Rail transit driving curve planning method and system based on multi-source perception data

Abstract

The application discloses a kind of track traffic driving curve planning method and system based on multi-source perception data, the application includes the travel line of train is divided into multiple intervals;During train operation, the current travel position of train is detected, if it is found that train is currently driven to the boundary of adjacent arbitrary interval q i , then the control sequence u2 (k) of interval q i And its real-time optimization energy consumption J si Are calculated based on the various perception data of train;Real-time optimization energy consumption J si And the preset experience optimization energy consumption J i Of interval q yi Are compared, if real-time optimization energy consumption J si Is better than experience optimization energy consumption J yi , then the control sequence u2 (k) is selected to control train travel in interval q i ; Otherwise, the preset control sequence u1 (k) corresponding to experience optimization energy consumption J yi Controls train travel in interval q i . The application can meet the actual energy consumption demand of train operation, reduce the performance energy consumption of train, reduce train operation cost, and has great significance for studying train energy-saving driving optimization.

Description

Technical Field

[0001] This invention relates to automatic driving technology for rail transit, specifically to a method and system for planning driving curves for rail transit based on multi-source perception data. Background Technology

[0002] In recent years, with the rapid development of society and the acceleration of urbanization, rail transit, due to its large capacity, high speed, safety, and punctuality, has gradually become one of the main modes of transportation. With the rapid development of rail transit, the energy consumption of rail transit systems is also increasing year by year. Therefore, reducing energy consumption while ensuring the normal operation of rail transit has become one of the main research directions for researchers. Rail transit energy consumption is mainly divided into operational energy consumption and running energy consumption. Operational energy consumption refers to the energy consumption generated by the vehicle's lighting and air conditioning systems, which is usually difficult to optimize directly. Running energy consumption refers to the energy consumed by traction during operation. This part of energy consumption can be optimized through methods such as optimizing driving strategies, achieving relatively ideal results. Currently, most researchers focus on offline optimization in their research on energy-saving driving, mainly focusing on energy-saving driving curve planning and automatic control. However, this method is often ineffective because the actual tracked speed curve and the theoretical simulation curve will have certain errors due to weather, train wear, and other factors. Therefore, a rail transit energy-saving driving curve planning technology based on multi-source sensing data is of great significance for addressing the shortcomings of offline optimization of driving curves and for researching energy-saving driving in rail transit. Summary of the Invention

[0003] The technical problem to be solved by this invention is to provide a method and system for planning rail transit driving curves based on multi-source sensing data, which can meet the actual energy consumption requirements of train operation, reduce the performance energy consumption of trains, and reduce the operating cost of trains. This invention is of great significance for the research on energy-saving driving optimization of trains.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0005] A method for planning rail transit driving curves based on multi-source sensing data includes:

[0006] S1 divides the train's route into multiple sections {q} i};

[0007] S2, during train operation, detects the train's current position. If it is found that the train is currently traveling to any adjacent interval q... i If the boundary is found, proceed to step S3; otherwise, proceed to step S2 again.

[0008] S3, calculates the interval q based on multiple sensing data of the train.i The control sequence u2(k) and its real-time optimized energy consumption J si ;

[0009] S4 will optimize energy consumption in real time. si and its interval q i Preset experience-optimized energy consumption J yi In comparison, if real-time energy consumption J is optimized... si Superior to empirically optimized energy consumption J yi Then, the control sequence u2(k) is selected to control the train in the interval q. i Otherwise, maintain experience-optimized energy consumption. yi The corresponding preset control sequence u1(k) controls the train in the interval q. i The driving.

[0010] Optionally, in step S1, the train's route is divided into multiple sections {q}. i} refers to dividing the train's route into multiple sections by using the locations of signal lights, junctions requiring track changes, and locations where the operating status changes as boundaries. i}, where changing the operating state refers to the train changing from a stopped state to a traction state, or from a traction state to a braking state.

[0011] Optionally, in step S2, the train is currently traveling to any adjacent interval q. i The boundary refers to the current driving position of the train, the section q i The distance between the boundaries is equal to the set value.

