Train automatic driving evaluation method and device

Abstract

The invention discloses an evaluation method and device for automatic driving of a train. The method comprises the following steps: acquiring first historical operation data of a train; according to the first historical operation data, evaluation indexes of the train in multiple dimensions are calculated; wherein the evaluation indexes of the train in multiple dimensions comprise a comfort index, an efficiency index, a stability index, an energy-saving index, an economic index, an accuracy index and a robustness index; and inputting the evaluation indexes of the train in the multiple dimensions into a pre-trained automatic driving evaluation model to obtain an evaluation result of automatic driving of the train. According to the scheme provided by the embodiment of the invention, the comprehensive evaluation capability and efficiency of the control effect of the ATO system in multiple dimensions can be improved.

Description

Technical Field

[0001] This article relates to the field of transportation engineering technology, and in particular to an evaluation method and device for automatic train driving. Background Art

[0002] The automatic train operation (ATO) system can automatically control complex parameters, which not only ensures the safety of train operation, but also improves performance indicators such as operation efficiency, passenger comfort, energy saving and parking accuracy. The train driving process is a multi-objective control process. So far, the research on the evaluation indicators and evaluation methods of the automatic train operation control system has not been perfected. Summary of the invention

[0003] The embodiments of the present application provide a method and device for evaluating automatic train driving, which can improve the comprehensive evaluation capability and efficiency of the control effects of the ATO system in multiple dimensions.

[0004] An embodiment of the present application provides a method for evaluating automatic train driving, including:

[0005] Obtaining the first historical running data of the train;

[0006] Calculate evaluation indicators of the train in multiple dimensions according to the first historical operation data; wherein the evaluation indicators of the train in multiple dimensions include: comfort index, efficiency index, stability index, energy saving index, economy index, accuracy index, and robustness index;

[0007] The evaluation indicators of the train in multiple dimensions are input into a pre-trained automatic driving evaluation model to obtain an evaluation result of the automatic driving of the train.

[0008] In an exemplary embodiment, the pre-trained autonomous driving evaluation model is trained according to the following method:

[0009] Acquire multiple sets of second historical running data of the train;

[0010] For each set of the second historical running data, the following steps are performed respectively until a preset training stop condition is met:

[0011] Calculating evaluation indicators of the train in multiple dimensions according to the set of the second historical operation data, and inputting the evaluation indicators into the automatic driving evaluation model to obtain a first evaluation result;

[0012] Obtaining N second evaluation results obtained after comprehensive evaluation by N experts in the industry on the evaluation indicators of the train corresponding to the group of the second historical operation data in multiple dimensions; taking the average value of the N second evaluation results as the third evaluation result; wherein N is an integer greater than or equal to 1;

[0013] According to the first evaluation result and the third evaluation result, parameters of the autonomous driving evaluation model are optimized.

[0014] In an exemplary embodiment, inputting the evaluation indicators of the train in multiple dimensions into a pre-trained automatic driving evaluation model to obtain the evaluation result of the automatic driving of the train includes:

[0015] Determining, according to the adjustment mode of the train and the operation schedule of the train, the coding type of the evaluation index of the train in the multiple dimensions;

[0016] The evaluation indicators of the train in the multiple dimensions are combined into a first input vector, and a corresponding encoding vector in the automatic driving evaluation model is selected according to the encoding type to encode the first input vector to obtain a second input vector;

[0017] The second input vector is input into the hidden layer of the neural network of the automatic driving evaluation model for forward propagation to obtain the evaluation result of the automatic driving of the train.

[0018] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include a comfort index, and the comfort index includes a first index, a second index, and a third index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0019] When the train is running in the stepless speed regulation mode, according to the actual speed v and the target speed v during the train running, target , preset speed difference v diff , the running process of the train is divided into acceleration process, deceleration process and uniform speed process; wherein, in the uniform speed process, the difference between the actual speed of the train and the target speed is less than v diff , when |vv target | <v diff When vv target ≥v diff When v target -v≥v diff When , it is the deceleration process;

[0020] Performing filtering processing on a plurality of acceleration values ​​collected during the acceleration process and the deceleration process to obtain a plurality of acceleration filter values; calculating an average value of the absolute values ​​of the plurality of acceleration filter values ​​to obtain an average acceleration value; determining an acceleration evaluation value as the first indicator according to the average acceleration value, the maximum acceleration value and the minimum acceleration value;

[0021] The operation data of the entire line is divided into segments according to the time interval of △t, and the acceleration impact rate of each time period is calculated respectively, and the average value of the absolute value of the acceleration impact rate of all time periods is calculated to obtain the average impact rate value; according to the average impact rate value, the acceleration impact rate evaluation value is determined as the second indicator;

[0022] When the train is running in a stepless speed regulation mode, for each running data sampling position, the level that satisfies the comfort level at the position is calculated according to the ramp acceleration at the position and the maximum tolerance value of the train acceleration under the ramp acceleration; the running time that meets the preset level switching conditions is counted, and the third indicator is determined according to the running time that meets the preset level switching conditions and the running time of the trains on the entire line;

[0023] The preset level switching conditions include: the level size of each position does not exceed L levels; the level maintenance time exceeds the time required for instruction transmission and instruction action; when switching levels, only changes to adjacent levels are allowed, and skipping levels are not allowed; in a continuous acceleration or deceleration process, the level change method should be to increase first and then decrease, and level repetition is not allowed.

[0024] In an exemplary embodiment, the evaluation indicators in the multiple dimensions include an efficiency indicator, and the efficiency indicator includes a fourth indicator, a fifth indicator, and a sixth indicator; and the calculating the evaluation indicators of the train in the multiple dimensions includes:

[0025] When there is an operating timetable, the fourth index is determined according to the ratio of the difference between the actual arrival time and the planned arrival time of the train and the preset threshold; the fifth index is determined according to the difference between the actual arrival time and the planned arrival time of the train and the difference between the actual arrival time and the planned arrival time of the train under the expected state;

[0026] When there is no operating timetable, an iterative method is used to respectively calculate the duration of the traction acceleration phase, the duration of the uniform speed driving phase and the duration of the deceleration driving phase when the train is running at the highest speed; the sum of the duration of the traction acceleration phase, the duration of the uniform speed driving phase and the duration of the deceleration driving phase is determined as the shortest driving duration;

[0027] Calculate the longest running time of the train under the most unfavorable conditions; determine the efficiency evaluation value when there is no operating timetable based on the shortest running time, the longest running time and the actual running time, as the sixth indicator.

