodiment of the first aspect, and/or perform the method according to the embodiment of the second aspect.
In a fifth aspect, embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, where the computer program is executed by a computer to perform the method according to the first aspect as described above and/or any one of the possible implementation modes in combination with the first aspect as described above, and/or to perform the method according to the second aspect as described above.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained from the drawings without inventive effort.
Fig. 1 is a schematic flow chart of a train safety protection method according to an embodiment of the present disclosure;
FIG. 2 is a schematic diagram of a safety speed protection curve according to an embodiment of the present application;
FIG. 3 is a schematic flow chart diagram illustrating a model training method according to an embodiment of the present disclosure;
FIG. 4 is a block diagram of a train safety device according to an embodiment of the present disclosure;
FIG. 5 is a block diagram of a model training apparatus according to an embodiment of the present disclosure;
fig. 6 is a block diagram illustrating a structure of an electronic device according to an embodiment of the present application;
fig. 7 is a schematic structural diagram of another electronic device according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.
It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures. Meanwhile, relational terms such as "first," "second," and the like may be used solely in the description herein to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
Further, the term "and/or" in the present application is only one kind of association relationship describing the associated object, and means that three kinds of relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist simultaneously, and B exists alone.
The technical solution of the present application will be clearly and completely described below with reference to the accompanying drawings.
Referring to fig. 1, fig. 1 is a schematic diagram illustrating a train safety protection method according to an embodiment of the present application, and steps included in the method will be described with reference to fig. 1.
S100: the method comprises the steps of obtaining first running data of a train within a preset time length.
The first running data of the train in the preset time length can be acquired in advance and stored in a database or a magnetic disk, and can be acquired directly when needed, or acquired in real time when needed.
The preset time length may be a fixed time length such as 10 seconds, 30 seconds, 50 seconds, 1 minute, 2 minutes, 3 minutes, or the like, or the preset time length may also be a time length required for the train to travel from the first preset location to the second preset location, for example, when the train needs 5 minutes to travel from the a station to the B station, the preset time length at this time is 5 minutes, and accordingly, the first travel data of the train within the preset time length is obtained, and the first travel data of the train within 5 minutes to travel from the a station to the B station is obtained.
The first travel data includes a first travel speed of the train at different times, a first stopping distance characterizing a distance of the train from a safe stopping point. Taking n seconds as the preset time length and 1 second as the interval between two adjacent moments as an example, at the 1 st second of the preset time length, the first running speed is v1, and the first stopping distance is s1; in the 2 nd second of the preset time length, the first running speed is v2, the first stopping distance is s2 … … in the jth second of the preset time length, the first running speed is vj, the first stopping distance is sj … … in the nth second of the preset time length, the first running speed is vn, and the first stopping distance is sn. The foregoing examples are merely for convenience of understanding, and the interval between two adjacent time instants can be set according to actual requirements, and is not limited herein.
Optionally, the first driving data may further include first accelerations of the train at different times, and positions of the train at different times.
S200: and obtaining the predicted running data of the train in the next control period based on the first running data and the pre-trained running data prediction model.
Wherein the next control period is a preset time length. Optionally, when the preset time length in S100 is a fixed time length, the next control period may be the same as the preset time length, for example, when the preset time length is n minutes, the next control period may also be n minutes. When the preset time length in S100 is a time length required for the operation from the first preset location to the second preset location, the next control cycle may be a time length required for the operation from the second preset location to the third preset location. Or, the next control period may also be set independently, without taking the preset time length in S100 as a reference, and at this time, the specific size of the next control period may be set according to actual requirements, and the specific setting manner is not limited herein.
Optionally, the preset time length in S100 may also be the current control cycle of the train, that is, the predicted running data of the train in the next control cycle is obtained by predicting the running data of the current control cycle and the running data prediction model trained in advance.
The specific form of the predicted travel data of the next control cycle is the same as that of the first travel data in S100, and is not described here again for the sake of brevity.