[0012] Optionally, step S3 includes:

[0013] S3.1 collects various sensor data of the train, including speed, acceleration, traction force and braking force;

[0014] S3.2, based on multiple sensing data of the train, perform data fusion and classification to determine the train's position in interval q. i The operating modes include four types: traction, cruise, coasting, and braking. Each operating mode corresponds to a gear curve, thereby obtaining the interval q. i The gear ratio curve is used as the interval q i The control sequence u2(k) is obtained; simultaneously, various sensing data of the train are input into the train's multi-objective energy consumption optimization model to obtain the train's energy consumption in interval q. i Real-time optimized energy consumption J si .

[0015] Optionally, the data fusion classification based on multiple train sensing data refers to inputting multiple train sensing data into a Bayesian classifier for data fusion classification to determine the train's position in interval q.i The operating mode.

[0016] Optionally, the functional expression of the multi-objective energy consumption optimization model is:

[0017]

[0018] In the above formula, y represents the optimization result of the multi-objective energy consumption optimization model, with the minimum optimization objective as the optimization objective; ω1, ω2, ω3, and ω4 are weighting factors, E is the total energy consumption, Δs is the parking error, and K... jerk For comfort, ΔT is the punctuality error, and y m (k) represents the velocity at any time k, v0 is the initial velocity, v(0) is the velocity at t=0, s0 is the initial displacement, s(0) is the displacement at t=0, and t=0 is the initial time, where:

[0019] Δs=|ss target |,

[0020] ΔT=|TT target |,

[0021] In the above formula, t is the number of intervals, and f i (s) represents the interval q i The traction force of the train at a mid-displacement s, ΔS i For traction force f i (s) The displacement produced by the following vehicle, ε B B is the product factor for the conversion of regenerative braking energy during train braking. i For the train in section q i Braking force, v i For the train in section q i The velocity in the interval q i Total energy consumption E minus interval q i-1 The difference in total energy consumption E is used as the train's energy consumption in interval q. i Real-time optimized energy consumption J si ; s represents the actual stopping position of the train, s target The target stopping position of the train is given, and the stopping error Δs is less than a preset threshold; a is the train's acceleration; t is time; T is the train's actual arrival time. target The target arrival time of the train is such that the on-time error ΔT is less than the preset value.

[0022] Optionally, in step S4, the train is controlled to move within section q. i The operation refers to controlling the train in section q according to the control sequence u2(k) or the control sequence u1(k). iThe control sequences u2(k) and u1(k) contain the correspondence between different displacements and their corresponding gear positions.

[0023] Optionally, before step S1, the pre-optimized speed curve of the train and the speed curve for each interval q are also set. i The preset control sequence u1(k) is set, and the corresponding empirically optimized energy consumption J is determined based on the train's pre-optimized speed curve. yi The pre-optimized speed curve includes a two-dimensional coordinate system curve of the train's displacement and speed, and the corresponding empirically optimized energy consumption J is determined based on the train's pre-optimized speed curve. yi This includes: based on the obtained pre-optimized speed curve, testing the train multiple times on the same line following the pre-optimized speed curve, and then taking q for each interval. i The average energy consumption is used as the interval q i Corresponding empirical optimization energy consumption J yi .

[0024] Furthermore, the present invention also provides a rail transit driving curve planning system based on multi-source sensing data, including a microprocessor and a memory interconnected thereto, wherein the microprocessor is programmed or configured to execute the steps of the rail transit driving curve planning method based on multi-source sensing data.

[0025] Furthermore, the present invention also provides a computer-readable storage medium storing a computer program for being programmed or configured by a microprocessor to perform the steps of the rail transit driving curve planning method based on multi-source sensing data.

[0026] Compared with existing technologies, this invention has the following main advantages: This invention involves dividing the train operating line into different sections, analyzing and processing real-time data collected by multi-source sensors to determine whether real-time speed curve planning is needed, comparing the energy consumption of the real-time planned curve for each section with the offline optimized section energy consumption, and selecting the curve with lower energy consumption for operation. This invention can meet the actual energy consumption requirements of train operation, reduce the performance energy consumption of the train, and reduce train operating costs, which is of great significance for research on energy-saving driving optimization of trains. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the basic process of the method in an embodiment of the present invention.