[0028] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include an accuracy index, and the accuracy index includes a seventh index, an eighth index, and a ninth index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0029] The parking error evaluation value of the train is obtained according to the following formula as the seventh index:

[0030] S parkmin =S Negative -S factmax

[0031] S max =S factmax +S Negative -v max ×T

[0032]

[0033] In the formula, S factmin ,S factmax are the minimum and maximum values ​​of the difference between the actual parking position and the parking point of the train; S parkmin ,S parkmax are the minimum and maximum values ​​of the difference range between the actual parking position and the estimated parking position of the train, which is determined by the train's benchmark speed; x1 is the error between the parking point and the actual parking position; x2 is the error between the estimated parking position and the actual parking position; S Negative is the size of the parking negative window; v max is the maximum benchmarking speed; T is the sampling period; R7 is the seventh index;

[0034] The benchmark speed evaluation value of the train is obtained according to the following formula as the eighth index:

[0035] v min =v low +v error

[0036]

[0037]

[0038] In the formula, v low is the minimum benchmarking speed supported by the system; v max is the maximum value of the benchmark speed; v min is the minimum value of the benchmark speed; verror is the error caused by speed measurement, distance measurement and line slope; T maxTrans is the maximum braking transmission delay; T minTrans is the minimum braking transmission delay; S error is the distance error caused by speed and distance measurement; a max is the maximum benchmark braking deceleration; a min is the minimum benchmark braking deceleration; R8 is the eighth index;

[0039] The train benchmarking state stability evaluation value is obtained according to the following formula as the ninth index:

[0040] t min2 =t action +t filter

[0041]

[0042] Where, t min2 is the shortest time required to maintain the state before benchmarking; t action is the time for braking and applying the brake before the train is unloaded; t filter is the time required for the filtered speed and acceleration to be the same as the actual situation of the train; R9 is the ninth indicator; the time the train maintains the state before benchmarking is t; t hold It is the maximum tolerable delay error in the line.

[0043] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include a stability index, and the stability index includes a tenth index and an eleventh index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0044] The train overshoot evaluation value is calculated as the tenth index according to the following steps:

[0045] The collected train speed is subjected to high-order delay filtering to obtain the filtered speed {v1,v2…v n};

[0046] The filtered velocity is sliced ​​with t1 as the slice length, {v1,v2…v n} is divided into {v 1,1 ,v 1,2 …v 1,k},{v 2,1 ,v 2,2 …v 2,k},…,{v m,1 ,v m,2 …v m,k};

[0047] For the fragment {v i,1 ,vi,2 …v i,k}, calculate the range r of the fragment i , if r i If it is less than r1, the i-th segment is a uniform speed stage;

[0048] Combine the time-continuous uniform speed segments and calculate the average speed of the uniform speed segments

[0049] Traverse forward from the start time of each uniform speed segment, if v j Greater than v j+1 and v j-1 , then v j is a local maximum if v j Less than v j+1 and v j-1 , then v j is a local minimum; for any local extremum, if v j and a i If the difference is greater than r2, then v j is the overshoot point; the overshoot amount is Where ai is a configuration parameter;

[0050] The calculation formula for the maximum and minimum values ​​of the reasonable range of overshoot is:

[0051]

[0052] The calculation formula of the tenth indicator is:

[0053]

[0054] In the formula, R 10 is the tenth indicator; a min3 , a max3 are the minimum and maximum braking deceleration values ​​allowed when the train reaches the target speed, t ease is the traction brake release time; v min3 , v max3 are the minimum and maximum values ​​of the reasonable range of overshoot respectively;

[0055] The train oscillation degree evaluation value is calculated as the eleventh indicator according to the following steps:

[0056] The collected train speed is subjected to high-order delay filtering to obtain the filtered speed {v1,v2…v n};

[0057] The filtered velocity is segmented with t2 as the segment length and t3 as the step length, and {v1,v2…v n} is divided into {v 1,1 ,v1,2 …v 1,k},{v 2,1 ,v 2,2 …v 2,k},…,{v m,1 ,v m,2 …v m,k};

[0058] For the fragment {v i,1 ,v i,2 …v i,k}, with time as the independent variable and speed as the dependent variable, a linear fit is performed to obtain the equation t = q i v+p i , and calculate the data range r of this segment i ;

[0059] If q i is less than a1, and r i If it is greater than r2, then the segment is a segment where oscillation occurs;

[0060] Merge time-continuous oscillation segments;

[0061] For the element v in the i-th oscillation segment j , if v j Greater than v j+1 and v j-1 , then v j is a local maximum if v j Less than v j+1 and v j-1 , then v j is the local minimum; the oscillation degree is determined as the segment range r i ;

[0062] Among them, t1, t2, t3, r1, r2, a1 are all configuration parameters;

[0063] The calculation formula of the eleventh indicator is:

[0064]

[0065] In the formula, R 11 is the eleventh indicator; v min4 , v max4 are the minimum and maximum values ​​of the reasonable range of oscillation respectively.

[0066] In an exemplary embodiment, the evaluation indicators in the multiple dimensions include energy saving indicators; and the step of calculating the evaluation indicators of the train in the multiple dimensions includes:

[0067] Traverse the historical operation data of the entire line and extract all local speed extreme points;

[0068] For the speeds v corresponding to any two consecutive local speed extreme points k and v m , if v k < v m , and k < m, then the region between these two extreme points is an acceleration region, where k and m are cycle numbers;

[0069] Calculate and accumulate the acceleration path integral values of all acceleration regions in the historical operation data of the entire line to obtain A total ;

[0070] Divide the line data according to the speed limit zones; extract the local speed extremes for each speed limit zone;

[0071] For any four consecutive local speed extremes, the corresponding speeds are v p , v q , v r , v h , if the speeds of v p and v r are both greater than v q and v h , then from cycle p to cycle h is a non - energy - saving operation segment, calculate the path integral value of the acceleration from cycle q to cycle r to obtain the loss value of this segment; where p, q, r, h are the corresponding cycle numbers;

[0072] Calculate and accumulate the loss values of all non - energy - saving regions in the historical operation data of the entire line to obtain A error ;

[0073] The path integral A of the acceleration ab The calculation formula is:

[0074]

[0075] In the formula, a and b are cycle numbers; a i is the acceleration; s i is the train position;

[0076] Calculate the energy - saving index according to the following formula:

[0077] R 12 =(A total -A error ) / A total

[0078] In the formula, R 12 is the energy - saving index.

[0079] In an exemplary embodiment, the evaluation indicators in the multiple dimensions include economic indicators; calculating the evaluation indicators of the train in multiple dimensions includes:

[0080] The deceleration of the train during braking is calculated according to the following formula:

[0081]

[0082] In the formula, a gradient is the ramp deceleration; a basic is the basic resistance deceleration; a ele The maximum regenerative braking deceleration that can be provided at this speed; the deceleration during the braking process is set to be no less than a min ; a is the deceleration of the train during braking;

[0083] Based on the speed and position of the train during deceleration, calculate the expected air braking at different positions:

[0084] a min =a+a gradient +a basic

[0085]

[0086] In the formula, a air To expect air brake;

[0087] If the actual air brake deceleration of the train at a certain position is less than the expected air brake a air , then it is considered that the position meets the economic requirements; the distance of all braking processes during the operation is S total , the braking distance that meets the economic requirement is S conform , the economic index is calculated according to the following formula:

[0088]

[0089] In the formula, R 13 is the economic index, S expect is the braking distance under the expected state.

