A trajectory prediction-based formation train safety protection method

Abstract

The application discloses a kind of based on trajectory prediction's platoon train safety protection method, comprising: obtaining the parameter information of platoon train in line and line operation parameter information, constructs the dynamics model of platoon train, the parameter of basic resistance in dynamics model is identified based on improved particle swarm algorithm, obtains single point platoon train dynamics model;According to the historical data of the preceding train in platoon train, the running curve of preceding train future time is predicted by LSTM-UKF trajectory fusion algorithm, the running curve of preceding train future time is solved according to single point platoon train dynamics model, obtains the safety protection curve of preceding train in platoon train, and the position relationship of preceding train and rear train in platoon train obtained according to tracking interval formula constructs the safety protection control model of rear train in platoon train;The safety protection control model of rear train is solved, and the operation of platoon train is realized according to the solution result.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of train automatic protection technology. More particularly, it relates to a train formation safety protection method based on trajectory prediction. BACKGROUND

[0002] At present, the urban rail transit system has the characteristics of large passenger turnover, punctuality, rapidity, safety, environmental protection and the like, and plays a vital role in alleviating urban traffic congestion. In recent years, with the continuous expansion of the city size in China, the tidal passenger flow caused by job and residence separation and the sudden passenger flow brought by large-scale holiday activities lead to station congestion and decrease of line network transportation capacity, thereby seriously affecting the operation quality and passenger experience of urban rail transit. At present, the common moving block train operation control system has reached the limit of its designed operation capacity, and cannot realize the rapid transfer of large-scale passenger flow by increasing the train density in the local area. How to realize the dynamic matching of transportation capacity and passenger flow and meet the rapidly changing passenger transport demand has become one of the key problems to improve the flexibility and service quality of rail transit transportation.

[0003] In recent years, in order to further improve the transportation capacity of the city, train formation technology has become a new research hotspot in the field of urban rail transit. Train formation technology, i.e. the trains in the formation realize information interaction between two trains through train-to-train communication, adopts the mode of relative braking distance, and realizes train formation operation. This technology allows the running trains to dynamically form and disband according to the passenger flow of the line, so that the trains run in formation with smaller running interval, reasonably match the traffic and passenger flow, and thus improve the utilization rate of the line and relieve the transportation pressure of tidal passenger flow. However, this new train operation mode puts forward new requirements for the train operation control method under the existing moving block. First, the trains run at a high speed, but the running interval is relatively smaller compared with the moving block, so higher precision and safety are required for the safety protection of the trains. Second, the traction and braking performance of the trains in the train formation is different, the running speed has certain difference, and there is a train-to-train communication delay problem, which will also affect the safe operation of the train formation. Finally, the train formation runs at a high speed with small interval and is subject to safety and real-time constraints, the performance difference between the trains needs to be considered, a train formation safety protection model based on the mode of relative braking distance is established, and a suitable algorithm is used to efficiently solve the safety protection curve, so as to realize the safe operation of the train formation. SUMMARY

[0004] The present application aims to provide a train formation safety protection method based on trajectory prediction to solve at least one of the problems in the prior art.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solution:

[0006] A first aspect of the present invention provides a vehicle platooning safety protection method based on trajectory prediction, comprising:

[0007] S1. Obtaining parameter information of train formations within the line and line operation parameter information;

[0008] S2. Constructing a dynamic model including basic train resistance, additional train resistance, and train traction and braking forces based on the parameter information of the train formation within the line and the line operation parameter information, and identifying the basic resistance parameters of the dynamic model using an improved particle swarm algorithm to obtain a single-point train formation dynamic model;

[0009] S3. Based on the historical data of the preceding train in the formation, the LSTM-UKF trajectory fusion algorithm is used to predict the future operation curve of the preceding train. The future operation curve of the preceding train is solved based on the single-particle formation dynamics model to obtain the safety protection curve of the preceding train in the formation.

[0010] S4. Determine a tracking interval formula between trains in a formation based on the train-to-train communication delay, differences in train traction and braking performance, and train speed differences; determine the positional relationship between the leading and trailing trains in the formation based on the tracking interval formula; and construct a safety protection control model for the trailing train in the formation based on the safety protection curve of the leading train in the formation and the positional relationship between the leading and trailing trains.

[0011] S5. Solve the safety protection control model of the following train based on the direct multi-point shooting method based on the distance interval, and realize the operation of the platoon train according to the solution.

[0012] Optionally,

[0013] Optionally, the single-point platoon dynamics model is

[0014]

[0015] Where M is the mass of the train; a(v x ) is the train speed v x The acceleration of the train at t x ) is the train speed v x traction or braking force when base (v x ) is the train speed v x The unit basic resistance when the train is at position x; W(x) is the total additional resistance when the train is at position x; a, b, and c are the basic resistance parameters of the train; where a is the friction resistance of the train, b is the resistance coefficient of the train, and c is the aerodynamic resistance coefficient of the train; W i Add resistance to the ramp; w iis the additional resistance per unit slope; g is the acceleration due to gravity, w r is the additional resistance of the unit curve; R is the curve radius of the line; A is an empirical constant; F(v m ) is the train speed v m Braking force at time v m 、v o and v x are the speeds of the train at the corresponding positions m, o, and x, respectively, and F m is the braking force of the train at the corresponding position m; F o is the braking force of the train at the corresponding position o.

