Method and system for designing a steering wheel form of a converted vehicle, storage medium and computer

Abstract

This invention provides a method, system, storage medium, and computer for designing the shape of modified vehicle steering wheels. The method includes: collecting steering wheel data from several vehicle databases for modified new energy vehicles, and constructing a multi-source original sensory database based on the steering wheel data; using a particle swarm optimization algorithm to automatically optimize and calculate the weights of a set of sensory evaluation terms, obtaining key sensory terms and their weight rankings; constructing and training a variable regression model based on the key sensory terms, their weights, and a set of steering wheel images to establish a mapping relationship between user sensory needs and steering wheel shape features, and generating multiple sets of steering wheel shape design schemes based on the mapping relationship; comprehensively calculating and ranking the various steering wheel shape design schemes based on experimental data generated by users in multimodal emotional experiments, and selecting the optimal design scheme set based on the comprehensive score. This invention can more accurately predict users' emotional preferences for the steering wheel shapes of NEVs.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method, system, storage medium, and computer for designing the shape of a modified vehicle steering wheel. Background Technology

[0002] As the global automotive industry transforms towards new energy and intelligent technologies, NEVs (New Energy Vehicles) have become an important vehicle for promoting sustainable development in the transportation sector. Against the backdrop of increasingly fierce competition in user experience, the emotional design of automotive interiors is gradually becoming a key factor influencing consumer purchasing decisions. Among these, the steering wheel, as the component with the most frequent interaction between the user and the vehicle, directly determines the user's subjective evaluation of the driving experience through its shape and visual style. However, current steering wheel designs still primarily focus on rational dimensions such as ergonomic optimization, material and tactile evaluation, and functional integration. While these studies have made significant progress in safety and handling, they have to varying degrees neglected the role of product form in stimulating users' emotional needs.

[0003] To address the aforementioned shortcomings, existing technologies employ methods for mapping product design to user perception. However, in the steering wheel field, previous research has largely relied on single psychological modality data such as semantic difference evaluation and questionnaire scoring, lacking cross-validation with objective physiological signals like ET and EEG, making it difficult to support design decisions under the premise of unified subjective and objective evidence. Furthermore, in the process of transforming and mapping sensory features, the methods still primarily rely on manually selecting words and using linear or shallow neural network modeling, which limits the ability to process high-dimensional, highly nonlinear, and highly coupled sensory data between modalities. Summary of the Invention

[0004] Based on this, the purpose of the present invention is to provide a method, system, storage medium and computer for designing the shape of a modified vehicle steering wheel, so as to at least solve the shortcomings of the above-mentioned technology.

[0005] This invention proposes a method for designing the shape of a modified vehicle steering wheel, comprising:

[0006] Steering wheel data of new energy modified vehicles are collected from several preset vehicle databases, and a corresponding multi-source original sensory database is constructed based on the steering wheel data. The steering wheel data includes a set of sensory evaluation terms and a set of steering wheel images.

[0007] The particle swarm optimization algorithm is used to automatically optimize and calculate the weights of the set of emotional evaluation words to obtain the key emotional words that reflect the user's emotional preferences and their weight ranking.

[0008] Based on the key emotional terms and their weights and the set of steering wheel images, a variable regression model is constructed and trained to build a mapping relationship between user emotional needs and steering wheel shape features, and multiple steering wheel shape design schemes are generated according to the mapping relationship.

[0009] The experiment data generated by users in conducting multimodal emotion experiments is obtained, and the steering wheel shape design schemes are comprehensively calculated and sorted based on the experimental data. The optimal design scheme set is selected based on the comprehensive score.

[0010] Furthermore, the steps of automatically optimizing and weighting the set of emotional evaluation words using the particle swarm optimization algorithm to obtain the key emotional words reflecting user sentiment preferences and their weight ranking include:

[0011] Based on user ratings of representative steering wheel images on the set of emotional evaluation terms, an emotional evaluation matrix is ​​constructed.

[0012] Define a fitness function to maximize the projection variance and control the sparsity of the weights. Iteratively optimize the weight vector using a particle swarm optimization algorithm to obtain the weight ranking of key perceptual words.

[0013] Furthermore, based on the key emotional terms and their weights, and the set of steering wheel images, the step of constructing and training a variable regression model to build a mapping relationship between user emotional needs and steering wheel morphological features includes:

[0014] Based on the key perceptual vocabulary and its weights as semantic feature inputs, and the morphological deconstruction features of the steering wheel image set as design feature outputs, a CPO-Transformer-GRU hybrid neural network model is constructed.