[0090] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include a robustness index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0091] Calculate the evaluation values ​​of the six indicators of the historical running data collected between different stations of the train; and calculate the robustness index using the following formula according to the standard deviation of the evaluation values ​​of the six indicators between different stations:

[0092]

[0093] In the formula, R 14 is the robustness index; σmin , σ max are the minimum and maximum values ​​of the preset reasonable range of robustness respectively; σ is the standard deviation of the evaluation values ​​of the six indicators between different stations; the six indicators include: comfort index, efficiency index, stability index, energy-saving index, economy index and accuracy index.

[0094] In another embodiment of the present application, a device for evaluating automatic train driving is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described above when executing the computer program.

[0095] The solution proposed in the embodiment of the present application can improve the comprehensive evaluation capability and efficiency of the control effect of the ATO system in multiple dimensions, intuitively reflect the current vehicle control performance of the system, and can more conveniently compare and analyze the control effects under different systems and different parameter settings.

[0096] Other features and advantages of the present application will be described in the following description, and partly become apparent from the description, or be understood by implementing the present application. Other advantages of the present application can be realized and obtained by the schemes described in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] The accompanying drawings are used to provide an understanding of the technical solution of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the technical solution of the present application and do not constitute a limitation on the technical solution of the present application.

[0098] Figure 1 A flow chart of a method for evaluating automatic train driving provided in an embodiment of the present application;

[0099] Figure 2 A schematic diagram of the structure of the autonomous driving evaluation model provided in the embodiment of the present application;

[0100] Figure 3 A schematic diagram of the calculation logic of the optimal efficiency in the automatic speed regulation stage provided in an embodiment of the present application;

[0101] Figure 4 A schematic diagram of the structure of an evaluation device for automatic train driving provided in an embodiment of the present application. DETAILED DESCRIPTION

[0102] The present application describes multiple embodiments, but the description is exemplary rather than restrictive, and it is obvious to those skilled in the art that there may be more embodiments and implementations within the scope of the embodiments described in the present application. Although many possible feature combinations are shown in the drawings and discussed in the specific embodiments, many other combinations of the disclosed features are also possible. Unless specifically limited, any feature or element of any embodiment may be used in combination with any other feature or element in any other embodiment, or may replace any other feature or element in any other embodiment.

[0103] The present application includes and contemplates combinations of features and elements known to those of ordinary skill in the art. The embodiments, features and elements disclosed in the present application may also be combined with any conventional features or elements to form a unique invention scheme. Any features or elements of any embodiment may also be combined with features or elements from other invention schemes to form another unique invention scheme. Therefore, it should be understood that any feature shown and / or discussed in the present application may be implemented individually or in any appropriate combination. Therefore, except for the limitations made according to the attached claims and their equivalents, the embodiments are not subject to other restrictions. In addition, various modifications and changes may be made within the scope of protection of the attached claims.

[0104] In addition, when describing representative embodiments, the specification may have presented the method and / or process as a specific sequence of steps. However, to the extent that the method or process does not rely on the specific order of the steps described herein, the method or process should not be limited to the steps of the specific order described. As will be understood by those of ordinary skill in the art, other sequences of steps are also possible. Therefore, the specific sequence of the steps set forth in the specification should not be interpreted as a limitation to the claims. In addition, the claims for the method and / or process should not be limited to the steps of performing them in the order written, and those skilled in the art can easily understand that these sequences can be changed and still remain within the spirit and scope of the embodiments of the present application.

[0105] This application proposes a comprehensive evaluation method for the automatic operation control effect of trains, which can intuitively reflect whether the control effect meets the design goals. It can also analyze the evaluation results, summarize the shortcomings of the current control strategy, and point out the direction for the next step of optimization of the control strategy.

[0106] Figure 1 A flow chart of a method for evaluating automatic train driving provided by an embodiment of the present application is shown. Figure 1 As shown, the method includes:

[0107] Step 100: Acquire first historical running data of the train;

[0108] Step 101: Calculating evaluation indicators of the train in multiple dimensions according to the first historical operation data;

[0109] Step 102: Input the evaluation indicators of the train in multiple dimensions into a pre-trained automatic driving evaluation model to obtain the evaluation result of the automatic driving of the train.

[0110] The solution proposed in the embodiment of the present application obtains the first historical operation data of the train; based on the first historical operation data, the evaluation indicators of the train in multiple dimensions are calculated; wherein the evaluation indicators of the train in multiple dimensions include: comfort index, efficiency index, stability index, energy saving index, economy index, accuracy index, and robustness index; the evaluation indicators of the train in multiple dimensions are input into a pre-trained automatic driving evaluation model to obtain the evaluation results of the automatic driving of the train, which can improve the comprehensive evaluation capability and efficiency of the control effect of the ATO system in multiple dimensions. By comprehensively evaluating the automatic driving system of the train, the current control performance of the system can be understood more intuitively and quickly, and the control effect of the train under different system configurations and parameter settings can be more conveniently compared and analyzed.

[0111] In an exemplary embodiment, the pre-trained autonomous driving evaluation model is trained according to the following method:

[0112] Acquire multiple sets of second historical running data of the train;

[0113] For each set of the second historical running data, the following steps are performed respectively until a preset training stop condition is met:

[0114] Calculating evaluation indicators of the train in multiple dimensions according to the set of the second historical operation data, and inputting the evaluation indicators into the automatic driving evaluation model to obtain a first evaluation result;

[0115] Obtaining N second evaluation results obtained after comprehensive evaluation of the evaluation indicators of the train corresponding to the group of the second historical operation data in multiple dimensions by N experts in the industry; taking the average value of the N second evaluation results as the third evaluation result; wherein N is an integer greater than or equal to 1;

[0116] According to the first evaluation result and the third evaluation result, parameters of the autonomous driving evaluation model are optimized.

[0117] In this embodiment, the autonomous driving evaluation model is a deep neural network model. The deep neural network model (DNN) imitates the structure and working principle of the human brain neural network and is a machine learning model composed of multiple neuron layers. Each neuron layer receives the output of the previous layer as input and calculates the output through a series of nonlinear transformations and weight adjustments.

[0118] The structure of the autonomous driving evaluation model provided in this embodiment is as follows: Figure 2 As shown, L is defined as the number of hidden layers, n i is the number of neurons in the hidden layer l, z l As the input of hidden layer l, the output y of hidden layer l can be l Described as:

[0119] y l =f(w l z l-1 +b l )

[0120] In the formula, is the weight matrix from layer l-1 to layer l, is the bias of the lth layer. 0 is the input of the first hidden layer, that is, the encoded input feature vector x. L is the output of the last hidden layer, i.e., the overall score of the train automatic operation control effect. f is the activation function, which can be Sigmoid or other functions.

[0121] In an exemplary embodiment, inputting the evaluation indicators of the train in multiple dimensions into a pre-trained automatic driving evaluation model to obtain the evaluation result of the automatic driving of the train includes:

[0122] Determining, according to the adjustment mode of the train and the operation schedule of the train, the coding type of the evaluation index of the train in the multiple dimensions;

[0123] The evaluation indicators of the train in the multiple dimensions are combined into a first input vector, and a corresponding encoding vector in the automatic driving evaluation model is selected according to the encoding type to encode the first input vector to obtain a second input vector;

[0124] The second input vector is input into the hidden layer of the neural network of the automatic driving evaluation model for forward propagation to obtain the evaluation result of the automatic driving of the train.