[0016] Optionally, the formula for identifying the parameters of the basic resistance in the dynamic model based on the improved particle swarm algorithm is:

[0017]

[0018] Where j is the particle index, H is the number of particles, is the position of the corresponding parameter of each particle in the particle swarm; is the speed of each particle in the particle swarm corresponding to the parameter; X min Represents the minimum position of each particle in the particle swarm corresponding to the parameter, X max Represents the maximum position of each particle in the particle swarm corresponding to the parameter, V min Represents the minimum velocity of each particle in the particle swarm corresponding to the parameter, V max Represents the maximum speed of each particle in the particle swarm corresponding to the parameter; rand(1,H) represents the generation of a random matrix with 1 row and H columns; and denote the velocity and position of the jth particle in the kth generation, respectively. represents the historical optimal solution searched by j particles in the kth generation, represents the optimal solution found by all particles in the kth generation; w is the inertia factor, c1 is the first learning factor, and c2 is the second learning factor; r1 and r2 are random numbers uniformly distributed between [0,1]; w min and w max are the minimum and maximum values ​​of the inertia factor, f min and f ave are the minimum fitness and average fitness, respectively.

[0019] Optionally, the S3 further includes

[0020] S31, pre-processing the historical data of the preceding train in the train formation;

[0021] S32. Based on the pre-processed historical data, use the LSTM-UKF algorithm to predict the future running curve of the leading train in the train formation;

[0022] S33. Using an inverse algorithm to solve the future time operation curve of the preceding train, obtain a safety protection curve of the preceding train in the formation.

[0023] Optionally, the safety protection curve of the preceding train is

[0024]

[0025] Where S and V are the distance and speed of the train under safe braking, respectively; γ is the rotation coefficient; u(v i ) is the speed v obtained by interpolation based on the traction braking characteristic curve i Unit braking force when R(v i ) is the train speed v i W(x) is the total unit basic resistance of the train at position x; i is the subscript of the time interval segment, i = 1, 2, 3...n; n is the number of time interval segments; Δv i The speed increment for each time interval; Δs i is the braking distance of each time interval; v0 is the initial speed; Δt i is the time interval; U is the unit conversion system.

[0026] Optionally, the position relationship between the preceding train and the following train is

[0027] d min +d breakL =d F_v +d breakF +d safe

[0028] Among them, d min is the minimum safe tracking distance before train braking, d F_v is the running distance of the following train under the communication delay and the reaction time of the preceding train, d breakF is the braking distance of the following train, d breakL is the braking distance of the preceding train, d safe This is the final safe stopping distance for the train.

[0029] Optionally, the safety protection control model for the rear train in the train formation is constructed based on the safety protection curve of the front train in the train formation, the single-particle train formation dynamics model, and the positional relationship between the front train and the rear train, including:

[0030] S421. Assume that the train formation in the safety protection control model for the following train includes a lead train, a following train, and a rear train, and that the lead train, the following train, and the rear train run in the same direction at the same speed within the section.

[0031] S422. Assume that the probability that the locomotive of the following train is always behind the tail of the leading train in the train formation is greater than the lower confidence limit; and under the dynamic behavior of the following train, the probability that the locomotive of the trailing train is always behind the tail of the following train is greater than the lower confidence limit, that is,

[0032]

[0033] Among them, x L is the position of the pilot train, x F is the position of the following train, x R is the position of the rear train, d is the safety interval, and p is the lower confidence limit;

[0034] S423. Construct a safety protection control model for the following train including dynamic relationship constraints, train traction constraints, braking performance constraints, line speed limit constraints, tracking constraints and optimization objectives.

[0035] Optionally, the optimization goal is

[0036] min(∫[x L (t)-x F (t)] 2 dt+∫[x F (t)-x R (t)] 2 dt)

[0037] Among them, x L (t) is the position of the pilot train at time t, x F (t) is the position of the following train at time t, x R (t) is the position of the following train at time t, where t is the time;

[0038] The dynamic relationship constraint is

[0039]

[0040] Where u(x) is the train control force at position x, R(v) is the total basic resistance of the train at speed v, and W(x) is the total additional resistance of the train at position x;

[0041] The train traction constraint is

[0042]

[0043] Among them, uL (x),u F (x) and u R (x) the tractive effort of the lead train, the following train and the rear train respectively; and are the minimum braking force and maximum braking force of the pilot train at speed v, respectively; and are the minimum braking force and maximum braking force of the following train at speed v, respectively; and are the minimum braking force and maximum braking force of the rear train at speed v, respectively;

[0044] The braking performance constraint is

[0045]

[0046] in, The minimum safe distance between the lead train and the following train. is the minimum safe interval between the following train and the rear train; v L is the speed of the pilot train, v F is the speed of the following train, v R is the speed of the rear train; t delay is the braking delay; d safe is the final safe stopping distance of the train; C is a constant; P L is twice the maximum acceleration of the pilot train; P F is twice the maximum acceleration of the following train; P R It is twice the maximum acceleration of the rear train; S F S is the braking reaction time of the following train; R Y is the braking reaction time of the rear train; F Y is the safe stopping distance deviation for following train; R The safe stopping distance deviation for the rear train;

[0047] The tracking constraints are

[0048]

[0049] Among them, x L (t) is the position of the pilot train at time t, x F (t) is the position of the following train at time t, x R (t) is the position of the rear train at time t;

[0050] The line speed limit is

[0051]

[0052] wherein X is a target tracking point; v L (X), v F (X), v R (X) are respectively the speed of the leading train, the following train and the tail train when reaching the respective target tracking point; L sm represents a target point speed limit; v limit (x) is a speed limit at position x.