[0028] Figure 2 In this embodiment of the invention, a Bayesian network is used to determine the train's position in interval q. i The flowchart of the operation mode. Detailed Implementation

[0029] like Figure 1 As shown, the rail transit driving curve planning method based on multi-source sensing data in this embodiment includes:

[0030] S1 divides the train's route into multiple sections {q} i};

[0031] S2, during train operation, detects the train's current position. If it is found that the train is currently traveling to any adjacent interval q... i If the boundary is found, proceed to step S3; otherwise, proceed to step S2 again.

[0032] S3, calculates the interval q based on multiple sensing data of the train. i The control sequence u2(k) and its real-time optimized energy consumption J si ;

[0033] S4 will optimize energy consumption in real time. si and its interval q i Preset experience-optimized energy consumption J yi In comparison, if real-time energy consumption J is optimized... si Superior to empirically optimized energy consumption J yi Then, the control sequence u2(k) is selected to control the train in the interval q. i Otherwise, maintain experience-optimized energy consumption. yi The corresponding preset control sequence u1(k) controls the train in the interval q. i The driving.

[0034] In this embodiment, step S1 divides the train's route into multiple intervals {q}. i} refers to dividing the train's route into multiple sections by using the locations of signal lights, junctions requiring track changes, and locations where the operating status changes as boundaries. i}, where changing the operating state refers to the train changing from a stopped state to a traction state (stop → traction), or from a traction state to a braking state (traction → braking). This section division method fully considers the differences in energy consumption between sections and the consistency of energy consumption within sections, which helps to reduce the overall energy consumption of train operation. In this embodiment, the variable t represents the total number of sections, thus obtaining the section set as {q1, q2...q...} t}

[0035] In this embodiment, in step S2, the train is currently traveling to any adjacent interval q. i The boundary refers to the current driving position of the train, the section q i The distance between the boundaries is equal to a set value, which can be set as needed. Moreover, considering the train's speed, the accuracy of this set value should not be too small.

[0036] In this embodiment, step S3 includes:

[0037] S3.1 collects various sensor data of the train, including speed, acceleration, traction force and braking force;

[0038] It should be noted that the technologies for obtaining speed, acceleration, traction, and braking force are known technologies. They can be obtained by using computer technology to comprehensively process different types of data collected by sensors (speed sensors, gravity sensors, temperature sensors, etc.) including radar (radar, camera), etc.

[0039] During data acquisition, data mining theory and filtering algorithms can be used to remove noisy data and comprehensively analyze incomplete information, thereby obtaining various sensory data reflecting the state and environmental information of objective objects. For example, data collected by multiple sensors may contain some noisy data, also known as outliers, which can affect subsequent simulation and optimization results. By referencing the data collected by each sensor and comparing their changes, some obviously unacceptable outliers can be removed, and appropriate values ​​can be inserted based on the preceding and following data. The method for identifying outliers is as follows: first, a threshold W is determined using the absolute value averaging method.

[0040]

[0041] Where p is an empirical parameter, n is the number of sampled data, and z i This is the sampled data. Then, when the collected data |z i When |≥W, then z is considered i Outliers in the collected data should be removed or replaced. Finally, by removing noise data, various sensor data of the train after noise removal are obtained, including speed, acceleration, traction, and braking force.

[0042] S3.2, based on multiple sensing data of the train, perform data fusion and classification to determine the train's position in interval q. i The operating modes include four types: traction, cruise, coasting, and braking. Each operating mode corresponds to a gear curve, thereby obtaining the interval q. i The gear ratio curve is used as the interval q i The control sequence u2(k) is obtained; simultaneously, various sensing data of the train are input into the train's multi-objective energy consumption optimization model to obtain the train's energy consumption in interval q. i Real-time optimized energy consumption J si .

[0043] Data fusion and classification based on multiple sensor data of trains can use feature-based fusion methods as needed. This involves using deep neural networks (DNNs) to learn new representations from the original features extracted from different datasets, obtaining corresponding feature data. By comprehensively analyzing and processing the feature data, the main features of the information are preserved, information can be compressed, and the fusion speed can be accelerated.

[0044] In this embodiment, considering that the information collected by multiple sensors and radar may have uncertain connections or incomplete information, the aforementioned data fusion classification based on multiple train sensing data refers to inputting multiple train sensing data into a Bayesian classifier for data fusion classification to determine the train's position in interval q. i The operating mode of Bayesian networks is to combine Bayesian parameter estimation methods with graph theory in a superior way. The relationship between random variables is represented by a directed graph, and this relationship is quantified by conditional probability.

[0045] Let the set of finite discrete random variables be:

[0046] X = {X1, X2, ..., X} n},

[0047] Where X1, X2, ..., X n These correspond to the nodes in a Bayesian classifier (Bayesian network).

[0048] The method for constructing a Bayesian network is as follows:

[0049] (1) Determine the correspondence between the characteristic attributes of the data received by the sensor and the nodes in the network;

[0050] (2) According to the probability multiplication formula If we construct an acyclic graph B, then the joint probability distribution of attribute X described by the Bayesian network B is: And P(X) is unique. Here, m represents the number of modalities in the data, Y represents the classification result, and C... i Represents attribute information X i A set;

[0051] (3) Determine the probability distribution of nodes: Fusion of multimodal sensing data based on Bayesian parameter estimation model: Assume that at a certain moment, the monitored samples have k classification modes, namely Y = {Y1, Y2, ..., Y...} K}, m-modal data indicates that there are m attribute information X = {X1, X2, ..., X} n Then the measurement model with attribute information is X = f(Y) + v noiseHere, f(Y) represents the functional relationship between attribute information X and classification information Y, v noise This represents random noise. Information fusion algorithms are based on m values ​​measured by sensors, using a certain estimation criterion function, from X1, X2...X... m Estimate the classification Y k The true value. The Bayesian classifier, based on the Bayesian network, is a classification model based on probabilistic statistical methods. It effectively combines the prior and posterior probabilities of an event, using prior information and sample data to calculate the posterior probability of the event, which is the probability that the final result belongs to a certain classification category. It utilizes Bayesian parameter estimation to fuse multiple feature attributes from sensors, constructs a Bayesian network of feature attributes and classification patterns, establishes a Bayesian classifier, and finally uses the maximum a posteriori probability criterion to obtain the classification result. The overall steps are as follows: Figure 2 As shown.

[0052] Specifically, multi-source sensing data fusion refers to using a pre-classified Bayesian classifier to classify the multi-source sensing data of a train to obtain fused features. Multi-source sensing data fusion is achieved by using Bayesian parameter estimation to fuse multiple feature attributes from sensors, constructing a Bayesian network of feature attributes and classification patterns, and finally using the maximum a posteriori probability criterion to obtain the classification result.

[0053] Suppose that at a certain moment, the monitored samples have k classification patterns, namely Y = {Y1, Y2, ..., Y...} K If a train operates in four modes—traction, cruise, coasting, and braking—it can be categorized into four types: Y1 = traction, Y2 = cruise, ... The m-modal data indicates that there are m attribute information X = {X1, X2, ..., X...} n}, such as speed, acceleration, traction force, braking force, etc., are attribute information. The information fusion algorithm is based on m values ​​measured by sensors, and uses a preset estimation criterion function as a basis to extract attribute feature information X1, X2...X n Estimate the classification Y K The actual value. The calculation steps are as follows: Let the attribute information be X. i The estimated value of category Y is At the same time, define a loss function:

[0054]

[0055] The minimum risk is used as the estimation criterion.

[0056] The optimal estimate of the maximum posterior probability of a single attribute. for:

[0057]

[0058] This transforms the multimodal information fusion problem into finding the posterior probability of classification Y:

[0059] P(Y|X):

[0060] This allows us to obtain the motion state at future moments.

[0061] Based on the results of multi-source sensing data fusion, a driving curve that needs to be planned in real time is obtained. Precise datasets collected by multi-source sensors, such as real-time speed, acceleration, traction, and braking force, are used as inputs to a multi-objective energy consumption optimization model to obtain the train's speed within the interval q. i Real-time optimized energy consumption J si The functional expression of the multi-objective energy consumption optimization model in this embodiment is:

[0062]

[0063] In the above formula, y represents the optimization result of the multi-objective energy consumption optimization model, with the minimum optimization objective as the optimization objective; ω1, ω2, ω3, and ω4 are weighting factors, E is the total energy consumption, Δs is the parking error, and K... jerk For comfort, ΔT is the punctuality error, and y m (k) represents the velocity at any time k, v0 is the initial velocity, v(0) is the velocity at t=0, s0 is the initial displacement, s(0) is the displacement at t=0, and t=0 is the initial time, where:

[0064] Δs=|ss target |,

[0065] ΔT=|TT target |,

[0066] In the above formula, t is the number of intervals, and f i (s) represents the interval q i The traction force of the train at a mid-displacement s, ΔS i For traction force f i (s) The displacement produced by the following vehicle, ε B B is the product factor for the conversion of regenerative braking energy during train braking. i For the train in section q i Braking force, v i For the train in section q i The velocity in the interval q i Total energy consumption E minus interval q i-1 The difference in total energy consumption E is used as the train's energy consumption in interval q.i Real-time optimized energy consumption J si ; s represents the actual stopping position of the train, s target The target stopping position of the train is given, and the stopping error Δs is less than a preset threshold; a is the train's acceleration; t is time; T is the train's actual arrival time. target The target arrival time of the train is such that the on-time error ΔT is less than the preset value.

[0067] In this embodiment, step S4 involves controlling the train in section q. i The operation refers to controlling the train in section q according to the control sequence u2(k) or the control sequence u1(k). i The control sequences u2(k) and u1(k) contain the correspondence between different displacements and their corresponding gear positions.

[0068] It should be noted that the intervals q i The preset control sequence u1(k) and its corresponding empirically optimized energy consumption J yi This can be set manually based on experience. As an optional implementation, in this embodiment, before step S1, the pre-optimized speed curve of the train and the speed of each interval q are further set. i The preset control sequence u1(k) is set, and the corresponding empirically optimized energy consumption J is determined based on the train's pre-optimized speed curve. yi The pre-optimized speed curve includes a two-dimensional coordinate system curve of the train's displacement and speed, and the corresponding empirically optimized energy consumption J is determined based on the train's pre-optimized speed curve. yi This includes: based on the obtained pre-optimized speed curve, testing the train multiple times on the same line following the pre-optimized speed curve, and then taking q for each interval. i The average energy consumption is used as the interval q i Corresponding empirical optimization energy consumption J yi In this embodiment, based on the obtained pre-optimized speed curve, PID control is used to test the train n times on the same line, following the pre-optimized speed curve, to obtain the values ​​of each interval {q1, q2, ..., q...}. t Energy consumption {J q1.1 J q1.2 , ..., J q1.n}, {J q2.1 J q2.2 , ..., J q2.n},…,{J q2.1 J q2.2 …J q2.n}, then take each interval q i The average energy consumption is used as the interval q i Corresponding empirical optimization energy consumption J yiTaking the first interval as an example, its calculation function expression is:

[0069]

[0070] In the above formula, J y1 The energy consumption is optimized based on experience for interval q1.

[0071] Furthermore, this embodiment also provides a rail transit driving curve planning system based on multi-source sensing data, including a microprocessor and a memory interconnected, wherein the microprocessor is programmed or configured to execute the steps of the rail transit driving curve planning method based on multi-source sensing data.

[0072] Furthermore, this embodiment also provides a computer-readable storage medium storing a computer program that is programmed or configured by a microprocessor to perform the steps of the rail transit driving curve planning method based on multi-source sensing data.

[0073] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0074] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for planning driving curves in rail transit based on multi-source sensing data, characterized in that, include: S1 divides the train's route into multiple sections. q i ; S2 detects the train's current position during operation. If it is found that the train is currently traveling to any adjacent section... q i If the boundary is found, proceed to step S3; otherwise, proceed to step S2 again. S3, a calculation range based on multiple sensing data of the train. q i control sequence u 2( k ) and its real-time optimized energy consumption J si ; S4 will optimize energy consumption in real time. J si and its interval q i Preset experience-optimized energy consumption J yi In comparison, if energy consumption is optimized in real time... J si Superior to experience-optimized energy consumption J yi Then select the control sequence. u 2( k Controlling trains in the section q i Drive; otherwise Maintain experience to optimize energy consumption J yi Corresponding preset control sequence u 1( k Controlling trains in the section q i The driving; Step S3 includes: S3.1 collects various sensor data of the train, including speed, acceleration, traction force and braking force; S3.2, Based on the multi-sensor data of the train, data fusion and classification are performed to determine the train's position within the section. q i The operating modes include four types: traction, cruise, coasting, and braking. Each operating mode corresponds to a gear curve, thereby obtaining the range. q i The gear shift curve is used as the interval q i control sequence u 2( k Simultaneously, various sensing data from the train are input into the train's multi-objective energy consumption optimization model to obtain the train's energy consumption during the interval. q i Real-time optimization of energy consumption J si The aforementioned data fusion classification based on multiple train sensing data refers to inputting multiple train sensing data into a Bayesian classifier for data fusion classification to determine the train's location within the interval. q i The operating mode; the functional expression of the multi-objective energy consumption optimization model is: ， In the above formula, The results represent the optimization of a multi-objective energy consumption optimization model, with the minimum optimization objective taken as the optimization objective. 、 、 、 These are the weighting factors, and E is the total energy consumption. For parking error, K jerk For comfort, To account for timeliness error, y m ( k () represents any time. k speed, v 0 represents the initial velocity. v (0) represents the velocity at t=0. s 0 represents the initial displacement. s (0) represents the displacement at time t=0, where t=0 is the initial time, and: ， ， ， ， In the above formula, t The number of intervals, f i ( s ) represents the interval q i The traction force of the train with a displacement of s. For traction force f i ( s The displacement produced by the following actions on the vehicle This is the product factor for the conversion of regenerative braking energy during train braking. For trains in the section Braking force in the middle, v i For trains in the section The speed in the middle, will be in any interval Total energy consumption E minus the interval The difference in total energy consumption E is used as the train's energy consumption in the interval. Real-time optimization of energy consumption J si ; This refers to the actual stopping position of the train. The target stopping position of the train, and the stopping error. Less than the preset threshold; For the train's acceleration, t For time; This refers to the train's actual arrival time. The on-time error is the train's target arrival time. Less than the preset value.

2. The rail transit driving curve planning method based on multi-source sensing data according to claim 1, characterized in that, In step S1, the train's route is divided into multiple sections. q i This refers to dividing a train's route into multiple sections by using the locations of signal lights, junctions requiring track switching, and locations where the operating status changes as boundaries. q i Changing the operating state refers to the train changing from a stopped state to a traction state, or from a traction state to a braking state.

3. The rail transit driving curve planning method based on multi-source sensing data according to claim 1, characterized in that, In step S2, the train is currently traveling to any adjacent section. q i The boundary refers to the current position and section of the train. q i The distance between the boundaries is equal to the set value.

4. The rail transit driving curve planning method based on multi-source sensing data according to claim 1, characterized in that, In step S4, the train is controlled within the section. q i Driving refers to driving according to the control sequence u 2( k ) or control sequence u 1( k Controlling trains in the section q i The control sequence refers to the gear positions at different displacements within the vehicle. u 2( k ) and control sequence u 1( k It includes the correspondence between different displacements and their corresponding gear positions.

5. The rail transit driving curve planning method based on multi-source sensing data according to any one of claims 1 to 4, characterized in that, Step S1 includes setting the pre-optimized speed curve of the train and the speed curves for each section. q i Setting the preset control sequence u 1( k Based on the train's pre-optimized speed curve, the corresponding empirically optimized energy consumption is determined. J yi The pre-optimized speed curve includes a two-dimensional coordinate system curve of the train's displacement and speed, and the corresponding empirically optimized energy consumption is determined based on the train's pre-optimized speed curve. J yi This includes: based on the obtained pre-optimized speed curve, testing the train multiple times on the same line following the pre-optimized speed curve, and then selecting each section... q i The average energy consumption of the interval q i Corresponding experience-optimized energy consumption J yi .

6. A rail transit driving curve planning system based on multi-source sensing data, comprising a microprocessor and a memory interconnected, characterized in that, The microprocessor is programmed or configured to perform the steps of the rail transit driving curve planning method based on multi-source sensing data as described in any one of claims 1 to 5.

7. A computer-readable storage medium storing a computer program, characterized in that, The computer program is used to be programmed or configured by a microprocessor to perform the steps of the rail transit driving curve planning method based on multi-source sensing data as described in any one of claims 1 to 5.

Images