[0125] In this embodiment, as shown in Table 1, the input parameters are divided into the following four categories according to the adjustment mode of the train and the train operation schedule, and different encoding vectors are used to process the input according to different categories:

[0126] x=R×P

[0127] In the formula, R is a vector composed of the evaluation results of different indicators, and P is the encoding vector.

[0128] Table 1 Input classification method

[0129] Classification The value of encoding P There is a step-by-step adjustment method and an operating timetable <![CDATA[P=[b,b,a,a,a,b,a,a,a,a,a,a,a,a] T ]]> There is a step-by-step adjustment method and no operating schedule <![CDATA[P=[b,b,a,b,b,a,a,a,a,a,a,a,a,a] T ]]> Stepless adjustment mode and operation schedule <![CDATA[P=[a,a,b,a,a,b,a,a,a,a,a,a,a,a] T ]]> Stepless adjustment mode, no operation schedule <![CDATA[P=[a,a,b,b,b,a,a,a,a,a,a,a,a,a] T ]]>

[0130] In the formula, the values ​​of a and b can be obtained through a certain calculation method. For example, a can be taken as 1 and b can be taken as 0.

[0131] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include a comfort index, and the comfort index includes a first index, a second index, and a third index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0132] When the train is running in the stepless speed regulation mode, according to the actual speed v and the target speed v during the train running, target , preset speed difference v diff , the running process of the train is divided into acceleration process, deceleration process and uniform speed process; wherein, in the uniform speed process, the difference between the actual speed of the train and the target speed is less than v diff , when |vv target | <v diff When vv target ≥v diff When v target -v≥v diff When , it is the deceleration process;

[0133] Performing filtering processing on a plurality of acceleration values ​​collected during the acceleration process and the deceleration process to obtain a plurality of acceleration filter values; calculating an average value of the absolute values ​​of the plurality of acceleration filter values ​​to obtain an average acceleration value; determining an acceleration evaluation value as the first indicator according to the average acceleration value, the maximum acceleration value and the minimum acceleration value;

[0134] The operation data of the entire line is divided into segments according to the time interval of △t, and the acceleration impact rate of each time period is calculated respectively, and the average value of the absolute value of the acceleration impact rate of all time periods is calculated to obtain the average impact rate value; according to the average impact rate value, the acceleration impact rate evaluation value is determined as the second indicator;

[0135] When the train is running in a stepless speed regulation mode, for each running data sampling position, the level that satisfies the comfort level at the position is calculated according to the ramp acceleration at the position and the maximum tolerance value of the train acceleration under the ramp acceleration; the running time that meets the preset level switching conditions is counted, and the third indicator is determined according to the running time that meets the preset level switching conditions and the running time of the trains on the entire line;

[0136] The preset level switching conditions include: the level size of each position does not exceed L levels; the level maintenance time exceeds the time required for instruction transmission and instruction action; when switching levels, only changes to adjacent levels are allowed, and skipping levels are not allowed; in a continuous acceleration or deceleration process, the level change method should be to increase first and then decrease, and level repetition is not allowed.

[0137] In this embodiment, the comfort under the stepless speed regulation mode is reflected by acceleration and acceleration impact rate. When evaluating acceleration, only the data of the acceleration process and the deceleration process are taken for evaluation; for the acceleration impact rate, the operation data of the entire line is taken for evaluation.

[0138] The acceleration data of the acceleration process and the deceleration process are filtered and then the absolute value is calculated. The average acceleration can be obtained by averaging the absolute values.

[0139] ″′

[0140] Assume that the original acceleration value is {a1, a2, …, a n}, the acceleration filter value is {a1,a2,…,a n}, the filter value is calculated as:

[0141]

[0142] Where k is the filtering parameter.

[0143] The average value of the absolute value of acceleration is recorded as The corresponding calculation formula is:

[0144]

[0145] When the average value of the absolute value of acceleration is less than a min It is considered to be the best state when the average value of the absolute value of acceleration is greater than a max It is considered to be the most unfavorable state.

[0146] The calculation formula of the first indicator is:

[0147]

[0148] In the formula, is the average acceleration value; R1 is the first index; a max is the maximum acceleration; a min is the minimum acceleration;

[0149] When evaluating the acceleration impact rate, the average value of the absolute value of the acceleration impact rate is calculated according to the fixed time interval △t. Assuming that the train acceleration measured at time t1 is a1, and the train acceleration measured at time t1+△t is a2, the average impact rate of the train during this period is I, then:

[0150]

[0151] The average impact rate value is:

[0152]

[0153] When the acceleration shock rate is less than J min When the acceleration impact rate is greater than J max It is considered to be the most unfavorable state.

[0154] The calculation formula of the second indicator is:

[0155]

[0156] In the formula, is the average impact rate value; R2 is the second index; J max is the maximum acceleration shock rate threshold; J min is the minimum acceleration shock rate threshold;

[0157] The evaluation of comfort under the stepped speed regulation mode is reflected by the degree of application of the steps and the way they change.

[0158] The calculation formula for the comfort level is:

[0159]

[0160] Where L is the extreme position of comfort; a L is the acceleration provided under L level; a L+1 is the acceleration provided at level L+1; a gradient is the ramp acceleration, a max2 It is the maximum tolerable acceleration value under the current ramp acceleration;

[0161] The calculation formula of the third indicator is:

[0162]

[0163] In the formula, t perfect To meet the running time of the preset level switching; ttotal is the running time of the train on the entire line; R3 is the third indicator.

[0164] In an exemplary embodiment, the evaluation indicators in the multiple dimensions include an efficiency indicator, and the efficiency indicator includes a fourth indicator, a fifth indicator, and a sixth indicator; and the calculating the evaluation indicators of the train in the multiple dimensions includes:

[0165] When there is an operating timetable, the fourth index is determined according to the ratio of the difference between the actual arrival time and the planned arrival time of the train and the preset threshold; the fifth index is determined according to the difference between the actual arrival time and the planned arrival time of the train and the difference between the actual arrival time and the planned arrival time of the train under the expected state;

[0166] When there is no operating timetable, an iterative method is used to respectively calculate the duration of the traction acceleration phase, the duration of the uniform speed driving phase and the duration of the deceleration driving phase when the train is running at the highest speed; the sum of the duration of the traction acceleration phase, the duration of the uniform speed driving phase and the duration of the deceleration driving phase is determined as the shortest driving duration;

[0167] Calculate the longest running time of the train under the most unfavorable conditions; determine the efficiency evaluation value when there is no operating timetable based on the shortest running time, the longest running time and the actual running time, as the sixth indicator.

[0168] In this embodiment, when there is an operating timetable, the efficiency is evaluated in two aspects: one is the difference between the actual arrival time of the train and the planned arrival time, and the other is the error between the set time error and the actual arrival time of the train.