[0053] Optionally, the safety protection control model of the following train is

[0054]

[0055] Optionally, the S5 further comprises

[0056] S51, dividing the train predicted parking position interval into the same distance interval;

[0057] initializing a set of vectors as the control variable values at N nodes; assuming that the train makes uniform acceleration linear motion on each distance interval, according to the control variable u(x) at each node; using Newton's law of motion and a single-particle dynamics model, the acceleration, speed and time of the train in the distance interval are solved;

[0058] obtaining the state variable value S(x) = [v(x), t(x)] at the node according to the acceleration, speed and time of the train in the distance interval T ,

[0059] The formula for solving the acceleration, speed and time of the train in the distance interval is

[0060]

[0061] wherein u(x) is the traction of the train at position x; v(x) is the speed of the train at position x, t(x) is the time of the train at position x, a i is the acceleration of the i-th time interval section; x i is the position of the i-th time interval section; v i is the speed of the i-th time interval section; u(x i ) is the train control force of the train at position x i ; R(v i ) is the total basic resistance of the train at speed v i ; W(x i ) is the total additional resistance of the train at position x i ; v(x i ) is the speed of the train at position x i ; v(x i+1 ) is the speed of the train at position x i+1The speed at the point; Δx is the distance interval; Δx=x i+1 -x i ; t(x i+1 ) is the train at position x i+1 The moment t(x i ) is the train at position x i moment;

[0062] S52. Bring the values ​​of the control variables and state variables at the nodes into the constraints and objective function, and transform the train safety protection control problem into a nonlinear programming problem.

[0063]

[0064] The objective function is

[0065]

[0066] The constraints are

[0067]

[0068] Among them, F(W) is the objective function; G(W) is the equality constraint; H(W) is the inequality constraint; s i ,q i , Δs i , Δq i is the optimization variable, where s i is the state variable, q i is the input variable, Δs i is the change of the state variable, Δq i is the change of the input variable; Q i 、S i 、R i are the first weighting matrix, the second weighting matrix and the third weighting matrix respectively; C i 、D i are the first matrix and the second matrix respectively; Δs0 is the change of the control variable in the initial state; is the initial condition; s N is the control variable of the terminal state; c N is the offset of the terminal state constraint;

[0069] S53. Use the interior point method to solve the nonlinear programming problem to obtain the safety protection control curves of the following train and the rear train.

[0070] The beneficial effects of the present invention are as follows:

[0071] The present invention focuses on the safety protection control problem of formation trains during operation, and considers line constraints, safety formation constraints, traction and braking constraints, etc., to establish a safety protection control model for the following trains with different traction and braking performance and speed. By comparing the inverse algorithms of time step, distance step and speed step, a solution method for the leading vehicle protection curve based on the time step is obtained. By analyzing the model, the direct multi-point shooting method and the interior point algorithm are selected to solve the model. Combining the actual line with the traction and braking characteristic curve, MATLAB is used to solve the protection curve. The solution results show that the requirements of safe operation can be met during operation and at the parking point. Compared with the formation train based on absolute braking distance, the minimum tracking interval between trains is reduced while ensuring the safe operation of the formation. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0073] Figure 1 A flow chart showing the safety protection method for platooning vehicles based on trajectory prediction of the present invention is shown.

[0074] Figure 2 A schematic diagram illustrating the use of LSTM-UKF to predict the running trajectory of a preceding train in an embodiment of the present invention is shown.

[0075] Figure 3 A schematic diagram of a time-step-based protection curve inverse algorithm in a time-step-based train protection curve calculation model according to an embodiment of the present invention is shown.

[0076] Figure 4 A schematic diagram of a collision under relative braking distance in an embodiment of the present invention is shown.

[0077] Figure 5 A schematic diagram of the safety protection curve of platooning vehicles under different braking performances in an embodiment of the present invention is shown.

[0078] Figure 6 A schematic diagram showing the principle of the platooning vehicle safety protection method based on trajectory prediction of the present invention. DETAILED DESCRIPTION

[0079] In order to more clearly illustrate the present invention, the present invention will be further described below in conjunction with the embodiments and drawings. Similar components in the drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the following specific description is illustrative rather than restrictive and should not be used to limit the scope of protection of the present invention.

[0080] The present invention provides a method for solving the safety protection control curve of platoon vehicles, which uses actual urban rail train operation data and line data to build a simulation test environment, and implements model construction and solution on the MATLAB platform.

[0081] like Figure 1 As shown, a platooning vehicle safety protection method based on trajectory prediction includes:

[0082] S1. Obtaining parameter information of train formations within the line and line operation parameter information;

[0083] S2. Constructing a dynamic model including basic train resistance, additional train resistance, and train traction and braking forces based on the parameter information of the train formation within the line and the line operation parameter information, and identifying the basic resistance parameters of the dynamic model using an improved particle swarm algorithm to obtain a single-point train formation dynamic model;

[0084] S3. Based on the historical data of the preceding train in the formation, the LSTM-UKF trajectory fusion algorithm is used to predict the future operation curve of the preceding train. The future operation curve of the preceding train is solved based on the single-particle formation dynamics model to obtain the safety protection curve of the preceding train in the formation.

