System and method for testing prompt function of visual heavy-load locomotive driving guidance device

Abstract

The invention discloses a system and a method for testing a prompt function of a visual heavy-load locomotive driving guidance device. The method comprises the following steps: S1, acquiring multi-dimensional operation data of a heavy-load locomotive and constructing a data set; s2, processing the data set, and generating an initial topological graph structure according to data relevance; s3, performing feature extraction on the initial topological graph structure by using a graph convolutional network to generate a high-dimensional feature vector; s4, constructing a causal relationship model through a causal inference method, and identifying key causal nodes and causal paths; s5, performing dynamic optimization on the initial topological graph structure by using an eagle group optimization algorithm; s6, dynamically verifying the optimized prompt function logic, and generating a prompt function test result and test feedback data; and S7, adjusting prompt function logic design and parameters according to test feedback data, and performing multi-round iterative optimization on the prompt function performance. According to the method, the graph convolutional network, causal inference and an eagle group optimization algorithm are combined to realize dynamic optimization and test of the prompt function of the heavy-load locomotive.

Description

Technical Field

[0001] The invention relates to the technical field of heavy-load locomotive driving guidance, and in particular to a testing system and method for a prompting function of a visual heavy-load locomotive driving guidance device. Background Art

[0002] As a key equipment in modern railway transportation, the driving safety and operating efficiency of heavy-load locomotives are of great significance to the smooth completion of transportation tasks. During the operation of heavy-load locomotives, the prompt function of the guidance device undertakes the important tasks of providing real-time driving information, optimizing operational decisions, improving operating efficiency and ensuring driving safety. However, the existing driving guidance device prompt function technology still has many limitations in terms of multi-dimensional operation data collection and analysis, adaptability to complex scenarios, optimization of prompt logic and test efficiency, and cannot fully meet the actual needs of modern railway transportation.

[0003] First, in terms of the collection and analysis of multi-dimensional operation data, existing technologies mainly rely on single sensors or simple signal fusion methods, resulting in fewer dimensions of operation data and weak correlation. The current prompt function design is usually based on fixed operation rules. This method cannot fully utilize the multi-dimensional dynamic data in heavy-load locomotive operation, such as speed, station information, line signals, cylinder pressure, speed limit, over-phase, weather information, etc. Due to insufficient data collection and limited analysis capabilities, the prompt function is prone to information omission or misjudgment when facing complex scenarios (such as steep slopes, sharp turns or bad weather), thereby affecting driving safety and operating efficiency.

[0004] Secondly, in terms of data processing and model building, existing technologies usually use traditional data cleaning and simple feature extraction methods, which have weak data correlation analysis capabilities and cannot accurately capture the complex relationships between multidimensional data. For example, the nonlinear correlation between running speed and slope is not fully explored in the prompt logic optimization, resulting in the prompt content being difficult to fully adapt to the actual scene. Especially in highly dynamic scenes, the feature extraction efficiency of existing methods is low, and the accuracy of the processing results is insufficient, which limits the real-time response capability of the prompt function.

[0005] In addition, in the process of prompt logic optimization, the existing technology mainly relies on rule-based decision-making models, which are difficult to dynamically adapt to complex operating scenarios. This fixed logic-based prompt design shows obvious lag and limitations when dealing with dynamically changing operating conditions, especially when facing continuous and changing slopes or sudden weather conditions. The prompt function easily deviates from actual needs and cannot accurately guide the driver to adjust the operating strategy. This not only reduces the safety of heavy-load locomotive operation, but may also lead to excessive energy consumption and increase transportation costs.

[0006] In the prompt function testing and verification stage, traditional technologies usually use artificial scenario simulation and static analysis methods, which are inefficient and inaccurate. Artificial simulation is difficult to cover all the complex operating scenarios of heavy-load locomotives, and the representativeness and universality of the test results are limited. In addition, due to the lack of a systematic feedback mechanism and intelligent optimization process, it is difficult for existing technologies to achieve automation and closed-loop optimization of prompt function testing, resulting in a long test cycle and the test results are difficult to directly use for the improvement and iteration of prompt logic.

[0007] Another important defect in the prior art is the lack of intelligent means of visualizing the performance of the prompt function. In the development and optimization process of the prompt function, the test results and optimization process are usually presented in the form of static data reports, which not only increases the difficulty of result analysis, but also limits the developer's intuitive understanding of the dynamic changes of the prompt logic. When dealing with complex operating scenarios, the lack of clear means of dynamic display of the prompt function further increases the difficulty of optimizing the prompt logic.

[0008] Therefore, how to provide a testing system and method for the visual prompting function of a heavy-load locomotive driving guidance device is an urgent problem to be solved by those skilled in the art. Summary of the invention

[0009] One purpose of the present invention is to propose a testing system and method for the prompt function of a visual heavy-load locomotive driving guidance device. The graph convolutional network, causal inference and eagle swarm optimization algorithm of the present invention realize the optimization and efficient testing of the prompt function of the heavy-load locomotive driving guidance device through multi-dimensional data analysis and dynamic verification technology. Specifically, the present invention can accurately identify key causal nodes and paths in complex operating scenarios, and dynamically adjust the prompt logic, significantly improving the accuracy, real-time and adaptability of the prompt function. In addition, combined with visualization technology to display topological changes and optimization processes in real time, a closed-loop feedback mechanism is formed, which further improves the test efficiency and the optimization effect of the prompt function, and is suitable for heavy-load locomotive operation guidance in a variety of complex scenarios.

[0010] The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to the embodiment of the present invention comprises the following steps:

[0011] S1. Collect multi-dimensional operation data of heavy-load locomotives through sensors and construct data sets;

[0012] S2, filtering noise, removing outliers and normalizing the data set, and generating an initial topological graph structure based on data correlation;

[0013] S3. Use graph convolutional networks to extract features from the initial topological graph structure, capture the multi-scale relationship between nodes and neighborhoods through multi-layer convolution, and generate high-dimensional feature vectors;

[0014] S4. Based on the high-dimensional feature vector, a causal relationship model is constructed through causal inference method to analyze the causal relationship between nodes and identify the key causal nodes and causal paths that have an important impact on the prompt function;

[0015] S5. Based on the key causal nodes and causal paths, the Eagle Swarm Optimization algorithm is used to dynamically optimize the initial topology graph structure. By combining global search with local optimization, node priorities and edge weights are adjusted to optimize the prompt function logic.

[0016] S6. Dynamically verify the optimized prompt function logic, generate prompt function test results and test feedback data by visually displaying the topological changes of the dynamic graph structure;

[0017] S7. Adjust the prompt function logic design and parameters according to the test feedback data, and optimize the prompt function performance through multiple rounds of iterations using the dynamically optimized topology structure.

[0018] Optionally, the multi-dimensional operating data specifically includes the heavy-load locomotive running speed, station information, line signal, cylinder pressure, speed limit and over-phase.

[0019] Optionally, the S2 specifically includes:

[0020] S21. Perform preliminary cleaning on the data set, identify noise data based on statistical distribution characteristics, and remove data that does not conform to normal operating conditions through filtering technology;

[0021] S22, remove outliers from the cleaned data set, identify extreme outliers by calculating the deviation and dispersion of the data distribution, and use interpolation to fill in missing or abnormal data;

[0022] S23, normalizing the multidimensional features in the data set, mapping the feature values ​​to a unified range, and eliminating the influence between different feature dimensions;

[0023] S24. Based on the normalized data set, the linear correlation between the running characteristics is calculated using the Pearson correlation coefficient:

[0024]

[0025] Among them, r ij represents the correlation coefficient between operation feature i and operation feature j, x ik represents the value of running feature i in the kth data, x jk represents the value of running feature j in the kth data, represents the mean of running feature i, represents the mean of running feature j, and n represents the total number of data samples;

[0026] S25. According to the results of the operation feature correlation analysis, define the relationship between nodes and edges, where nodes represent operation features and edges represent the correlation between operation features:

[0027]

[0028] Among them, w ij represents the weight between node i and node j, |r ij | represents the absolute value of the correlation coefficient between operation feature i and operation feature j, and τ represents the set correlation threshold;

[0029] S26. Based on the definitions of the operation feature nodes and edges and in combination with the correlation analysis results, an initial topological graph structure is generated, wherein the layout of the nodes and edges dynamically reflects the associations between the operation features.

[0030] Optionally, the S3 specifically includes:

[0031] S31. Input the generated initial topological graph structure into the graph convolutional network and initialize the node feature matrix:

[0032]

[0033] Among them, H (0) represents the node feature matrix, represents the initial feature vector of node i, and N represents the total number of nodes;

[0034] S32, normalize the adjacency matrix of the initial topology graph, add self-loops to enhance the connectivity of the graph, and calculate the normalized adjacency matrix:

[0035]

[0036] in, represents the normalized adjacency matrix, A represents the original adjacency matrix, I represents the identity matrix, represents the degree matrix of A+I;

[0037] S33. In each layer of graph convolution, the variable topology adjustment mechanism is used to perform convolution operations on node features:

[0038]

[0039] Among them, H (l+1) represents the feature matrix of the l+1th layer node, σ represents the activation function, N(i) represents the set of neighboring nodes of node i, Represents the weight value in the normalized adjacency matrix, W (l) represents the training weight matrix of layer 1, c ij represents the normalization factor, α represents the dynamic adjustment coefficient, Represents the feature vector of node j in layer 1;

[0040] S34, perform multi-scale fusion on the output features of multi-layer convolution, integrate features at different levels through splicing operations, and generate multi-scale features;

[0041] S35. Dynamically adjust the weight contribution of each layer of features according to the sparsity of multi-scale features, and introduce a sparsity-driven weighting mechanism:

[0042]

[0043] Among them, H (weighted) represents the feature representation after weighted fusion, L represents the total number of layers, exp represents the exponential function, β represents the sparse adjustment coefficient, and H (l) represents the node feature matrix of the first layer, ∥H (l) ∥1 represents the sparsity of the node feature matrix of the first layer, ∥H (m) ∥1 represents the sparsity of the node feature matrix of the mth layer;

[0044] S36. Perform linear transformation on the fused weighted features through a fully connected layer or a mapping matrix to generate a high-dimensional feature vector.

[0045] Optionally, the S4 specifically includes:

[0046] S41. Based on the high-dimensional feature vector, a causal relationship model G = (V, E, W) is constructed, where V is a node set, representing features related to the prompt function; E is an edge set, representing the causal relationship between features; W is an edge weight set, used to characterize the strength of the causal influence between features;

[0047] S42. Quantify the causal strength of each edge in the edge set E and calculate the mutual information entropy of the causal strength based on the Bayesian network inference:

[0048]

[0049] Among them, C pq represents the causal strength between node p and node q, P(x p ,x q ) represents the joint probability distribution of node p and node q, P(x p ) represents the marginal probability distribution of node p, P(x q ) represents the edge probability distribution of node q;

[0050] S43. In causal path optimization, the path set is dynamically optimized in combination with the inference results of the Bayesian network:

[0051]

[0052] Among them, F path represents the path optimization objective function, P represents the set of candidate causal paths, Ω pq represents the correlation weight between node p and node q, exp represents the exponential function, γ represents the degree attenuation coefficient, and d q represents the degree value of node q, Δ pq represents the path length between node p and node q, and λ represents the penalty factor of the path length;

[0053] S44. Normalize the overall weight of the causal path:

[0054]

[0055] Among them, W pq represents the normalized causal weight matrix, N(p) represents the set of neighboring nodes of node p;

[0056] S45. Prioritize the input node features of the prompt function according to the normalized causal weight matrix, and prioritize high-weight nodes into the optimization process:

[0057]

[0058] Among them, P p represents the priority of node p, Φ(x q ) represents the feature importance score of node q;

[0059] S46. Output the key causal nodes and causal paths that have an important impact on the prompt function.

[0060] Optionally, the S5 specifically includes:

[0061] S51, initializing the initial topological graph structure according to the key causal nodes and causal paths, generating a population P = {P (1) ,P (2) ,…,P (M)},in m represents the individual number, M represents the number of individuals in the population, V represents the node set, and E represents the edge set. represents the edge weight set, Represents node v in individual m p With node v q The edge weights between represents the node priority set, Represents node v in individual m p Priority;

[0062] S52. In the global search phase, the strong individuals of the eagle swarm optimization algorithm are introduced to define the global objective function:

[0063]

[0064] in, represents the global objective function, N(v p ) represents node v p The domain node set, C pq Represents node v p and node v q The causal strength between Represents node v p The degree value, Represents node v q The degree value, μ represents the degree attenuation factor;

[0065] S53. In the local optimization stage, the weak individuals of the Eagle Swarm Optimization algorithm are introduced to fine-tune the edge weights according to the global search results of the strong individuals:

[0066]

[0067] in, represents the updated edge weight, δ represents the edge weight update coefficient, tanh represents the hyperbolic tangent function, ζ represents the priority influencing factor, Represents node v in individual m q The priority of Represents node v in individual m r Priority;

[0068] S54, adaptively adjust the node priority and update the node priority according to the topology change:

[0069]

[0070] in, represents the updated node priority, exp represents the exponential function, η represents the priority adjustment coefficient, |N(v p )| represents node v p The number of domain nodes;

[0071] S55. Evaluate the fitness of individuals:

[0072]

[0073] in, represents the fitness function, θ1, θ2 and θ3 represent weight parameters;

[0074] S56. The eagle group is updated through multiple rounds of iterations. In each round of iteration, the individual with the highest fitness is selected as the new center, and the node priority and edge weight distribution are passed to the next generation of individuals. By repeating the global search and local optimization process, the optimized topological structure is finally obtained.

[0075] Optionally, the S6 specifically includes:

[0076] S61. Dynamically verify the optimized prompt function logic, set different operation scenarios, including normal operation, ramp, sharp curve and bad weather, and record the triggering time, triggering conditions and prompt content of the prompt function in each scenario;

[0077] S62. In each operation scenario, collect signal data and test data output by the prompt function, including the accuracy, real-time and adaptability of the prompt signal to the operation status;

[0078] S63. Dynamically analyze the test data, evaluate the performance of the prompt function in different operation scenarios by comparing the prompt signal with the actual operation status, and calibrate the error range and potential optimization space of the prompt signal;

[0079] S64. Use visualization technology to display the topological changes of dynamic graph structures, including node priority, edge weight adjustment, and dynamic changes of causal paths;

[0080] S65, generating prompt function test results, including the accuracy of trigger time, the adaptability of prompt content, and the real-time performance of prompt response;

[0081] S66. Output the test results and scenario analysis data as test feedback data, and record the visual log generated during the test process.

[0082] The testing system of the prompting function of the visual heavy-load locomotive driving guidance device according to the embodiment of the present invention comprises the following modules:

[0083] Multi-dimensional operation data acquisition module, used to collect multi-dimensional operation data of heavy-load locomotives;

[0084] A data preprocessing module is used to process multi-dimensional operation data and generate an initial topological graph structure;

[0085] The feature extraction module is used to extract features from the initial topological graph structure through a graph convolutional network to generate a high-dimensional feature vector;

[0086] The causal relationship analysis module is used to build a causal relationship model based on high-dimensional feature vectors and identify key causal nodes and causal paths;

[0087] The optimization calculation module is used to dynamically optimize the initial topology structure using the Eagle Swarm Optimization algorithm, adjust node priorities and edge weights, and optimize the prompt function logic;

[0088] Dynamic verification module, used to dynamically verify the optimized prompt function logic, collect test data and generate test results;

[0089] Visual display module, used to display the topological changes of dynamic graph structure and generate test feedback data of prompt function;

[0090] The feedback and storage module is used to store and record test feedback data and support multiple rounds of iterative optimization of the prompt function logic.

[0091] The beneficial effects of the present invention are:

[0092] First, in terms of the collection and processing of multi-dimensional operation data, the present invention realizes the comprehensive collection of multi-dimensional dynamic data such as speed, station information, line signal, cylinder pressure, speed limit and over-equalization, and combines noise filtering, outlier removal and normalization processing to generate the initial topological graph structure. This method not only improves the quality of the data, but also explores the deep correlation between various operation characteristics, providing strong support for subsequent feature extraction and model construction.

[0093] Secondly, the present invention uses a graph convolutional network to extract features from the initial topological graph structure, captures the multi-scale relationship between nodes and neighborhoods through multi-layer convolution operations, and combines multi-scale feature fusion and sparsity-driven weighting mechanisms to effectively solve the problem that traditional methods are insufficient in handling complex nonlinear associations. This feature extraction method significantly enhances the adaptability of the prompt logic to complex operating scenarios, while improving the global perception ability and local optimization effect of multi-dimensional dynamic data, making the prompt function more accurate and efficient.

[0094] In addition, the present invention constructs a causal relationship model based on the causal inference method, identifies key causal nodes and paths that have an important impact on the prompt logic, and optimizes the decision logic of the prompt function by dynamically adjusting the causal weight and priority allocation. Combined with the Eagle Swarm Optimization Algorithm, the combination of global search and local optimization further improves the dynamic adaptability of the prompt function, especially in complex scenarios such as steep slopes, sharp turns or severe weather conditions. The optimized prompt logic can adjust the node priority and edge weight in real time, providing prompt content that better meets actual needs.

[0095] Finally, in terms of the testing and verification of the prompt function, the present invention sets up a variety of operating scenarios to perform comprehensive dynamic verification of the optimized prompt logic, and combines visualization technology to display the topological change process of the dynamic graph structure in real time, including node priority adjustment, edge weight optimization, and dynamic changes in causal paths. By generating traceable test feedback data and optimization logs, a closed-loop optimization mechanism is formed, which not only improves the efficiency of the prompt function test, but also provides data support and intuitive display for the continuous improvement of the prompt logic, solving the inefficiency of traditional testing methods and the lack of dynamic visualization. BRIEF DESCRIPTION OF THE DRAWINGS

[0096] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0097] Figure 1 It is an overall flow chart of the testing method of the prompting function of the visual heavy-load locomotive driving guidance device proposed by the present invention;

[0098] Figure 2 This is a structural schematic diagram of a testing system for the prompting function of a visual heavy-load locomotive driving guidance device proposed by the present invention. DETAILED DESCRIPTION

[0099] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0100] refer to Figure 1 , a test method for the prompt function of a visual heavy-load locomotive driving guidance device includes the following steps:

[0101] S1. Collect multi-dimensional operation data of heavy-load locomotives through sensors and construct data sets;

[0102] S2, filtering noise, removing outliers and normalizing the data set, and generating an initial topological graph structure based on data correlation;

[0103] S3. Use graph convolutional networks to extract features from the initial topological graph structure, capture the multi-scale relationship between nodes and neighborhoods through multi-layer convolution, and generate high-dimensional feature vectors;

[0104] S4. Based on the high-dimensional feature vector, a causal relationship model is constructed through causal inference method to analyze the causal relationship between nodes and identify the key causal nodes and causal paths that have an important impact on the prompt function;

[0105] S5. Based on the key causal nodes and causal paths, the Eagle Swarm Optimization algorithm is used to dynamically optimize the initial topology graph structure. By combining global search with local optimization, node priorities and edge weights are adjusted to optimize the prompt function logic.

[0106] S6. Dynamically verify the optimized prompt function logic, generate prompt function test results and test feedback data by visually displaying the topological changes of the dynamic graph structure;

[0107] S7. Adjust the prompt function logic design and parameters according to the test feedback data, and optimize the prompt function performance through multiple rounds of iterations using the dynamically optimized topology structure.

[0108] In this implementation manner, the multi-dimensional operation data specifically includes the heavy-load locomotive running speed, station information, line signal, cylinder pressure, speed limit and over-phase.

[0109] In this implementation, S2 specifically includes:

[0110] S21. Perform preliminary cleaning on the data set, identify noise data based on statistical distribution characteristics, and remove data that does not conform to normal operating conditions through filtering technology;

[0111] S22, remove outliers from the cleaned data set, identify extreme outliers by calculating the deviation and dispersion of the data distribution, and use interpolation to fill in missing or abnormal data;

[0112] S23, normalizing the multidimensional features in the data set, mapping the feature values ​​to a unified range, and eliminating the influence between different feature dimensions;

[0113] S24. Based on the normalized data set, the linear correlation between the running characteristics is calculated using the Pearson correlation coefficient:

[0114]

[0115] Among them, r ij represents the correlation coefficient between operation feature i and operation feature j, x ik represents the value of running feature i in the kth data, x jk represents the value of running feature j in the kth data, represents the mean of running feature i, represents the mean of running feature j, and n represents the total number of data samples;

[0116] S25. According to the results of the operation feature correlation analysis, define the relationship between nodes and edges, where nodes represent operation features and edges represent the correlation between operation features:

[0117]

[0118] Among them, w ij represents the weight between node i and node j, |r ij | represents the absolute value of the correlation coefficient between operation feature i and operation feature j, and τ represents the set correlation threshold;

[0119] S26. Based on the definitions of the operation feature nodes and edges and in combination with the correlation analysis results, an initial topological graph structure is generated, wherein the layout of the nodes and edges dynamically reflects the associations between the operation features.

[0120] In this implementation, S3 specifically includes:

[0121] S31. Input the generated initial topological graph structure into the graph convolutional network and initialize the node feature matrix:

[0122]

[0123] Among them, H (0) represents the node feature matrix, represents the initial feature vector of node i, and N represents the total number of nodes;

[0124] S32, normalize the adjacency matrix of the initial topology graph, add self-loops to enhance the connectivity of the graph, and calculate the normalized adjacency matrix:

[0125]

[0126] in, represents the normalized adjacency matrix, A represents the original adjacency matrix, I represents the identity matrix, represents the degree matrix of A+I;

[0127] S33. In each layer of graph convolution, the variable topology adjustment mechanism is used to perform convolution operations on node features:

[0128]

[0129] Among them, H (l+1) represents the feature matrix of the l+1th layer node, σ represents the activation function, N(i) represents the set of neighboring nodes of node i, Represents the weight value in the normalized adjacency matrix, W (l) represents the training weight matrix of layer 1, c ij represents the normalization factor, α represents the dynamic adjustment coefficient, Represents the feature vector of node j in layer l;

[0130] S34, perform multi-scale fusion on the output features of multi-layer convolution, integrate features at different levels through splicing operations, and generate multi-scale features;

[0131] S35. Dynamically adjust the weight contribution of each layer of features according to the sparsity of multi-scale features, and introduce a sparsity-driven weighting mechanism:

[0132]

[0133] Among them, H (weighted) represents the feature representation after weighted fusion, L represents the total number of layers, exp represents the exponential function, β represents the sparse adjustment coefficient, and H (l) represents the node feature matrix of the first layer, ∥H (l) ∥1 represents the sparsity of the node feature matrix of the first layer, ∥H (m) ∥1 represents the sparsity of the node feature matrix of the mth layer;

[0134] S36. Perform linear transformation on the fused weighted features through a fully connected layer or a mapping matrix to generate a high-dimensional feature vector.

[0135] In this implementation manner, the S4 specifically includes:

[0136] S41. Based on the high-dimensional feature vector, a causal relationship model G = (V, E, W) is constructed, where V is a node set, representing features related to the prompt function; E is an edge set, representing the causal relationship between features; W is an edge weight set, used to characterize the strength of the causal influence between features;

[0137] S42. Quantify the causal strength of each edge in the edge set E and calculate the mutual information entropy of the causal strength based on the Bayesian network inference:

[0138]

[0139] Among them, C pq represents the causal strength between node p and node q, P(x p ,x q ) represents the joint probability distribution of node p and node q, P(x p ) represents the marginal probability distribution of node p, P(x q ) represents the edge probability distribution of node q;

[0140] S43. In causal path optimization, the path set is dynamically optimized in combination with the inference results of the Bayesian network:

[0141]

[0142] Among them, F path represents the path optimization objective function, P represents the set of candidate causal paths, Ω pqrepresents the correlation weight between node p and node q, exp represents the exponential function, γ represents the degree attenuation coefficient, and d q represents the degree value of node q, Δ pq represents the path length between node p and node q, and λ represents the penalty factor of the path length;

[0143] S44. Normalize the overall weight of the causal path:

[0144]

[0145] Among them, W pq represents the normalized causal weight matrix, N(p) represents the set of neighboring nodes of node p;

[0146] S45. Prioritize the input node features of the prompt function according to the normalized causal weight matrix, and prioritize high-weight nodes into the optimization process:

[0147]

[0148] Among them, P p represents the priority of node p, Φ(x q ) represents the feature importance score of node q;

[0149] S46. Output the key causal nodes and causal paths that have an important impact on the prompt function.

[0150] In this implementation manner, S5 specifically includes:

[0151] S51, initializing the initial topological graph structure according to the key causal nodes and causal paths, generating a population P = {P (1) ,P (2) ,…,P (M)},in m represents the individual number, M represents the number of individuals in the population, V represents the node set, and E represents the edge set. represents the edge weight set, Represents node v in individual m p With node v q The edge weights between represents the node priority set, Represents node v in individual m p Priority;

[0152] S52. In the global search phase, the strong individuals of the eagle swarm optimization algorithm are introduced to define the global objective function:

[0153]

[0154] in, represents the global objective function, N(v p ) represents node v p The domain node set, C pq Represents node v p and node v q The causal strength between Represents node v p The degree value, Represents node v q The degree value, μ represents the degree attenuation factor;

[0155] S53. In the local optimization stage, the weak individuals of the Eagle Swarm Optimization algorithm are introduced to fine-tune the edge weights according to the global search results of the strong individuals:

[0156]

[0157] in, represents the updated edge weight, δ represents the edge weight update coefficient, tanh represents the hyperbolic tangent function, ζ represents the priority influencing factor, Represents node v in individual m q The priority of Represents node v in individual m r Priority;

[0158] S54, adaptively adjust the node priority and update the node priority according to the topology change:

[0159]

[0160] in, represents the updated node priority, exp represents the exponential function, η represents the priority adjustment coefficient, |N(v p )| represents node v p The number of domain nodes;

[0161] S55. Evaluate the fitness of individuals:

[0162]

[0163] in, represents the fitness function, θ1, θ2 and θ3 represent weight parameters;

[0164] S56. The eagle group is updated through multiple rounds of iterations. In each round of iteration, the individual with the highest fitness is selected as the new center, and the node priority and edge weight distribution are passed to the next generation of individuals. By repeating the global search and local optimization process, the optimized topological structure is finally obtained.

[0165] In this implementation manner, S6 specifically includes:

[0166] S61. Dynamically verify the optimized prompt function logic, set different operation scenarios, including normal operation, ramp, sharp curve and bad weather, and record the triggering time, triggering conditions and prompt content of the prompt function in each scenario;

[0167] S62. In each operation scenario, collect signal data and test data output by the prompt function, including the accuracy, real-time and adaptability of the prompt signal to the operation status;

[0168] S63. Dynamically analyze the test data, evaluate the performance of the prompt function in different operation scenarios by comparing the prompt signal with the actual operation status, and calibrate the error range and potential optimization space of the prompt signal;

[0169] S64. Use visualization technology to display the topological changes of dynamic graph structures, including node priority, edge weight adjustment, and dynamic changes of causal paths;

[0170] S65, generating prompt function test results, including the accuracy of trigger time, the adaptability of prompt content, and the real-time performance of prompt response;

[0171] S66. Output the test results and scenario analysis data as test feedback data, and record the visual log generated during the test process.

[0172] refer to Figure 2 , the test system of the prompt function of the visual heavy-load locomotive driving guidance device includes the following modules:

[0173] Multi-dimensional operation data acquisition module, used to collect multi-dimensional operation data of heavy-load locomotives;

[0174] A data preprocessing module is used to process multi-dimensional operation data and generate an initial topological graph structure;

[0175] The feature extraction module is used to extract features from the initial topological graph structure through a graph convolutional network to generate a high-dimensional feature vector;

[0176] The causal relationship analysis module is used to build a causal relationship model based on high-dimensional feature vectors and identify key causal nodes and causal paths;

[0177] The optimization calculation module is used to dynamically optimize the initial topology structure using the Eagle Swarm Optimization algorithm, adjust node priorities and edge weights, and optimize the prompt function logic;

[0178] Dynamic verification module, used to dynamically verify the optimized prompt function logic, collect test data and generate test results;

[0179] Visual display module, used to display the topological changes of dynamic graph structure and generate test feedback data of prompt function;

[0180] The feedback and storage module is used to store and record test feedback data and support multiple rounds of iterative optimization of the prompt function logic.

[0181] Embodiment 1:

[0182] In order to verify the feasibility of the present invention in implementation, the present invention is applied to a heavy-duty railway transport line with a total length of 120 kilometers. The line is complex and changeable, including 20 kilometers of continuous steep slopes, 12 sharp bends and 4 common bad weather sections (foggy, rainy, etc.). During operation, the locomotive needs to optimize traction energy consumption while ensuring driving safety in a complex environment, and provide accurate prompt information in real time to support the driver's operational decisions. The traditional prompt function shows problems of insufficient adaptability and delayed prompt response in this scenario. It cannot effectively deal with steep slopes and sharp bends. The prompt error rate is as high as 18%, and the prompt delay time is 2.5 seconds on average, which seriously affects the driver's operating efficiency and driving safety.

[0183] In this embodiment, the test system of the present invention is used to dynamically optimize and verify the prompt function. First, during the operation of the heavy-loaded locomotive, the speed, station information, line signal, cylinder pressure, speed limit and over-equality data are acquired in real time through the multi-dimensional operation data acquisition module, and recorded at a frequency of milliseconds. The data preprocessing module performs noise filtering and outlier removal on these data, and generates an initial topological graph structure after normalization. Each operating feature is mapped to a node in the graph, and the correlation between features is defined as the edge weight. Subsequently, the generated initial topological graph is input into the feature extraction module, and the node and its neighborhood features are extracted at multiple levels using a graph convolutional network. The generated high-dimensional feature vector provides high-quality input data for the optimization of the prompt logic.

[0184] Based on feature extraction, the causal analysis module constructs a causal model through causal inference methods and identifies the key causal nodes and paths of the prompt function. For example, in steep slope areas, traction, speed and slope are identified as the nodes that have the greatest impact on the prompt content, while in sharp bends, acceleration, braking force and weather information have higher causal weights. The optimization calculation module combines the Eagle Swarm Optimization algorithm to dynamically optimize the prompt logic, adjust the priority of key nodes and redistribute edge weights to ensure that the prompt function can quickly respond to changes in operating status in complex scenarios.

[0185] In order to verify the optimized prompt function logic, the dynamic verification module sets up a variety of scenarios such as continuous steep slopes, sharp turns and bad weather, and collects the triggering time, prompt content and degree of matching with the actual operating status of the prompt function in these scenarios. The visualization display module dynamically displays the topological changes of the graph structure, including the priority adjustment of key nodes and the optimization process of edge weights, and generates detailed prompt function test feedback data.

[0186] The experimental results show that the optimized prompt function performs better than the traditional method in complex scenarios. For example, in the continuous steep slope area, the accuracy of the prompt content is improved to 98%, and the prompt delay time is reduced to 0.8 seconds; in the sharp bend area, the prompt error rate is reduced from the original 22% to 6%; in the bad weather section, the adaptability of the prompt content is significantly enhanced, and the driver's satisfaction with the prompt information is increased to 92%. In addition, the optimized prompt function significantly reduces traction energy consumption, saving about 8% of energy consumption during the entire operation process. The following Table 1 shows the specific experimental data.

[0187] Table 1 Comparison of prompt function performance before and after optimization and analysis of improvement effect

[0188]

[0189] It can be seen from the effect analysis of this embodiment that the present invention has achieved remarkable results in solving the technical bottleneck of the traditional prompt function. For complex scenes such as continuous steep slopes, sharp bends and bad weather, the optimized prompt function of the present invention shows extremely high adaptability and real-time performance. In the continuous steep slope scene, the prompt accuracy rate is increased from 85% of the traditional method to 98%, the prompt delay time is significantly shortened to 0.8 seconds, and the prompt error rate is reduced to 4%, providing drivers with more accurate operation suggestions and effectively improving driving safety. In the sharp bend scene, the prompt error rate dropped from 22% to 6%, and the prompt adaptability was improved by 13 percentage points, which fully demonstrates that the present invention can dynamically adapt to the operating requirements of complex bends. Under bad weather conditions, the driver's satisfaction with the prompt information increased from 68% to 92%, further verifying the robustness and reliability of the present invention under complex operating conditions. In addition, the present invention reduces energy consumption by optimizing the prompt logic, saving about 8% of traction energy, showing strong economic and environmental benefits.

[0190] This example fully proves that the present invention effectively solves the problems of insufficient adaptability, high error, delayed response, and lack of energy consumption optimization of traditional prompt functions in complex scenarios. By comprehensively improving the accuracy, real-time and adaptability of the prompt function, it further provides technical support for the intelligence, economy and safety of heavy-duty locomotive driving guidance devices, showing a wide range of application prospects in railway transportation.

[0191] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for testing the prompt function of a visual heavy-load locomotive driving guidance device, characterized in that: The steps include: S1. Collect multi-dimensional operation data of heavy-load locomotives through sensors and construct data sets; S2, filtering noise, removing outliers and normalizing the data set, and generating an initial topological graph structure based on data correlation; S3. Use graph convolutional networks to extract features from the initial topological graph structure, capture the multi-scale relationship between nodes and neighborhoods through multi-layer convolution, and generate high-dimensional feature vectors; S4. Based on the high-dimensional feature vector, a causal relationship model is constructed through causal inference method to analyze the causal relationship between nodes and identify the key causal nodes and causal paths that have an important impact on the prompt function; S5. Based on the key causal nodes and causal paths, the Eagle Swarm Optimization algorithm is used to dynamically optimize the initial topology graph structure. By combining global search with local optimization, node priorities and edge weights are adjusted to optimize the prompt function logic. S6. Dynamically verify the optimized prompt function logic, generate prompt function test results and test feedback data by visually displaying the topological changes of the dynamic graph structure; S7. Adjust the prompt function logic design and parameters according to the test feedback data, and optimize the prompt function performance through multiple rounds of iterations using the dynamically optimized topology structure.

2. The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to claim 1 is characterized in that: The multi-dimensional operation data specifically include the heavy-load locomotive running speed, station information, line signal, cylinder pressure, speed limit and over-phase.

3. The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to claim 1 is characterized in that: The S2 specifically includes: S21. Perform preliminary cleaning on the data set, identify noise data based on statistical distribution characteristics, and remove data that does not conform to normal operating conditions through filtering technology; S22, remove outliers from the cleaned data set, identify extreme outliers by calculating the deviation and dispersion of the data distribution, and use interpolation to fill in missing or abnormal data; S23, normalizing the multidimensional features in the data set, mapping the feature values ​​to a unified range, and eliminating the influence between different feature dimensions; S24. Based on the normalized data set, the linear correlation between the running characteristics is calculated using the Pearson correlation coefficient: Among them, r ij represents the correlation coefficient between operation feature i and operation feature j, x ik represents the value of running feature i in the kth data, x jk represents the value of running feature j in the kth data, represents the mean of running feature i, represents the mean of running feature j, and n represents the total number of data samples; S25. According to the results of the operation feature correlation analysis, define the relationship between nodes and edges, where nodes represent operation features and edges represent the correlation between operation features: Among them, w ij represents the weight between node i and node j, |r ij | represents the absolute value of the correlation coefficient between operation feature i and operation feature j, and τ represents the set correlation threshold; S26. Based on the definitions of the operation feature nodes and edges and in combination with the correlation analysis results, an initial topological graph structure is generated, wherein the layout of the nodes and edges dynamically reflects the associations between the operation features.

4. The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to claim 1 is characterized in that: The S3 specifically includes: S31. Input the generated initial topological graph structure into the graph convolutional network and initialize the node feature matrix: Among them, H (0) represents the node feature matrix, represents the initial feature vector of node i, and N represents the total number of nodes; S32, normalize the adjacency matrix of the initial topology graph, add self-loops to enhance the connectivity of the graph, and calculate the normalized adjacency matrix: in, represents the normalized adjacency matrix, A represents the original adjacency matrix, I represents the identity matrix, represents the degree matrix of A+I; S33. In each layer of graph convolution, the variable topology adjustment mechanism is used to perform convolution operations on node features: Among them, H (l+1) represents the feature matrix of the l+1th layer node, σ represents the activation function, N(i) represents the set of neighboring nodes of node i, Represents the weight value in the normalized adjacency matrix, W (l) represents the training weight matrix of layer 1, c ij represents the normalization factor, α represents the dynamic adjustment coefficient, Represents the feature vector of node j in layer 1; S34, perform multi-scale fusion on the output features of multi-layer convolution, integrate features at different levels through splicing operations, and generate multi-scale features; S35. Dynamically adjust the weight contribution of each layer of features according to the sparsity of multi-scale features, and introduce a sparsity-driven weighting mechanism: Among them, H (weighted) represents the feature representation after weighted fusion, L represents the total number of layers, exp represents the exponential function, β represents the sparse adjustment coefficient, and H (l) represents the node feature matrix of the first layer, ||H (l) ||1 represents the sparsity of the node feature matrix of the first layer, ||H (m) ||1 represents the sparsity of the node feature matrix of the mth layer; S36. Perform linear transformation on the fused weighted features through a fully connected layer or a mapping matrix to generate a high-dimensional feature vector.

5. The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to claim 1 is characterized in that: The S4 specifically includes: S41. Based on the high-dimensional feature vector, a causal relationship model G = (V, E, W) is constructed, where V is a node set, representing features related to the prompt function; E is an edge set, representing the causal relationship between features; W is an edge weight set, used to characterize the strength of the causal influence between features; S42. Quantify the causal strength of each edge in the edge set E and calculate the mutual information entropy of the causal strength based on the Bayesian network inference: Among them, C pq represents the causal strength between node p and node q, P(x p ,x q ) represents the joint probability distribution of node p and node q, P(x p ) represents the marginal probability distribution of node p, P(x q ) represents the edge probability distribution of node q; S43. In causal path optimization, the path set is dynamically optimized in combination with the inference results of the Bayesian network: Among them, F path represents the path optimization objective function, P represents the set of candidate causal paths, Ω pq represents the correlation weight between node p and node q, exp represents the exponential function, γ represents the degree attenuation coefficient, and d q represents the degree value of node q, Δ pq represents the path length between node p and node q, and λ represents the penalty factor of the path length; S44. Normalize the overall weight of the causal path: Among them, W pq represents the normalized causal weight matrix, N(p) represents the set of neighboring nodes of node p; S45. Prioritize the input node features of the prompt function according to the normalized causal weight matrix, and prioritize high-weight nodes into the optimization process: Among them, P p represents the priority of node p, Φ(x q ) represents the feature importance score of node q; S46. Output the key causal nodes and causal paths that have an important impact on the prompt function.

6. The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to claim 1 is characterized in that: The S5 specifically includes: S51, initializing the initial topological graph structure according to the key causal nodes and causal paths, generating a population P = {P (1) ,P (2) ,…,P (M) },in m represents the individual number, M represents the number of individuals in the population, V represents the node set, and E represents the edge set. represents the edge weight set, Represents node v in individual m p With node v q The edge weights between represents the node priority set, Represents node v in individual m p Priority; S52. In the global search phase, the strong individuals of the eagle swarm optimization algorithm are introduced to define the global objective function: in, represents the global objective function, N(v p ) represents node v p The domain node set, C pq Represents node v p and node v q The causal strength between Represents node v p The degree value, Represents node v q The degree value, μ represents the degree attenuation factor; S53. In the local optimization stage, the weak individuals of the Eagle Swarm Optimization algorithm are introduced to fine-tune the edge weights according to the global search results of the strong individuals: in, represents the updated edge weight, δ represents the edge weight update coefficient, tanh represents the hyperbolic tangent function, ζ represents the priority influencing factor, Represents node v in individual m q The priority of Represents node v in individual m r Priority; S54, adaptively adjust the node priority and update the node priority according to the topology change: in, represents the updated node priority, exp represents the exponential function, η represents the priority adjustment coefficient, |N(v p )| represents node v p The number of domain nodes; S55. Evaluate the fitness of individuals: in, represents the fitness function, θ1, θ2 and θ3 represent weight parameters; S56. The eagle group is updated through multiple rounds of iterations. In each round of iteration, the individual with the highest fitness is selected as the new center, and the node priority and edge weight distribution are passed to the next generation of individuals. By repeating the global search and local optimization process, the optimized topological structure is finally obtained.

7. The method for testing the prompt function of the visual heavy-load locomotive driving guidance device according to claim 1 is characterized in that: The S6 specifically includes: S61. Dynamically verify the optimized prompt function logic, set different operation scenarios, including normal operation, ramp, sharp curve and bad weather, and record the triggering time, triggering conditions and prompt content of the prompt function in each scenario; S62. In each operation scenario, collect signal data and test data output by the prompt function, including the accuracy, real-time and adaptability of the prompt signal to the operation status; S63. Dynamically analyze the test data, evaluate the performance of the prompt function in different operation scenarios by comparing the prompt signal with the actual operation status, and calibrate the error range and potential optimization space of the prompt signal; S64. Use visualization technology to display the topological changes of dynamic graph structures, including node priority, edge weight adjustment, and dynamic changes of causal paths; S65, generating prompt function test results, including the accuracy of trigger time, the adaptability of prompt content, and the real-time performance of prompt response; S66. Output the test results and scenario analysis data as test feedback data, and record the visual log generated during the test process.

8. A test system for the prompt function of a visual heavy-load locomotive driving guidance device, which executes the test method for the prompt function of a visual heavy-load locomotive driving guidance device according to any one of claims 1 to 7, characterized in that: Includes the following modules: Multi-dimensional operation data acquisition module, used to collect multi-dimensional operation data of heavy-load locomotives; A data preprocessing module is used to process multi-dimensional operation data and generate an initial topological graph structure; The feature extraction module is used to extract features from the initial topological graph structure through a graph convolutional network to generate a high-dimensional feature vector; The causal relationship analysis module is used to build a causal relationship model based on high-dimensional feature vectors and identify key causal nodes and causal paths; The optimization calculation module is used to dynamically optimize the initial topology structure using the Eagle Swarm Optimization algorithm, adjust node priorities and edge weights, and optimize the prompt function logic; Dynamic verification module, used to dynamically verify the optimized prompt function logic, collect test data and generate test results; Visual display module, used to display the topological changes of dynamic graph structure and generate test feedback data of prompt function; The feedback and storage module is used to store and record test feedback data and support multiple rounds of iterative optimization of the prompt function logic.

Images