[0169] The difference between the actual arrival time and the planned arrival time of the train should be within a preset threshold, for example, the preset threshold is ±30s. When the difference is greater than 60s, it is considered an unacceptable state.

[0170] The calculation formula of the fourth indicator is:

[0171]

[0172] The fifth indicator calculation formula is:

[0173]

[0174] In the formula, t diff is the difference between the actual arrival time and the planned arrival time of the train; t1 is the preset threshold; t expectdiff is the difference between the actual arrival time and the planned arrival time of the train under the expected state; R4 is the fourth indicator; R5 is the fifth indicator.

[0175] When there is no running time requirement, the train should run at the highest possible speed, and the time required is the shortest time t min .

[0176] The shortest running time is calculated in three parts, such as Figure 3 As shown, it includes a traction acceleration phase AC, a constant speed driving phase CD, and a speed reduction driving phase DF.

[0177] The train in section AB runs at maximum traction, but the maximum traction acceleration of the train changes with the train speed. Therefore, the AB section is calculated using an iterative method with a calculation period of T. Assume that the maximum traction acceleration of the train is a tra , the line slope acceleration is a gra , v o is the initial speed of this cycle, S o is the initial position of this cycle, then the speed v and position S of each cycle are:

[0178] v=v o +(a tra +a gra )×T

[0179]

[0180] In the BC segment train traction relief phase, the time t of the BC segment can be obtained by using the maximum acceleration impact rate J and the safety difference between the maximum running speed of the train and the ATP allowed speed, Δv. BC :

[0181]

[0182] v c =v B +(a tra +a gra )×t=v ATP -Δv

[0183]

[0184] In the formula, t BC is the duration of the train traction relief phase; S BC is the length of the train traction relief phase; J1 is the maximum acceleration impact rate; v C , v B are the speeds at the start and end of the train traction relief phase; Δv is the maximum running speed of the train and the ATP allowed speed v ATP Safety difference between

[0185] CD segment is the uniform speed time, and this segment time t CD for:

[0186]

[0187] In the formula, S CD is the length of the uniform speed driving phase; t CD is the duration of the uniform speed driving stage;

[0188] The DF segment is the process of following the ATP deceleration. The calculation process of this segment is:

[0189]

[0190]

[0191]

[0192] In the formula, t DF is the duration of the deceleration phase; v M , v N , S MN They are the speed at the starting point and end point of the ATP protection curve and the speed reduction distance; S DF is the length of the deceleration phase; v D , v F are the speeds at the start and end of the deceleration phase in the best case, respectively;

[0193] Considering the operational needs, the acceptable time waste in the most unfavorable case is set as α%, and the time required for train operation in the most unfavorable efficiency case is obtained as t max , the calculation formula for the longest driving time is:

[0194]

[0195] In the formula, t max is the maximum driving time; t min is the sum of the duration of all traction acceleration phases, uniform speed driving phases and deceleration driving phases during the running of the train; α is the percentage of wasted time that can be tolerated under the most unfavorable situation;

[0196] The calculation formula of the sixth index is:

[0197]

[0198] In the formula, t fact is the actual travel time of the train; R6 is the sixth indicator.

[0199] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include an accuracy index, and the accuracy index includes a seventh index, an eighth index, and a ninth index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0200] The parking error evaluation value of the train is obtained according to the following formula as the seventh index:

[0201] S parkmin =S Negative -S factmax

[0202] S max =S factmax +S Negative -v max ×T

[0203]

[0204] In the formula, S factmin ,S factmax are the minimum and maximum values ​​of the difference between the actual parking position and the parking point of the train; S parkmin ,S parkmax are the minimum and maximum values ​​of the difference range between the actual parking position and the estimated parking position of the train, which is determined by the train's benchmark speed; x1 is the error between the parking point and the actual parking position; x2 is the error between the estimated parking position and the actual parking position; S Negative is the size of the parking negative window; v max is the maximum benchmarking speed; T is the sampling period; R7 is the seventh index;

[0205] The benchmark speed evaluation value of the train is obtained according to the following formula as the eighth index:

[0206] v min =v low +v error

[0207]

[0208] In the formula, v low is the minimum benchmarking speed supported by the system; v max is the maximum value of the benchmark speed; v min is the minimum value of the benchmark speed; v error is the error caused by speed measurement, distance measurement and line slope; T maxTrans is the maximum braking transmission delay; T minTrans is the minimum braking transmission delay; S error is the distance error caused by speed and distance measurement; a max is the maximum benchmark braking deceleration; a min is the minimum benchmark braking deceleration; R8 is the eighth index;

[0209] The train benchmarking state stability evaluation value is obtained according to the following formula as the ninth index:

[0210] tmin2 =t action +t filter

[0211]

[0212] In the formula, t min2 is the shortest time required to maintain the state before benchmarking; t action is the time for braking and applying the brake before the train is unloaded; t filter is the time required for the filtered speed and acceleration to be the same as the actual situation of the train; R9 is the ninth indicator; the time the train maintains the state before benchmarking is t; t hold It is the maximum tolerable delay error in the line.

[0213] In this embodiment, the accuracy evaluation is divided into three aspects: the evaluation value of the train's parking error, the evaluation value of the train's benchmarking speed, and the evaluation value of the train's state stability during benchmarking.

[0214] The parking error of the train needs to meet two requirements: first, the difference between the actual parking position of the train and the parking point is within the parking window range [S factmin ,S factmax ] to ensure that the doors can be opened normally. This range is determined by the size of the platform and the train doors. On the other hand, the difference between the actual parking position and the estimated parking position of the train is within a reasonable range [S parkmin ,S parkmax ], which is determined by the train’s benchmark speed.

[0215] For the minimum target speed of the train, the influence of speed and distance measurement errors and line slopes is considered. error , to avoid the train from sliding and stopping.

[0216] For the train benchmarking state stability evaluation value, in order to ensure the accurate calculation of the braking distance, it is required that when the train applies the parking brake, the braking percentage / brake level output by the train should remain unchanged, so that the train can maintain a relatively stable train control state. Therefore, the state holding time before benchmarking needs to meet the time t of the train braking and applying the brake before unloading. action The time required for the filtered speed and acceleration to be the same as the actual situation of the train is t filter .