[0085] S4. Determine a tracking interval formula between trains in a formation based on the train-to-train communication delay, differences in train traction and braking performance, and train speed differences; determine the positional relationship between the leading and trailing trains in the formation based on the tracking interval formula; and construct a safety protection control model for the trailing train in the formation based on the safety protection curve of the leading train in the formation and the positional relationship between the leading and trailing trains.

[0086] S5. Solve the safety protection control model of the following train based on the direct multi-point shooting method based on the distance interval, and realize the operation of the platoon train according to the solution.

[0087] The present invention focuses on the safety protection control problem of formation trains during operation, and considers line constraints, safety formation constraints, traction and braking constraints, etc., to establish a safety protection control model for the following trains with different traction and braking performance and speed. By comparing the inverse algorithms of time step, distance step and speed step, a solution method for the leading vehicle protection curve based on the time step is obtained. By analyzing the model, the direct multi-point shooting method and the interior point algorithm are selected to solve the model. Combining the actual line with the traction and braking characteristic curve, MATLAB is used to solve the protection curve. The solution results show that the requirements of safe operation can be met during operation and at the parking point. Compared with the formation train based on absolute braking distance, the minimum tracking interval between trains is reduced while ensuring the safe operation of the formation.

[0088] In one embodiment, step S2 constructs a dynamic model including basic train resistance, additional train resistance, and train traction and braking forces based on a parameter-identified single-particle train formation dynamics model. Basic resistance parameters are identified using an improved particle swarm optimization algorithm. Step S2 forms the basis for calculating the train resultant force in steps S3-S5 and further includes the following sub-steps:

[0089] S21. Constructing a single-mass dynamic model for train formations. In urban rail transit, trains generally adopt a distributed power formation method to ensure good consistency in traction and braking. Therefore, a single-mass model can be used when modeling the dynamics of urban rail transit trains.

[0090] The single-particle platooning dynamics model is:

[0091]

[0092] Where M is the mass of the train; a(v x ) is the train speed v x The acceleration of the train at time t, in m / s 2 ;F(v x ) is the train speed v x The traction or braking force at the time, the unit is KN; f base (v x ) is the train speed v x The unit basic resistance when the train is at position x is in N / KN; W(x) is the total additional resistance of the train at position x, in KN; a, b, and c are the basic resistance parameters of the train; where a is the friction resistance of the train, b is the resistance coefficient of the train, and c is the aerodynamic resistance coefficient of the train; W i is the additional resistance of the ramp, the unit is N / KN; w i is the additional resistance per unit slope, in N / KN, g is the acceleration due to gravity, w r is the additional resistance of the unit curve, in KN; R is the curve radius of the line, in m; A is an empirical constant; F(v m ) is the train speed v m Braking force at time v m 、v o and v x are the speeds of the train at the corresponding positions m, o, and x, respectively, and F m is the braking force of the train at the corresponding position m; F o is the braking force of the train at the corresponding position o.

[0093] S22. Method for identifying parameters of basic resistance model of train formation based on improved particle swarm algorithm. The basic resistance of train is highly random and varies at different running speeds, so it is generally expressed by empirical formula. For multi-dimensional parameter identification and solving the contradiction between local optimal solution and global optimal solution, an improved particle swarm algorithm is proposed to identify basic resistance parameters. The formula for identifying parameters of basic resistance in the dynamic model based on the improved particle swarm algorithm is:

[0094]

[0095] Where j is the particle index, H is the number of particles, is the position of the corresponding parameter of each particle in the particle swarm; is the speed of each particle in the particle swarm corresponding to the parameter; X min Represents the minimum position of each particle in the particle swarm corresponding to the parameter, X max Represents the maximum position of each particle in the particle swarm corresponding to the parameter, V min Represents the minimum velocity of each particle in the particle swarm corresponding to the parameter, V max Represents the maximum speed of each particle in the particle swarm corresponding to the parameter; rand(1,H) represents the generation of a random matrix with 1 row and H columns; and denote the velocity and position of the jth particle in the kth generation, respectively. represents the historical optimal solution searched by j particles in the kth generation, represents the optimal solution found by all particles in the kth generation; w is the inertia factor, c1 is the first learning factor, and c2 is the second learning factor; r1 and r2 are random numbers uniformly distributed between [0,1]; w min and w max are the minimum and maximum values ​​of the inertia factor, f min and f ave are the minimum fitness and average fitness, respectively.

[0096] In one embodiment, Figure 2 As shown in the figure, the structure diagram of using LSTM-UKF to predict the running trajectory of the leading train; step S3 uses the LSTM-UKF trajectory fusion prediction algorithm to obtain the future ATO (Automatic Train Operation) running curve of the train based on the historical data of the leading train in the train formation, and uses the inverse algorithm to solve the safety protection curve of the leading train in the train formation, further including the following sub-steps:

[0097] S31 performs data preprocessing on historical data of the front train in the platoon train, including data conversion, data cleaning and data normalization processing; wherein the data conversion refers to converting the train head SEG and the train head offset into the cumulative running distance of the train in combination with the line map; the data cleaning refers to eliminating and regenerating unreasonable running data; the normalization processing refers to converting the speed, control current and position into the same dimension;

[0098]

[0099] wherein p_x represents a parameter such as speed / control current / position that needs to be normalized, and p_x' is the normalized parameter;