The pre-trained driving data prediction model may be trained by a third party, or may be trained by the present application. In one embodiment, the training process for training the driving data prediction model may be: the method comprises the steps of firstly obtaining a training data set, wherein the training data set comprises N historical driving data of different control periods, each historical driving data comprises the driving speed of a train at different time and the stopping distance of the train at different time, then training a driving data prediction model by using the training data set, updating parameters of the driving data prediction model by using a back propagation method and a gradient descent method during training until an error value of the ith historical driving data corresponding to the predicted driving data and the (i + 1) th historical driving data meets a preset condition, and obtaining the trained driving data prediction model. The specific form of the historical travel data for each control cycle coincides with the form of the first travel data in S100, and is not described here again for the sake of brief description.
Wherein the preset condition may be presetAnd setting an error threshold, and when the error value of the predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data is smaller than the preset error threshold, determining that the driving data prediction model is trained. Or, the preset condition may be that the error index corresponding to the driving data prediction model is smaller than a preset error threshold, and J represents the error index, J =1/2[y (i) -y (i + 1)] 2 And y (i) is an error value between the predicted driving data and the i +1 th historical driving data corresponding to the ith historical driving data, and y (i + 1) is an error value between the predicted driving data and the i +2 th historical driving data corresponding to the i +1 th historical driving data, and when J is smaller than a preset error threshold value, the driving data prediction model is considered to be trained.
Optionally, before the driving data prediction model is trained by using the acquired training data set, initial historical driving data of the train in N different control periods is acquired, then wavelet denoising preprocessing is performed on each initial historical driving data, and the N historical driving data after wavelet denoising preprocessing are used as the training data set.
In order to further reduce the error between the predicted travel data and the actual travel data predicted by the travel data prediction model, in one embodiment, the actual travel data of the train in the same period as the predicted travel data may be obtained, the prediction error value may be determined based on the actual travel data and the predicted travel data, and the error compensation correction may be performed on the output value of the travel data prediction model trained in advance based on the prediction error value.
For example, when the travel speed of the actual travel data at the k-th time is represented by v (k) and the travel speed of the predicted travel data at the k-th time is represented by v (k) 'with respect to the data at the same time in the predicted travel data and the actual travel data, v (k) = v (k)' + e v (k) Wherein e is v (k) Indicates an error between the predicted travel speed and the actual travel speed at the k-th time by the trained travel data prediction model, and if the trained travel data prediction model predicts again the predicted travel speed at the k-th time, v (k)' = v (k) is obtained) 1 ″+e v (k) Wherein v (k) 1 "indicates the predicted traveling speed at the k-th time, and v (k)" indicates the predicted traveling speed at the k-th time after the correction. The manner of performing compensation correction on other types of data included in the predicted traveling data is the same as the manner of performing compensation correction on the predicted traveling data exemplified herein, and for brevity, details thereof are not repeated herein.
Optionally, the driving data prediction model may be a fuzzy neural network model, for example, a fuzzy neural network model including a 5-layer grid structure, the 1 st layer is an input layer for completing input of driving data, the 2 nd layer is a fuzzy layer, fuzzy membership of each input quantity is calculated according to a result of a clustering algorithm by using gaussian membership, the 3 rd layer is a fuzzy calculation layer for completing calculation of a front part parameter of a rule, the 4 th layer performs fuzzy inference and calculates output of a fuzzy rule, and completes calculation of a back part parameter of the rule, and the 5 th layer is an output layer for completing output of a control quantity in a sharpening process, that is, outputting the predicted driving data.
Optionally, the optimal control is used to achieve the effect of following the objective function through rolling optimization. In order to obtain the control law, the sum of the minimized performance index, namely the error, and the control quantity added with the weight factor is used as the performance index J. The performance index J can be obtained by calculating in the following way:
wherein the closer the evaluation index J is to 1, the higher the accuracy of the evaluation result is; otherwise, the lower. di is the target distance calculated at the ith time, and dv is the velocity of the ith particle.
For the optimal value of the target distance at the ith time, d error The parking accuracy is obtained.