[0217] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include a stability index, and the stability index includes a tenth index and an eleventh index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0218] The train overshoot evaluation value is calculated as the tenth index according to the following steps:

[0219] The collected train speed is subjected to high-order delay filtering to obtain the filtered speed {v1,v2…v n};

[0220] The filtered velocity is sliced ​​with t1 as the slice length, {v1,v2…v n} is divided into {v 1,1 ,v 1,2 …v 1,k},{v 2,1 ,v 2,2 …v 2,k},…,{v m,1 ,v m,2 …v m,k};

[0221] For the fragment {v i,1 ,v i,2 …v i,k}, calculate the range r of the fragment i , if r i If it is less than r1, the i-th segment is a uniform speed stage;

[0222] Combine the time-continuous uniform speed segments and calculate the average speed of the uniform speed segments

[0223] Traverse forward from the start time of each uniform speed segment, if v j Greater than v j+1 and v j-1 , then v j is a local maximum if v j Less than v j+1 and v j-1 , then v j is a local minimum; for any local extremum, if v j and a i If the difference is greater than r2, then v j is the overshoot point; the overshoot amount is Where ai is a configuration parameter;

[0224] The calculation formula for the maximum and minimum values ​​of the reasonable range of overshoot is:

[0225]

[0226] The calculation formula of the tenth indicator is:

[0227]

[0228] In the formula, R 10 is the tenth indicator; a min3 , a max3are the minimum and maximum braking deceleration values ​​allowed when the train reaches the target speed, t ease is the traction brake release time; v min3 , v max3 are the minimum and maximum values ​​of the reasonable range of overshoot respectively;

[0229] The train oscillation degree evaluation value is calculated as the eleventh indicator according to the following steps:

[0230] The collected train speed is subjected to high-order delay filtering to obtain the filtered speed {v1,v2…v n};

[0231] The filtered velocity is segmented with t2 as the segment length and t3 as the step length, and {v1,v2…v n} is divided into {v 1,1 ,v 1,2 …v 1,k},{v 2,1 ,v 2,2 …v 2,k},…,{v m,1 ,v m,2 …v m,k};

[0232] For the fragment {v i,1 ,v i,2 …v i,k}, with time as the independent variable and speed as the dependent variable, a linear fit is performed to obtain the equation t = q i v+p i , and calculate the data range r of this segment i ;

[0233] If q i is less than a1, and r i If it is greater than r2, then the segment is a segment where oscillation occurs;

[0234] Merge time-continuous oscillation segments;

[0235] For the element v in the i-th oscillation segment j , if v j Greater than v j+1 and v j-1 , then v j is a local maximum if v j Less than v j+1 and v j-1 , then v j is the local minimum; the oscillation degree is determined as the segment range r i ;

[0236] Among them, t1, t2, t3, r1, r2, and a1 are all configuration parameters;

[0237] The calculation formula of the eleventh index is:

[0238]

[0239] In the formula, R 11 is the eleventh index; v min4 , v max4 are respectively the minimum value and the maximum value of the reasonable range of oscillation.

[0240] In this embodiment, the stability index is reflected by the overshoot and the degree of oscillation during the train operation. The overshoot phenomenon refers to the situation where the speed deviates from the target speed before the train enters the uniform speed process. Specifically, in the train operation, it is manifested as exceeding the target speed during the acceleration process and being lower than the target speed during the deceleration process. The oscillation phenomenon is caused by the train's inability to apply the traction force required for uniform train operation, and is specifically manifested as the periodic up and down fluctuation of the train speed around the target speed.

[0241] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include an energy-saving index; calculating the evaluation indexes of the train in the multiple dimensions includes:

[0242] Traverse the historical operation data of the whole line and extract all local speed extreme points;

[0243] For the speeds v k and v m corresponding to any two consecutive local speed extreme points, if v k < v m , and k < m, then the area between these two extreme points is an acceleration area, where k and m are cycle numbers;

[0244] Calculate and accumulate the acceleration path integral values of all acceleration areas in the historical operation data of the whole line to obtain A total ;

[0245] Divide the line data according to the speed limit areas; extract local speed extremes for each speed limit area;

[0246] For any four consecutive local speed extremes, the corresponding speeds are v p , v q , v r , v h , if the speeds of v p and v r are both greater than v q and v h, then the period from period p to period h is a non-energy-saving operation segment, and the path integral value of the acceleration from period q to period r is calculated to obtain the loss value of this segment; where p, q, r, h are the corresponding period numbers;

[0247] Calculate the loss values ​​of all non-energy-saving areas in the historical operation data of the entire line and add them up to get A error ;

[0248] The path integral A of the acceleration ab The calculation formula is:

[0249]

[0250] In the formula, a and b are period numbers; a i is the acceleration; s i is the train position;

[0251] The energy saving index is calculated according to the following formula:

[0252] R 12 =(A total -A error ) / A total

[0253] In the formula, R 12 is the energy saving index.

[0254] In this embodiment, the energy saving situation is reflected by the energy waste during the train operation. The energy saving evaluation calculation model needs to extract the acceleration energy consumption E during the whole line operation. total and additional energy consumption E error , and calculate the evaluation result R 12 :

[0255]

[0256] Assume that the train is in an accelerating state when it moves from position s1 to position s2. For any position s during the movement, the traction force output by the train is f(s), and the distance from point s1 to point s2 is The energy consumption corresponding to this distance is:

[0257]

[0258] Among them, m is the mass of the train, and the value A is the distance integral corresponding to the acceleration. And similarly, we can get:

[0259] E total =mA total

[0260] E error =mA error

[0261]

[0262] Therefore, the energy saving index can be calculated based on R 12 =(A total -A error ) / A total Calculated.

[0263] In an exemplary embodiment, the evaluation indicators in the multiple dimensions include an economic index; and the step of calculating the evaluation indicators of the train in the multiple dimensions includes:

[0264] The deceleration of the train during braking is calculated according to the following formula:

[0265]

[0266] In the formula, a gradient is the ramp deceleration; a basic is the basic resistance deceleration; a ele The maximum regenerative braking deceleration that can be provided at this speed; the deceleration during the braking process is set to be no less than a min ; a is the deceleration of the train during braking;

[0267] Based on the speed and position of the train during deceleration, calculate the expected air braking at different positions:

[0268] a min =a+a gradient +a basic

[0269]

[0270] In the formula, a air To expect air brake;

[0271] If the actual air brake deceleration of the train at a certain position is less than the expected air brake a air , then it is considered that the position meets the economic requirements; the distance of all braking processes during the operation is S total , the braking distance that meets the economic requirement is S conform , the economic index is calculated according to the following formula:

[0272]

[0273] In the formula, R 13 is the economic index, S expect is the braking distance under the expected state.

[0274] In this embodiment, the braking force of the train is provided by regenerative braking and air braking. Air braking will cause wear of the brake disc. Reasonable control of the application of braking force can reduce wear of the brake disc, reduce replacement frequency, and save operating costs.

[0275] According to the train braking performance, at high speeds, the degree of braking force applied can be controlled to avoid the train using air braking. air The train only has air braking.

[0276] To ensure the braking efficiency of the train, the deceleration during the braking process should not be less than a min , then the braking distance S under the desired state is obtained by iteration expect for:

[0277]

[0278] In the formula, a gradient is the ramp deceleration; a basic is the basic resistance deceleration; a ele is the maximum regenerative braking deceleration available at this speed; v i+1 v is the deceleration start speed; i+1 The deceleration end speed.