[0100] S32 uses the LSTM-UKF algorithm to predict the future time running curve of the front train in the platoon train according to the historical data after data preprocessing; uses the speed, control current and cumulative running distance as the input time sequence, and the distance as the output time sequence, and uses the final displacement error (FDE), the average absolute displacement error (ABDE) and the root mean square error (RMSE) to evaluate the performance of the improved algorithm; the trajectory of the platoon train is predicted by the LSTM trajectory prediction model, and a more reasonable prediction of the running trajectory of the platoon train is realized. However, due to the inevitable noise interference in the complex environment of train operation, and the trajectory after the LSTM algorithm trajectory prediction will deviate, it is necessary to optimize and adjust the trajectory of the platoon train after the LSTM prediction, so as to improve the prediction accuracy of the trajectory prediction algorithm result, so as to ensure the accurate parking of the front train of the platoon train, facilitate the accurate control of the rear platoon train, and ensure the safe operation of the platoon.

[0101] S33 obtains the safety protection curve of the front train in the platoon train according to the future time running curve of the front train by using the inverse algorithm.

[0102] Specifically, as shown in Figure 3 the time step-based protection curve inverse algorithm principle diagram; according to the predicted parking position of the front train of the platoon, the position difference between the known train and the target parking point or deceleration point, the current speed of the train and the braking characteristic curve of the train, then segment by time, assuming that the train maintains uniform variable speed motion within the smallest segment interval, then the acceleration can be obtained by interpolation, and then the current speed can be calculated from the target parking point in reverse, to obtain the safety protection curve of the front train in the platoon train.

[0103] In one embodiment, the safety protection curve of the front train is

[0104]

[0105] Where S and V are the distance and speed of the train under safe braking, respectively; γ is the rotation coefficient; u(v i ) is the speed v obtained by interpolation based on the traction braking characteristic curve i The unit braking force is KN; R(v i ) is the train speed v i The total unit basic resistance at the time, in KN, W(x) is the total additional resistance of the train at position x, in KN; i is the subscript of the time interval segment, i = 1, 2, 3...n; n is the number of time interval segments; Δv i The speed increment for each time interval; Δs i is the braking distance of each time interval; v0 is the initial speed; Δt i is the time interval. U is the unit conversion system, which is a constant, for example, the value is 1000.

[0106] In one embodiment, in step S4, the positional relationship between the leading train and the trailing train in the train formation is obtained based on the train-to-train communication delay, the train traction and braking performance difference, the train speed difference, and the tracking interval formula between the trains in the train formation;

[0107] In one embodiment, the position relationship between the preceding train and the following train is

[0108] d min +d breakL =d F_v +d breakF +d safe

[0109] Among them, d min is the minimum safe tracking distance before train braking, d F_v is the running distance of the following train under the communication delay and the reaction time of the preceding train, d breakF is the braking distance of the following train, d breakL is the braking distance of the preceding train, d safe This is the final safe stopping distance for the train.

[0110] In one embodiment, to ensure the safety of the rear vehicle in the train formation during braking, the following vehicle interval under the most unfavorable conditions is used as the safe tracking interval of the train formation. The tracking interval between trains in the train formation is

[0111]

[0112] Among them, d F_v It is the distance that the following train runs after receiving the braking command. The braking distance of the following train under the most unfavorable conditions requires consideration of five parts: A, B, C, D, and E are the five stages of braking, respectively. d represents the distance. Stage A is the reaction stage of the on-board ATP device, stage B is the traction removal stage, stage C is the pre-braking stage, stage D is the emergency braking stage, and stage E is the emergency braking stage. is the braking distance of the preceding train, with the maximum braking deceleration as the deceleration of the process; d safe_end It is the final stopping distance between the preceding train and the following train.

[0113] In one embodiment, in step S4, the safety protection control model for the rear train in the train formation is constructed based on the safety protection curve of the front train in the train formation and the positional relationship between the front train and the rear train, including:

[0114] S421. Assume that the train formation in the safety protection control model for the following train includes a lead train, a following train, and a rear train, and that the lead train, the following train, and the rear train run in the same direction at the same speed within the section.

[0115] S422. Assume that the probability that the locomotive of the following train is always behind the tail of the leading train in the train formation is greater than the lower confidence limit; and under the dynamic behavior of the following train, the probability that the locomotive of the trailing train is always behind the tail of the following train is greater than the lower confidence limit, that is,

[0116]

[0117] Among them, x L is the position of the pilot train, x F is the position of the following train, x R is the position of the rear train, d is the safety interval, and p is the lower confidence limit;

[0118] S423. Construct a safety protection control model for the following train including dynamic relationship constraints, train traction constraints, braking performance constraints, line speed limit constraints, tracking constraints and optimization objectives.

[0119] When generating the train safety protection curve, the leading train, following train and tail train further shorten the running interval under the premise of ensuring the safe braking distance, which is the optimization goal.

