Cooperative control method and system under optimal group train operation index combination

Abstract

The invention belongs to the technical field of train operation control, and provides a cooperative control method and system under the optimal group train operation index combination. The method comprises the following steps: acquiring a target position and speed curve, running state data and group line data of a target train, and calculating a target control quantity; acquiring the position of the target train in the group train, and judging a comprehensive index mode according to the position to obtain an index optimal weight combination; based on a target control quantity and an index optimal weight combination, dividing and setting a sub-parameter control target and an optimal weight through a main controller; calculating a distribution control law based on the sub-parameter control target and the optimal weight, and obtaining control output; and according to the position of the target train, redundancy comparison is carried out based on train data of a forward train, and if a preset condition is met, control smooth correction is carried out on control output. According to the cooperative control method under the optimal group train operation index combination, the real-time performance and reliability of the control effect are improved.

Description

Technical Field

[0001] This application belongs to the technical field of train operation control, and particularly relates to a cooperative control method and system under optimal combination of train operation indicators for a group of trains. Background Technique

[0002] The train operation control system is the core technology to ensure railway operation safety and improve transportation efficiency. It ensures the safety and smoothness of trains in complex operation environments by real-time controlling the train speed, tracking the target curve, and dynamically adjusting the operation interval. Currently, the widely used China Train Control System (CTCS) adopts a hierarchical technical system (CTCS-0 level to CTCS-3 level) to achieve speed and interval control through vehicle-ground communication. However, with the continuous growth of railway transport capacity requirements, traditional train control systems face challenges in scenarios such as compressed tracking intervals and high-density operations. For this reason, the industry has proposed advanced technologies such as CTCS-N, moving block, virtual coupling, and train group control. Among them, the train group operation control system is based on vehicle-vehicle and vehicle-ground cooperative communication, and realizes multi-train dynamic cooperation through a speed-distance two-dimensional control mode, further shortening the tracking interval and improving the section passing capacity.

[0003] At the control algorithm level, traditional train operation control mostly uses the PID control method, which has a simple structure and rapid response, but is insufficient in adapting to non-linear and time-varying working conditions (such as slope changes, load fluctuations, communication delays, etc.), and it is difficult to meet the precise control requirements in high-dynamic scenarios. In recent years, Model Predictive Control (MPC) has become a research hotspot due to its advantages in dealing with multi-variable constraints, time delays, and non-linear problems. MPC can dynamically adjust the control input through rolling horizon optimization and feedback correction, but it has problems such as increased computational cost due to multi-step prediction and corresponding prediction optimization processes, which affect control real-time performance, and too many assumptions and constraints are required, resulting in a low degree of freedom of the entire algorithm. Once relevant assumptions are violated, it may lead to the collapse of the algorithm performance. Summary of the Invention

[0004] In view of the deficiencies in the prior art, the present application provides a cooperative control method and system under the optimal combination of group train operation indicators. By adding a comprehensive index mode determination module before the execution of the control algorithm, using a neural network to calculate the optimal combination of the control target index weights of the train's real-time operation, designing the controller as a main controller and a distributed controller, dividing the control target and weights by the main controller, then selecting an appropriate distributed control law for calculation, and combining the control information obtained from the state information of the second train ahead for smoothing correction, and finally outputting the control, a real-time stable and smooth control effect more suitable for the operation environment and the train operation requirements can be obtained, improving the real-time performance and reliability of the control effect.

[0005] The present application is realized through the following technical solutions:

[0006] Obtain the target position, speed curve, operation state data and group line data of the target train, and calculate the target control quantity;

[0007] Obtain the position of the target train in the group train, and based on the position, determine the comprehensive index mode to obtain the optimal weight combination of the index;

[0008] Based on the target control quantity and the optimal weight combination of the index, divide and set the sub-parameter control target and the optimal weight by the main controller;

[0009] Based on the sub-parameter control target and the optimal weight, calculate the distributed control law to obtain the control output;

[0010] According to the position of the target train, based on the train data of the train ahead for redundancy comparison, and combine the redundancy comparison result to judge whether it meets the preset conditions. If it meets the preset conditions, perform control smoothing correction on the control output.

[0011] Optionally,

[0012] The determining the comprehensive index mode according to the position to obtain the optimal weight combination of the index includes:

[0013] If the target train is the leading train in the group, determine the comprehensive index mode based on the train data of the target train to obtain the optimal weight combination of the index;

[0014] If the target train is not the leading train in the group, receive the train data of the train ahead according to the preset communication topology, and determine the comprehensive index mode based on the train data of the target train and the train data of the train ahead to obtain the optimal weight combination of the index;

[0015] Among them, the train data includes: operation state data and control data.