[0015] The CPO-Transformer-GRU hybrid neural network model is trained using labeled training samples to establish an end-to-end high-dimensional nonlinear mapping relationship between emotional needs and steering wheel morphological features.

[0016] Furthermore, the CPO-Transformer-GRU hybrid neural network model sequentially includes an embedding layer, a Transformer encoder layer, a gated recurrent unit layer, and a regression output layer. The CPO optimization algorithm runs before model training to automatically search for and determine the hyperparameter combination of the Transformer encoder layer and the gated recurrent unit layer.

[0017] Furthermore, the steps of acquiring experimental data generated by users in conducting multimodal emotion experiments, comprehensively calculating and ranking the various steering wheel shape design schemes based on the experimental data, and selecting the optimal design scheme based on the comprehensive score include:

[0018] A multimodal decision matrix incorporating eye movement, electroencephalography, and subjective ratings was constructed based on the experimental data.

[0019] Distinguish between benefit-based and cost-based indicators, and normalize all indicators to ensure they are aligned in direction;

[0020] The group utility value, individual regret value and compromise ranking index of each steering wheel shape design scheme were calculated using the VIKOR algorithm, and the corresponding relative closeness was calculated using the TOPSIS algorithm.

[0021] The compromise ranking index and the relative closeness are linearly weighted and fused to obtain the comprehensive utility score of each steering wheel shape design scheme. The steering wheel shape design schemes are then ranked according to the comprehensive utility score, and the optimal design scheme is selected based on the comprehensive score.

[0022] This invention also proposes a system for designing the shape of a modified vehicle steering wheel, comprising:

[0023] The database construction module is used to collect steering wheel data of new energy modified vehicles from several preset vehicle databases, and construct a corresponding multi-source original sensory database based on the steering wheel data. The steering wheel data includes a set of sensory evaluation terms and a set of steering wheel images.

[0024] The data optimization module is used to automatically optimize and calculate the weights of the set of emotional evaluation words using the particle swarm optimization algorithm, so as to obtain the key emotional words that reflect the user's emotional preferences and their weight ranking.

[0025] The model building module is used to build and train a variable regression model based on the key emotional words and their weights and the set of steering wheel images, so as to build a mapping relationship between user emotional needs and steering wheel shape features, and generate multiple steering wheel shape design schemes according to the mapping relationship.

[0026] The decision filtering module is used to acquire experimental data generated by users in conducting multimodal emotion experiments, and to perform comprehensive calculations and sorting of each steering wheel shape design scheme based on the experimental data, and to filter out the optimal design scheme set based on the comprehensive score.

[0027] Furthermore, the data optimization module is specifically used for:

[0028] Based on user ratings of representative steering wheel images on the set of emotional evaluation terms, an emotional evaluation matrix is ​​constructed.

[0029] Define a fitness function to maximize the projection variance and control the sparsity of the weights. Iteratively optimize the weight vector using a particle swarm optimization algorithm to obtain the weight ranking of key perceptual words.

[0030] Furthermore, the model building module is specifically used for:

[0031] Based on the key perceptual vocabulary and its weights as semantic feature inputs, and the morphological deconstruction features of the steering wheel image set as design feature outputs, a CPO-Transformer-GRU hybrid neural network model is constructed.

[0032] The CPO-Transformer-GRU hybrid neural network model is trained using labeled training samples to establish an end-to-end high-dimensional nonlinear mapping relationship between emotional needs and steering wheel morphological features.

[0033] Furthermore, the decision filtering module is specifically used for:

[0034] A multimodal decision matrix incorporating eye movement, electroencephalography, and subjective ratings was constructed based on the experimental data.

[0035] Distinguish between benefit-based and cost-based indicators, and normalize all indicators to ensure they are aligned in direction;

[0036] The group utility value, individual regret value and compromise ranking index of each steering wheel shape design scheme were calculated using the VIKOR algorithm, and the corresponding relative closeness was calculated using the TOPSIS algorithm.

[0037] The compromise ranking index and the relative closeness are linearly weighted and fused to obtain the comprehensive utility score of each steering wheel shape design scheme. The steering wheel shape design schemes are then ranked according to the comprehensive utility score, and the optimal design scheme is selected based on the comprehensive score.

[0038] The present invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the above-described method for designing the shape of a modified vehicle steering wheel.

[0039] The present invention also proposes a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for designing the shape of a modified vehicle steering wheel.