TABLE 1
As shown in table 1, the area difference and the curve difference of the curve (speed distance curve) corresponding to the driving data predicted by the prediction model in the prior art are both greater than the area difference and the curve difference of the curve (speed distance curve) corresponding to the driving data prediction model in the present embodiment, and the evaluation index J of the driving data prediction model in the present embodiment is 0.968759, which is also greater than the evaluation index 0.685991 in the prior art.
S300: and if the predicted running data triggers a preset safe speed protection curve of the train, acquiring the target predicted train speed and the target predicted stopping distance when the safe speed protection curve of the train is triggered for the first time in the predicted running data.
Wherein, the safety speed protection curve is a relation curve of speed and distance, as shown in fig. 2, which represents that the parking distance is d 1 When the maximum running speed of the train is v 1 If the running speed of the train exceeds the maximum running speed v 1 And determining that the train safety speed protection curve is triggered, namely determining that the coordinates of the running speed and the stopping distance in the safety speed protection curve are positioned outside the range enclosed by the safety speed protection curve and the horizontal and vertical coordinate axes, and determining that the train safety speed protection curve is triggered. For example, the predicted travel data includes t 1 、t 2 、t 3 、t 4 、t 5 、t 6 、t 7 、t 8 、t 9 The predicted travel speed and the predicted stopping distance at these 9 times, when t 5 Predicted travel speed corresponding to timeDegree v 5 Corresponding to a predicted stopping distance of s 5 If the safety speed protection curve is neutralized with s 5 The maximum running speed of the train corresponding to the equal stopping distance d is v, and v is less than the predicted running speed v 5 Then confirm the target predicted train speed as v 5 Target predicted stopping distance is s 5 。
S400: and determining a set of target time series data comprising target predicted train speed and target predicted stopping distance from a preset protection curve time series set.
And the protection curve time sequence number group comprises a plurality of groups of second train speeds and second stopping distances of the trains at different moments. The data representation form of the second train speed and the second stopping distance of each group of trains at different time is the same as the first running data in S100, and for the sake of brief description, details are not repeated here.
In one embodiment, the specific process of S400 may be that, first, at least one set of first time series data in which the second parking distance is equal to the target predicted parking distance is obtained from the protection curve time series group; and then determining at least one set of second time series data, from the first time series data, of which the error between the second train speed corresponding to a second parking distance equal to the target predicted parking distance and the target predicted train speed is within a preset threshold value, and determining the target time series data from the second time series data.
For example, when the target predicts the stopping distance as s m Target predicted travel speed is v m The preset threshold value is E, the protection curve time sequence group comprises 10 groups of sequence data, each group of sequence data comprises a second train speed and a second stopping distance of the train at different moments, and 4 groups of sequence data comprise a target predicted stopping distance s m Equal second stopping distance, s in the first sequence data of 4 sequence data m The second train speed corresponding to the equal second stopping distance is v 1m In the second set of sequence data with s m The second train speed corresponding to the equal second stopping distance is v 2m Second set of sequence numbersAccording to neutralization of s m The second train speed corresponding to the equal second stopping distance is v 3m In the second set of sequence data with s m The second train speed corresponding to the equal second stopping distance is v 4m Wherein, | v m -v 1m |>E，|v m -v 2m |<E，|v m -v 3m |＝E，|v m -v 4m |>E, the second time series data includes v 2m And v 3m And corresponding sequence data, and determining a set of sequence data from the second time-series data as target time-series data.
Determining the target time-series data from the second time-series data may be selecting, as the target time-series data, any one set of the second time-series data, or may be selecting, as the target time-series data, sequence data in which a second train speed corresponding to a second stopping distance equal to the target predicted stopping distance is smaller than the target predicted train speed, from the second time-series data.
Optionally, if there is sequence data in which a second train speed corresponding to a second stopping distance equal to the target predicted stopping distance is equal to the target predicted train speed in the first time-series data, the sequence data is regarded as target time-series data. Alternatively, the first time-series data may be set to include, as the target time-series data, sequence data in which an error between the target predicted train speed and a second train speed corresponding to a second stopping distance equal to the target predicted stopping distance is minimized.