[0016] Optionally,

[0017] The train data-based determination comprehensive index mode for the target train to obtain the optimal weight combination of the index includes: based on the operation status data of the target train, combined with the group scheduling plan, obtaining the optimal weight combination of the index of the target train through neural network learning;

[0018] The train data-based determination comprehensive index mode for the target train and the train data of the forward train to obtain the optimal weight combination of the index includes: based on the operation status data of the target train and the operation status data of the forward train, combined with the group scheduling plan, obtaining the optimal weight combination of the index of the target train through neural network learning.

[0019] Optionally,

[0020] Based on the target control quantity and the optimal weight combination of the index, dividing and setting the sub-parameter control target and the optimal weight by the main controller includes:

[0021] According to the target control quantity and the optimal weight combination of the index, using the MPC main controller to design an error cost function, and dividing and setting the sub-parameter control target and the optimal weight.

[0022] Optionally,

[0023] Calculating the distributed control law based on the sub-parameter control target and the optimal weight includes:

[0024] Based on the sub-parameter control target and the optimal weight, calculating the distributed control law according to the non-linearity degree of the sub-parameter control target to obtain the control output.

[0025] Optionally,

[0026] Based on the position of the target train, performing redundancy comparison based on the train data of the forward train, and combining the redundancy comparison result to determine whether a preset condition is satisfied. If the preset condition is satisfied, performing control smoothing correction on the control output includes:

[0027] If the target train is not the first train or the second train in the group, performing redundancy comparison based on the control data of the forward train;

[0028] Determining whether the preset condition is satisfied. If the preset condition is satisfied, performing control smoothing correction on the control output;

[0029] Wherein, the preset condition includes: the comparison result is that the deviation does not exceed the preset threshold, and the received train data contains additional information data.

[0030] This application also provides a cooperative control system under the optimal combination of group train operation indexes for implementing the foregoing method. The system includes:

[0031] The dynamic analysis module is used to obtain the target position, speed curve, operating state data and group line data of the target train, and calculate the target control quantity;

[0032] The comprehensive index mode determination module is used to obtain the position of the target train in the group of trains, and based on the position, receive the train data of the forward train, determine the comprehensive index mode, and obtain the optimal weight combination of the index;

[0033] The main controller module is used to divide and set the sub-parameter control target and the optimal weight based on the target control quantity and the optimal weight combination of the index;

[0034] The distributed controller module is used to calculate the distributed control law based on the sub-parameter control target and the optimal weight, and obtain the control output;

[0035] The smoothing correction module is used to perform redundant comparison based on the train data of the forward train according to the position of the target train, and combine the redundant comparison result to judge whether the preset condition is satisfied. If the preset condition is satisfied, the control output is corrected for control smoothing.

[0036] Optionally,

[0037] The comprehensive index mode determination module is further configured to:

[0038] If the target train is the leading train in the group, determine the comprehensive index mode based on the train data of the target train, and obtain the optimal weight combination of the index;

[0039] If the target train is not the leading train in the group, receive the train data of the forward train according to the preset communication topology, and determine the comprehensive index mode based on the train data of the target train and the train data of the forward train, and obtain the optimal weight combination of the index;

[0040] Wherein, the train data includes: operating state data and control data.

[0041] Optionally,

[0042] The main controller module is further configured to:

[0043] According to the target control quantity and the optimal weight combination of the index, use the MPC main controller to design the error cost function, and divide and set the sub-parameter control target and the optimal weight.

[0044] Optionally,

[0045] The distributed controller module is further configured to:

[0046] Based on the sub-parameter control target and the optimal weight, calculate the distributed control law according to the non-linear degree of the sub-parameter control target, and obtain the control output.

[0047] Optionally,

[0048] the smoothing correction module is further configured to:

[0049] if the target train is neither the first train of the group nor the second train, perform redundant comparison based on the control data of the forward train;

[0050] judge whether a preset condition is satisfied. If the preset condition is satisfied, perform control smoothing correction on the control output;

[0051] wherein, the preset condition includes: the comparison result is that the deviation does not exceed a preset threshold, and the received train data includes additional information data.

[0052] The present application also provides a computer-readable storage medium storing one or more programs, which when executed, can implement the foregoing collaborative control method under the optimal combination of group train operation indexes.

[0053] The present application also provides a device, including a processor, a communication interface, a computer-readable storage medium, and a communication bus; wherein, the processor, the communication interface, and the computer-readable storage medium communicate with each other through the communication bus;

[0054] the processor is used to execute the programs stored in the computer-readable storage medium.