[0040] The method, system, storage medium, and computer for designing the shape of a modified vehicle steering wheel in this invention incorporate eye-tracking behavior, neurophysiological responses, and subjective emotional evaluation into a unified assessment framework. This framework performs an integrated subjective and objective evaluation of the generated steering wheel designs, achieving a comprehensive and accurate assessment of user emotional responses. This enhances the comprehensiveness, accuracy, and reliability of emotional measurement, providing a scientific basis for design decisions. Furthermore, the VIKOR-TOPSIS integrated decision-making method is introduced to weight, integrate, and rank the multimodal quantitative data of the generated candidate steering wheel samples, obtaining a multimodal comprehensive score. The top three steering wheel designs are then selected as the final recommended solutions. Through the synergistic application of multimodal emotional data and the CPO-Transformer-GRU model, user emotional preferences for NEVs steering wheel designs are more accurately predicted, providing quantitative and interpretable decision-making basis for design generation and solution selection. This significantly improves the user's emotional experience and visual satisfaction in actual driving scenarios. Attached Figure Description

[0041] Figure 1 This is a flowchart of the method for designing the shape of a modified vehicle steering wheel in the first embodiment of the present invention;

[0042] Figure 2 This is a schematic diagram of the structure of the CPO-Transformer-GRU model in the first embodiment of the present invention;

[0043] Figure 3 This is a structural block diagram of the modified vehicle steering wheel shape design system in the second embodiment of the present invention;

[0044] Figure 4 This is a structural block diagram of the computer in the third embodiment of the present invention.

[0045] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0046] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0048] Example 1

[0049] Please see Figure 1 The figure shows a method for designing the shape of a modified vehicle steering wheel according to the first embodiment of the present invention. The method specifically includes steps S101 to S104:

[0050] S101, collect steering wheel data of new energy modified vehicles from several preset vehicle databases, and construct a corresponding multi-source original sensory database based on the steering wheel data, wherein the steering wheel data includes a set of sensory evaluation terms and a set of steering wheel images;

[0051] In practice, steering wheel data of new energy modified vehicles is collected from relevant databases in the automotive field using methods such as web crawling. A corresponding multi-source original sensory database is constructed based on the obtained steering wheel data. The relevant databases in the automotive field include automotive forums, e-commerce platforms, and design databases. The steering wheel data includes a set of sensory evaluation terms (e.g., "stylish", "smooth", "exquisite") and a set of steering wheel images (in this embodiment, the set of steering wheel images contains several high-resolution images). This data comes from the manufacturer's official website and professional design databases, covering a variety of shapes and materials.

[0052] S102, The particle swarm optimization algorithm is used to automatically optimize and calculate the weights of the set of emotional evaluation words to obtain the key emotional words that reflect the user's emotional preferences and their weight ranking.

[0053] Furthermore, step S102 specifically includes steps S1021 to S1022:

[0054] S1021, Based on the user's rating data for representative steering wheel images on the set of emotional evaluation terms, construct an emotional evaluation matrix;

[0055] S1022, define a fitness function to maximize the projection variance and control the sparsity of the weights. Iteratively optimize the weight vector through the particle swarm optimization algorithm to obtain the weight ranking of the key perceptual words.

[0056] In practical implementation, the PSO algorithm is used to obtain the priority ranking of emotional words. To identify the potential emotional characteristics of users' perception of steering wheel design, a multi-dimensional database consisting of a set of emotional evaluation words and steering wheel images is constructed. Based on web crawling of e-commerce platforms and professional forum comments, and subsequent cleaning, a large-scale, timely, and realistic emotional corpus is formed to reduce insufficient samples and subjective bias. Semantically, the KJ clustering method is used to establish a clear and interpretable vocabulary system. Furthermore, PSO is introduced to globally optimize and rank the weights of the emotional dimensions, reducing expert weighting bias and providing more accurate feature weight support for design element mapping and model training. The specific PSO algorithm is shown below:

[0057] 1. Initialization

[0058] Let the intuition weight vector to be optimized be... Preprocessed feature matrix L2 normalization is performed on the weights inside the objective function:

[0059]

[0060] in, ; Indicates the number of samples. Represents the preprocessed feature matrix, the first... Behavior No. one sample Dimensional features; Represents the norm, Indicates to conduct Normalized weights The projected scalar sequence of the samples;

[0061] The sample variance of a projection is defined as:

[0062]

[0063] In the formula, Indicates the first ; This represents the mean of the projected values; This represents the sample variance statistical function, used to measure the dispersion of the projection results.

[0064] Define the entropy of the normalized weights ( (where the numerical stability constant is used)

[0065]

[0066] in, , represents the normalized feature weight; used to characterize the relative importance of the feature in the overall projection; Indexed by feature dimensions, This represents the total dimension of the features.

[0067] 2. Objective function

[0068] Box constraints Minimize:

[0069]

[0070] in, .