S500: and outputting the running data of the train in the next control period according to the target time sequence data and the current state data of the train.
And obtaining operation data through the target time sequence data and the current state data of the train, so that the train operates in the next period according to the operation data, and the safe operation of the train is guaranteed.
In one embodiment, the operation data of the train in the next control cycle is obtained based on the target time series data, the current state data of the train and a pre-trained operation data prediction model after the target time series data and the current state data of the train are obtained.
The training process of the trained operation data prediction model may be that, first, a training sample set is obtained, where the training sample set includes target time series data of multiple control cycles and current state data of a train, and actual operation data of a next control cycle corresponding to the target time series data of each cycle and the current state data of the train, the operation data prediction model is trained by using the training sample set, during training, the target time series data of the same control cycle and the current state data of the train are input into the operation data prediction model to obtain corresponding predicted operation data, based on an error between the predicted operation data and the actual operation data corresponding to the control cycle, parameters of the operation data prediction model are updated by using a back propagation method and a gradient descent method, until an error between the predicted operation data and the actual operation data corresponding to the predicted operation data is smaller than a preset threshold, the preset threshold may be set according to an actual demand, where no limitation is made on a specific numerical value thereof, but it should be understood that a smaller value of the preset threshold is, a higher accuracy of the predicted operation data is.
Based on the same inventive concept, the present application also provides a model training method, which will be described with reference to fig. 3, where fig. 3 is a schematic flow chart of the model training method.
S600: acquiring a training data set, wherein the training data set comprises historical driving data of N different control cycles, and each historical driving data comprises the driving speed of a train at different time and the stopping distance of the train at different time.
S700: and training the driving data prediction model by using the training data set, and updating parameters of the driving data prediction model by using a back propagation method and a gradient descent method during training until an error value of the predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain the trained driving data prediction model.
The specific process and principle of the model training method are described clearly in the foregoing, and are not repeated here for the sake of brief description.
The implementation principle and the generated technical effect of the model training method provided in the embodiment of the present application are the same as those of the model training process described in the embodiment of the train safety protection method, and for brief description, reference may be made to corresponding contents in the embodiment of the train safety protection method where no part of the embodiment of the model training method is mentioned.
Referring to fig. 4, fig. 4 is a schematic diagram of a train safety protection device 100 according to an embodiment of the present disclosure, where the train safety protection device 100 includes an obtaining module 110 and a processing module 120.
The obtaining module 110 is configured to obtain first driving data of the train within a preset time length, where the first driving data includes first driving speeds of the train at different times and a first stopping distance representing a distance between the train and a safe stopping point.
And the processing module 120 is configured to obtain predicted running data of the train in a next control period based on the first running data and a pre-trained running data prediction model, where the next control period is a preset time length.
The processing module 120 is further configured to, if the predicted travel data triggers a preset safe speed protection curve of the train, obtain a target predicted train speed and a target predicted stopping distance when the safe speed protection curve of the train is triggered for the first time in the predicted travel data.
The processing module 120 is further configured to determine a set of target time series data including the target predicted train speed and the target predicted stopping distance from a preset protection curve time series group, where the protection curve time series group includes a plurality of sets of second train speeds and second stopping distances of the train at different times.
The processing module 120 is further configured to output operation data of the train in the next control period according to the target time series data and the current state data of the train, so that the train operates according to the operation data.
The processing module 120 is specifically configured to obtain at least one set of first time series data in the protection curve time series group, where a second parking distance is equal to the target predicted parking distance; determining at least one group of second time series data of which the error between a second train speed corresponding to a second parking distance equal to the target predicted parking distance and the target predicted train speed is within a preset threshold value from the first time series data; determining target time-series data from the second time-series data.
The processing module 120 is further configured to obtain actual driving data of the train in the same period as the predicted driving data; determining a prediction error value based on the actual driving data and the predicted driving data; and carrying out error compensation correction on the output value of the pre-trained driving data prediction model based on the prediction error value.