[0055] Compared with the prior art, the present application has the following advantages:

[0056] 1. The collaborative control method under the optimal combination of group train operation indexes proposed by the present application calculates the optimal combination of the control target index weights of the real-time train operation by adding a comprehensive index mode determination module before the execution of the control algorithm, designs the controller as a main controller and a distributed controller, divides the control target and weights by the main controller, then selects a suitable distributed control law for calculation, performs smoothing correction on the control information obtained by combining the state information of the forward second train, and finally outputs the control, so as to obtain a real-time stable and smooth control effect that is more suitable for the operating environment and train operation requirements, and improves the real-time performance and reliability of the control effect.

[0057] 2. Through the comprehensive index mode determination module, combining the motion state information of the current train and the forward train, and statistically analyzing the operation state data such as line data, historical operation records, and dispatching plans, the train operation control target under the input conditions of the corresponding train motion state variables is obtained through neural network learning. That is, by learning and comprehensively judging, under the conditions of different operation states in the area where the current train is located, the optimal weight combination of the control optimization indexes in terms of low energy consumption, high efficiency, smooth control curve, fast response speed, short stable convergence time, and strong robustness is obtained. The control domain is estimated based on the spatio-temporal domain where the vehicle is located, so as to better meet the actual operation requirements in different environments. According to the spatio-temporal domain requirements such as operation plans and line data, a comprehensive result of multi-index optimization of the control target is given, eliminating the phenomenon that the optimization of a single index in ordinary control methods may cause the deterioration of other indexes, and improving the control effect.

[0058] 3. Through the mode of "master controller + multiple distributed controllers", combining the target optimization characteristics of MPC in the master controller, the target control value and the optimal weight are assigned to the subsequent distributed controllers. In the distributed controllers, the corresponding distributed control law is adopted according to the control target value, so as to obtain the optimal combined stable cooperative control effect. In terms of the amount of calculation, on the one hand, only the set value of the control target is iteratively optimized in the MPC master controller, and on the other hand, after the parameter optimization, the overall complex matrix is also dispersed into multiple small distributed matrices, greatly reducing the amount of calculation and further ensuring the real-time performance and reliability of the control effect.

[0059] 4. Data is transmitted according to the preset communication topology. Each train in the group sends its own data to the two adjacent trains in the rear, and redundant comparison is carried out in combination with the information data of the forward train. When the preset conditions are met, the control output is smoothly corrected, which can reduce the accumulation of chain errors in the train group operation scenario and improve the train operation safety and control smoothness.

[0060] Other features and advantages of the present application will be described in the subsequent specification, and some of them will become obvious from the specification, or can be understood by implementing the present application. The objectives and other advantages of the present application can be achieved and obtained through the structures pointed out in the specification, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0062] Figure 1Shows the schematic flow chart of the cooperative control method under the optimal combination of group train operation indicators;

[0063] Figure 2 Shows the schematic block diagram of the structure of the cooperative control system under the optimal combination of group train operation indicators;

[0064] Figure 3 Shows the schematic flow chart of the cooperative control under the optimal combination of group train operation indicators in the embodiment of the present application;

[0065] Figure 4 Shows the schematic diagram of the cooperative control principle under the optimal combination of group train operation indicators in the embodiment of the present application;

[0066] Figure 5 Shows the schematic diagram of the group information transmission communication topology in the embodiment of the present application;

[0067] Figure 6 Shows the schematic diagram of the neural network model for learning the optimal weight combination of the comprehensive index mode in the embodiment of the present application;

[0068] Figure 7 Is the schematic diagram of the structure of a device in the embodiment of the present application. Specific embodiments

[0069] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Apparently, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0070] See the appendix Figure 1 , the method of the present application includes:

[0071] S1. Obtain the target position, speed curve, operation state data, and group line data of the target train, and calculate the target control quantity;

[0072] S2. Obtain the position of the target train in the group of trains, and determine the comprehensive index mode according to the position to obtain the optimal weight combination of the indicators;

[0073] Wherein, if the target train is the leading train of the group, determine the comprehensive index mode based on the train data of the target train to obtain the optimal weight combination of the indicators;

[0074] If the target train is not the leading train of the group, receive the train data of the forward train according to the preset communication topology, and determine the comprehensive index mode based on the train data of the target train and the train data of the forward train to obtain the optimal weight combination of the indicators;

[0075] Among them, the train data includes: operation status data and control data.

[0076] Among them, the comprehensive index mode for determining the optimal weight combination based on the train data of the target train includes: obtaining the optimal weight combination of the target train through neural network learning according to the operation status data of the target train in combination with the group scheduling plan.

[0077] The comprehensive index mode for determining the optimal weight combination based on the train data of the target train and the train data of the forward train includes: obtaining the optimal weight combination of the target train through neural network learning according to the operation status data of the target train and the operation status data of the forward train in combination with the group scheduling plan.

