Operation behavior detection method, device and equipment based on heavy-load train simulation platform, storage medium and product

Abstract

The invention relates to an operation behavior detection method and device based on a heavy-load train simulation platform, equipment, a storage medium and a product. Operation logic detection is carried out on operation information of a heavy-load train in a simulation transportation scene of a heavy-load train simulation platform based on an object, an operation logic detection result is determined, and abnormal response detection is carried out on the operation information of the heavy-load train in a simulation abnormal scene of the heavy-load train simulation platform based on the object. And determining an operation behavior detection result of the object in combination with the logic detection result and the abnormal response detection result. Compared with a traditional method of detecting the operation behavior according to a preset scoring rule, the method has the advantages that the heavy-load train simulation platform is constructed, various simulation scenes are provided for the detected object to perform simulation operation, whether the operation behavior of the object is abnormal or not can be detected more accurately, and the accuracy of operation behavior detection is improved.

Description

Technical Field

[0001] This application relates to the field of train safety technology, and in particular to a method, device, computer equipment, computer-readable storage medium, and computer program product for detecting operational behavior based on a heavy-haul train simulation platform. Background Technology

[0002] With the development of transportation technology, heavy-haul trains have become the main mode of railway transportation. Heavy-haul trains are technically complex, and to ensure the safety of their transportation and operation, timely and accurate monitoring of the operational behaviors of relevant personnel is necessary. Currently, the method for monitoring operational behaviors typically involves evaluating certain pre-set behaviors according to preset scoring rules. However, relying on preset scoring rules for operational behavior monitoring leads to a rigid monitoring method and reduces the accuracy of the monitoring.

[0003] Therefore, current methods for detecting the operational behavior of heavy-haul trains suffer from low accuracy. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for detecting operational behavior based on a heavy-haul train simulation platform, which can improve the detection accuracy of the above-mentioned technical problems.

[0005] Firstly, this application provides a method for detecting operational behavior based on a heavy-haul train simulation platform, including:

[0006] Obtain the first operation information corresponding to the object to be detected; the first operation information represents the operation information of the object on the heavy-haul train in the simulated transportation scenario of the heavy-haul train simulation platform; the heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-car simulation control unit and the braking strategy simulation unit;

[0007] Based on the first operation information, an operation logic detection is performed to obtain the operation logic detection result;

[0008] Obtain the second operation information corresponding to the object; the second operation information represents the operation information of the object on the heavy-haul train in the simulated abnormal scenario of the heavy-haul train simulation platform; the simulated abnormal scenario is matched with the abnormal recovery capability of the object;

[0009] Anomaly response detection is performed based on the second operation information to obtain the anomaly response detection result;

[0010] Based on the logic detection results and the abnormal response detection results, the operation behavior detection results corresponding to the object are obtained.

[0011] Secondly, this application also provides an operational behavior detection device based on a heavy-haul train simulation platform, comprising:

[0012] The first acquisition module is used to acquire the first operation information corresponding to the object to be detected; the first operation information represents the operation information of the object on the heavy-haul train in the simulated transportation scenario of the heavy-haul train simulation platform; the heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-car simulation control unit and the braking strategy simulation unit;

[0013] The first detection module is used to perform operation logic detection based on the first operation information and obtain the operation logic detection result.

[0014] The second acquisition module is used to acquire second operation information corresponding to the object; the second operation information represents the operation information of the object on the heavy-haul train in the simulated abnormal scenario of the heavy-haul train simulation platform; the simulated abnormal scenario is matched with the abnormal recovery capability of the object.

[0015] The second detection module is used to perform abnormal response detection based on the second operation information and obtain the abnormal response detection result.

[0016] The third detection module is used to obtain the operation behavior detection result corresponding to the object based on the logical detection result and the abnormal response detection result.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method.

[0018] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0019] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0020] The aforementioned method, apparatus, computer equipment, computer-readable storage medium, and computer program product for detecting operational behavior based on a heavy-haul train simulation platform perform operational logic detection on the operational information of a heavy-haul train within a simulated transportation scenario on the platform, determining the operational logic detection result. It also performs abnormal response detection on the operational information of the heavy-haul train within a simulated abnormal scenario on the same platform, obtaining the abnormal response result. Combining the logic detection result and the abnormal response detection result, the operational behavior detection result of the object is determined. Compared to traditional operational behavior detection based on preset scoring rules, this solution, by constructing a heavy-haul train simulation platform and providing multiple simulated scenarios for the object to perform simulated operations, can more accurately detect whether the object's operational behavior is abnormal, thus improving the accuracy of operational behavior detection. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating an operational behavior detection method based on a heavy-haul train simulation platform in one embodiment.

[0023] Figure 2 This is a schematic diagram of the structure of a heavy-haul train simulation platform in one embodiment;

[0024] Figure 3 This is a structural block diagram of an operation behavior detection device based on a heavy-haul train simulation platform in one embodiment;

[0025] Figure 4 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0027] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0028] In related technologies, railway simulation systems have shortcomings. First, they lack simulation of the specific operating conditions of heavy-haul trains, such as braking control of long train formations, longitudinal impulse mitigation, and wireless multiple-connection synchronization, making it difficult to meet the detection needs of simulated operational behaviors in heavy-haul transport stations. Second, in terms of operational behavior detection, reliance on preset scoring rules fails to deeply analyze operational intentions, and the fault handling behavior detection mode is too rigid, unable to flexibly adapt to the learning pace and ability improvement of different participants. Third, for the detection of multi-role collaborative operational behaviors, the collaborative logic of positions such as station duty officers, dispatchers, and drivers differs significantly from actual scenarios, and there is a lack of content related to dispatching and command processes specific to heavy-haul transport, resulting in an overall gap between the effectiveness of operational behavior detection and actual work requirements.

[0029] Based on this, this application constructs a heavy-haul train simulation platform, providing various simulation scenarios for the objects to be tested to perform simulated operations, which can more accurately detect whether the operation behavior of the objects is abnormal and improve the accuracy of operation behavior detection.

[0030] In one embodiment, such as Figure 1 As shown, an operational behavior detection method based on a heavy-haul train simulation platform is provided. This embodiment illustrates the application of this method to a terminal. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server, including the following steps S202 to S210. Wherein:

[0031] Step S202: Obtain the first operation information corresponding to the object to be detected; the first operation information represents the operation information of the object on the heavy-haul train in the simulated transportation scenario of the heavy-haul train simulation platform; the heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-vehicle simulation control unit and the braking strategy simulation unit.

[0032] The subjects to be tested can be personnel related to heavy-haul trains, such as drivers and station staff. To ensure the operational safety of heavy-haul trains, detailed testing of the operational behaviors of these subjects is required. This testing can be performed on a heavy-haul train simulation platform, which can be installed on a terminal. To ensure that the heavy-haul train simulation platform is consistent with the real heavy-haul train environment, the terminal can be combined with a heavy-haul train dynamics simulation unit, a multi-car simulation control unit, and a braking strategy simulation unit to construct the heavy-haul train simulation platform.

[0033] The heavy-haul train dynamics simulation unit is designed for longitudinal train dynamics, simulating air braking wave propagation delay and inter-car impulse transmission effects under different train formation lengths. The multi-car simulation control unit simulates abnormal fault scenarios such as interruption of wireless communication between master and slave locomotives and asynchronous traction, and integrates BeiDou positioning correction technology to improve positioning accuracy. The braking strategy simulation unit dynamically generates energy consumption and braking curves based on three-dimensional terrain data of the track, specifically for typical long gradients on heavy-haul railways.

[0034] After constructing a heavy-haul train simulation platform, the terminal can provide simulated operations related to heavy-haul trains to the object being monitored through the platform. For example, the terminal can obtain the object's operation information on the heavy-haul train within the simulated transportation scenario of the heavy-haul train simulation platform, as the first operation information. The simulated transportation scenario can be a scenario provided by the heavy-haul train simulation platform that simulates the transportation process of heavy-haul trains in the real world. The object can perform corresponding operations in this simulated transportation scenario, thereby allowing the terminal to obtain the corresponding first operation information.

[0035] Step S204: Perform operation logic detection based on the first operation information mentioned above to obtain the operation logic detection result.

[0036] The first operation information may include a series of operations related to heavy-haul train transportation performed by the aforementioned object in the simulated transportation scenario. The terminal can perform operation logic detection based on the first operation information to obtain operation logic detection results. This operation logic detection can be related to the operation logic of the object during heavy-haul train transportation, such as detecting whether the object's operation sequence is incorrect or whether any operations are omitted.

[0037] Step S206: Obtain the second operation information corresponding to the above object; the second operation information represents the operation information of the above object on the above heavy-haul train in the simulated abnormal scenario of the above heavy-haul train simulation platform; the simulated abnormal scenario matches the abnormal recovery capability of the above object.

[0038] The aforementioned heavy-haul train simulation platform also includes simulated abnormal scenarios. These simulated abnormal scenarios represent real-world situations where heavy-haul trains experience abnormal conditions. To ensure the safety of these heavy-haul trains, the ability of the aforementioned objects to handle abnormalities needs to be tested. The terminal can obtain the operational information of the aforementioned objects regarding the heavy-haul train in the simulated abnormal scenarios of the heavy-haul train simulation platform, thus obtaining second operational information. The operational difficulty of the simulated abnormal scenarios can be adjusted according to the abnormal recovery capabilities of the aforementioned objects, ensuring a match between the simulated abnormal scenarios and the objects' abnormal recovery capabilities. For example, the stronger the abnormal recovery capability of the object, the greater the difficulty of recovering from the simulated abnormal scenario; conversely, the weaker the abnormal recovery capability of the object, the easier the recovery from the simulated abnormal scenario, avoiding misjudgments of the object's detection results due to a mismatch in difficulty. The abnormal recovery capability represents the aforementioned objects' ability to restore the abnormal scenario of the heavy-haul train to normal, and this capability can be determined based on the object's historical second operational information.

[0039] Step S208: Perform anomaly response detection based on the second operation information described above, and obtain the anomaly response detection result.

[0040] The second operation information may include the object's response actions to the simulated abnormal scenario. The terminal can perform abnormal response detection on the object based on this second operation information to obtain the abnormal response detection result. The abnormal response detection indicates the object's ability to respond to abnormal situations involving heavy-load trains in the simulated abnormal scenario. Examples include the time it takes for the object to initiate recovery operations after an anomaly occurs, the time it takes for the heavy-load train to return to normal, and the specific response operations performed by the object in the simulated abnormal scenario.

[0041] Step S210: Based on the above logic detection results and the above abnormal response detection results, obtain the operation behavior detection results corresponding to the above object.

[0042] After obtaining the aforementioned logic detection results and anomaly response detection results, the terminal can derive the operation behavior detection results corresponding to the aforementioned object based on these results. For example, the logic detection results may include whether the object's operation conforms to the target operation logic corresponding to the simulated transportation scenario; the anomaly response detection results may include whether the object's operation conforms to the target operation corresponding to the simulated anomaly scenario; and whether the object's operation meets the target time conditions. If the terminal determines that the logic detection results indicate that the object's operation conforms to the target operation logic corresponding to the simulated transportation scenario, conforms to the target operation corresponding to the simulated anomaly scenario, and meets the target time conditions, then the terminal can determine that the operation behavior detection result for the aforementioned object is passed. Otherwise, if the terminal detects that any one of these conditions is not met, the operation behavior detection result can be considered failed.

[0043] The aforementioned operational behavior detection method based on a heavy-haul train simulation platform involves performing operational logic detection on the object's operational information regarding the heavy-haul train in a simulated transportation scenario within the platform, determining the operational logic detection result, and performing abnormal response detection on the object's operational information regarding the heavy-haul train in a simulated abnormal scenario within the same platform, obtaining the abnormal response result. The operational behavior detection result is then determined by combining the logic detection result and the abnormal response detection result. Compared to traditional operational behavior detection based on preset scoring rules, this solution, by constructing a heavy-haul train simulation platform and providing multiple simulated scenarios for the object to perform simulated operations, can more accurately detect whether the object's operational behavior is abnormal, thus improving the accuracy of operational behavior detection.

[0044] In one embodiment, the method further includes: constructing a heavy-haul train dynamics simulation unit based on the air braking propagation delay simulation information and the inter-car impulse transmission simulation information of the heavy-haul train; constructing a multi-car simulation control unit based on the wireless communication interruption simulation information, the traction force difference simulation information between cars, and the positioning simulation information of the heavy-haul train; constructing a braking strategy simulation unit based on the terrain data simulation information of the line where the heavy-haul train is located, the dynamic change simulation information, resistance simulation information, energy consumption simulation information, and the target braking curve of the heavy-haul train; and constructing a heavy-haul train simulation platform based on the heavy-haul train dynamics simulation unit, the multi-car simulation control unit, and the braking strategy simulation unit.

[0045] In this embodiment, the terminal can pre-build a heavy-haul train simulation platform. The terminal can construct the heavy-haul train simulation platform by building multiple simulation units. For example, the terminal can build a heavy-haul train dynamics simulation unit based on the simulated information of air braking propagation delay and the simulated information of impulse transmission between carriages. The terminal can build a multi-car simulation control unit based on the simulated information of wireless communication interruption, the simulated information of traction force differences between carriages, and the simulated information of positioning. Furthermore, the terminal can build a braking strategy simulation unit based on the simulated information of terrain data of the line where the heavy-haul train is located, the simulated information of dynamic changes, resistance, energy consumption, and the target braking curve. Thus, the terminal can construct a heavy-haul train simulation platform based on the aforementioned heavy-haul train dynamics simulation unit, multi-car simulation control unit, and braking strategy simulation unit.

[0046] Specifically, for the heavy-haul train dynamics simulation unit, the terminal can perform air brake propagation delay simulation. For example, the terminal can simulate the air brake propagation delay based on the number of vehicles in the train formation (n) and the braking signal propagation speed (v). s and vehicle spacing d iA model is established to calculate the time delay of the braking signal propagating from the lead vehicle to subsequent vehicles; for the i-th vehicle, the time delay t for receiving the braking signal is calculated. d,i Represented as: Wherein: d k v represents the distance between the k-th car and the (k+1)-th car. s This refers to the propagation speed of the braking signal in the vehicle's brake lines. Its value is determined based on the actual vehicle braking system parameters, generally between 30 m / s and 50 m / s. When the train length changes, i.e., n changes, the terminal adjusts the spacing d between each vehicle. i To adapt to different formation scenarios, thereby achieving accurate simulation of the propagation delay of air braking waves under different formation lengths.

[0047] The terminal can also simulate the transmission of impulse between carriages. For example, the terminal considers the vehicle mass m. i Stiffness k of the connection device between vehicles c Damping coefficient c c Including the vehicle's operating speed v, a longitudinal dynamics model of the vehicle is established; for the i-th vehicle, its longitudinal motion equation is: m i (dv i / dt)=F i-1,i -F i,i+1 , where: F i-1,i F represents the longitudinal force exerted by the (i-1)th vehicle on the ith vehicle. i,i+1 This represents the longitudinal force exerted by vehicle i on vehicle i+1; the terminal can convert the force F according to the characteristics of the vehicle connection device. i,i+1 Represented as: F i,i+1 =k c (x i -x i+1 )+c c (dx i / dt-dx i+1 / dt), where: x i x i+1 Let k be the longitudinal displacement of the i-th vehicle and the (i+1)-th vehicle, respectively. c and c c These are the stiffness coefficient and damping coefficient of the vehicle connection device, respectively, which are determined based on the physical parameters of the actual vehicle connection device. The terminal can solve the above set of equations to obtain the impulse force transmission process between vehicles under braking or traction conditions, and thus simulate the impulse force transmission effect between carriages under different train formation lengths.

[0048] For multi-vehicle simulation control units, the terminal can simulate wireless communication interruptions. For example, the terminal constructs a communication interruption model based on the reliability parameter p of the communication link and the interruption duration distribution parameter λ. The reliability parameter p represents the probability that the communication link remains normal per unit time, and its value is between 0 and 1, usually determined based on the performance statistics of the actual wireless reconnection communication system. The interruption duration distribution parameter λ is used to describe the exponential distribution of the communication interruption duration, i.e., the interruption duration t. int The probability density function is f(t) int )=λe -λtint The value of λ is obtained by fitting historical communication interruption data. During the simulation, the terminal simulates the occurrence of communication interruption events through a random number generator. When a communication interruption is detected, the interruption duration is randomly determined according to the set interruption duration distribution, and the data transmission between the master locomotive and the slave locomotive is cut off during the time period, thereby realizing the simulation of wireless reconnection communication interruption scenario.

[0049] The terminal can also simulate traction force differences between carriages. For example, based on the differences in traction control systems between the master locomotive and slave locomotive of a heavy-haul train, as well as communication delays, the terminal establishes a traction force asynchrony model; assuming the traction force command of the master locomotive is F... head The traction command received from the locomotive is F slave Due to communication delay t delay and the response time t of the controlled locomotive response The existence of the locomotive's actual traction output F slave_actual There is a delay and deviation relative to the main control locomotive, which is expressed as: F slave_actual (t)=F slave (tt delay -t response In some embodiments, the terminal may also introduce a traction fluctuation coefficient k. fluctuate To simulate the instability of traction output, its value randomly varies between 0.9 and 1.1, indicating that the traction force may fluctuate within ±10% of the commanded value. Therefore, the actual traction output of the slave locomotive can be further expressed as: F slave_actual (t)= k fluctuate F slave (tt delay -t response Thus, the terminal can adjust the communication delay t. delay Response time t of the controlled locomotive response and traction fluctuation coefficient k fluctuate The parameter values ​​can simulate different degrees of asynchronous traction forces, where t delay and t response The value of t is determined based on the performance of the actual communication system and locomotive control system.delay Between 0.1 and 0.5 seconds, t response Between 0.2 and 1 second.

[0050] The terminal can also perform positioning simulation. For example, the multi-vehicle simulation control unit of the terminal includes a positioning system, and the positioning is based on BeiDou positioning. In the locomotive positioning system, the terminal incorporates BeiDou satellite positioning data and correction algorithms to improve positioning accuracy. Specifically, the terminal sets the initial positioning coordinates of the locomotive as (x... init ,y init The positioning coordinates obtained by receiving BeiDou satellite signals are (x bd ,y bd Due to factors such as satellite signal errors and environmental obstructions, positioning deviations exist. To correct these deviations, the terminal employs a Kalman filter algorithm to fuse the positioning data. The state vector x of the Kalman filter... k Including the locomotive's actual position coordinates (x true ,y true ) and velocity (v) x ,v y Its state transition equation is: x k =Ax k-1 +w k-1 Where: A is the state transition matrix, expressed as: . t is the sampling time interval, w k-1 The process noise is assumed to be Gaussian white noise with a covariance matrix of Q; the observation vector z k For BeiDou positioning coordinates (x bd ,y bd The observation equation is: z k =Hx k +v k Where H is the observation matrix, expressed as: Among them, v k For the observation noise, it is also assumed to be Gaussian white noise with a covariance matrix of R; the terminal can obtain the corrected locomotive positioning coordinates (x, y, y) by iteratively performing the prediction and update steps of Kalman filtering. correct ,y correct This improves positioning accuracy; parameters in the Kalman filter algorithm. The value of t is determined based on the sampling frequency of the positioning system, and is generally between 0.1 and 1 second. The values ​​of the process noise covariance matrix Q and the observation noise covariance matrix R are adjusted according to the statistical characteristics of the actual positioning error in order to optimize the filtering effect.

[0051] For the braking strategy simulation unit, the terminal can perform terrain data simulation. For example, the terminal can acquire three-dimensional terrain data of a typical long gradient on a heavy-haul railway, including gradient information i(x), curve radius R(x), and the relationship between altitude h(x) and mileage x. This data can be obtained through surveying technology or railway line databases, represented in the form of discrete points, and converted into a continuous function form through interpolation algorithms (such as cubic spline interpolation) for subsequent calculations.

[0052] The terminal can also simulate the dynamic changes of heavy-haul trains. For example, based on the train's mass m, speed v(x), gravitational acceleration g, and track gradient i(x), the terminal calculates the changes in the train's kinetic energy KE(x) and potential energy PE(x) on a slope. The expressions for kinetic energy and potential energy are: KE(x) = 1 / 2mv(x). 2 ; PE(x) = mgh(x) = mg The equation is: ∫ix (i(x')dx'). Where: v(x) is the train speed at mileage x, which can be determined based on the train's operating conditions and control strategy; i(x) is the gradient function, representing the gradient value of the line at mileage x, in dimensionless decimal form (e.g., a 1% gradient is represented as 0.01); g is the gravitational acceleration, approximately 9.81 m / s². The terminal can assess the train's energy conversion on the gradient by calculating the changes in kinetic and potential energy at different mileages.

[0053] The terminal can also simulate the resistance of heavy-haul trains. For example, the terminal calculates the resistance experienced by a train running on a slope, including the basic resistance F. b , ramp resistance F g and curve resistance F c The basic resistance is mainly caused by factors such as wheel-rail friction and air resistance, and its calculation formula is: F b =av+bv 2 +c, where: a, b, and c are resistance coefficients related to train shape, vehicle type, etc., determined through experiments or empirical formulas; v is the train speed; gradient resistance is caused by the track gradient, expressed as: Fg=mgi(x); curve resistance is generated when the train passes through a curve section, and its calculation formula is: F c =(d / R(x))mg. Where: d is a coefficient related to the train bogie parameters, R(x) is the curve radius; total resistance F t The sum of the three resistances mentioned above is expressed as: F t =F b +F g +F c .

[0054] The terminal can also perform energy consumption simulations. For example, the terminal simulates energy consumption based on the train's traction force F. 牵引 Total resistance Ft Given the train speed v(x), calculate the train's energy consumption E(x) on the slope; where the work done by the traction force to overcome the total resistance is the energy consumption, expressed as: E(x) = (F 牵引 -F t )v(x')dx', where: F 牵引 It is the traction force provided by the train traction system, which can be determined based on the relationship between the train's power and speed, i.e., F. 牵引 =P 牵引 / v(x), P 牵引 This refers to the train's traction power. The terminal then calculates the energy consumption over different mileage ranges by integrating the data, thus obtaining a curve showing the change in energy consumption over mileage.

[0055] The terminal can also simulate the target braking curve. For example, the terminal generates a braking curve based on the train speed v(x), braking deceleration a, and braking distance s; during braking, the relationship between train speed and distance can be described by a kinematic equation: v(x) 2 =v0 2 -2a 制动 (x-x0), where v0 is the initial velocity at the start of braking, x0 is the initial position at the start of braking, and a 制动 The train's deceleration is determined by factors such as the performance of the braking system and the train's load; simultaneously, the braking distance s 制动 It can be represented as: s 制动 =v0 2 / 2a 制动 The terminal can adjust the braking strategy based on the track gradient and train status to adapt to braking requirements under different gradient conditions, thereby generating a braking curve corresponding to the energy consumption curve and realizing dynamic optimization of the cyclic braking strategy. Furthermore, through the braking strategy simulation unit, the terminal can dynamically generate energy consumption and braking curves based on the three-dimensional terrain data of typical long gradients on heavy-haul railways. This provides a simulation scenario close to actual operating conditions for the object under test, helping to master the application of the cyclic braking strategy and improve energy utilization efficiency and train operation safety.

[0056] Through this embodiment, the terminal can combine a dynamics simulation unit, a multi-vehicle simulation control unit, and a braking strategy simulation unit to build a heavy-haul train simulation platform, thereby improving the simulation accuracy of the heavy-haul train simulation platform for heavy-haul trains.

[0057] In one embodiment, performing operation logic detection based on the first operation information to obtain an operation logic detection result includes: inputting the first operation information into a logic recognition model; the logic recognition model is used to identify each semantic vector corresponding to the first operation information based on an attention mechanism; determining each operation rule corresponding to the first operation information based on the matching of each semantic vector with a preset logic rule base; determining operation logic anomaly information in the first operation information based on the matching of each operation rule with a preset logic anomaly rule base; and obtaining the operation logic detection result based on the operation logic anomaly information.

[0058] In this embodiment, the terminal can perform operation logic detection on the first operation information through a logic recognition model. Specifically, the terminal can input the first operation information into the logic recognition model. The logic recognition model identifies the semantic vectors corresponding to the first operation information based on an attention mechanism, and determines the operation rules corresponding to the first operation information based on the matching of each semantic vector with a preset logic rule base. Thus, the logic recognition model can match each operation rule with a preset logic anomaly rule base, and determine the operation logic anomaly information in the first operation information based on the matching results. The terminal can then obtain the operation logic detection result based on the operation logic anomaly information.

[0059] Specifically, the aforementioned logic recognition model can be an attention mechanism model. The first operation information includes the voice command input by the object. The terminal parses the voice command using natural language processing technology, combines it with the operation timing logic to detect compliance, and identifies implicit errors such as "departing without confirming the train's integrity," thus obtaining the operation logic detection result.

[0060] The terminal can convert the input voice commands into a text sequence T={t1,t2,…,t3} via a voice recognition system. i}, where t i Let A represent the i-th word in the text, and let A be the accuracy of the speech recognition. r The accuracy should reach over 90% to ensure the effectiveness of subsequent semantic analysis. The terminal can input text sequences into a logic recognition model based on a self-attention mechanism, which includes a multi-head self-attention mechanism and a feedforward neural network; model parameters include the embedding dimension d. model (Usually set to 512 dimensions), number of attention heads h (usually 8), and hidden layer dimension d of the feedforward network. ff (Can be 2048-dimensional) and the word embedding matrix E belongs to R V×dmodel , where V is the vocabulary size.

[0061] The terminal uses a logical recognition model to generate a context-dependent semantic vector representation h for each word in the text sequence. i Belongs to Rdmodel And perform multi-head self-attention calculation. For example, for each word t i Calculate the query vector, key vector K, and value vector: Query = h i W Q Key=h j W K Value=h j W V Among them, W Q W K W V It is the weight matrix, W Q W K W V Belongs to R dmodel×dk d k =d model / h; The terminal can calculate the AttentionScore. i,j AttentionScore i,j =QueryKey T / √d k Then, the attention score is normalized and represented as: Softmax(AttentionScore) i,j )=exp(AttentionScore i,j ) / exp(AttentionScore i,j ), final updated word t i The semantic vector is represented as: h i new = Softmax (AttentionScore) i,j V j .

[0062] When a terminal detects operational logic using a logic recognition model, it can do so based on a pre-defined logic rule base. The terminal can pre-set this rule base. The rule base is represented as R={r1,r2,…,r...} k}. Where each rule r k This is represented as an operation sequence pattern and its compliance requirements. For example, rule r1 indicates that "confirming train integrity" must be completed before the "departure" operation. The logic recognition model uses a detection algorithm to scan operation keywords in the text sequence through a sliding window and maps them to a rule base for matching. The terminal can also pre-set a preset logic exception rule base, which includes rules for multiple operation logic errors. The preset logic exception rule base can be represented as E={e1,e2,…,e…} l}. Wherein, each pattern e l Let e1 represent a possible implicit error scenario, such as "departing without confirming train integrity," which can be represented as the operation sequence {departing without confirming train integrity}. The logical recognition model identifies implicit errors by calculating the similarity between the text sequence and each pattern in the pattern library. The similarity calculation can use cosine similarity, which can be expressed as: Similarity(T, e) l )= h i e l,j / (√ ||h i || 2 √ ||e l,j || 2 ), where e l,j It is mode e l The semantic vector of the j-th word in n l It is mode e l If the similarity exceeds a threshold, for example, if the threshold is set to 0.7, then the operation logic detection result is determined to be that there is a corresponding hidden error.

[0063] In this embodiment, the terminal accurately parses the voice commands of the object based on the logic recognition model, combines the operation timing logic to detect compliance, and effectively identifies hidden errors, providing targeted simulated operation feedback for the object to be detected, thereby improving the accuracy of operation logic detection.

[0064] In one embodiment, before performing anomaly response detection based on the aforementioned second operation information and obtaining the anomaly response detection result, the method includes: acquiring historical second operation information of the object for historical simulated anomaly scenarios; inputting the historical simulated anomaly scenarios and the historical second operation information into a trained deep Q-network model; the deep Q-network model being used to adjust the anomaly level of the historical simulated anomaly scenarios based on the historical simulated anomaly scenarios and the historical second operation information to obtain an adjusted simulated anomaly scenario; the anomaly level representing the severity of the anomaly in the simulated anomaly scenario; the anomaly level of the adjusted simulated anomaly scenario matching the anomaly recovery capability of the object; and acquiring the second operation information corresponding to the adjusted simulated anomaly scenario for the object.

[0065] In this embodiment, the terminal can adjust the difficulty of the simulated abnormal scenario in real time by combining the abnormal recovery capability of the object to be detected. For example, the terminal can obtain historical second operation information of the object in response to historical simulated abnormal scenarios; input the historical simulated abnormal scenario and the historical second operation information into a trained deep Q-network model. The deep Q-network model determines the current abnormal recovery capability of the object based on the historical simulated abnormal scenario and the historical second operation information, and adjusts the abnormality level of the historical simulated abnormal scenario to obtain the adjusted simulated abnormal scenario. The abnormality level represents the severity of the abnormality in the simulated abnormal scenario, such as the degree of damage to various components in a heavy-duty train. The adjusted abnormality level of the simulated abnormal scenario matches the abnormal recovery capability of the object, thereby enabling the terminal to obtain the second operation information of the object in response to the adjusted simulated abnormal scenario and perform abnormal response detection.

[0066] Specifically, taking a simulated abnormal scenario as an example of a fault scenario, the terminal can construct corresponding simulated abnormal scenarios by combining faults of heavy-load trains. For example, the terminal learns the optimal fault level adjustment strategy using a deep Q-network algorithm. The core of the deep Q-network is to use a neural network Q(s,a,θ) to approximate the Q-value function, where θ is the network parameter, the network input is the state s, and the output is the Q-value of all possible actions a. The specific steps include: Network structure design: The deep Q-network includes an input layer, two hidden layers, and an output layer. Each hidden layer has 64 neurons, and the activation function is the Rectified Linear Unit (ReLU). The input layer receives the state vector s, and the output layer outputs the Q-value corresponding to action a. Experience replay buffer: The terminal sets up an experience replay buffer of size N (e.g., 10000) to store the most recent N experiences (s,a,r,s'), where s is the current state, a is the action performed, r is the reward obtained, and s' is the new state transitioned to. Training process: In each training step of the deep Q-network model, the terminal randomly draws a mini-batch of empirical samples (e.g., 32 samples) from the empirical replay buffer and updates the network using the following loss function: Where M is the number of mini-batch samples, y i The target Q value can be represented as: y i =r i +γmax a’ Q(s i ',a',θ - ), where γ is the discount factor with a value of 0.99, and θ -These are the parameters of the target network, updated every C steps (e.g., 1000 steps) to synchronize them with the current network parameters θ. Exploration strategy: The terminal uses an ε-greedy strategy for action selection. The initial ε value is set to 1.0, representing completely random exploration. As training progresses, the ε value gradually decays linearly to 0.1, representing partly random exploration and partly utilizing learned knowledge, to balance the relationship between exploration and utilization.

[0067] Thus, the terminal can utilize a trained deep Q-network model to adjust for simulated abnormal scenarios. For example, during detection, the terminal monitors the object's operational performance and capability level changes in real time, dynamically adjusting the fault level, including: State update: At each time step, the latest capability parameter L' and the current fault scenario parameter F' of the object are obtained, and the state s'=(L',F') is updated. Action selection: The terminal uses the trained deep Q-network model to select the optimal action a' based on the current state s', i.e., the fault level adjustment amount (g1',g2',…,g...). i Fault Combination Update: Based on action a', the terminal adjusts the fault level, generates a new fault combination, and updates the fault scenario parameter F''. The specific adjustment formula is: f j ''=f j '+g j '。 Among them, f j ' is the current level of the j-th type of fault, g j 'This is the corresponding action adjustment amount, ensuring the adjusted fault level f' j "Within a reasonable range (e.g., level 1 to 5)."

[0068] Through this embodiment, the terminal can combine a deep Q-network to adjust the difficulty of simulated abnormal scenarios in real time. The difficulty level of the abnormal scenario is adjusted in real time according to the actual performance of the object, so that the content of detection always matches the ability level of the object, thereby enhancing the detection effect and the object's emergency response capability.

[0069] In one embodiment, anomaly response detection based on the second operation information to obtain anomaly response detection results includes: determining the post-operation simulated anomaly scenario corresponding to the simulated anomaly scenario based on the second operation information; determining the change in operation accuracy, change in response time, and change in the number of correct operands corresponding to the object based on the scenario comparison results between the simulated anomaly scenario and the post-operation simulated anomaly scenario; the change in the number of correct operands represents the change in the number of operations that are conducive to the recovery of the simulated anomaly scenario; and obtaining the anomaly response detection results based on the weighted sum of the change in operation accuracy, change in response time, and change in the number of correct operands.

[0070] In this embodiment, the terminal can perform abnormal response detection on the second operation information of the object to be detected, thereby determining the object's ability to handle simulated abnormal scenarios. For example, the terminal can determine the simulated abnormal scenario after the operation corresponding to the simulated abnormal scenario based on the second operation information. The terminal can then compare the simulated abnormal scenario with the simulated abnormal scenario after the operation to obtain a scenario comparison result, and determine the change in operation accuracy, response time, and number of correct operations corresponding to the object based on the scenario comparison result.

[0071] The change in the number of correct operands represents the change in the number of operations that are beneficial to the recovery from the simulated abnormal scenario. Therefore, the terminal can obtain the abnormal response detection result based on the weighted sum of the changes in operation accuracy, response time, and correct operations.

[0072] Specifically, taking a simulated abnormal scenario as an example of a fault scenario, the terminal can define a simulated abnormal scenario and combine it with a reward function to achieve abnormal response detection. For example, the terminal can define a reinforcement learning framework. Here, the terminal models the fault combination generation and adjustment problem as a Markov decision process, defining the state space. Action space The reward function R(s,a) and the transition probability P(s'|s,a) are defined. The state space includes the trainee's ability parameters L=[l1,l2,…,l...]. n (e.g., operation accuracy, response time, number of correct operations) and current fault scenario parameters (e.g., triggered fault type, fault duration, fault complexity, etc.), specifically represented as s=(L,F).

[0073] The number of correct operations represents the object's level of knowledge mastery. The action space includes actions for generating new fault combinations and adjusting fault levels; specific actions are represented as a = (g1, g2, ..., g...). i ), where g i This represents the adjustment amount for the i-th type of fault level, with values ​​ranging from -2 (decrease by two levels), -1 (decrease by one level), 0 (remain unchanged), +1 (increase by one level), and +2 (increase by two levels), thus ensuring that the fault level changes within a reasonable range. The reward function can be expressed as the terminal calculating the reward based on the trainee's performance and skill improvement in the current fault scenario. The reward function is defined as: R(s,a)=w1 P+w2 T+w3 K-w4CZ, where... P is the change in the student's operational accuracy, representing the increase or decrease in the student's operational accuracy in the current scenario compared to the previous scenario, and its value ranges from [-1, 1]. T is the change in student response time, representing the decrease or increase in student response time in the current scenario compared to the previous scenario. The value range is [-1, 1] (negative values ​​indicate a decrease in response time, and positive values ​​indicate an increase in response time). K represents the change in the learner's knowledge level, indicating the increase or decrease in the learner's knowledge level in the current scenario compared to the previous scenario, with a value range of [-1, 1]. CZ represents the matching degree between the complexity of the fault scenario and the learner's ability. When the fault scenario complexity matches the learner's ability well, the value is 0; when the fault scenario complexity is too high or too low, the value is positive or negative, with an absolute value range of [0, 2]. w1, w2, w3, and w4 are weighting coefficients, represented as 0.4, 0.3, 0.2, and 0.8 respectively, used to balance the influence of different factors on the reward. The transition probability represents the probability of transitioning to state s' after performing action a in state s. This probability is estimated based on historical data and the object model, with the initial value set to a uniform distribution and updated through continuous learning.

[0074] After the object completes the operation of the current simulated fault scenario, the terminal can calculate the reward R(s',a'), perform reward calculation and feedback, and determine the abnormal response detection result. The experience (s',a',R(s',a'),s'') is stored in the experience replay buffer for further training of the network, where s'' is the new state of the simulated abnormal scenario of the heavy-load train simulation platform after the object completes the operation.

[0075] Through this embodiment, the terminal can determine the abnormal response detection result by comparing the simulated abnormal scenario with the simulated abnormal scenario after the object operation, and combining multiple parameters corresponding to the object operation, thereby improving the accuracy of abnormal response detection.

[0076] In one embodiment, obtaining the second operation information corresponding to the above-mentioned object includes: obtaining the scene feature vector of the above-mentioned simulated abnormal scenario; determining the abnormal case attribute corresponding to the above-mentioned simulated abnormal scenario based on the similarity between the above-mentioned scene feature vector and the preset scene feature vector of each preset abnormal scenario in the preset abnormal case library; obtaining the operation information input by the above-mentioned object based on the above-mentioned abnormal case attribute to obtain the second operation information.

[0077] In this embodiment, the aforementioned second operation information can be an operation performed by the object under the condition that the terminal recommends a corresponding preset abnormal scenario. This enables the detection of the object's ability to analyze the combination of preset abnormal scenarios and simulated abnormal scenarios. Specifically, the terminal can determine the preset abnormal scenario that matches the simulated abnormal scenario. For example, the terminal can obtain the scene feature vector of the simulated abnormal scenario. Based on the similarity between the scene feature vector and the preset scene feature vectors of various preset abnormal scenarios in the preset abnormal case library, the terminal determines the abnormal case attribute corresponding to the simulated abnormal scenario. The abnormal case attribute represents information such as the scene description and fault type of the preset abnormal scenario. The terminal can obtain the operation information input by the object based on the abnormal case attribute to obtain the second operation information.

[0078] Specifically, the aforementioned pre-defined exception case library can be obtained by constructing a knowledge graph based on graph database technology. The pre-defined exception scenarios are represented as nodes and edges in the knowledge graph. Each pre-defined exception scenario corresponds to a node named Code, and the node's attributes include a node identifier Code. id Scene Title Code title Scenario Description Code desc The scenario includes the scene, the action taken, the expected result, the relevant fault type (F), and the scene keyword (Kw). The fault type (F) and keyword (Kw) can be used as edges (E) to connect other nodes, establishing relationships between scenes. The nodes and edges in the pre-defined exception case library can be represented as: G = (Set, E), where Set represents the set of nodes and E represents the set of associated edges.

[0079] The terminal can calculate scene similarity, enabling matching of simulated abnormal scenarios with preset abnormal scenarios. Specifically, when an emergency scenario occurs during the simulation, such as a simulated abnormal scenario, the terminal extracts the scene feature vector S=[s1,s2,…,s…]. i ], where s i This represents the i-th feature of the scenario (such as fault type, environmental conditions, etc.); the terminal can calculate the scenario feature vector and the case feature vector c of the case node in the knowledge base. j =[c j1 ,c j2 ,…,c jn The similarity sim(s,cj) between ] is expressed using the cosine similarity formula as follows:

[0080] .

[0081] The terminal can sort nodes based on their relevance to the scene. For example, the terminal can sort nodes by similarity and select the top m cases with the highest similarity as recommended cases, i.e., outputting cases with the aforementioned abnormal case attributes. The sorting algorithm can be quicksort, with a time complexity of O(nlogn); the terminal can also combine the historical usage frequency f of the node's scene. j and the number of times it was successfully resolved (s) j The comprehensive relevance score of the computing nodes j =αsim(s,c j )+βf j / f max +γs j / s max Where α=0.6, β=0.2, γ=0.2 are weighting coefficients, and f max and s max These represent the maximum usage frequency and maximum number of successful resolutions for all cases in the preset abnormal case library. The terminal can perform a final ranking of cases based on a comprehensive relevance score, ensuring that the pushed cases best meet the needs of the current exercise scenario. This is done in real-time. For example, during the simulation, the terminal monitors scenario changes in real time and triggers a case recommendation process every T seconds (e.g., T=5 seconds), pushing the ranked recommended cases to the aforementioned objects in a list format. Each case displays key information such as abnormal case attributes, including scenario title, brief scenario description, and related fault types. Simultaneously, it obtains information on the objects' usage of the recommended abnormal case attributes, such as viewing time and whether the case was adopted, to update the historical usage frequency and number of successful resolutions of the cases, thereby optimizing the subsequent case recommendation effect.

[0082] Through this embodiment, the terminal can obtain the second operation information of the object input under the condition of recommending abnormal case attributes, so that the simulated abnormal scenario is more in line with the handling process when an abnormality occurs in the real world, thereby improving the realism of the simulated abnormal scenario.

[0083] In one embodiment, before obtaining the second operation information corresponding to the aforementioned object, the method further includes: obtaining the positions and speeds of each heavy-haul train in the heavy-haul train simulation platform; for each heavy-haul train in the heavy-haul train simulation platform, determining the travel distance threshold and interval distance threshold of the heavy-haul train based on the position and speed of the preceding heavy-haul train; if the travel distance of the heavy-haul train is greater than or equal to the travel distance threshold, or the train interval between the heavy-haul train and the preceding heavy-haul train is less than or equal to the interval distance threshold, then outputting distance anomaly information; and constructing the simulated anomaly scenario based on the distance anomaly information.

[0084] In this embodiment, the simulated abnormal scenario can include a scenario of simulated distance abnormality corresponding to a heavy-haul train. The distance abnormality includes abnormal travel distance of the heavy-haul train and abnormal interval distance between two heavy-haul trains. The heavy-haul train simulation platform can include multiple heavy-haul trains. The terminal can acquire the position and speed of each heavy-haul train in the simulation platform. For each heavy-haul train in the simulation platform, the terminal can determine the travel distance threshold and interval distance threshold of the heavy-haul train based on the position and speed of the preceding heavy-haul train. The travel distance and interval distance of the heavy-haul train are then compared with the corresponding thresholds. If the travel distance of the heavy-haul train is greater than or equal to the travel distance threshold, or the train interval between the heavy-haul train and the preceding heavy-haul train is less than or equal to the interval distance threshold, the terminal can output distance abnormality information and construct the simulated abnormal scenario based on this information.

[0085] Specifically, the heavy-haul train simulation platform may also include a train coordination and linkage unit at the dispatching station. This unit comprises a radio block center simulation subunit and a virtual role behavior modeling subunit. The terminal calculates the clearance for heavy-haul trains and provides conflict warnings under moving block conditions through the radio block center simulation subunit. Specifically, the terminal acquires the position and speed of the heavy-haul train through the radio block center simulation subunit. For example, the terminal can acquire the real-time position x of the heavy-haul train. i (t) and velocity v i (t) information, where i represents the i-th train and t represents time; position information is obtained through track circuits or a train positioning system, and speed information is obtained through a train speed sensor; the update frequency of position and speed information is f. u Hz, f u It can be 5Hz. The terminal then performs moving block safety interval calculations. For example, the terminal calculates the safety interval distance D under moving block based on train speed and braking performance. safe (t), the safe separation distance includes the train's braking distance and safety margin, which can be specifically expressed as: D safe (t)=(v i (t) 2 / 2a max )+d safe , where: a max This is the train's maximum braking deceleration; for heavy-haul trains, we can take 'a'. max =0.5 m / s². d safe This is a safety margin, with a value of d. safe =50 meters.

[0086] The terminal can perform train permission calculations. For example, the terminal calculates the train's movement authorization limit, i.e., the maximum distance L that the train is allowed to travel. MA (t), the movement authorization limit is dynamically adjusted according to the position and speed of the heavy-haul train ahead, which can be specifically expressed as: L MA (t)=x i+1 (t)-x i (t)-D safe (t), where x i+1 (t) is the position of the (i+1)th train ahead. MA If (t)≤0, it means the train is not allowed to move forward. The terminal can also perform conflict warning condition detection to determine whether to generate a simulated abnormal scenario. For example, when the terminal detects that the actual distance between trains is less than the safe interval, a conflict warning is triggered. The specific condition is: x i+1 (t)-x i (t)<D safe (t), when the above conditions are met, the wireless block center simulation subunit sends a conflict warning signal to the relevant heavy-load trains to construct a simulated abnormal scenario, and at the same time feeds back the warning information to the train coordination unit of the dispatching station.

[0087] When generating early warning information, the conflict warning information includes the warning level, the reason for the warning, and recommended measures. The warning level is divided into three levels: low warning (distance greater than 0.8D). safe Intermediate warning (distance within 0.5D) safe up to 0.8D safe (between) and advanced warning (distance less than 0.5D) safe The warning information is determined by the following formula: Warning level = {(low-level warning, if D...}} safe (t)≤x i+1 (t)-x i (t) < 0.8D safe (Intermediate warning, if 0.5D) safe ≤x i+1 (t)-x i (t) < 0.8D safe (t)); (Advanced warning, if x i+1 (t)-x i (t) < 0.5D safe (t))}. The aforementioned warning information is pushed to the entities to be monitored, such as train drivers and dispatchers, in text and audio format, to help them take timely measures to avoid conflicts and simulate abnormal scenarios.

[0088] Through this embodiment, the terminal can accurately calculate the clearance of heavy-haul trains under moving block signaling and issue timely warnings when conflict risks occur, thereby improving the realism of the heavy-haul train simulation platform. Furthermore, the calculation accuracy of the wireless block center simulation subunit meets relevant railway industry standards, and the conflict warning response time does not exceed 0.5 milliseconds.

[0089] In one embodiment, the method further includes: inputting the historical train status and historical route information of the heavy-haul train in the heavy-haul train simulation platform into an instruction generation model; the instruction generation model is used to determine the various predicted operation information of the heavy-haul train at each time based on the historical train status and historical route information; determine the predicted speed and predicted position of the heavy-haul train based on the predicted operation information at the current time; obtain the comparison result of the current speed and current position of the heavy-haul train with the predicted speed and predicted position; adjust the predicted operation information at the current time using a proportional-integral-differential algorithm and the comparison result of the operating state to obtain the adjusted predicted operation information; and obtain the first operation information and / or the second operation information input by the object based on the adjusted predicted operation information.

[0090] In this embodiment, to improve the realism of the heavy-haul train simulation platform, the terminal can be configured with virtual characters to provide corresponding auxiliary information and detect the first and second operation information of the aforementioned object on the heavy-haul train under the condition of the auxiliary information. The terminal can pre-build an instruction generation model. This instruction generation model can be a neural network model that predicts the state of the heavy-haul train and outputs corresponding predicted operation information.

[0091] In some embodiments, the method further includes: acquiring historical operation information for the heavy-haul train; extracting features from the historical operation information to obtain various operation information features; and determining the instruction generation model based on the Markov chain and the various operation information features.

[0092] In this embodiment, the terminal can generate an instruction model by combining historical operation information of heavy-haul trains. For example, the terminal extracts features from the aforementioned historical operation information to obtain features of each operation information, and determines the instruction generation model based on the Markov chain and each of the aforementioned operation information features.

[0093] The terminal can input the historical train status and historical route information of the aforementioned heavy-haul trains from the heavy-haul train simulation platform into the instruction generation model. Based on this historical train status and route information, the instruction generation model can predict the operational information of the heavy-haul trains, determining the predicted operational information for each heavy-haul train at different times. Therefore, the terminal can determine the predicted speed and predicted position of the heavy-haul train based on the predicted operational information at the current time.

[0094] The terminal can also compare the current speed and position of the aforementioned heavy-haul train with the predicted speed and position. Using a proportional-integral-differential algorithm and the comparison results, it adjusts the predicted operation information at the current moment to obtain adjusted predicted operation information. Thus, the terminal can obtain the first operation information and / or the second operation information input by the aforementioned object based on the adjusted predicted operation information, and determine whether the object should perform the corresponding operation according to the predicted operation information.

[0095] Specifically, the generation of the aforementioned predictive operation information can be executed by the virtual character behavior modeling unit in the heavy-haul train simulation platform. The virtual character behavior modeling unit can construct a virtual assistant that simulates the operating habits of heavy-haul trains based on historical data, outputting predictive operation information and reducing reliance on manual intervention. The terminal can collect and preprocess historical data. For example, the terminal collects historical operation data from heavy-haul train drivers, including information on train speed v(t), traction force F(t), braking command B(t), track gradient i(t), and train position x(t), with a data sampling frequency of f. s Hz can be f s =10Hz; The terminal can clean and normalize the collected data, mapping all data to the [0,1] interval to eliminate differences between different units. The normalization formula is: y norm =(yy min ) / (y max -y min Where y is the original data, y min and y max These are the historical minimum and maximum values ​​of the data, y. norm This is the normalized data. The terminal can also extract driving habit features. For example, the terminal can extract key features of the driver's driving habits, including the rate of change of acceleration a(t), braking deceleration b(t), and traction adjustment frequency f. F Braking command advance t B The calculation formulas are as follows: a(t) = d / dt(dv / dt), b(t) = dv / dt (when the train is braking), f F =N F / T,t B =t B，actual -t B，ideal , where N F The number of traction force adjustments is t, where T is the statistical time window. B，actual t is the actual time when the braking command is issued. B，ideal This refers to the time when a braking command is issued under ideal conditions.

[0096] The terminal can perform maneuvering behavior modeling based on Markov chains. For example, the terminal represents the driver's maneuvering behavior as a Markov chain, with the state space including different operating states of the train, such as the acceleration state S. a Uniform speed state S c Braking state S b The state transition probability matrix P is defined as follows:

[0097] ;

[0098] Where, p ij The probability of transitioning from state i to state j can be obtained through historical data statistics; the terminal can predict the next state S(t+j) based on the current state S(t) and the state transition probability matrix. t), and generate corresponding manipulation instructions. The state transition prediction formula is: S(t+ t)=argmax s属于状态空间 P(S(t)→s). The terminal can also generate corresponding traction or braking commands based on the predicted train status and current track conditions (such as gradient and position). Traction command F cmd and braking command B cmd The calculation formula is: F cmd =F base +k F a target B cmd =B base +k B b target Among them, F base and B base These are the base values ​​for traction force and braking command, k, respectively. F and k B a is the adjustment coefficient. target and b target These are the target acceleration rate of change and the target braking deceleration, respectively. These parameters are obtained by fitting historical data; for example, k can be taken as... F =0.8 and k B =0.6.

[0099] During the simulated operation of heavy-haul trains, the terminal can adjust and optimize the predicted operation information in real time. For example, during the simulation, the terminal monitors the deviation between the actual response of the train and the predicted operation information generated by the virtual assistant in real time, and adjusts the command parameters through a feedback control mechanism. The deviation is defined as: e(t) = y actual (t)-y cmd (t), where y actual (t) represents the actual operating parameters of the train, such as speed and position; y cmd(t) represents the desired command value; the terminal can use a proportional-integral-derivative (PID) controller to adjust the deviation and update the control command parameters. The PID control formula is: Wherein: K p =0.5, K i =0.1, K d =0.2 represents the proportional, integral, and derivative control coefficients, respectively.

[0100] Through the above embodiments, the terminal accurately simulates the operating habits of heavy-haul train drivers based on historical data, generates reasonable predictive operation information, reduces reliance on human assistance during the simulation process, and improves the efficiency of the simulation operation.

[0101] In one exemplary embodiment, such as Figure 2 As shown, Figure 2 This is a schematic diagram of the heavy-haul train simulation platform in one embodiment. In this embodiment, it includes a train dynamics and control simulation unit, an operation behavior analysis and scenario generation unit, and a dispatching station train coordination and linkage unit. The train dynamics and control simulation unit provides basic train dynamics data and fault scenario support to the operation behavior analysis and scenario generation unit. The operation behavior analysis and scenario generation unit performs operation analysis and scenario generation based on the provided data, and feeds back the analysis results and generated scenarios to the dispatching station train coordination and linkage unit to achieve collaborative work between the units and jointly complete the simulation of the dispatching and receiving of trains by staff at the heavy-haul transport station. The train dynamics and control simulation unit includes a train dynamics simulation subunit, a multi-car simulation control unit, and a braking strategy simulation subunit. The operation behavior analysis and scenario generation unit includes an operation behavior semantic analysis subunit, a dynamic scenario generation subunit, and a knowledge graph-assisted decision-making subunit. The dispatching station train coordination and linkage unit includes a radio block center simulation subunit and a virtual role behavior modeling subunit.

[0102] Through the above embodiments, the terminal constructs a heavy-haul train simulation platform, providing various simulated scenarios for the objects to perform simulated operations. This enables more accurate detection of whether the object's operational behavior is abnormal, improving the accuracy of operational behavior detection. Furthermore, the terminal, through a dynamics simulation subunit, constructs simulations based on factors such as the number of train formations, vehicle type, train speed, and track gradient. This accurately simulates the longitudinal dynamic behavior of the train under different operating conditions. The simulation accuracy of air brake propagation delay reaches within 0.5 seconds, and the simulation accuracy of the inter-carriage impulse transmission effect meets the actual operating condition error range of ±10%. Moreover, through the natural language processing technology of the operational behavior semantic analysis subunit, a self-attention mechanism architecture is used to parse the voice commands of the aforementioned objects in real time. Operational timing logic detection covers key operational nodes throughout the entire train arrival and departure process. Fault level adjustment is based on a comprehensive judgment of the object's operational performance in historical training records, knowledge mastery, and the complexity of the current simulation scenario. The matching degree of the pushed cases reaches over 70%.

[0103] In addition, the terminal constructs a longitudinal dynamics model of a 10,000-ton combined train to meticulously simulate unique working conditions such as braking control and longitudinal impulse mitigation of long train formations. This provides the object under test with a simulation scenario that is close to reality, making the simulation content more focused on the key operations of heavy-haul trains and improving the object's ability to cope with complex working conditions.

[0104] The terminal utilizes natural language processing technology to parse voice commands and combines it with operational timing logic detection. This not only identifies explicit errors but also uncovers implicit errors such as "departing without confirming train integrity," helping users standardize their operational procedures. Simultaneously, a dynamic scenario generation engine based on reinforcement learning breaks away from traditional fixed fault patterns, adjusting fault scenarios in real time according to the user's capabilities, making simulations more challenging and adaptable. The knowledge graph-assisted decision-making function can quickly push relevant cases, shortening decision-making time and improving emergency response efficiency. Furthermore, through a three-level linkage design for trains at dispatching stations, and integrating a wireless block signaling center simulation unit, automatic calculation and conflict warning are achieved, accurately controlling train operation permits and ensuring operational safety. Virtual character behavior modeling of the driver's virtual assistant, based on historical data to simulate operating habits, reduces reliance on manual intervention. This makes the collaborative logic of station staff, dispatchers, and drivers more aligned with real-world scenarios, fully integrating into the unique dispatching and command processes of heavy-haul transportation. This lays a solid foundation for the arrival and departure work of heavy-haul transport stations and has immeasurable value in ensuring the safe, stable, and efficient operation of railway heavy-haul transportation.

[0105] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0106] Based on the same inventive concept, this application also provides an operation behavior detection device based on a heavy-haul train simulation platform for implementing the above-mentioned operation behavior detection method based on a heavy-haul train simulation platform. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations of one or more operation behavior detection device embodiments based on a heavy-haul train simulation platform provided below can be found in the limitations of the operation behavior detection method based on a heavy-haul train simulation platform described above, and will not be repeated here.

[0107] In one exemplary embodiment, such as Figure 3 As shown, an operational behavior detection device based on a heavy-haul train simulation platform is provided, comprising: a first acquisition module 500, a first detection module 502, a second acquisition module 504, a second detection module 506, and a third detection module 508, wherein:

[0108] The first acquisition module 500 is used to acquire the first operation information corresponding to the object to be detected; the first operation information represents the operation information of the object on the heavy-haul train in the simulated transportation scenario of the heavy-haul train simulation platform; the heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-vehicle simulation control unit and the braking strategy simulation unit.

[0109] The first detection module 502 is used to perform operation logic detection based on the first operation information mentioned above, and obtain the operation logic detection result.

[0110] The second acquisition module 504 is used to acquire the second operation information corresponding to the above-mentioned object; the second operation information represents the operation information of the above-mentioned object on the above-mentioned heavy-haul train in the simulated abnormal scenario of the above-mentioned heavy-haul train simulation platform; the simulated abnormal scenario is matched with the abnormal recovery capability of the above-mentioned object.

[0111] The second detection module 506 is used to perform abnormal response detection based on the above-mentioned second operation information and obtain the abnormal response detection result.

[0112] The third detection module 508 is used to obtain the operation behavior detection result corresponding to the above object based on the above logical detection result and the above abnormal response detection result.

[0113] In one embodiment, the apparatus further includes: a construction module, configured to construct a heavy-haul train dynamics simulation unit based on the air braking propagation delay simulation information and the inter-car impulse transmission simulation information of the heavy-haul train; construct a multi-car simulation control unit based on the wireless communication interruption simulation information, the traction force difference simulation information between cars, and the positioning simulation information of the heavy-haul train; construct a braking strategy simulation unit based on the terrain data simulation information of the line where the heavy-haul train is located, the dynamic change simulation information, resistance simulation information, energy consumption simulation information, and the target braking curve of the heavy-haul train; and construct a heavy-haul train simulation platform based on the heavy-haul train dynamics simulation unit, the multi-car simulation control unit, and the braking strategy simulation unit.

[0114] In one embodiment, the first detection module 502 is used to input the first operation information into a logic recognition model; the logic recognition model is used to identify each semantic vector corresponding to the first operation information based on an attention mechanism, determine each operation rule corresponding to the first operation information based on the matching of each semantic vector with a preset logic rule library, determine operation logic anomaly information in the first operation information based on the matching of each operation rule with a preset logic anomaly rule library, and obtain operation logic detection results based on the operation logic anomaly information.

[0115] In one embodiment, the apparatus further includes: an adjustment module, configured to acquire historical second operation information of the object in response to a historical simulated abnormal scenario; input the historical simulated abnormal scenario and the historical second operation information into a trained deep Q-network model; the deep Q-network model is configured to adjust the anomaly level of the historical simulated abnormal scenario based on the historical simulated abnormal scenario and the historical second operation information to obtain an adjusted simulated abnormal scenario; the anomaly level characterizes the severity of the anomaly in the simulated abnormal scenario; the adjusted anomaly level of the simulated abnormal scenario matches the anomaly recovery capability of the object; and acquire second operation information of the object in response to the adjusted simulated abnormal scenario.

[0116] In one embodiment, the second detection module 506 is configured to: determine the post-operation simulated abnormal scenario corresponding to the simulated abnormal scenario based on the second operation information; determine the change in operation accuracy, change in response time, and change in the number of correct operands corresponding to the object based on the scenario comparison result between the simulated abnormal scenario and the post-operation simulated abnormal scenario; the change in the number of correct operands represents the change in the number of operations that are conducive to the recovery of the simulated abnormal scenario; and obtain the abnormal response detection result based on the weighted sum of the change in operation accuracy, change in response time, and change in the number of correct operands.

[0117] In one embodiment, the second acquisition module 504 is used to acquire the scene feature vector of the simulated abnormal scenario; determine the abnormal case attribute corresponding to the simulated abnormal scenario based on the similarity between the scene feature vector and the preset scene feature vector of each preset abnormal scenario in the preset abnormal case library; and acquire the operation information input by the object based on the abnormal case attribute to obtain the second operation information.

[0118] In one embodiment, the apparatus further includes: an anomaly simulation module, configured to acquire the positions and speeds of each heavy-haul train in the heavy-haul train simulation platform; for each heavy-haul train in the heavy-haul train simulation platform, determine a travel distance threshold and an interval distance threshold for the heavy-haul train based on the position and speed of the preceding heavy-haul train; if the travel distance of the heavy-haul train is greater than or equal to the travel distance threshold, or the train interval between the heavy-haul train and the preceding heavy-haul train is less than or equal to the interval distance threshold, then output distance anomaly information; and construct the simulated anomaly scenario based on the distance anomaly information.

[0119] In one embodiment, the apparatus further includes: a prediction module, configured to input the historical train status and historical route information of the heavy-haul train in the heavy-haul train simulation platform into an instruction generation model; the instruction generation model, configured to determine the various prediction operation information of the heavy-haul train at each time point based on the historical train status and historical route information; determine the prediction speed and prediction position of the heavy-haul train based on the prediction operation information at the current time point; obtain a comparison result of the current speed and current position of the heavy-haul train with the predicted speed and prediction position; adjust the prediction operation information at the current time point using a proportional-integral-differential algorithm and the comparison result of the operating states to obtain adjusted prediction operation information; and obtain the first operation information and / or the second operation information input by the object based on the adjusted prediction operation information.

[0120] In one embodiment, the above-mentioned apparatus further includes: a generation module, configured to acquire historical operation information for the heavy-haul train; extract features from the historical operation information to obtain various operation information features; and determine the instruction generation model based on the Markov chain and the various operation information features.

[0121] Each module in the aforementioned operational behavior detection device based on a heavy-haul train simulation platform can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the operations corresponding to each module.

[0122] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When executed by the processor, the computer program implements an operational behavior detection method based on a heavy-load train simulation platform. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0123] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0124] In one exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described operation behavior detection method based on a heavy-load train simulation platform.

[0125] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described method for detecting operational behavior based on a heavy-load train simulation platform.

[0126] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method for detecting operational behavior based on a heavy-load train simulation platform.

[0127] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0128] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0129] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0130] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for detecting operational behavior based on a heavy-haul train simulation platform, characterized in that, The method includes: Obtain the first operation information corresponding to the object to be detected; the first operation information represents the operation information of the object on the heavy-haul train in the simulated transportation scenario of the heavy-haul train simulation platform; the heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-car simulation control unit and the braking strategy simulation unit; Based on the first operation information, an operation logic detection is performed to obtain the operation logic detection result; Obtain the second operation information corresponding to the object; the second operation information represents the operation information of the object on the heavy-haul train in the simulated abnormal scenario of the heavy-haul train simulation platform; the simulated abnormal scenario is matched with the abnormal recovery capability of the object; Anomaly response detection is performed based on the second operation information to obtain the anomaly response detection result; Based on the logic detection results and the abnormal response detection results, the operation behavior detection results corresponding to the object are obtained.

2. The method according to claim 1, characterized in that, The method further includes: Based on the simulation information of air brake propagation delay and the simulation information of impulse transmission between carriages of the heavy-haul train, a dynamic simulation unit for the heavy-haul train is constructed. Based on the wireless communication interruption simulation information, the traction force difference simulation information between carriages, and the positioning simulation information of the heavy-haul train, a multi-car simulation control unit is constructed. Based on the terrain data simulation information of the line where the heavy-haul train is located, the dynamic change simulation information of the heavy-haul train, the resistance simulation information, the energy consumption simulation information, and the target braking curve, a braking strategy simulation unit is constructed. A heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-vehicle simulation control unit, and the braking strategy simulation unit.

3. The method according to claim 1, characterized in that, The step of performing operation logic detection based on the first operation information to obtain the operation logic detection result includes: The first operation information is input into a logic recognition model; the logic recognition model is used to identify each semantic vector corresponding to the first operation information based on an attention mechanism, determine each operation rule corresponding to the first operation information based on the matching of each semantic vector with a preset logic rule library, and determine operation logic anomaly information in the first operation information based on the matching of each operation rule with a preset logic anomaly rule library. Based on the abnormal operation logic information, the operation logic detection result is obtained.

4. The method according to claim 1, characterized in that, Before performing anomaly response detection based on the second operation information and obtaining the anomaly response detection result, the following steps are included: Obtain the historical second operation information of the object for the historical simulated abnormal scenario; The historical simulated abnormal scenarios and the historical second operation information are input into a trained deep Q-network model; the deep Q-network model is used to adjust the anomaly level of the historical simulated abnormal scenarios based on the historical simulated abnormal scenarios and the historical second operation information to obtain an adjusted simulated abnormal scenario; the anomaly level represents the severity of the anomaly of the simulated abnormal scenario; the adjusted anomaly level of the simulated abnormal scenario matches the anomaly recovery capability of the object; Obtain the second operation information of the object corresponding to the adjusted simulated abnormal scenario.

5. The method according to claim 1, characterized in that, The step of performing abnormal response detection based on the second operation information to obtain abnormal response detection results includes: Based on the second operation information, determine the operation corresponding to the simulated abnormal scenario and then simulate the abnormal scenario; Based on the comparison results between the simulated abnormal scenario and the simulated abnormal scenario after the operation, the changes in operation accuracy, response time, and number of correct operands corresponding to the object are determined; the change in the number of correct operands represents the change in the number of operations that are conducive to the recovery of the simulated abnormal scenario. The abnormal response detection result is obtained by weighting the change in operation accuracy, the change in response time, and the change in correct operation.

6. The method according to claim 1, characterized in that, The step of obtaining the second operation information corresponding to the object includes: Obtain the scene feature vector of the simulated abnormal scenario; Based on the similarity between the scene feature vector and the preset scene feature vector of each preset abnormal scene in the preset abnormal case library, the abnormal case attribute corresponding to the simulated abnormal scene is determined. Obtain the operation information input by the object based on the abnormal case attributes to obtain the second operation information.

7. The method according to claim 1, characterized in that, Before obtaining the second operation information corresponding to the object, the method further includes: Obtain the position and speed of each heavy-haul train in the heavy-haul train simulation platform; For each heavy-haul train in the heavy-haul train simulation platform, the travel distance threshold and interval distance threshold of the heavy-haul train are determined based on the position and speed of the preceding heavy-haul train. If the travel distance of the heavy-haul train is greater than or equal to the travel distance threshold, or the train interval between the heavy-haul train and the preceding heavy-haul train is less than or equal to the interval distance threshold, then output distance anomaly information; Based on the distance anomaly information, the simulated anomaly scenario is constructed.

8. The method according to any one of claims 1 to 7, characterized in that, The method further includes: The historical train status and historical route information of the heavy-haul train in the heavy-haul train simulation platform are input into the instruction generation model; the instruction generation model is used to determine the various predicted operation information of the heavy-haul train at each time based on the historical train status and historical route information. Based on the prediction operation information at the current moment, determine the predicted speed and predicted position of the heavy-haul train; The current speed and current position of the heavy-haul train are compared with the predicted speed and predicted position to obtain the running status results. By comparing the proportional-integral-differential algorithm with the running state, the predicted operation information at the current moment is adjusted to obtain the adjusted predicted operation information; Obtain the first operation information and / or the second operation information input by the object based on the adjusted predicted operation information.

9. The method according to claim 8, characterized in that, The method further includes: Obtain historical operation information for the heavy-haul train; Feature extraction is performed on the historical operation information to obtain the features of each operation information; The instruction generation model is determined based on the Markov chain and the characteristics of each operation information.

10. An operational behavior detection device based on a heavy-haul train simulation platform, characterized in that, The device includes: The first acquisition module is used to acquire the first operation information corresponding to the object to be detected; the first operation information represents the operation information of the object on the heavy-haul train in the simulated transportation scenario of the heavy-haul train simulation platform; the heavy-haul train simulation platform is constructed based on the heavy-haul train dynamics simulation unit, the multi-car simulation control unit and the braking strategy simulation unit; The first detection module is used to perform operation logic detection based on the first operation information and obtain the operation logic detection result. The second acquisition module is used to acquire second operation information corresponding to the object; the second operation information represents the operation information of the object on the heavy-haul train in the simulated abnormal scenario of the heavy-haul train simulation platform; the simulated abnormal scenario is matched with the abnormal recovery capability of the object. The second detection module is used to perform abnormal response detection based on the second operation information and obtain the abnormal response detection result. The third detection module is used to obtain the operation behavior detection result corresponding to the object based on the logical detection result and the abnormal response detection result.

11. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

Images