In one embodiment, the train safety protection device 100 further includes a training module, where the training module is configured to obtain a training data set, where the training data set includes historical driving data of N different control cycles, and each of the historical driving data includes a driving speed of a train at different time and a stopping distance of the train at different time; the training module is further configured to train the driving data prediction model by using the training data set, and during training, update parameters of the driving data prediction model by using a back propagation method and a gradient descent method until an error value between predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain a trained driving data prediction model.
The train safety protection device 100 provided in the embodiment of the present application has the same implementation principle and technical effect as those of the train safety protection method embodiment described above, and for brief description, reference may be made to corresponding contents in the train safety protection method embodiment described above where no mention is made in the device embodiment.
Referring to fig. 5, fig. 5 is a block diagram of a model training apparatus 300 according to an embodiment of the present disclosure, which includes a training data obtaining module 310 and a training module 320.
The training data acquiring module 310 is configured to acquire a training data set, where the training data set includes historical driving data of N different control cycles, and each historical driving data includes a driving speed of a train at different time and a stopping distance of the train at different time.
The training module 320 is configured to train the driving data prediction model by using the training data set, and during training, update parameters of the driving data prediction model by using a back propagation method and a gradient descent method until an error value between predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain a trained driving data prediction model.
The model training apparatus 300 according to the embodiment of the present application has the same implementation principle and technical effect as those of the foregoing model training method embodiment, and for brief description, reference may be made to the corresponding contents in the foregoing model training method embodiment for the part of the apparatus embodiment that is not mentioned.
Please refer to fig. 6, which is an electronic device 200 according to an embodiment of the present disclosure. The electronic device 200 includes: a transceiver 210, a memory 220, a communication bus 230, and a processor 240.
The elements of the transceiver 210, the memory 220, and the processor 240 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, the components may be electrically coupled to each other via one or more communication buses 230 or signal lines. The transceiver 210 is used for transceiving data. The memory 220 is used to store a computer program, such as the software functional module shown in fig. 4, i.e., the train safety guard 100. The train safety protection device 100 includes at least one software function module, which may be stored in the memory 220 in the form of software or firmware (firmware) or solidified in an Operating System (OS) of the electronic device 200. The processor 240 is configured to execute executable modules stored in the memory 220, such as software functional modules or computer programs included in the train safety device 100.
At this time, the processor 240 is configured to obtain first driving data of the train within a preset time length, where the first driving data includes a first driving speed of the train at different times and a first stopping distance representing a distance between the train and a safe stopping point; obtaining the predicted running data of the train in the next control period based on the first running data and a pre-trained running data prediction model, wherein the next control period is a preset time length; if the predicted running data triggers a preset safe speed protection curve of the train, acquiring a target predicted train speed and a target predicted stopping distance when the safe speed protection curve of the train is triggered for the first time in the predicted running data; determining a set of target time sequence data comprising the target predicted train speed and the target predicted stopping distance from a preset protection curve time sequence set, wherein the protection curve time sequence set comprises a plurality of sets of second train speeds and second stopping distances of trains at different moments; and outputting the operation data of the train in the next control period according to the target time sequence data and the current state data of the train so as to enable the train to operate according to the operation data.
The Memory 220 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.
Optionally, the processor may receive Radio Frequency Identification information of an RFID (Radio Frequency Identification) and information of a GNSS (global navigation satellite system) navigation positioning module through a UART (Universal Asynchronous Receiver/Transmitter), where the Radio Frequency Identification information enables the processor to achieve accurate positioning of the control terminal and non-contact data communication.
In one embodiment, the electronic device 200 further includes an Input/Output (I/O) signal processing module and a display module.