[0078] S3. Based on the target control quantity and the optimal weight combination of the index, divide and set the sub-parameter control target and the optimal weight through the main controller.

[0079] Among them, according to the target control quantity and the optimal weight combination of the index, design an error cost function using the MPC main controller, and divide and set the sub-parameter control target and the optimal weight.

[0080] S4. Based on the sub-parameter control target and the optimal weight, calculate the distributed control law and obtain the control output.

[0081] Among them, based on the sub-parameter control target and the optimal weight, calculate the distributed control law according to the non-linear degree of the sub-parameter control target and obtain the control output.

[0082] S5. According to the position of the target train, perform redundancy comparison based on the train data of the forward train, and combine the redundancy comparison result to determine whether the preset condition is satisfied. If the preset condition is satisfied, perform control smoothing correction on the control output.

[0083] Among them, if the target train is not the first train or the second train in the group, perform redundancy comparison based on the control data of the forward train.

[0084] Determine whether the preset condition is satisfied. If the preset condition is satisfied, perform control smoothing correction on the control output.

[0085] Among them, the preset condition includes: the comparison result is that the deviation does not exceed the preset threshold, and the received train data contains additional information data.

[0086] Specifically,

[0087] The embodiment of the present application discloses a cooperative control method under the optimal combination of group train operation indicators, including the following steps:

[0088] Step 1: Obtain the target position, speed curve, operation status data, and group line data of the target train as input data.

[0089] In this embodiment, according to the operation plan and dispatching data of the group trains, the target position and speed curve of the target train at the current time are obtained, that is, the target speed and position that the train needs to finally reach through control outputs such as continuous traction and braking. At the same time, the operation status data and group line data of the target train are read, and the target position and speed curve, train operation status, train real-time position information, and group line data are used as input variables together.

[0090] Among them, the operation status data includes: historical operation data and real-time motion status data.

[0091] Step 2: Calculate the target control quantity to be achieved according to the target position and speed curve, operation status data, and group line data of the target train.

[0092] Substitute the target position and speed curve, operation status data, and group line data of the target train into the longitudinal dynamics model of the group trains to calculate the target control quantity, including:

[0093] Compare the target position and speed curve with the operation status data and group line data to obtain the real-time position and speed deviation of the target train;

[0094] According to the real-time position and speed deviation of the target train, determine the target acceleration, and combine it with the longitudinal dynamics model of the train to calculate the target control quantity, that is, the traction input required by the target train.

[0095] In this embodiment, the expression of the longitudinal dynamics model of the group trains is as follows:

[0096]

[0097] Where p i , v i , m i are the displacement, speed, and mass of the i-th train respectively; is the second derivative of the position, that is, the acceleration; u i is the traction input of the i-th train; k i , d i are the elastic coefficient and damping coefficient between trains respectively; a, b are the empirical coefficients of the Davis equation; θ i , D i are the gradient angle and curvature of the running line of the i-th train.

[0098] In this embodiment, the target position and speed curve at the current time are used as known conditions for the entire control method, and the control quantity is output to try to reach the target position and speed curve as much as possible. The establishment and output of the longitudinal dynamics model are based on a detailed force analysis of the train, and the target control quantity is calculated as the target to be achieved by the next control law.

[0099] Step 3: Obtain the position of the target train in the group of trains.

[0100] In this embodiment, after determining the initial state of the target train and collecting the operation state data, the position of the target train in the group of trains is obtained, and data transmission is performed according to the preset communication topology:

[0101] If the target train is the leading train in the group (i = 1), the train data of the target train is sent to the two adjacent trains in the backward direction;

[0102] If the target train is not the leading train in the group (i = 2, 3... N, where N is the number of trains in this group), the train data of the forward train is received according to the preset communication topology, and the train data of the target train is sent to the two adjacent trains in the backward direction.

[0103] Among them, the train data includes: operation state data and control data.

[0104] In this embodiment, receiving the train data of the forward train according to the preset communication topology includes:

[0105] If the target train is the second train in the group, the train data of the leading train is received;

[0106] If the target train is the train at the third and subsequent positions in the group (i = 3, 4... N, where N is the number of trains in this group), the train data of the immediately preceding train and the second forward train is received.

[0107] In this embodiment, the train data is transmitted through wireless communication, and the transmitted train data includes but is not limited to: the real-time target position and position curve of the forward train, the real-time position, speed, acceleration of the train, and other operation control-related information such as historical operation data and dispatching information.

[0108] Step 4: According to the position of the target train, combined with the group scheduling plan, line data, and train data, determine the comprehensive index mode to obtain the optimal weight combination of the indexes.

[0109] According to the position of the target train determined in Step 3:

[0110] If the target train is the leading train in the group, according to the operation state data of the target train, combined with the group scheduling plan, line data, historical operation data, and dispatching information and other train data, the optimal weight combination of the indexes of the target train is obtained through neural network learning;

[0111] If the target train is not the first train in the group, based on the operation status data of the target train and the forward train, combined with train data such as the group scheduling plan, line data, historical operation data, and scheduling information, the optimal weight combination of the indicators of the target train is obtained through neural network learning.

[0112] In this embodiment, if the target train is not the first train in the group, based on the operation status data of the target train and the forward train, combined with train data such as the group scheduling plan, line data, historical operation data, and scheduling information, the optimal weight combination of the indicators of the target train is obtained through neural network learning, including:

[0113] Using the target train motion state, forward train motion state, line data, historical operation data, and scheduling plan, etc. as the input layer data; through the hidden layer algorithm, the optimal weight combination of the output layer (the sum of all weights is 1) is obtained, which respectively represents the weights corresponding to energy consumption, control efficiency, control curve smoothness, response speed, convergence time, and robustness, that is, the importance of the target.

[0114] In this embodiment, if the target train is not the first train in the group, according to the motion state information of this train and the forward train, combined with the group scheduling plan, and by counting train data such as line data, historical operation records, and scheduling information, the train operation control target under the input conditions of the corresponding train motion state variables is obtained through neural network learning, that is, by learning and comprehensively judging the train in the current area and different operation state conditions within the area, the optimal weight combination of the control optimization indicators in terms of low energy consumption, high efficiency, smooth control curve, fast response speed, short stable convergence time, and strong robustness is obtained, and the control domain is estimated based on the space-time domain where the vehicle is located, so as to better meet the actual operation requirements in different environments.

[0115] Step 5: Based on the target control quantity and the optimal weight combination of the indicators, design an error cost function, and divide and set the sub-parameter control target and the optimal weight.

[0116] In this embodiment, according to the target control quantity and the optimal weight combination of the indicators, an error cost function is designed using the MPC main controller, and the sub-parameter control target and the optimal weight are divided and set.

[0117] Combined with the target control quantity and the optimal weight combination of the indicators, an error cost function is designed through the MPC main controller and the control law is executed. The error cost function is as follows:

[0118]

[0119] where y(t), are the system state and expected state at time t, respectively; J is the objective function, which indicates the degree of closeness between the system state and the expected state in the next n time steps; c is the optimal weight combination of indicators; c j is the optimal weight of the jth indicator, y j is the state of the jth indicator at time t (j=1, 2, 3...m); the state errors corresponding to m different types of control optimization indicators are weighted and combined to find the minimum value to obtain the corresponding control setting value; k is the prediction step size.

[0120] Different from the common MPC control law, the control target of the MPC master controller in this embodiment is the control setting value assigned according to the combination of indicators, which provides control target segmentation and sub-parameter optimal weight calculation for the next step of calculating the distributed control law, and performs prediction and iterative optimization according to the common MPC algorithm. This process divides the relevant matrix calculation in the entire MPC control law from a total matrix into multiple small matrices, and only calculates the target setting and weight allocation of the sub-parameters, which greatly reduces the matrix dimension and computational complexity of the current common MPC method, enhances real-time performance, and avoids the risk of the MPC control algorithm itself collapsing due to too many conditional assumptions and constraints.

[0121] Step 6: Based on the sub-parameter control target and the optimal weight, the distributed control law is calculated according to the nonlinear degree of the sub-parameter control target to obtain the control output.

[0122] In this embodiment, based on the sub-parameter control targets and optimal weights divided and set by the MPC master controller, the distributed control law is calculated according to the nonlinearity of the sub-parameter control targets:

[0123] (1) For parameters with low nonlinearity, the traditional PID control algorithm can be used, which has fast response speed and excellent performance. The distributed control law expression is as follows:

[0124]

[0125] Where u(t) is the controller output; e(t) is the system error; K p , K i , K d They are proportional coefficient, integral coefficient and differential coefficient respectively; by adjusting the three coefficients of proportional P, integral I and differential D, safe and stable closed-loop regulation performance can be obtained.

[0126] (2) For parameters with obvious nonlinearity, a distributed control algorithm based on fixed time or preset performance conditions can be used. The control output can converge within a fixed time and does not exceed the upper and lower bounds, thereby improving the efficiency and real-time performance of the overall control method. The distributed control law expression is:

[0127]

[0128] Among them, u i is the output of the controller of the i-th vehicle; are respectively the differences between the position, speed and expected value of the i-th vehicle; s1, s2, μ1, μ2, α1, α2 are constants; a i,i-1 is the coupling relationship connection weight between the train and the immediately preceding train; a i0 is the connection weight between the train and the leading vehicle of the group; R(v i ) is the resistance suffered by the i-th vehicle.

[0129] Step 7: According to the position of the target train, perform redundancy comparison based on the train data of the forward train, and judge whether the preset conditions are met in combination with the redundancy comparison result. If the preset conditions are met, perform control smoothing correction on the control output.

[0130] In this embodiment, if the target train is not the leading vehicle of the group and the second train, perform redundancy comparison based on the control data of the forward train;

[0131] Judge whether the preset conditions are met. If the preset conditions are met, perform control smoothing correction on the control output;

[0132] Among them, the preset conditions include: the comparison result is that the deviation does not exceed the preset threshold, and the train data contains additional information data.

[0133] In this embodiment, judge whether the target train is the leading vehicle of the group and the second train: If the target train is not the leading vehicle of the group and the second train, then make a redundancy comparison between the control law and prediction calculated by the forward second train main controller received according to the communication topology and the control law and prediction calculated by the immediately preceding train main controller. If the result of the redundancy comparison is that the deviation exceeds the preset threshold, prompt an alarm for further processing to reduce the communication transmission delay and chain control error caused by the increase in the number of trains in the group, reduce the accidental risk caused by unknown special situations, and ensure the train operation safety;

[0134] If the result of the redundancy comparison is that the deviation does not exceed the preset threshold, and at the same time the train data contains additional information such as the reliability of communication quality and the error and covariance of the train operation control state, perform control smoothing correction according to the train data of the forward train to ensure the smooth operation and operation safety of the train.

[0135] In this embodiment, the formula for control smoothing correction is as follows:

[0136]

[0137] Among them, u' i is the output of the smoothing correction controller of the i-th vehicle; is the output of the j-th distributed controller of the i-th vehicle; is the output of the j-th distributed controller of the i-th vehicle after predicting k steps; ε and ε k are respectively the weight value of the output of the distributed controller of the i-th vehicle and the weight value corresponding to the predicted k steps, which can be set according to the specific operating conditions and prediction steps.

[0138] In this embodiment, the target train receives the operation state data of the forward train based on the communication topology, and obtains a more reliable operation state and change trend through the control output data of the immediately preceding train and the second forward train, so as to output a more smoothly changing control information, ensure the stability and safety of the train operation state, and reduce the energy consumption caused by sudden train traction and braking.

[0139] In this embodiment, if the target train is the leading vehicle and the second vehicle in the group, redundant comparison and control smoothing correction are not required.

[0140] Step 8: Based on the control output / smoothed corrected control output, output a control signal, and feedback the control output / smoothed corrected control output and the real-time state data of the target train to the reference input end, and continue the calculation for the next moment.

[0141] The present application will be further elaborated in detail below with reference to the accompanying drawings and specific embodiments.

[0142] See the attached Figure 2 shows the structure of a cooperative control system under the optimal combination of group train operation indexes for implementing the above method, including: a dynamic analysis module, a comprehensive index mode determination module, a main controller module, a distributed controller module, and a smoothing correction module.

[0143] The dynamic analysis module is used to obtain the target position and speed curve, operation state data, and group line data of the target train, and calculate the target control quantity;

[0144] The comprehensive index mode determination module is used to obtain the position of the target train in the group train, and determine the comprehensive index mode according to the position to obtain the optimal weight value combination of the indexes;

[0145] The main controller module is used to divide and set the sub-parameter control target and the optimal weight value through the main controller based on the target control quantity and the optimal weight value combination of the indexes;

[0146] The distributed controller module is used to calculate the distributed control law based on the sub-parameter control target and the optimal weight value, and obtain the control output;

[0147] A smoothing correction module, which is used to perform redundant comparison based on the train data of the leading train according to the position of the target train, and determine whether a preset condition is satisfied in combination with the redundant comparison result. If the preset condition is satisfied, control smoothing correction is performed on the control output.

[0148] Figure 3 This is a schematic diagram of the cooperative control process under the optimal combination of group train operation indicators in the embodiment of the present application. In this embodiment, the specific process includes:

[0149] Read the target position, speed curve, operation status data and group line data of the target train;

[0150] Determine whether the target train is the leading train of the group: If it is not the leading train of the group, read the train data of the leading train according to the communication topology;

[0151] According to the operation section of the target train, obtain the optimal combination of conditional indicators and the corresponding weights through the experience library and deep learning;

[0152] Design an error cost function according to the optimal combination of indicators and the corresponding weights, and divide the main set value and the optimal weights of the sub-parameters;

[0153] Design a distributed control law according to the sub-parameter target set value and the optimal weights of the sub-parameters provided by the main controller;

[0154] Determine whether the target train is the leading train or the second train of the group: If it is not the leading train or the second train of the group, perform redundant comparison in combination with the control data of the leading train and further smooth and correct the control output;

[0155] Obtain the final control output.

[0156] Figure 4 This is a schematic diagram of the cooperative control principle under the optimal combination of group train operation indicators in the embodiment of the present application. In this embodiment, the target position, speed curve and operation status data are used as reference inputs and input into the longitudinal dynamics model to obtain the target control quantity; through comprehensive index mode learning and determination, the optimal weight combination of indicators is obtained; the target control quantity and the optimal weight combination of indicators are input into the MPC main controller to obtain the main set value and the optimal weights of the sub-parameters, and then input into the distributed controller; perform control smoothing correction on the control output calculated by the distributed controller to obtain the final control output and input it into the control platform.

[0157] Among them, the MPC main controller performs iterative optimization of the control parameters through the model and prediction; the final control output is fed back to the reference input end.

[0158] Figure 5 This is a schematic diagram of the group information transmission communication topology in the embodiment of the present application. In this embodiment, each train in the group receives and transmits data according to the communication topology:

[0159] Each train sends its own train data to the two adjacent trains in the backward direction.

[0160] Specifically, if the target train is the first train in the group (i = 1), it sends its own train data to the two adjacent trains in the backward direction;

[0161] If the target train is the second train in the group (i = 2), it sends its own train data to the two adjacent trains in the backward direction and receives the train data of the first train;

[0162] If the target train is neither the first train nor the second train in the group (i = 3, 4... N, where N is the number of trains in this group), it sends its own train data to the two adjacent trains in the backward direction and receives the train data of the train immediately in front and the second train in the forward direction.

[0163] Figure 6 It is a schematic diagram of the neural network model for the optimal weight combination learning of the comprehensive index mode in the embodiment of this application. In this embodiment, the neural network model includes an input layer, a hidden layer, and an output layer. Among them, the input layer includes: the motion state of the target train, the motion state of the forward train, line data, historical operation data, and scheduling plans, etc.; the output layer includes: the optimal weights respectively representing the importance degrees corresponding to energy consumption, control efficiency, control curve smoothness, response speed, convergence time, and robustness.

[0164] In addition, the embodiment of this application also provides a cooperative control device under the optimal combination of group train operation indicators, including:

[0165] A dynamics analysis module, which acquires the target position and speed curve, operation state data, and group line data of the target train, and calculates the target control quantity;

[0166] A comprehensive index mode determination module, which acquires the position of the target train in the group of trains, and determines the comprehensive index mode according to the position to obtain the optimal weight combination of the indicators;

[0167] A main controller module, which divides and sets the sub-parameter control target and the optimal weight through the main controller based on the target control quantity and the optimal weight combination of the indicators;

[0168] A distributed controller module, which calculates the distributed control law based on the sub-parameter control target and the optimal weight to obtain the control output;

[0169] A smoothing correction module, which based on the position of the target train, makes a redundant comparison based on the train data of the forward train, and combines the redundant comparison result to judge whether the preset condition is satisfied. If the preset condition is satisfied, it performs control smoothing correction on the control output.

[0170] Based on the same application concept, the present application further provides a computer-readable storage medium storing one or more programs, which, when executed, can implement the collaborative control method under the optimal combination of the above-mentioned group train operation indicators.

[0171] As Figure 7 shown, the embodiment of the present application further provides a device, including a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory communicate with each other through the communication bus.

[0172] The memory is a computer-readable storage medium for storing one or more programs.

[0173] The processor is configured to execute the programs stored in the computer-readable storage medium.

[0174] This computer-readable storage medium may be included in the device / apparatus described in the above embodiments; or it may exist alone without being assembled into the device / apparatus.

[0175] Although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A coordinated control method for optimal combination of group train operation indicators, characterized in that: include: Obtain the target position and speed curve, running status data and group line data of the target train, and calculate the target control amount; Obtaining the position of the target train in the group of trains, determining the comprehensive index mode according to the position, and obtaining the optimal weight combination of the indexes; Based on the target control quantity and the optimal weight combination of the indicators, the main controller is used to divide and set the sub-parameter control targets and the optimal weights; Based on the sub-parameter control objectives and the optimal weights, a distributed control law is calculated to obtain a control output; According to the position of the target train, a redundant comparison is performed based on the train data of the forward train, and whether a preset condition is met is determined in combination with the redundant comparison result. If the preset condition is met, a control smoothing correction is performed on the control output.

2. The method according to claim 1, characterized in that: The step of determining the comprehensive indicator mode according to the position and obtaining the optimal weight combination of indicators includes: If the target train is the first train of the group, the comprehensive index mode is determined based on the train data of the target train to obtain the optimal weight combination of the indexes; If the target train is not the first train of the group, the train data of the preceding train is received according to the preset communication topology, and the comprehensive index mode is determined based on the train data of the target train and the train data of the preceding train to obtain the optimal weight combination of the indexes; The train data includes: running status data and control data.

3. The method according to claim 2, characterized in that The determining of the comprehensive index mode based on the train data of the target train to obtain the optimal weighted combination of the indexes includes: obtaining the optimal weighted combination of the indexes of the target train through neural network learning according to the running status data of the target train and in combination with the group scheduling plan; The comprehensive indicator mode is determined based on the train data of the target train and the train data of the preceding train to obtain the optimal weight combination of indicators, including: according to the running status data of the target train and the running status data of the preceding train, combined with the group scheduling plan, obtaining the optimal weight combination of indicators of the target train through neural network learning.

4. The method according to claim 1, characterized in that: Based on the target control amount and the optimal weight combination of the index, the main controller is used to divide and set the sub-parameter control targets and the optimal weights, including: According to the target control quantity and the optimal weight combination of the index, the error cost function is designed by using the MPC main controller, and the sub-parameter control targets and the optimal weights are divided and set.

5. The method according to claim 1, characterized in that The calculating of the distributed control law based on the sub-parameter control target and the optimal weight includes: Based on the sub-parameter control target and the optimal weight, the distributed control law is calculated according to the nonlinear degree of the sub-parameter control target to obtain the control output.

6. The method according to any one of claims 1 to 5, characterized in that: The method of performing redundant comparison based on the position of the target train and the train data of the forward train, judging whether a preset condition is met in combination with the redundant comparison result, and performing control smoothing correction on the control output if the preset condition is met, comprises: If the target train is not the first train or the second train of the group, performing a redundancy comparison based on the control data of the forward train; Determining whether a preset condition is met, and if the preset condition is met, performing control smoothing correction on the control output; The preset conditions include: the comparison result is that the deviation does not exceed a preset threshold, and the received train data contains additional information data.

7. A coordinated control system for optimal combination of group train operation indicators, characterized in that: The system comprises: Dynamic analysis module, used to obtain the target position and speed curve of the target train, running status data and group line data, and calculate the target control amount; A comprehensive index mode determination module is used to obtain the position of the target train in the group of trains, determine the comprehensive index mode according to the position, and obtain the optimal weight combination of the indexes; A main controller module, used for dividing and setting sub-parameter control targets and optimal weights based on the target control amount and the optimal weight combination of indicators; A distributed controller module, used to calculate the distributed control law based on the sub-parameter control target and the optimal weight to obtain the control output; The smoothing correction module is used to perform redundant comparison based on the train data of the forward train according to the position of the target train, and judge whether the preset conditions are met in combination with the redundant comparison results. If the preset conditions are met, the control output is controlled to be smoothed and corrected.

8. The system according to claim 7, characterized in that The comprehensive indicator mode determination module is also configured to: If the target train is the first train of the group, the comprehensive index mode is determined based on the train data of the target train to obtain the optimal weight combination of the indexes; If the target train is not the first train of the group, the train data of the preceding train is received according to the preset communication topology, and the comprehensive index mode is determined based on the train data of the target train and the train data of the preceding train to obtain the optimal weight combination of the indexes; The train data includes: running status data and control data.

9. The system according to claim 7, characterized in that The main controller module is also configured to: According to the target control quantity and the optimal weight combination of the index, the error cost function is designed by using the MPC main controller, and the sub-parameter control targets and the optimal weights are divided and set.

10. The system according to claim 7, characterized in that The distributed controller module is further configured to: Based on the sub-parameter control target and the optimal weight, the distributed control law is calculated according to the nonlinear degree of the sub-parameter control target to obtain the control output.

11. The system according to any one of claims 7 to 10, characterized in that: The smoothing correction module is further configured to: If the target train is not the first train or the second train of the group, performing a redundancy comparison based on the control data of the forward train; Determining whether a preset condition is met, and if the preset condition is met, performing control smoothing correction on the control output; The preset conditions include: the comparison result is that the deviation does not exceed a preset threshold, and the received train data contains additional information data.

12. A computer-readable storage medium storing one or more programs, characterized in that: When the one or more programs are executed, the coordinated control method under the optimal combination of group train operation indicators as described in any one of claims 1 to 6 can be implemented.

13. An electronic device comprising a processor, a communication interface, the computer-readable storage medium of claim 12, and a communication bus; wherein: The processor, the communication interface, and the computer-readable storage medium communicate electronically with each other via a communication bus; characterized in that: The processor is configured to execute a program stored in a computer-readable storage medium.

Images