[0279] In an exemplary embodiment, the evaluation indexes in the multiple dimensions include a robustness index; and the step of calculating the evaluation indexes of the train in the multiple dimensions includes:

[0280] Calculate the evaluation values ​​of the six indicators of the historical running data collected between different stations of the train; and calculate the robustness index using the following formula according to the standard deviation of the evaluation values ​​of the six indicators between different stations:

[0281]

[0282] In the formula, R 14 is the robustness index; σ min , σ max are the minimum and maximum values ​​of the preset reasonable range of robustness respectively; σ is the standard deviation of the evaluation values ​​of the six indicators between different stations; the six indicators include: comfort index, efficiency index, stability index, energy-saving index, economy index and accuracy index.

[0283] Figure 4 A schematic diagram of the structure of an evaluation device for automatic train driving provided by another embodiment of the present application is shown. Figure 4As shown, it includes a memory 400, a processor 401, and a computer program stored in the memory and executable on the processor. When the processor 401 executes the computer program, the steps of the method described above are implemented.

[0284] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In hardware implementations, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or temporary medium). As known to those skilled in the art, the term "computer storage medium" includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

Claims

1. A method for evaluating automatic train driving, characterized in that: include: Obtaining the first historical running data of the train; Calculate evaluation indicators of the train in multiple dimensions according to the first historical operation data; wherein the evaluation indicators of the train in multiple dimensions include: comfort index, efficiency index, stability index, energy saving index, economy index, accuracy index, and robustness index; The evaluation indicators of the train in multiple dimensions are input into a pre-trained automatic driving evaluation model to obtain an evaluation result of the automatic driving of the train.

2. The method for evaluating automatic train driving according to claim 1, characterized in that: The pre-trained autonomous driving evaluation model is trained according to the following method: Acquire multiple sets of second historical running data of the train; For each set of the second historical running data, the following steps are performed respectively until a preset training stop condition is met: Calculating evaluation indicators of the train in multiple dimensions according to the set of the second historical operation data, and inputting the evaluation indicators into the automatic driving evaluation model to obtain a first evaluation result; Obtaining N second evaluation results obtained after comprehensive evaluation by N experts in the industry on the evaluation indicators of the train corresponding to the group of the second historical operation data in multiple dimensions; taking the average value of the N second evaluation results as the third evaluation result; wherein N is an integer greater than or equal to 1; According to the first evaluation result and the third evaluation result, parameters of the autonomous driving evaluation model are optimized.

3. The method for evaluating automatic train driving according to claim 2, characterized in that: The step of inputting the evaluation indicators of the train in multiple dimensions into a pre-trained automatic driving evaluation model to obtain the evaluation result of the automatic driving of the train includes: Determining, according to the adjustment mode of the train and the operation schedule of the train, the coding type of the evaluation index of the train in the multiple dimensions; The evaluation indicators of the train in the multiple dimensions are combined into a first input vector, and a corresponding encoding vector in the automatic driving evaluation model is selected according to the encoding type to encode the first input vector to obtain a second input vector; The second input vector is input into the hidden layer of the neural network of the automatic driving evaluation model for forward propagation to obtain the evaluation result of the automatic driving of the train.

4. The method for evaluating automatic train driving according to claim 3, characterized in that: The evaluation indexes in the multiple dimensions include a comfort index, and the comfort index includes a first index, a second index, and a third index; and the calculation of the evaluation indexes of the train in the multiple dimensions includes: When the train is running in the stepless speed regulation mode, according to the actual speed v and the target speed v during the train running, target , preset speed difference v diff , the running process of the train is divided into acceleration process, deceleration process and uniform speed process; wherein, in the uniform speed process, the difference between the actual speed of the train and the target speed is less than v diff , when |vv target | <v diff When vv target ≥v diff When v target -v≥v diff When , it is the deceleration process; Performing filtering processing on a plurality of acceleration values ​​collected during the acceleration process and the deceleration process to obtain a plurality of acceleration filter values; calculating an average value of the absolute values ​​of the plurality of acceleration filter values ​​to obtain an average acceleration value; determining an acceleration evaluation value as the first indicator according to the average acceleration value, the maximum acceleration value and the minimum acceleration value; The operation data of the entire line is divided into segments according to the time interval of △t, and the acceleration impact rate of each time period is calculated respectively, and the average value of the absolute value of the acceleration impact rate of all time periods is calculated to obtain the average impact rate value; according to the average impact rate value, the acceleration impact rate evaluation value is determined as the second indicator; When the train is running in a stepless speed regulation mode, for each running data sampling position, the level that satisfies the comfort level at the position is calculated according to the ramp acceleration at the position and the maximum tolerance value of the train acceleration under the ramp acceleration; the running time that meets the preset level switching conditions is counted, and the third indicator is determined according to the running time that meets the preset level switching conditions and the running time of the trains on the entire line; The preset level switching conditions include: the level size of each position does not exceed L levels; the level maintenance time exceeds the time required for instruction transmission and instruction action; when switching levels, only changes to adjacent levels are allowed, and skipping levels are not allowed; in a continuous acceleration or deceleration process, the level change method should be to increase first and then decrease, and level repetition is not allowed.

5. The method for evaluating automatic train driving according to claim 3, characterized in that: The evaluation indicators in the multiple dimensions include an efficiency indicator, and the efficiency indicator includes a fourth indicator, a fifth indicator, and a sixth indicator; and the calculation of the evaluation indicators of the train in the multiple dimensions includes: When there is an operating timetable, the fourth index is determined according to the ratio of the difference between the actual arrival time and the planned arrival time of the train and the preset threshold; the fifth index is determined according to the difference between the actual arrival time and the planned arrival time of the train and the difference between the actual arrival time and the planned arrival time of the train under the expected state; When there is no operating timetable, an iterative method is used to respectively calculate the duration of the traction acceleration phase, the duration of the uniform speed driving phase and the duration of the deceleration driving phase when the train is running at the highest speed; the sum of the duration of the traction acceleration phase, the duration of the uniform speed driving phase and the duration of the deceleration driving phase is determined as the shortest driving duration; Calculate the longest running time of the train under the most unfavorable conditions; determine the efficiency evaluation value when there is no operating timetable based on the shortest running time, the longest running time and the actual running time, as the sixth indicator.

6. The method for evaluating automatic train driving according to claim 3, characterized in that: The evaluation indicators on multiple dimensions include an accuracy indicator, and the accuracy indicator includes a seventh indicator, an eighth indicator, and a ninth indicator; and the calculation of the evaluation indicators on multiple dimensions of the train includes: The parking error evaluation value of the train is obtained according to the following formula as the seventh index: S parkmin =S Negative -S factmax S max =S factmax +S Negative -v max ×T In the formula, S factmin ,S factmax are the minimum and maximum values ​​of the difference between the actual parking position and the parking point of the train; S parkmin ,S parkmax are the minimum and maximum values ​​of the difference range between the actual parking position and the estimated parking position of the train, which is determined by the train's benchmark speed; x1 is the error between the parking point and the actual parking position; x2 is the error between the estimated parking position and the actual parking position; S Negative is the size of the parking negative window; v max is the maximum benchmarking speed; T is the sampling period; R7 is the seventh index; The benchmark speed evaluation value of the train is obtained according to the following formula as the eighth index: v min =v low +v error In the formula, v low is the minimum benchmarking speed supported by the system; v max is the maximum value of the benchmark speed; v min is the minimum value of the benchmark speed; v error is the error caused by speed measurement, distance measurement and line slope; T maxTrans is the maximum braking transmission delay; T minTrans is the minimum braking transmission delay; S error is the distance error caused by speed and distance measurement; a max is the maximum benchmark braking deceleration; a min is the minimum benchmark braking deceleration; R8 is the eighth index; The train benchmarking state stability evaluation value is obtained according to the following formula as the ninth index: t min2 =t action +t filter Where, t min2 is the shortest time required to maintain the state before benchmarking; t action is the time for braking and applying the brake before the train is unloaded; t filter is the time required for the filtered speed and acceleration to be the same as the actual situation of the train; R9 is the ninth indicator; the time the train maintains the state before benchmarking is t; t hold It is the maximum tolerable delay error in the line.