[0071] That is, after L2 normalization, the projection variance is maximized, while the entropy is heavily penalized (to promote sparsity and highlight keywords).

[0072] 3. PSO update equation

[0073] Record No. Each particle is iterating Position, speed Individual historical optimum and global optimum are Generate dimensionally independent random factors. ,but:

[0074] ;

[0075]

[0076] 4. Algorithm Flow

[0077] enter: Population size Maximum iteration Inertia weight Learning factors , (Optional speed) ).

[0078] Output: Final weights (Non-negative, sum to 1).

[0079] (1) Initialization: Randomly generated , , place , .

[0080] (2) Fitness assessment: Calculations for each particle

[0081] ;

[0082] (3) Update individual / global optimum: if ,but ; and then Optimal update .

[0083] (4) Velocity and position update: Iterate according to formulas (5)-(6) and perform boundary truncation to cut off the velocity in each dimension. ).

[0084] (5) Termination and Output: When Stop when the tolerance threshold is met; output the optimal solution. Post-processing: Cube "sharpening" and normalization:

[0085]

[0086] The final weights used for ranking and downstream modeling are obtained. .

[0087] S103, Based on the key emotional words and their weights and the set of steering wheel images, construct and train a variable regression model to build a mapping relationship between user emotional needs and steering wheel shape features, and generate multiple steering wheel shape design schemes according to the mapping relationship.

[0088] Furthermore, step S103 specifically includes steps S1031 to S1032:

[0089] S1031, Based on the key perceptual words and their weights as semantic feature inputs, and using the morphological deconstruction features of the steering wheel image set as design feature outputs, a CPO-Transformer-GRU hybrid neural network model is constructed.

[0090] S1032, The CPO-Transformer-GRU hybrid neural network model is trained using labeled training samples to establish an end-to-end high-dimensional nonlinear mapping relationship between emotional needs and steering wheel morphological features.

[0091] In practical implementation, to achieve a high-dimensional nonlinear mapping relationship between the sensory needs of automotive steering wheels and product design features, a dual optimization mechanism of GRU and CPO is introduced on the Transformer backbone structure to construct a CPO-Transformer-GRU sensory mapping model. This model integrates the global semantic modeling capability of Transformer, the dynamic memory mechanism of GRU, and the global parameter optimization performance of CPO, thereby achieving high-precision learning and prediction of sensory features under limited sample conditions. For ease of understanding, the model structure is visualized; please refer to [link / reference]. Figure 2 .

[0092] Furthermore, the CPO-Transformer-GRU hybrid neural network model sequentially includes an embedding layer, a Transformer encoder layer, a gated recurrent unit layer, and a regression output layer. The CPO optimization algorithm runs before model training to automatically search for and determine the hyperparameter combination of the Transformer encoder layer and the gated recurrent unit layer.

[0093] Specifically, the overall structure of the model mainly consists of the four parts mentioned above:

[0094] 1. Input layer: Receives the normalized perceptual evaluation matrix. ,in For the sample size, For the dimension of emotional vocabulary.

[0095] 2. Transformer backbone network: Utilizes a self-attention mechanism to extract higher-order semantic dependencies between intuitive words. The calculation formula is as follows:

[0096]

[0097] in, These are query, key, and value matrices, respectively. This is a scaling factor for the feature dimension. This layer can effectively capture the interactive features between different perceptual dimensions, enabling global feature modeling.

[0098] 3. GRU Optimization Layer: To address the shortcomings of the Transformer in representing local and temporal information, a GRU is introduced to dynamically filter and model the time dependencies of its output features. Its update mechanism is as follows:

[0099]

[0100]

[0101]

[0102] in, t This represents the index of the time step in the input sequence. To update the door, This represents the Sigmoid activation function, used to map the gated output to the [0,1] interval; , , These are the learnable weight matrices corresponding to the update gate, reset gate, and candidate hidden states, respectively. To reset the door, This indicates that the Transformer module is at time step The output feature vector, and These represent the GRU at time step. and Hidden state vector The time step index is used to characterize the temporal positional relationships within sequence data. This represents the Hadamard product. Through a gating mechanism, GRU can enhance the Transformer's sensitivity to temporal features and avoid information redundancy and gradient vanishing.

[0103] 4. CPO Optimization Module: CPO is used to globally optimize the hyperparameters and weight distribution of the Transformer-GRU model. This algorithm simulates the dynamic behavior of a porcupine quill formation's defense and cooperation, achieving a dynamic balance between global exploration and local development through a "quill formation expansion-retraction" mechanism. The optimization objective of CPO is to minimize the predicted output. Compared with the true value Mean square error between:

[0104]

[0105] in, For the set of model parameters, The number of samples; For the first The true label of each sample; For the model in parameters Next to the The predicted output for each sample; This is the fitness function value.