The I/O signal processing module comprises an STM32F407 chip, and the chip reads AI (direct current analog quantity input) and DI (switching value input) signals of 4-20 MA (milliampere) of the train sensor which is subjected to protective measures such as detection and isolation through a transmission mode of SPI (Serial Peripheral Interface), and reflects whether the state of the switching value is in a split state or in a combined state. And a DO (switching value output) command is transmitted to the train through safety protection circuits such as feedback, self-checking and isolation.
The display module comprises an S3C2440 chip for operating the linux system, and the chip exchanges information with the processor 240 through an Ethernet (ETH), so that the functions of displaying system operation information, manually setting, overspeed alarm and the like are completed. For example, manual setting can be realized by setting a key, an overspeed alarm function can be realized by a buzzer and a sound alarm module, and a USB interface can be further arranged, so that data exchange with other equipment is facilitated.
In order to facilitate understanding of the connection relationship among the processor 240, the I/O signal processing module, and the display module, an implementation of an electronic device including the processor 240, the I/O signal processing module, and the display module is shown in fig. 7. The processor 240 and the I/O signal processing module perform data interaction through two CANs (Controller Area networks). The EMI (electromagnetic interference) protection module is used for solving the problem of electromagnetic interference of electronic equipment. The DC (power supply) module is used to supply power to the electronic device, and the electronic device 200 may be provided with a battery to supply power.
In the GPS measurement, for example, the static, fast static and dynamic measurement need to be solved afterwards to obtain the centimeter-level precision. The hardware part of the invention carries an RTK (Real-TIme kinematic) radio station, and by a Real-TIme differential positioning method, centimeter-level positioning accuracy measurement can be obtained outdoors in Real TIme, the invention is not limited to indoor operation, and the outdoor operation efficiency is greatly improved.
The electronic device 200 includes, but is not limited to, a personal computer, a server, and the like.
The embodiment of the present application further provides a computer-readable storage medium (hereinafter, referred to as a storage medium), where a computer program is stored on the storage medium, and when the computer program is run by the electronic device 200 as described above, the train safety protection method described above is executed.
The computer-readable storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, an optical disk, or other various media capable of storing program codes.
The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made to the present application by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (7)
1. A train safety protection method is characterized by comprising the following steps:
acquiring first running data of a train in a preset time span, wherein the first running data comprises first running speeds of the train at different moments and first stopping distances representing distances between the train and a safe stopping point;
obtaining the predicted running data of the train in the next control period based on the first running data and a pre-trained running data prediction model, wherein the next control period is a preset time length;
if the predicted running data triggers a preset safe speed protection curve of the train, acquiring a target predicted train speed and a target predicted stopping distance when the safe speed protection curve of the train is triggered for the first time in the predicted running data;
determining a set of target time sequence data comprising the target predicted train speed and the target predicted stopping distance from a preset protection curve time sequence set, wherein the protection curve time sequence set comprises a plurality of sets of second train speeds and second stopping distances of trains at different moments;
outputting operation data of the train in the next control cycle according to the target time-series data and the current state data of the train so that the train operates according to the operation data,
the determining a set of target time series data including the target predicted train speed and the target predicted stopping distance from a preset protection curve time series set includes:
acquiring at least one group of first time sequence data with a second parking distance equal to the target predicted parking distance in the protection curve time sequence group;
determining at least one set of second time series data, from the first time series data, of which the error between a second train speed corresponding to a second parking distance equal to the target predicted parking distance and the target predicted train speed is within a preset threshold value;
determining target time-series data from the second time-series data.
2. The method of claim 1, further comprising:
acquiring a training data set, wherein the training data set comprises historical driving data of N different control periods, and each historical driving data comprises the driving speed of a train at different time and the stopping distance of the train at different time;
and training the driving data prediction model by using the training data set, and updating parameters of the driving data prediction model by using a back propagation method and a gradient descent method during training until an error value of the predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain the trained driving data prediction model.
3. The method of claim 1, further comprising:
acquiring actual running data of the train in the same period as the predicted running data;
determining a prediction error value based on the actual driving data and the predicted driving data;
and carrying out error compensation correction on the output value of the pre-trained running data prediction model based on the prediction error value.