7. The method for evaluating automatic train driving according to claim 1, characterized in that: The evaluation indicators in the multiple dimensions include a stability indicator, and the stability indicator includes a tenth indicator and an eleventh indicator; and the calculating of the evaluation indicators of the train in the multiple dimensions includes: The train overshoot evaluation value is calculated as the tenth index according to the following steps: The collected train speed is subjected to high-order delay filtering to obtain the filtered speed {v1,v2…v n }; The filtered velocity is sliced ​​with t1 as the slice length, and {v1,v2…v n } is divided into {v 1,1 ,v 1,2 …v 1,k },{v 2,1 ,v 2,2 …v 2,k },…,{v m,1 ,v m,2 …v m,k }; For the fragment {v i,1 ,v i,2 …v i,k }, calculate the range r of the segment i , if r i If it is less than r1, the i-th segment is a uniform speed stage; Combine the time-continuous uniform speed segments and calculate the average speed of the uniform speed segments Traverse forward from the start time of each uniform speed segment, if v j Greater than v j+1 and v j-1 , then v j is a local maximum, if v j Less than v j+1 and v j-1 , then v j is a local minimum; for any local extremum, if v j and a i If the difference is greater than r2, then v j is the overshoot point; the overshoot amount is where a i is a configuration parameter; The calculation formula for the maximum and minimum values ​​of the reasonable range of overshoot is: The calculation formula of the tenth indicator is: In the formula, R 10 is the tenth indicator; a min3 , a max3 are the minimum and maximum braking deceleration values ​​allowed when the train reaches the target speed, t ease is the traction brake release time; v min3 , v max3 are the minimum and maximum values ​​of the reasonable range of overshoot respectively; The train oscillation degree evaluation value is calculated as the eleventh indicator according to the following steps: The collected train speed is subjected to high-order delay filtering to obtain the filtered speed {v1,v2…v n }; The filtered velocity is segmented with t2 as the segment length and t3 as the step length, and {v1,v2…v n } is divided into {v 1,1 ,v 1,2 …v 1,k },{v 2,1 ,v 2,2 …v 2,k },…,{v m,1 ,v m,2 …v m,k }; For the fragment {v i,1 ,v i,2 …v i,k }, with time as the independent variable and speed as the dependent variable, a linear fit is performed to obtain the equation t = q i v+p i , and calculate the data range r of this segment i ; If q i is less than a1, and r i If it is greater than r2, then the segment is a segment where oscillation occurs; Merge time-continuous oscillation segments; For the element v in the i-th oscillation segment j , if v j Greater than v j+1 and v j-1 , then v j is a local maximum if v j Less than v j+1 and v j-1 , then v j is the local minimum; the oscillation degree is determined as the segment range r i ; Among them, t1, t2, t3, r1, r2, a1 are all configuration parameters; The calculation formula of the eleventh indicator is: In the formula, R 11 is the eleventh indicator; v min4 , v max4 are the minimum and maximum values ​​of the reasonable range of oscillation respectively.

8. The method for evaluating automatic train driving according to claim 1, characterized in that: The evaluation indicators in multiple dimensions include energy-saving indicators; and the calculation of the evaluation indicators of the train in multiple dimensions includes: Traverse the historical operation data of the entire line and extract all local speed extreme points; For the velocities v corresponding to any two consecutive local velocity extreme points k and v m , if v k < v m , and k < m, then the region between these two extreme points is an acceleration region, where k and m are cycle numbers; Calculate the acceleration path integral value of all acceleration areas in the historical operation data of the entire line and accumulate them to obtain A total ; Divide the route data according to the speed limit zone; extract the local speed extreme value for each speed limit zone; For any four consecutive local velocity extremes, the corresponding velocities are v p ,v q ,v r ,v h , if v p and v r The speed is greater than v q and v h , then the period from period p to period h is a non-energy-saving operation segment, and the path integral value of the acceleration from period q to period r is calculated to obtain the loss value of this segment; where p, q, r, h are the corresponding period numbers; Calculate the loss values ​​of all non-energy-saving areas in the historical operation data of the entire line and add them up to get A error ; The path integral A of the acceleration ab The calculation formula is: In the formula, a and b are period numbers; a i is the acceleration; s i is the train position; The energy saving index is calculated according to the following formula: R 12 =(A total -A error ) / A total In the formula, R 12 is the energy saving index.

9. The method for evaluating automatic train driving according to claim 1, characterized in that: The evaluation indicators in multiple dimensions include economic indicators; the calculation of the evaluation indicators of the train in multiple dimensions includes: The deceleration of the train during braking is calculated according to the following formula: In the formula, a gradient is the ramp deceleration; a basic is the basic resistance deceleration; a ele The maximum regenerative braking deceleration that can be provided at this speed; the deceleration during the braking process is set to be no less than a min ; a is the deceleration of the train during braking; Based on the speed and position of the train during deceleration, calculate the expected air braking at different positions: a min =a+a gradient +a basic In the formula, a air To expect air brake; If the actual air brake deceleration of the train at a certain position is less than the expected air brake a air , then it is considered that the position meets the economic requirements; the distance of all braking processes during the operation is S total , the braking distance that meets the economic requirement is S conform , the economic index is calculated according to the following formula: In the formula, R 13 is the economic index, S expect is the braking distance under the expected state.

10. The method for evaluating automatic train driving according to any one of claims 1 to 8, characterized in that: The evaluation indicators in the multiple dimensions include a robustness indicator; the calculation of the evaluation indicators of the train in the multiple dimensions includes: Calculate the evaluation values ​​of the six indicators of the historical running data collected between different stations of the train; and calculate the robustness index using the following formula according to the standard deviation of the evaluation values ​​of the six indicators between different stations: In the formula, R 14 is the robustness index; σ min , σ max are the minimum and maximum values ​​of the preset reasonable range of robustness respectively; σ is the standard deviation of the evaluation values ​​of the six indicators between different stations; the six indicators include: comfort index, efficiency index, stability index, energy-saving index, economy index and accuracy index.

11. An evaluation device for automatic train driving, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 10 when executing the computer program.

Images