[0106] CPO Global Optimization Mechanism: In each iteration, CPO treats the model parameters as the position vectors of individual porcupines, and performs parameter search by dynamically adjusting the "porcupine radius." Its core update rule is:

[0107]

[0108]

[0109] in, t The first step in the particle swarm optimization process is... t Next iteration step For inertial weights, For the first Individuals are iterating t The position vector; It is the velocity vector; For an individual, the best historical position; The globally optimal position; , As a learning factor, The disturbance coefficient is... The "spike radius" for the current individual. The term represents a Gaussian perturbation. Through fitness function evaluation and dynamic weight adjustment, CPO can balance global exploration and local convergence in a high-dimensional parameter space, making the model training process more stable and significantly reducing the risk of overfitting.

[0110] Further, model output and performance evaluation:

[0111] The output layer uses the Sigmoid activation function to map the prediction results to the interval [0, 1]:

[0112]

[0113] in, The model predicts the output; h is the sigmoid activation function; h is the hidden state vector output from the previous layer (GRU / fusion layer); and These are the output layer weight matrix and the bias term (both are learnable parameters).

[0114] Model performance is measured by mean squared error (MSE) and coefficient of determination (COP). ) and Pearson correlation coefficient ( A comprehensive evaluation will be conducted.

[0115]

[0116] in, Indicates the number of samples being evaluated. This corresponds to the predicted value; For the true value The sample mean, For the first The true value of each sample; cov The covariance between the actual value and the predicted value; and These are the sample standard deviations of the true values ​​and the predicted values, respectively. MSE measures the mean squared deviation of the prediction error. Reflecting the goodness of fit of the model, It characterizes the strength of the linear correlation between the predicted results and the actual values.

[0117] Understandably, when the sample size is limited or the data is noisy, the CPO-Transformer-GRU model is an intuitive mapping framework that combines interpretability and prediction, while still maintaining high prediction accuracy and stability. The model performs well on the intuitive dataset of car steering wheel design and can provide scientific modeling tools and quantitative basis for intuitive design.

[0118] S104: Obtain experimental data generated by the user in the multimodal emotion experiment, and perform comprehensive calculation and sorting on each steering wheel shape design scheme according to the experimental data, and select the optimal design scheme set according to the comprehensive score.

[0119] Furthermore, step S104 specifically includes steps S1041 to S1044:

[0120] S1041, Construct a multimodal decision matrix that includes eye movement, electroencephalography and subjective rating based on the experimental data;

[0121] S1042, differentiate between benefit-type indicators and cost-type indicators, and normalize all indicators to ensure they are aligned in the same direction;

[0122] S1043, respectively, the VIKOR algorithm is used to calculate the group utility value, individual regret value and compromise ranking index of each steering wheel shape design scheme, and the TOPSIS algorithm is used to calculate the corresponding relative closeness.

[0123] S1044, perform linear weighted fusion of the compromise ranking index and the relative closeness to obtain the comprehensive utility score of each steering wheel shape design scheme, and rank each steering wheel shape design scheme according to the comprehensive utility score, and select the optimal design scheme according to the comprehensive score.

[0124] In practice, experimental data generated by users in conducting multimodal emotion experiments are obtained. These multimodal emotion experiments include eye-tracking experiments, electroencephalogram (EEG) experiments, and self-assessment scales (subjective ratings).

[0125] The eye-tracking experiment aims to use eye-tracking data as behavioral evidence to explain the temporal progression and spatial distribution of EEG signals, and to provide attentional priors for subsequent emotion assessment. The experimental procedure comprises three main steps. First, basic preparations are completed, including setting up a constant lighting and low-noise environment, ensuring a fixed distance between the subject and the screen, and connecting the Tobii Pro Fusion eye tracker with a maximum sampling rate of 250 Hz. Second, the experimental procedure is designed: NEVs steering wheel images generated by SDM based on key sensory words and processed to a standardized format are presented as stimulus material in a random order. Data is collected and exported using Tobii Pro Lab (the latest version). The experiment specifically includes five-point calibration and drift correction, followed by 1000 ms fixation point orientation, 3000 ms single image presentation, and empty screen transition. After debugging to ensure the procedure is error-free, the experimental data is finally collected and analyzed to obtain the characteristics of the user's attention process under different stimulus conditions. All eye-tracking events are recorded and saved via Tobii Pro Lab and synchronized with the EEG device to the same timeline, ensuring trial-by-trial alignment between modalities.