4. A train safety device, comprising:
the system comprises an acquisition module, a storage module and a control module, wherein the acquisition module is used for acquiring first running data of a train in a preset time length, and the first running data comprises first running speeds of the train at different moments and a first stopping distance representing the distance between the train and a safe stopping point;
the processing module is used for obtaining the predicted running data of the train in the next control period based on the first running data and a pre-trained running data prediction model, wherein the next control period is a preset time length;
the processing module is further configured to obtain a target predicted train speed and a target predicted stopping distance when the train safety speed protection curve is triggered for the first time in the predicted running data if the predicted running data triggers a preset safety speed protection curve of the train;
the processing module is further configured to determine a set of target time series data including the target predicted train speed and the target predicted stopping distance from a preset protection curve time series group, where the protection curve time series group includes a plurality of sets of second train speeds and second stopping distances of trains at different times;
the processing module is further configured to output operation data of the train in the next control cycle according to the target time series data and the current state data of the train, so that the train operates according to the operation data;
the processing module is specifically configured to acquire at least one set of first time series data in the protection curve time series group, where a second parking distance is equal to the target predicted parking distance; determining at least one set of second time series data, from the first time series data, of which the error between a second train speed corresponding to a second parking distance equal to the target predicted parking distance and the target predicted train speed is within a preset threshold value; and determining target time-series data from the second time-series data.
5. The train safety device of claim 4, further comprising:
the training module is used for acquiring a training data set, wherein the training data set comprises historical driving data of N different control periods, and each historical driving data comprises the driving speed of a train at different time and the stopping distance of the train at different time;
the training module is further configured to train the driving data prediction model by using the training data set, and during training, update parameters of the driving data prediction model by using a back propagation method and a gradient descent method until an error value between predicted driving data corresponding to the ith historical driving data and the (i + 1) th historical driving data meets a preset condition, so as to obtain a trained driving data prediction model.
6. An electronic device, comprising: a memory and a processor, the memory and the processor connected;
the memory is used for storing programs;
the processor to invoke a program stored in the memory to perform the method of any of claims 1-3.
7. A computer-readable storage medium, having stored thereon a computer program which, when executed by a computer, performs the method of any one of claims 1-3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210248260.5A CN114559992B (en) | 2022-03-14 | 2022-03-14 | Train safety protection method, model training method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210248260.5A CN114559992B (en) | 2022-03-14 | 2022-03-14 | Train safety protection method, model training method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114559992A CN114559992A (en) | 2022-05-31 |
| CN114559992B true CN114559992B (en) | 2022-12-02 |
Family
ID=81719565
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210248260.5A Active CN114559992B (en) | 2022-03-14 | 2022-03-14 | Train safety protection method, model training method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114559992B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115219250A (en) * | 2022-06-23 | 2022-10-21 | 中车大连电力牵引研发中心有限公司 | Motor train unit power vehicle IO loop test method |
| CN115257865B (en) * | 2022-07-25 | 2023-09-08 | 青岛海信微联信号有限公司 | Train control method, equipment and device |
| CN115424368A (en) * | 2022-08-25 | 2022-12-02 | 武汉迪昌科技有限公司 | Unpowered grouping test method and device for motor train unit |