[0120] min(∫[x L (t)-x F (t)] 2 dt+∫[x F (t)-x R (t)] 2 dt)

[0121] Among them, x L(t) is the position of the pilot train at time t, x F (t) is the position of the following train at time t, x R (t) is the position of the following train at time t, where t is the time;

[0122] The train's dynamic equations are established with position as the independent variable, as the dynamic relationship constraint

[0123]

[0124] Where u(x) is the train control force at position x, R(v) is the total basic resistance of the train at speed v, and W(x) is the total additional resistance of the train at position x;

[0125] The traction and braking force must meet the constraints of the train's maximum traction and maximum braking force, that is, the train traction constraint is

[0126]

[0127] Among them, u L (x),u F (x) and u R (x) the tractive effort of the lead train, the following train and the rear train respectively; and are the minimum braking force and maximum braking force of the pilot train at speed v, respectively; and are the minimum braking force and maximum braking force of the following train at speed v; and are the minimum braking force and maximum braking force of the rear train at speed v, respectively;

[0128] Considering that the traction braking performance of the three trains is different, it is necessary to calculate the safe interval between the lead train and the following train, and between the following train and the rear train to ensure safety. This interval can ensure that under the most unfavorable conditions, the leading train immediately starts emergency braking, and the rear train can also operate safely after a delay. In other words, the braking performance constraint is

[0129]

[0130] in, The minimum safe distance between the lead train and the following train. is the minimum safe interval between the following train and the rear train; v L is the speed of the pilot train, v F is the speed of the following train, v R is the speed of the rear train; t delay is the braking delay; d safeis the final safe stopping distance of the train; C is a constant, with a value of 3.6; P L It is twice the maximum acceleration of the pilot train and is set to 2.28; P F It is twice the maximum acceleration of the following train, and its value is 2; P R It is twice the maximum acceleration of the rear train, and its value is 1.8; S F The braking reaction time of the following train is 2.3; S R is the braking reaction time of the rear train, which is 2.33; Y F Y is the safe stopping distance deviation of the following train, which is set to 0.415; R is the safe stopping distance deviation of the rear train, and its value is 0.4325; that is,

[0131]

[0132] like Figure 4 As shown in Figure 1, in order to avoid rear-end collisions between the front and rear trains during braking, it is necessary to ensure that the front end of the rear train does not exceed the rear end of the front train during operation. Therefore, the position difference between the front and rear trains at any time should not be less than the safety margin d. safe According to the principle of relative braking distance mode, the following tracking constraints can be obtained:

[0133]

[0134] Among them, x L (t) is the position of the following train at time t, x F (t) is the position of the pilot train at time t, x R (t) is the position of the rear train at time t;

[0135] The pilot train, the following train and the tail train must meet the following requirements during operation: the speed when they reach their respective target tracking points is 0, and they must meet the line speed limit requirements during operation, that is, the line speed limit is

[0136]

[0137] Among them, X is the target tracking point; v L (X), v F (X), v R (X) are the speeds of the lead train, the following train and the tail train when they reach their respective target tracking points; L sm Indicates the speed limit of the target point; v limit (x) is the speed limit at position x.

[0138] In one embodiment, the safety protection control model of the following train is

[0139]

[0140] Among them, u L (x) and u F (x) The constraints of the train's maximum traction and maximum braking force must be met.

[0141] In one embodiment, step S5 is to efficiently solve the safety protection control curve based on the direct multi-point shooting method based on the distance, so as to achieve the safe operation of the platoon; Figure 5 A schematic diagram of the safety protection curve of three-car platoons under different braking performances in an embodiment of the present invention is shown.

[0142] Specifically, the calculation of the train formation safety protection curve needs to be based on the dynamic behavior of the leading vehicle, taking into account the random errors in train speed measurement, and the constraints of the train state and control force; therefore, it can be considered that the train safety protection control based on the relative braking distance of the train is a nonlinear optimal control problem that considers the dual constraints of state and control variables. Therefore, a distance-based direct multi-point shooting method is used to solve the safety protection control curve of the following train.

[0143] In one embodiment, step S5 further includes

[0144] S51, dividing the train's expected stopping position interval into equal distance intervals;

[0145] Initialize a set of vectors as the control variable values ​​at N nodes;

[0146] Assuming that the train moves in a uniformly accelerated straight line at each distance interval, based on the control variable u(x) at each node, use Newton's laws of motion and the single-particle dynamics model to solve for the train's acceleration, velocity, and time in the distance interval.

[0147] The state variable value S(x)=[v(x),t(x)] at the node is obtained based on the acceleration, speed and time of the train in the distance interval. T ,

[0148] The formula for solving the acceleration, speed and time of the train in the distance interval is:

[0149]

[0150] Among them, u(x) is the traction force of the train at position x; v(x) is the speed of the train at position x, t(x) is the time when the train is at position x, and a i is the acceleration of the ith time interval; x i is the position of the i-th time interval segment; v i is the speed of the i-th time interval segment; u(x i ) is the train at position xi The train control force at R(v i ) is the train speed v i The total basic resistance when W(x i ) is the train at position x i The total additional resistance at v(x i ) is the train at position x i The velocity at v(x i+1 ) is the train at position x i+1 The speed at the point; Δx is the distance interval; Δx=x i+1 -x i ; t(x i+1 ) is the train at position x i+1 The moment t(x i ) is the train at position x i moment;

[0151] S52. Bring the values ​​of the control variables and state variables at the nodes into the constraints and objective function, and transform the train safety protection control problem into a nonlinear programming problem.

[0152]

[0153] The objective function is

[0154]

[0155] The constraints are

[0156]

[0157] Where F(W) is the objective function, which represents the cost or goal to be minimized; is to take the minimum value of all objective functions; W is a variable; G(W) is an equality constraint, which represents a strict condition that the system must satisfy; H(W) is an inequality constraint, which represents a restriction in the system; s i ,q i , Δs i , Δq i is the optimization variable, where s i is the state variable, q i is the input variable, Δs i is the change of the state variable, Δq i is the change of the input variable; Q i 、S i 、R i They are the first weighting matrix, the second weighting matrix and the third weighting matrix, which are used to define the weights of different items in the objective function, involving the trade-off between state and control; C i 、D irespectively, are the first and second matrices representing the coefficients involved in the inequality constraints; and is the variation of the control variable of the initial state; is the initial condition, representing the initial state value of the system; and N is the control variable of the terminal state; and N is the offset of the terminal state constraint; and T is the transpose;

[0158] S53, using the interior point method to solve the nonlinear programming problem obtains the safety protection control curve of the following train and the tail train.

[0159] As Figure 6 The principle diagram of the train formation safety protection method based on trajectory prediction of the application is shown, the improved particle swarm algorithm is used for accurate identification of the basic resistance parameters of the single-particle train formation dynamics model; after the identification, the historical data of the preceding train in the train formation is used to predict the future ATO operation curve of the train through the LSTM-UKF trajectory fusion algorithm, and the inverse algorithm is used to solve the safety protection curve of the preceding train in the train formation; the tracking interval formula between the train formations is proposed based on the communication delay between the trains, the difference in the train traction and braking performance and the speed difference between the trains, the safety protection control model of the following train in the train formation is constructed under the safe tracking interval and considering the train performance difference and the speed difference, the direct multi-point shooting method based on the distance interval is used to solve the safety protection control model of the following train, and the safe operation of the train formation is realized according to the solving result.

[0160] The present application is around the safety protection control problem of the train formation in the running process, considering the line constraint, the safety formation constraint, the traction and braking constraint and the like, the safety protection control model of the multi-train formation with different traction and braking performance and speed is established. Then, the inverse algorithm based on the time step, the distance step and the speed step is compared, and the preceding train protection curve solving mode based on the time step is obtained. Then, the model is analyzed, the direct multi-point shooting method and the interior point algorithm are selected to solve the model. Finally, the protection curve is solved by using MATLAB in combination with the actual line and the traction and braking characteristic curve, and the solving result shows that the safety operation demand can be met in the running process and the stopping point. Compared with the train formation based on the absolute braking distance, the minimum tracking interval between the trains is reduced under the premise of ensuring the safe operation of the train formation.

[0161] In the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper" and "lower" is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be internal communication between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances.

[0162] It should also be noted that, in the description of the present invention, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0163] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in this field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. A platooning vehicle safety protection method based on trajectory prediction, characterized in that: include: S1. Obtaining parameter information of train formations within the line and line operation parameter information; S2. Constructing a dynamic model including basic train resistance, additional train resistance, and train traction and braking forces based on the parameter information of the train formation within the line and the line operation parameter information, and identifying the basic resistance parameters of the dynamic model using an improved particle swarm algorithm to obtain a single-point train formation dynamic model; S3. Based on the historical data of the preceding train in the formation, the LSTM-UKF trajectory fusion algorithm is used to predict the future operation curve of the preceding train. The future operation curve of the preceding train is solved based on the single-particle formation dynamics model to obtain the safety protection curve of the preceding train in the formation. S4. Determine a tracking interval formula between trains in a formation based on the train-to-train communication delay, differences in train traction and braking performance, and train speed differences. Determine the positional relationship between the leading and trailing trains in the formation based on the tracking interval formula. Construct a safety protection control model for the trailing train in the formation based on the safety protection curve for the leading train in the formation, a single-particle formation dynamics model, and the positional relationship between the leading and trailing trains. S5. Solve the safety protection control model of the following train using a direct multi-point shooting method based on the distance interval, and realize the operation of the platoon according to the solution; The single-particle platooning dynamics model is: in, M is the train mass; The train speed is The acceleration of the train at ; The train speed is traction or braking force when The train speed is Unit basic resistance when For the train x Total additional resistance in position; a 、 b 、 c is the basic resistance parameter of the train; among them, a is the friction resistance of the train, b is the drag coefficient of the train, c is the aerodynamic drag coefficient of the train; Add resistance to the ramp; is the additional resistance per unit slope; g is the acceleration due to gravity, is the additional resistance of the unit curve; is the curve radius of the line; A is an empirical constant; For trains at speed The braking force when 、 and The trains are in the corresponding m 、 o 、 x Position velocity, For trains at corresponding positions m Braking force at For trains at corresponding positions o Braking force at The safety protection curve of the preceding train is: in, S and V are the distance and speed of the train under safe braking, respectively, and γ is the slewing coefficient; The speed is obtained by interpolation based on the traction braking characteristic curve. Unit braking force at 1 hour; For trains at speed Total unit basic resistance when ; W ( x ) is the train position x The total additional resistance at i is the subscript of the time interval segment, i =1,2,3... n ; n The number of segments into which the time interval is segmented; The speed increment for each time interval segment; The braking distance for each time interval segment; is the initial velocity; is the time interval; Converts the unit to a base.

2. The method according to claim 1, characterized in that The formula for identifying the parameters of the basic resistance in the dynamic model based on the improved particle swarm algorithm is: in, is the particle index, H is the number of particles, is the position of the corresponding parameter of each particle in the particle swarm; is the speed of each particle in the particle swarm corresponding to the parameter; Represents the minimum position of the corresponding parameter of each particle in the particle swarm, Represents the maximum position of the corresponding parameter of each particle in the particle swarm, Represents the minimum speed of each particle in the particle swarm corresponding to the parameter, Represents the maximum speed of each particle in the particle swarm corresponding to the parameter; rand(1, H ) means generate 1 line H Random matrix of columns; and Respectively represent j The particle in k The speed and position of the generation, express j The particle in k The historical optimal solution found by the search, Indicates that all particles in k The optimal solution found by the search; w is the inertia factor, The first learning factor, is the second learning factor; and are random numbers uniformly distributed between [0, 1] are the minimum and maximum values ​​of the inertia factor, are the minimum fitness and average fitness, respectively.

3. The method according to claim 2, characterized in that Said S3 further comprises S31, preprocessing historical data of the preceding train in the train formation; S32. Based on the pre-processed historical data, use the LSTM-UKF algorithm to predict the future running curve of the leading train in the train formation; S33. Using an inverse algorithm to solve the future time operation curve of the preceding train, obtain a safety protection curve of the preceding train in the formation.

4. The method according to claim 1, wherein The positional relationship between the preceding train and the following train is: in, is the minimum safe tracking distance before the train brakes. is the running distance of the following train under the communication delay and the reaction time of the preceding train, is the braking distance of the following train, is the braking distance of the preceding train, This is the final safe stopping distance for the train.

5. The method according to claim 4, characterized in that The safety protection control model for the rear train in the train formation is constructed based on the safety protection curve of the front train in the train formation, the single-particle train formation dynamics model, and the positional relationship between the front train and the rear train. S421. Assume that the train formation in the safety protection control model for the following train includes a lead train, a following train, and a rear train, and that the lead train, the following train, and the rear train run in the same direction at the same speed within the section. S422. Assume that the probability that the locomotive of the following train is always behind the tail of the leading train in the train formation is greater than the lower confidence limit; and under the dynamic behavior of the following train, the probability that the locomotive of the trailing train is always behind the tail of the following train is greater than the lower confidence limit, that is, in, The position of the pilot train, To follow the train's position, is the position of the rear train, For safety intervals, is the lower confidence limit; S423. Construct a safety protection control model for the following train including dynamic relationship constraints, train traction constraints, braking performance constraints, line speed limit constraints, tracking constraints and optimization objectives.

6. The method according to claim 5, characterized in that The optimization goal is in, For the pilot train The location at the moment, To follow the train The location at the moment, To follow the train The location at the moment, t For the moment; The dynamic relationship constraint is in, For trains in position x The train control force at For trains at speed The total basic resistance when For trains in position x The total additional resistance at The train traction constraint is in, 、 and are the traction forces of the lead train, following train and tail train respectively; and The pilot train is at speed The minimum braking force and maximum braking force when and Following train at speed The minimum braking force and maximum braking force when and The speed of the rear train is The minimum braking force and maximum braking force when The braking performance constraint is in, The minimum safe distance between the lead train and the following train. The minimum safe interval between the following train and the rear train; is the speed of the pilot train, To follow the speed of the train, is the speed of the rear train; For braking delay; It is the final safe stopping distance for the train; is a constant; twice the maximum acceleration of the pilot train; It is twice the maximum acceleration of the following train; It is twice the maximum acceleration of the rear train; Braking reaction time for following train; is the braking reaction time of the rear train; The deviation in stopping distance for following train safety; The safe stopping distance deviation for the rear train; The tracking constraints are in, for t The position of the pilot train at all times, for t Always follow the train's position. for t The position of the last train at the time; The line speed limit is in, X is the target tracking point; are the speeds of the lead train, following train and tail train when they reach their respective target tracking points; Indicates the speed limit at the target point; In position x The speed limit at the place.

7. The method according to claim 6, characterized in that The safety protection control model of the following train is: 。 8. The method according to claim 7, characterized in that The S5 further includes S51, dividing the train's expected stopping position interval into equal distance intervals; Initialize a set of vectors as the control variable values ​​at N nodes; assume that the train moves in a straight line with uniform acceleration at each distance interval, according to the control variables at each node u ( x ); using Newton's laws of motion and a single-particle dynamics model, solving for the acceleration, velocity, and time of the train in the distance interval; The state variable value at the node is obtained according to the acceleration, speed and time of the train in the distance interval , The formula for solving the acceleration, speed and time of the train in the distance interval is: in, u ( x ) is the train position x traction; For trains in position x speed, For trains in position x moment, For the i The acceleration of each time interval; For the i The position of each time interval segment; For the i The speed of each time interval segment; For trains in position Train control capability at For trains at speed Total basic resistance when For trains in position The total additional resistance at For trains in position The speed at which For trains in position The speed at which is the distance interval; ; For trains in position moment; For trains in position moment; S52. Bring the values ​​of the control variables and state variables at the nodes into the constraints and objective function, and transform the train safety protection control problem into a nonlinear programming problem. The objective function is The constraints are in, is the objective function; G ( W ) is an equality constraint; is an inequality constraint; 、 is the optimization variable, where is the state variable, is the input variable, is the change of the state variable, is the change in the input variable; 、 、 are respectively the first weighting matrix, the second weighting matrix and the third weighting matrix; 、 are the first matrix and the second matrix respectively; is the change of the control variable in the initial state; is the initial condition; is the control variable of the terminal state; is the offset of the terminal state constraint; S53. Use the interior point method to solve the nonlinear programming problem to obtain the safety protection control curves of the following train and the rear train.

Images