[0126] Specifically, the EEG experiment served as the main focus, aiming to characterize the subjects' perception and evaluation process of the steering wheel stimulation generated by the SDM using high temporal resolution neural signals. The experiment employed the Emotiv Flex 2 device. After setting up the environment, connecting the device, and using conductive gel to reduce resistance with wet electrodes, the stimulation trigger signal was written by EmotivPRO and synchronized with eye-tracking recordings at the millisecond level. Formal EEG data was collected after the process stabilized.

[0127] Data analysis was performed in MATLAB. Preprocessing included channel-by-channel detrending, zero-phase-shift bandpass filtering (0.5-45Hz), and slicing around the trigger point. 0.2,0.8][-0.2,0.8][ 0.2,0.8] s and in [ 0.2,0][-0.2,0][ Baseline correction is performed on the interval [0.2, 0]. For individual segments lacking external labels, only equally spaced spare events are generated for visualization inspection; conditional comparisons and statistics are based solely on real events or onsets derived from label intervals, and the entire epoch is guaranteed to belong to the same condition to avoid cross-condition contamination.

[0128] Furthermore, feature extraction was conducted at three levels: frequency domain, time-frequency domain, and ERP. The Welch method was used to estimate the power spectrum from 0 to 25 Hz and to plot scalp topography at typical frequencies to describe the overall power and spatial distribution. Morlet wavelet time-frequency analysis was performed on the posterior-top priority channel to obtain the power and inter-trial phase consistency (ITC) index. Peak amplitudes within predefined time windows of N200, P200, and P300 were extracted from the front, middle, and back clusters of channels. All EEG features were Z-normalized within the participants and aligned with the corresponding eye movement and SAM indices at the sample level for correlation testing and multimodal fusion analysis. Components such as N200, P200, and P300 were used as core physiological indicators of user emotional representation, providing key data support for a multimodal emotional evaluation system.

[0129] The SAM scale is a recognized classic nonverbal quantification tool in the field of emotion assessment. Given the assessment scenario of the car steering wheel's styling and its visual elements (shape and form) in this embodiment, the SAM scale is used to quantify the user's emotional response to the steering wheel's design. The immediate subjective emotion of the subjects to the visual stimuli is obtained within a multimodal evaluation system: Valence, Arousal, and Dominance are all assessed using a 9-point Likert scale with uniform positive coding (1 for low, 9 for high), ultimately yielding the subjects' subjective evaluation of the NEVs steering wheel rendering scheme. The stimulus material consists of steering wheel images standardized for parameters such as size, brightness, and background. Subjects immediately complete a three-dimensional rating after each stimulus image disappears.

[0130] Furthermore, cost-related indicators such as first-gaze time are reoriented, while other benefit-related indicators remain positive. Range standardization is used to uniformly map these indicators to a range where "larger values ​​indicate better performance." Equal weights are applied between and within modes, treating the three types of indicators as complementary information sources. Subsequently, the VIKOR-TOPSIS fusion model is used to calculate the overall deviation and maximum regret to obtain a compromise ranking index, and the TOPSIS model is used to calculate the proximity coefficient. After 0-1 linear normalization, both are weighted and synthesized into a comprehensive utility score. The optimal design scheme is then selected based on the comprehensive score. Understandably, the VIKOR-TOPSIS fusion model can improve the discriminative power of steering wheel samples, providing a clear quantitative basis for subsequent styling optimization.

[0131] In summary, the modified vehicle steering wheel design method in the above embodiments of the present invention incorporates eye-tracking behavior, neurophysiological responses, and subjective emotional evaluation into a unified evaluation framework. This allows for an integrated subjective and objective assessment of the generated steering wheel designs, achieving a comprehensive and accurate evaluation of user emotional responses. This enhances the comprehensiveness, accuracy, and reliability of emotional measurement, providing a scientific basis for design decisions. Furthermore, the introduction of the VIKOR-TOPSIS integrated decision-making method weights, integrates, and ranks the multimodal quantitative data of the generated candidate steering wheel samples to obtain a multimodal comprehensive score. The top three steering wheel designs are then selected as the final recommended solutions. Through the synergistic application of multimodal emotional data and the CPO-Transformer-GRU model, the method more accurately predicts user emotional preferences for NEVs steering wheel designs, providing quantitative and interpretable decision-making basis for design generation and solution selection. This significantly improves the user's emotional experience and visual satisfaction in actual driving scenarios.

[0132] Example 2

[0133] In another aspect, this invention also proposes a system for designing the shape of a modified vehicle steering wheel; please refer to [link / reference needed]. Figure 3 The figure shows a modified vehicle steering wheel shape design system according to a second embodiment of the present invention, the system comprising:

[0134] The database construction module 11 is used to collect steering wheel data of new energy modified vehicles from several preset vehicle databases, and construct a corresponding multi-source original sensory database based on the steering wheel data. The steering wheel data includes a set of sensory evaluation terms and a set of steering wheel images.

[0135] Data optimization module 12 is used to automatically optimize and calculate the weights of the set of emotional evaluation words using the particle swarm optimization algorithm, so as to obtain the key emotional words that reflect the user's emotional preferences and their weight ranking.

[0136] The model building module 13 is used to build and train a variable regression model based on the key emotional words and their weights and the set of steering wheel images, so as to build a mapping relationship between user emotional needs and steering wheel shape features, and generate multiple steering wheel shape design schemes according to the mapping relationship.

[0137] The decision filtering module 14 is used to acquire experimental data generated by users in conducting multimodal emotion experiments, and to perform comprehensive calculations and sorting of each steering wheel shape design scheme based on the experimental data, and to filter out the optimal design scheme set based on the comprehensive score.

[0138] Furthermore, the data optimization module 12 is specifically used for:

[0139] Based on user ratings of representative steering wheel images on the set of emotional evaluation terms, an emotional evaluation matrix is ​​constructed.

[0140] Define a fitness function to maximize the projection variance and control the sparsity of the weights. Iteratively optimize the weight vector using a particle swarm optimization algorithm to obtain the weight ranking of key perceptual words.

[0141] Furthermore, the model building module 13 is specifically used for:

[0142] Based on the key perceptual vocabulary and its weights as semantic feature inputs, and the morphological deconstruction features of the steering wheel image set as design feature outputs, a CPO-Transformer-GRU hybrid neural network model is constructed.

[0143] The CPO-Transformer-GRU hybrid neural network model is trained using labeled training samples to establish an end-to-end high-dimensional nonlinear mapping relationship between emotional needs and steering wheel morphological features.

[0144] Furthermore, the decision filtering module 14 is specifically used for:

[0145] A multimodal decision matrix incorporating eye movement, electroencephalography, and subjective ratings was constructed based on the experimental data.

[0146] Distinguish between benefit-based and cost-based indicators, and normalize all indicators to ensure they are aligned in direction;

[0147] The group utility value, individual regret value and compromise ranking index of each steering wheel shape design scheme were calculated using the VIKOR algorithm, and the corresponding relative closeness was calculated using the TOPSIS algorithm.

[0148] The compromise ranking index and the relative closeness are linearly weighted and fused to obtain the comprehensive utility score of each steering wheel shape design scheme. The steering wheel shape design schemes are then ranked according to the comprehensive utility score, and the optimal design scheme is selected based on the comprehensive score.

[0149] The functions or operation steps implemented by the above modules and units are largely the same as those in the above method embodiments, and will not be repeated here.

[0150] The modified vehicle steering wheel shape design system provided in this embodiment of the invention has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0151] Example 3

[0152] This invention also proposes a computer, please refer to [link / reference]. Figure 4 The computer shown in the third embodiment of the present invention includes a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, it implements the above-described method for designing the shape of a modified vehicle steering wheel.

[0153] The memory 10 includes at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10 can be an internal storage unit of a computer, such as the computer's hard disk. In other embodiments, the memory 10 can be an external storage device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Furthermore, the memory 10 can include both internal and external storage units of the computer. The memory 10 can be used not only to store application software and various types of data installed on the computer, but also to temporarily store data that has been output or will be output.

[0154] In some embodiments, the processor 20 may be an electronic control unit (ECU), a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip, used to run program code stored in the memory 10 or process data, such as executing access restriction programs.

[0155] It should be pointed out that, Figure 4 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0156] This invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the modified vehicle steering wheel shape design method described above.

[0157] Those skilled in the art will understand that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0158] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0159] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0160] 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 specification.

[0161] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. 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 patent application should be determined by the appended claims.

Claims

1. A method for designing the shape of a modified vehicle steering wheel, characterized in that, include: Steering wheel data of new energy modified vehicles are collected from several preset vehicle databases, and a corresponding multi-source original sensory database is constructed based on the steering wheel data. The steering wheel data includes a set of sensory evaluation terms and a set of steering wheel images. The particle swarm optimization algorithm is used to automatically optimize and calculate the weights of the set of emotional evaluation words to obtain the key emotional words that reflect the user's emotional preferences and their weight ranking. Based on the key emotional terms and their weights and the set of steering wheel images, a variable regression model is constructed and trained to build a mapping relationship between user emotional needs and steering wheel shape features, and multiple steering wheel shape design schemes are generated according to the mapping relationship. The experiment data generated by users in conducting multimodal emotion experiments is obtained, and the steering wheel shape design schemes are comprehensively calculated and sorted according to the experiment data. The optimal design scheme set is selected based on the comprehensive score. The step of automatically optimizing and calculating the weights of the set of emotional evaluation words using the particle swarm optimization algorithm to obtain the key emotional words reflecting user emotional preferences and their weight ranking includes: Based on user ratings of representative steering wheel images on the set of emotional evaluation terms, an emotional evaluation matrix is ​​constructed. Define a fitness function to maximize the projection variance and control the sparsity of the weights. Iteratively optimize the weight vector using a particle swarm optimization algorithm to obtain the weight ranking of key perceptual words. The step of constructing and training a variable regression model based on the key emotional terms and their weights and the set of steering wheel images to build a mapping relationship between user emotional needs and steering wheel morphological features includes: Based on the key perceptual vocabulary and its weights as semantic feature inputs, and the morphological deconstruction features of the steering wheel image set as design feature outputs, a CPO-Transformer-GRU hybrid neural network model is constructed. The CPO-Transformer-GRU hybrid neural network model is trained using labeled training samples to establish an end-to-end high-dimensional nonlinear mapping relationship between emotional needs and steering wheel morphological features.

2. The method for designing the shape of a modified vehicle steering wheel according to claim 1, characterized in that, The CPO-Transformer-GRU hybrid neural network model sequentially includes an embedding layer, a Transformer encoder layer, a gated recurrent unit layer, and a regression output layer. The CPO optimization algorithm runs before model training to automatically search for and determine the hyperparameter combination of the Transformer encoder layer and the gated recurrent unit layer.

3. The method for designing the shape of a modified vehicle steering wheel according to claim 1, characterized in that, The steps of acquiring experimental data generated by users in conducting multimodal emotion experiments, comprehensively calculating and ranking the various steering wheel shape design schemes based on the experimental data, and selecting the optimal design scheme based on the comprehensive score include: A multimodal decision matrix incorporating eye movement, electroencephalography (EEG), and subjective ratings was constructed based on the experimental data. Distinguish between benefit-based and cost-based indicators, and normalize all indicators to ensure they are aligned in direction; The group utility value, individual regret value and compromise ranking index of each steering wheel shape design scheme were calculated using the VIKOR algorithm, and the corresponding relative closeness was calculated using the TOPSIS algorithm. The compromise ranking index and the relative closeness are linearly weighted and fused to obtain the comprehensive utility score of each steering wheel shape design scheme. The steering wheel shape design schemes are then ranked according to the comprehensive utility score, and the optimal design scheme is selected based on the comprehensive score.

4. A system for designing the shape of a modified vehicle steering wheel, characterized in that, include: The database construction module is used to collect steering wheel data of new energy modified vehicles from several preset vehicle databases, and construct a corresponding multi-source original sensory database based on the steering wheel data. The steering wheel data includes a set of sensory evaluation terms and a set of steering wheel images. The data optimization module is used to automatically optimize and calculate the weights of the set of emotional evaluation words using the particle swarm optimization algorithm, so as to obtain the key emotional words that reflect the user's emotional preferences and their weight ranking. The model building module is used to build and train a variable regression model based on the key emotional words and their weights and the set of steering wheel images, so as to build a mapping relationship between user emotional needs and steering wheel shape features, and generate multiple steering wheel shape design schemes according to the mapping relationship. The decision filtering module is used to acquire experimental data generated by users in conducting multimodal emotion experiments, and to perform comprehensive calculations and sorting of each steering wheel shape design scheme based on the experimental data, and to filter out the optimal design scheme set based on the comprehensive score. Specifically, the data optimization module is used for: Based on user ratings of representative steering wheel images on the set of emotional evaluation terms, an emotional evaluation matrix is ​​constructed. Define a fitness function to maximize the projection variance and control the sparsity of the weights. Iteratively optimize the weight vector using a particle swarm optimization algorithm to obtain the weight ranking of key perceptual words. Specifically, the model building module is used for: Based on the key perceptual vocabulary and its weights as semantic feature inputs, and the morphological deconstruction features of the steering wheel image set as design feature outputs, a CPO-Transformer-GRU hybrid neural network model is constructed. The CPO-Transformer-GRU hybrid neural network model is trained using labeled training samples to establish an end-to-end high-dimensional nonlinear mapping relationship between emotional needs and steering wheel morphological features.

5. A readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method for designing the shape of a modified vehicle steering wheel as described in any one of claims 1 to 3.

6. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for designing the shape of a modified vehicle steering wheel as described in any one of claims 1 to 3.

Images