| CN116015785B (en) * | 2022-12-14 | 2024-04-30 | 中国联合网络通信集团有限公司 | Information security protection method, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109625033A (en) * | 2018-12-13 | 2019-04-16 | 交控科技股份有限公司 | ATO controls accurate shutdown phase stopping accuracy prediction technique and device when vehicle |
| CN109783843A (en) * | 2018-12-05 | 2019-05-21 | 交控科技股份有限公司 | Analogue system vehicle dynamic model speed predicting method based on LSTM |
| CN112198799A (en) * | 2020-10-28 | 2021-01-08 | 北京交通大学 | High-speed train parking control method and system based on deep learning |
| CN112590871A (en) * | 2020-12-23 | 2021-04-02 | 交控科技股份有限公司 | Train safety protection method, device and system |
| CN113762464A (en) * | 2021-07-30 | 2021-12-07 | 北京交通大学 | Train operation reference curve dynamic generation method based on learning |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016035597A1 (en) * | 2014-09-05 | 2016-03-10 | 三菱電機株式会社 | Automated train operation system |
-
2022
- 2022-03-14 CN CN202210248260.5A patent/CN114559992B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109783843A (en) * | 2018-12-05 | 2019-05-21 | 交控科技股份有限公司 | Analogue system vehicle dynamic model speed predicting method based on LSTM |
| CN109625033A (en) * | 2018-12-13 | 2019-04-16 | 交控科技股份有限公司 | ATO controls accurate shutdown phase stopping accuracy prediction technique and device when vehicle |
| CN112198799A (en) * | 2020-10-28 | 2021-01-08 | 北京交通大学 | High-speed train parking control method and system based on deep learning |
| CN112590871A (en) * | 2020-12-23 | 2021-04-02 | 交控科技股份有限公司 | Train safety protection method, device and system |
| CN113762464A (en) * | 2021-07-30 | 2021-12-07 | 北京交通大学 | Train operation reference curve dynamic generation method based on learning |
Non-Patent Citations (2)
| Title |
|---|
| 基于ARM与多模型实现的试运线列车超速防护系统;姜俊彤 等;《重庆理工大学学报（自然科学）》;20210131;第35卷(第1期);第180-187页 * |
| 姜俊彤 等.基于ARM与多模型实现的试运线列车超速防护系统.《重庆理工大学学报（自然科学）》.2021,第35卷(第1期),第180-187页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114559992A (en) | 2022-05-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114559992B (en) | Train safety protection method, model training method and device and electronic equipment | |
| CN109597864A (en) | Instant positioning and map constructing method and the system of ellipsoid boundary Kalman filtering | |
| CN103152826A (en) | Moving target tracking method based on NLOS (non line of sight) state inspection compensation | |
| CN110471055B (en) | Method and device for predicting flight path, readable storage medium and terminal equipment | |
| CN113326964B (en) | Data-based composite multi-stage comprehensive pipe gallery risk management and control method | |
| CN112763995A (en) | Radar calibration method and device, electronic equipment and road side equipment | |
| CN102981160B (en) | Method and device for ascertaining aerial target track | |
| CN113203419A (en) | Indoor inspection robot correction positioning method based on neural network | |
| CN105424043A (en) | Motion state estimation method based on maneuver judgment | |
| Yan | Positioning of logistics and warehousing automated guided vehicle based on improved LSTM network | |
| Fan et al. | Cueing Flight Object Trajectory and Safety Prediction Based on SLAM Technology | |
| Papathanasopoulou et al. | Flexible car-following models incorporating information from adjacent lanes | |
| CN112037559A (en) | Indoor parking dynamic accurate navigation system based on wireless positioning technology | |
| CN111954188B (en) | Robot control method, device, terminal and medium | |
| CN109379716A (en) | A kind of indoor orientation method and system for safety monitoring project | |
| CN114510037A (en) | Intelligent road condition analysis method and system based on unmanned driving | |
| CN110837903A (en) | Rail transit short-term passenger flow prediction method and system and electronic equipment | |
| CN109831737B (en) | Bluetooth positioning method, device, equipment and system based on confidence degree | |
| CN113848828B (en) | Automatic obstacle avoidance control method, device, equipment and medium for AGV trolley | |
| Zhang et al. | The research for a kind of information fusion model based on BP neural network with multi position sources and big data selection | |
| CN115598592B (en) | Time-frequency difference joint positioning method, system, electronic equipment and storage medium | |
| Azenha et al. | A neural network approach for AGV localization using trilateration | |
| CN115508773B (en) | Multi-station passive positioning method and system by time difference method, electronic equipment and storage medium | |
| WO2024012006A1 (en) | Positioning precision estimation method and apparatus, electronic device and storage medium | |
| CN117406754B (en) | Logistics robot environment sensing and obstacle avoidance method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |