Vehicle comfort evaluation method, device, electronic device and storage medium

Abstract

The embodiments of the present invention disclose a vehicle comfort assessment method, device, electronic device, and storage medium. The method includes: obtaining the vehicle's current driving characteristic information and user characteristic information, and obtaining the vehicle's driving characteristic information and in-vehicle environment information within a preset time period; inputting the driving characteristic information and in-vehicle environment information within the preset time period into a time series characteristic assessment model of the corresponding category to obtain the single characteristic assessment results corresponding to the driving characteristic information and in-vehicle environment information respectively; inputting the single characteristic assessment results corresponding to the driving characteristic information and in-vehicle environment information, the driving characteristic information, and the user characteristic information into a multi-task assessment model to obtain the comprehensive characteristic assessment results of the vehicle. The method of the embodiment of the present invention can comprehensively consider various factors that affect the user's comfort in riding in the vehicle, improve the accuracy of the comprehensive characteristic assessment results, and narrow the gap between the comprehensive assessment results and the user's actual feelings.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of vehicle ride comfort control, and in particular to a vehicle comfort evaluation method, device, electronic device, and storage medium. Background Art

[0002] In recent years, with the development of the economy, the number of cars has continued to surge, and more and more people are riding or driving cars. When users are driving or riding in a car, the ride comfort will greatly affect the user experience.

[0003] Existing vehicle ride comfort assessment methods evaluate vehicle comfort based solely on acceleration and steering angle information, providing an incomplete assessment of ride comfort. This results in inaccurate results that differ significantly from the user's actual experience while driving or riding the vehicle, failing to meet user needs and reducing the user experience. Summary of the Invention

[0004] The present invention provides a vehicle comfort evaluation method, device, electronic equipment and storage medium, which can accurately and comprehensively predict the user's evaluation result on vehicle comfort.

[0005] In a first aspect, an embodiment of the present invention provides a vehicle comfort evaluation method, the method comprising:

[0006] Acquiring the current driving characteristic information and user characteristic information of the vehicle, and acquiring the driving characteristic information and in-vehicle environment information of the vehicle within a preset time period;

[0007] Inputting the driving characteristic information and the in-vehicle environment information within the preset time period into a time series feature evaluation model of a corresponding category, obtaining single feature evaluation results corresponding to the driving characteristic information and the in-vehicle environment information, respectively, wherein the single feature evaluation results are used to represent user comfort evaluation results of the corresponding features;

[0008] The single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the driving feature information and the user feature information are input into a multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle, and the comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle.

[0009] In a second aspect, an embodiment of the present invention further provides a vehicle comfort evaluation device, the device comprising:

[0010] An information acquisition module is used to obtain the current driving characteristic information and user characteristic information of the vehicle, and to obtain the driving characteristic information and in-vehicle environment information of the vehicle within a preset time period;

[0011] a single result determination module, configured to input the driving characteristic information and the in-vehicle environment information within the preset time period into a time series feature evaluation model of a corresponding category, and obtain single feature evaluation results corresponding to the driving characteristic information and the in-vehicle environment information, respectively, wherein the single feature evaluation results are used to represent user comfort evaluation results of the corresponding features;

[0012] A comprehensive result determination module is used to input the single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the driving feature information and the user feature information into a multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle, and the comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle.

[0013] In a third aspect, an embodiment of the present invention further provides an electronic device, comprising:

[0014] one or more processors;

[0015] a memory for storing one or more programs;

[0016] When the one or more programs are executed by the one or more processors, the one or more processors implement the vehicle comfort evaluation method provided by any embodiment of the present invention.

[0017] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the vehicle comfort evaluation method provided by any embodiment of the present invention.

[0018] In this embodiment of the present invention, current vehicle driving characteristic information and user characteristic information are obtained, as well as driving characteristic information and in-vehicle environment information for a preset duration. The driving characteristic information and in-vehicle environment information for the preset duration are input into a time-series feature evaluation model of corresponding categories to obtain individual feature evaluation results corresponding to the driving characteristic information and in-vehicle environment information, respectively. The individual feature evaluation results are used to represent the user comfort evaluation results for the corresponding features. The individual feature evaluation results, driving characteristic information, and user characteristic information corresponding to the driving characteristic information and in-vehicle environment information are input into a multi-task evaluation model to obtain a comprehensive feature evaluation result for the vehicle. The comprehensive feature evaluation result is used to represent the user comfort evaluation result for the vehicle. In other words, in this embodiment of the present invention, by obtaining the vehicle's driving characteristic information, user characteristic information, driving characteristic information, and in-vehicle environment information, various factors that affect a user's comfort in the vehicle are comprehensively considered. Using the time-series feature evaluation model, the user's comfort evaluation results for individual features can be accurately predicted, narrowing the gap between the comprehensive evaluation results and the user's actual experience. This further improves the accuracy of the comprehensive feature evaluation results predicted by the multi-task model. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flow chart of a vehicle comfort evaluation method provided by an embodiment of the present invention;

[0020] Figure 2 1 is a flow chart of obtaining a comprehensive feature evaluation result of a vehicle according to an embodiment of the present invention;

[0021] Figure 3 is a flow chart of a method for obtaining a time series feature evaluation model provided by an embodiment of the present invention;

[0022] Figure 4 is a flowchart of a method for obtaining a multi-task evaluation model provided by an embodiment of the present invention;

[0023] Figure 5 is a schematic structural diagram of model parameters of the initial task model provided by an embodiment of the present invention;

[0024] Figure 6 is a flow chart of another vehicle comfort evaluation method provided by an embodiment of the present invention;

[0025] Figure 7 This is an overall flow chart of the vehicle comfort evaluation method provided by an embodiment of the present invention;

[0026] Figure 8 1 is a schematic structural diagram of a vehicle comfort evaluation device provided by an embodiment of the present invention;

[0027] Figure 9 It is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The present invention will be further described in detail below with reference to the accompanying drawings and examples. It will be understood that the specific embodiments described herein are intended only to illustrate the present invention and are not intended to limit the present invention. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions relevant to the present invention, not all structures.

[0029] Figure 1 This is a flow chart of a vehicle comfort assessment method provided by an embodiment of the present invention. This method can accurately and comprehensively predict a user's assessment of vehicle comfort. This method can be performed by a vehicle comfort assessment device provided by an embodiment of the present invention. This device can be integrated into an electronic device, such as a server. The method can be implemented using software and / or hardware. The vehicle comfort assessment method provided by this embodiment specifically includes the following steps:

[0030] Step 101: Acquire the current driving characteristic information and user characteristic information of the vehicle, and acquire the driving characteristic information and in-vehicle environment information of the vehicle within a preset time period.

[0031] Among them, the vehicle's current driving characteristic information is the information generated by the user during the process of driving or riding the vehicle, such as the current time, driving duration, and driving history information. The vehicle's current user characteristic information is the information of the user who is currently driving or riding the vehicle, such as the user's gender, age, and other information. Among them, the preset time length can be set according to specific needs and actual environment. The vehicle's driving characteristic information includes the maximum longitudinal acceleration, maximum longitudinal deceleration, maximum longitudinal impact degree, maximum lateral acceleration, and acceleration and deceleration frequency information within the preset time length. The in-vehicle environment information includes the humidity and temperature information of the vehicle cabin within the preset time length.

[0032] Specifically, after a user begins driving or riding in a vehicle, information such as the vehicle's current time, driving duration, and driving history is obtained. User profile information, such as age and gender, is received through a mobile app or vehicle screen, and the vehicle's current user profile information is obtained. User profile information can also be obtained through a dealer management system (DMS) or an operational method sheet (OMS). The vehicle's speed sensor determines the vehicle's maximum longitudinal acceleration, maximum longitudinal deceleration, maximum longitudinal impact, maximum lateral acceleration, and acceleration / deceleration frequency in real time. For example, within a preset time period, longitudinal acceleration information is extracted every 5 seconds to obtain the maximum longitudinal acceleration value (positive value) for each cycle. If the maximum acceleration value within a cycle is negative, the maximum longitudinal acceleration within that cycle is set to 0. Similarly, longitudinal deceleration information is extracted every 5 seconds to obtain the maximum longitudinal deceleration value (negative value) for each cycle. If the maximum deceleration value within a cycle is positive, the maximum longitudinal deceleration within that cycle is set to 0. The longitudinal acceleration information within the preset time length is extracted with a period of 5 seconds, the derivative of the longitudinal acceleration with respect to time in each period is obtained, the longitudinal impact degree is calculated, and then the longitudinal impact degree is estimated using the Kalman filter algorithm. Using the Kalman filter algorithm to estimate the longitudinal impact degree can reduce the influence of high-frequency noise generated during the calculation process on the calculation results. Similarly, the lateral acceleration information is extracted with a period of 5 seconds to obtain the maximum lateral acceleration (the maximum lateral acceleration in absolute value) value in each period. The number of acceleration and deceleration conversions within the preset time length is extracted with a period of 5 seconds to obtain the number of acceleration and deceleration conversions in each period. Among them, before extracting the data sent by the speed sensor, the data sent by the speed sensor needs to be eliminated of outliers and smoothed by median filtering.

[0033] Specifically, the vehicle's temperature sensor obtains the average cabin temperature, while the humidity sensor obtains the average cabin humidity. For example, cabin temperature information is extracted over a preset time period at a 30-second interval to obtain the average cabin temperature. Similarly, cabin humidity information is extracted over a preset time period at a 30-second interval to obtain the average cabin humidity. Before extracting the data from the temperature and humidity sensors, outliers must be removed and median-filtered to obtain the desired in-vehicle environmental information.

[0034] Step 102: Input the driving characteristic information and the in-vehicle environment information within a preset time period into the time series feature evaluation model of the corresponding category to obtain the single feature evaluation results corresponding to the driving characteristic information and the in-vehicle environment information respectively. The single feature evaluation results are used to represent the user comfort evaluation results of the corresponding features.

[0035] Among them, the temporal feature evaluation model includes a reset gate and an update gate. The reset gate can be used to combine the current input information in the model with the previous input information, and the update gate can be used to set the time length information between the previous input information and the current input information that needs to be saved. The temporal feature evaluation model includes the GRU model, LSTM model, RNN model, etc. Taking the GRU model as an example, the GRU model can better capture the temporal relationship and dependency between driving feature information and in-vehicle environment information. The single feature evaluation result is the corresponding user comfort evaluation result predicted by the GRU model based on the driving feature information and in-vehicle environment information. The user comfort evaluation result is the user's current evaluation result of the comfort level of riding or driving the vehicle.

[0036] Specifically, each type of driving characteristic information and in-vehicle environment information has its own corresponding time series feature evaluation model. The driving characteristic information and in-vehicle environment information within a preset time period are used as input data. After inputting this input data into the corresponding GRU model, the GRU model can predict the user comfort evaluation result corresponding to each feature information in the input data. For example, the driving characteristic information is obtained by extracting and processing speed information with a 5-second period, and the in-vehicle environment information is obtained by extracting and processing cabin temperature and humidity information with a 30-second period. After obtaining the driving characteristic information and in-vehicle environment information, the driving characteristic information and in-vehicle environment information are input into the GRU model with a 5-minute period (the time interval between the previous input information and the current input information needs to be saved). The GRU model performs calculations and inferences on the input data, predicting the user's comfort evaluation result for each type of driving characteristic information and in-vehicle environment information.

[0037] Step 103: Input the individual feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the travel feature information, and the user feature information into a multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle. The comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle.

[0038] The Multi-Task Learning (MTL) model is a predictive model based on a neural network architecture. Sample data is shared within the MTL model, effectively handling differences in data distribution and relationships between tasks. Based on the one-way feature evaluation results of driving characteristics and in-vehicle environment information, as well as driving characteristics and user characteristics, the MTL model can calculate and infer the user's assessment of vehicle comfort.

[0039] Specifically, the multi-task evaluation model uses individual evaluation results of driving characteristics and in-vehicle environment information, along with driving characteristics and user characteristics, as input data. The model then predicts the user's comfort rating based on the input data. For example, the model receives individual evaluation results of driving characteristics and in-vehicle environment information, along with driving characteristics and user characteristics, over a five-minute period. The model then calculates and infers the received data to predict the user's comfort rating for the vehicle.

[0040] Figure 2 FIG. 1 is a flow chart of obtaining a comprehensive feature evaluation result of a vehicle according to an embodiment of the present invention. Figure 2 As shown in the figure, driving feature information and in-vehicle environment information are input into the corresponding GRU model to obtain the individual feature evaluation results corresponding to each category of driving feature information and in-vehicle environment information. The individual feature evaluation results, driving feature information, and user feature information are input into the multi-task evaluation model to ultimately obtain the comprehensive feature evaluation results of the vehicle.

[0041] The solution of the embodiment of the present invention obtains the vehicle's current driving characteristic information and user characteristic information, and obtains the vehicle's driving characteristic information and in-vehicle environment information within a preset time period. The driving characteristic information and in-vehicle environment information within the preset time period are input into the corresponding category of time series feature evaluation model to obtain the individual feature evaluation results corresponding to the driving characteristic information and in-vehicle environment information, respectively. The individual feature evaluation results are used to represent the user comfort evaluation results of the corresponding features. The individual feature evaluation results corresponding to the driving characteristic information and in-vehicle environment information, the driving characteristic information, and the user characteristic information are input into the multi-task evaluation model to obtain the comprehensive feature evaluation results of the vehicle. The comprehensive feature evaluation results are used to represent the user comfort evaluation results of the vehicle. By obtaining user characteristic information, this solution pays attention to the impact of different users' physiological states on the user comfort evaluation results. Using different models to process different types of data can comprehensively and accurately predict the user's comfort evaluation results of the vehicle, further improving the prediction accuracy of the vehicle comfort evaluation results.

[0042] Figure 3 This is a flow chart of a method for obtaining a time series feature evaluation model provided by an embodiment of the present invention. This embodiment further refines the method for obtaining a time series feature evaluation model, such as Figure 3 As shown in FIG, the method for obtaining the time series feature evaluation model after refinement mainly includes the following steps:

[0043] Step 201: Obtain a first training data set, where each sample in the first training data set includes sample data and a sample label.

[0044] The first training dataset is used to train the initial time series feature model. The sample data includes historical driving feature information and each type of feature information in the in-vehicle environment information. The sample labels include the comfort evaluation results of each type of feature information provided by the experimenter based on the sample data.

[0045] Specifically, corresponding sample labels can be collected for driving characteristic information and in-vehicle environment information. For example, to obtain a training dataset corresponding to the maximum longitudinal acceleration characteristic information within the driving characteristic information, an experimenter can accelerate the vehicle at specified acceleration levels at different speeds. After the vehicle is accelerated, the experimenter's comfort assessment results for the acceleration process are recorded. The comfort assessment results range from 0 to 5, with 0 representing extreme discomfort and 5 representing extreme comfort. During the collection of the training dataset, different experimenters may have different comfort assessment results for the same acceleration process. Therefore, to improve the generalization of the time series feature evaluation model, multiple experimenters can simultaneously evaluate comfort during the acceleration process, obtaining multiple comfort assessment results. After obtaining multiple comfort assessment results, the consistency of these multiple comfort assessment results can be calculated using the Kendall coefficient. If the consistency of the multiple comfort assessment results exceeds a preset consistency level, the average of the multiple comfort assessment results is calculated, and the average is rounded up to obtain the final comfort assessment result. Suppose five experimenters simultaneously evaluate the comfort of an acceleration process of 60 km / h, and their comfort evaluation results are: 3, 2, 3, 3, and 2. Calculation of the Kendall coefficient shows that the consistency of these five comfort evaluation results is greater than the preset consistency. The average of these five comfort evaluation results is 2.6, and the final comfort evaluation result is rounded up to 3. The final data set is: sample data: vehicle acceleration 60 km / h, sample label: user comfort evaluation result is 3 points. Furthermore, training data for each type of feature information in the historical driving feature information and in-vehicle environment information is obtained to obtain the first training data set.

[0046] Step 202: Input the sample data into the initial time series feature model for training to obtain output data of the initial time series feature model.

[0047] Among them, the initial time series feature model is a time series feature evaluation model that has not been trained. Each type of feature information in the historical driving feature information and the in-vehicle environment information has its corresponding initial time series feature model. Each type of feature information is input into the corresponding initial time series feature model in chronological order, and the comfort evaluation result for each type of feature information feedback output by the initial time series feature model can be obtained. For example, with a period of 5 minutes, each type of feature information is input into the corresponding initial time series feature model in chronological order to obtain the output data of the initial time series feature model. When the initial time series feature model has not been trained, the comfort evaluation result it outputs for sample data feedback is quite different from the actual user comfort evaluation result. Therefore, the initial time series feature model needs to be optimized based on the output data and the corresponding sample labels (the comfort evaluation results fed back by the experimenter for the sample data).

[0048] Step 203: Determine a first loss function based on the output data and the corresponding sample labels.

[0049] Among them, the output data is the comfort evaluation result output by the initial time series feature model, and the sample label is the comfort evaluation result fed back by the experimenter for the sample data in the first training data set. The loss function is a function that maps the value of a random event or its related random variables to a non-negative real number to represent the "risk" or "loss" of the random event. For example, the mean square error loss function (MSE), the cross entropy loss function (CEL), etc. The first loss function is the loss function of the time series feature evaluation model in this scheme. For example, the initial time series feature evaluation model is a GRU model, and the first loss function is the MSE function. Then, the first loss function is determined according to the output data and the corresponding sample label as follows:

[0050]

[0051] Among them, n represents the number of categories of feature information in the sample data, y i represents the sample label, Represents the output data of the initial time series feature model, 1≤i≤n.

[0052] Step 204: Inversely optimize the model parameters of the initial time series feature model based on the first loss function to obtain a time series feature evaluation model.

[0053] Among them, model parameters are configuration variables within the model, and the values of model parameters can be adjusted according to the loss function. When the time series feature evaluation model receives input data, it needs to calculate and infer the prediction results corresponding to the input data based on the model parameters. Therefore, appropriate model parameter values can make the prediction results of the time series feature evaluation model closer to the user's evaluation results of vehicle comfort. Specifically, the initial time series feature model is continuously optimized according to the first loss function, and the model parameters of the initial time series feature model are adjusted so that the gap between the output data of the initial time series feature model and the sample label is continuously narrowed, and finally the time series feature evaluation model is obtained.

[0054] The vehicle comfort evaluation method provided by an embodiment of the present invention can obtain a first training data set, and each sample in the first training data set includes sample data and a sample label. The sample data is input into the initial time series feature model for training to obtain the output data of the initial time series feature model. The first loss function is determined based on the output data and the corresponding sample label. The model parameters of the initial time series feature model are reversely optimized based on the first loss function to obtain a time series feature evaluation model. In the embodiment of the present invention, each type of feature information in the historical driving feature information and the in-vehicle environment information is used as sample data, and the time series feature evaluation model is obtained based on the output data and sample labels of the initial time series feature model, so that the single feature evaluation result is closer to the user's comfort evaluation result, thereby improving the accuracy of the single feature evaluation result.

[0055] Figure 4 This is a flow chart of a method for obtaining a multi-task evaluation model provided by an embodiment of the present invention. This embodiment further refines the method for obtaining a multi-task evaluation model, such as Figure 4 As shown in Figure 2, the refined method for obtaining the multi-task evaluation model mainly includes the following steps:

[0056] Step 301: Obtain a second training data set, where each sample in the second training data set includes sample data and a sample label.

[0057] The second training dataset is used to train the initial task model. The sample data includes comfort assessment results corresponding to each type of historical driving characteristics and in-vehicle environment information, as well as each item of driving characteristics and user characteristics within a preset time period. Sample labels include the comfort assessment results provided by the experimenter for the sample data. Specifically, corresponding sample labels can be collected for each historical driving characteristics and user characteristics.

[0058] For example, the experimenters involved in the sample label collection process include the driver and passengers. Before the vehicle starts driving, user profile information, such as gender and age, is obtained and recorded. The driver can drive the vehicle according to their own driving habits, and the passengers can choose their seats according to their preferences. While the vehicle is driving, passengers can provide feedback on their comfort assessment results every 5 minutes via a mobile app. The comfort assessment results range from 0 to 5 points, with 0 representing extreme discomfort and 5 representing very comfortable. During the vehicle's driving process, each item of historical driving profile information, in-vehicle environment information, historical driving profile information, and user profile information is recorded simultaneously. The comfort evaluation results provided by the passengers are recorded and organized to ultimately generate sample labels.

[0059] Step 302: Input the sample data into the initial task model for training to obtain output data of the initial task model.

[0060] The initial task model is an untrained multi-task evaluation model. By inputting sample data into the initial task model, the initial task model outputs a comfort evaluation result based on the sample data feedback. For example, by inputting sample data into the initial task model over a 5-minute period, the initial task model outputs the output data. When the initial task model is untrained, the comfort evaluation results it outputs based on the sample data feedback often differ significantly from the actual user comfort evaluation results. Therefore, the initial task model needs to be optimized based on the output data and the corresponding sample labels (the comfort evaluation results provided by the experimenter based on the sample data feedback).

[0061] Step 303: Determine a second loss function based on the output data and the corresponding sample labels.

[0062] The output data is the comfort assessment result output by the initial task model, and the sample labels are the comfort assessment results provided by the experimenters for the sample data in the second training dataset. A loss function maps the values of a random event or its related random variables to non-negative real numbers to represent the "risk" or "loss" of the random event. The second loss function is the loss function of the multi-task model in this solution. The second loss function, determined based on the output data and the corresponding sample labels, can represent the "gap" between the output data of the initial task model and the sample labels.

[0063] Step 304: Inversely optimize the model parameters of the initial task model based on the second loss function to obtain a multi-task evaluation model.

[0064] The model parameters are configuration variables within the model, and their values can be adjusted based on the data. In this embodiment, optionally, the model parameters of the initial task model are reversely optimized based on the second loss function, including: reversely optimizing the model parameters of the shared layer and the model parameters of the personalized layer of the initial task model based on the second loss function.

[0065] In a multi-task model, the model parameters of the shared layer are shared across multiple tasks. This helps effectively handle the relationships between tasks, improving the learning efficiency and quality of each task. The model parameters of the personalized layer can be set differently for different tasks (different categories of feature information) based on different user characteristics. These parameters are not shared, effectively handling the variability between sample data.

[0066] Figure 5 is a schematic diagram of the structure of the model parameters of the initial task model provided by the embodiment of the present invention, such as Figure 5 As shown in Figure 1, the output data and sample labels of the initial task model are used as input data. The model parameters of the shared layer and the model parameters of the personalized layer are adjusted based on the input data. The model parameters of the shared layer are shared by all tasks. The model parameters in the personalized layer are model parameters set by different users.

[0067] Specifically, multiple tasks in the shared layer use a single loss function (the second loss function), and model parameters can be adjusted based on the second loss function. However, the "loss" between the output data and sample labels obtained by different tasks through the loss function is different, and the importance of their "loss" for adjusting the shared layer parameters is also different. Therefore, the weights of each task can be adjusted using a dynamic weighted average method:

[0068]

[0069] Where L is the total loss, ω i (t) represents the weight of different task losses in each round of training, L i is the loss of different tasks in each round of training, ω i (t) is calculated as follows:

[0070]

[0071]

[0072] Among them, N represents the number of tasks, L n Represents the loss of different tasks, r nrepresents the training speed of different tasks, t represents the number of task training steps, T is a constant, and n represents the total number of task training rounds, where 1≤i≤n. Furthermore, the initial task model is continuously optimized based on the second loss function, adjusting its parameters to narrow the gap between the initial task model's output data and the sample labels, ultimately resulting in a multi-task evaluation model.

[0073] The model parameters of the shared layer and the personalized layer are optimized through the second loss function, which can focus on the correlation between multiple tasks in the sample data, and at the same time handle the differences between the sample data well, thereby improving the accuracy of the comprehensive feature evaluation results output by the multi-task evaluation model.

[0074] The vehicle comfort evaluation method provided by the embodiment of the present invention can obtain a second training data set, and each sample in the second training data set includes sample data and sample labels. The sample data is input into the initial task model for training to obtain the output data of the initial task model. The second loss function is determined based on the output data and the corresponding sample label. The model parameters of the initial task model are reversely optimized based on the second loss function to obtain a multi-task evaluation model. In the embodiment of the present invention, each item of driving characteristic information and user characteristic information within a preset time length is used as sample data, and a multi-task evaluation model is obtained based on the output data and sample labels of the initial task model, thereby improving the comprehensiveness and generalization of the multi-task evaluation model. The model parameters of the shared layer and the personalized layer are optimized by the second loss function, that is, it is possible to pay attention to the correlation between multiple tasks in the sample data, and at the same time, it can also handle the differences between the sample data well, thereby improving the accuracy of the comprehensive feature evaluation results output by the multi-task evaluation model, so that the comprehensive feature evaluation results are closer to the user's comfort evaluation results.

[0075] Figure 6 This is a flow chart of another vehicle comfort evaluation method provided by an embodiment of the present invention. This embodiment further refines the vehicle comfort evaluation method, such as Figure 6 As shown in Figure 2, the refined vehicle comfort evaluation method mainly includes the following steps:

[0076] Step 401: Acquire the current driving characteristic information and user characteristic information of the vehicle, and acquire the driving characteristic information and in-vehicle environment information of the vehicle within a preset time period.

[0077] Step 402: Input the driving characteristic information and the in-vehicle environment information within a preset time period into the time series feature evaluation model of the corresponding category to obtain the single feature evaluation results corresponding to the driving characteristic information and the in-vehicle environment information respectively. The single feature evaluation results are used to represent the user comfort evaluation results of the corresponding features.

[0078] Step 403: Input the individual feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the travel feature information, and the user feature information into the multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle. The comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle.

[0079] Step 404: Determine prompt control information based on the comprehensive feature evaluation result and the individual feature evaluation result, and display the prompt control information.

[0080] The comprehensive feature evaluation result is the user's vehicle comfort assessment calculated and inferred by the multi-task evaluation model based on the unidirectional feature evaluation results of driving characteristics and in-vehicle environment information, driving characteristics information, and user characteristics. The individual feature evaluation result is the user comfort assessment result predicted by the time series feature evaluation model based on driving characteristics and in-vehicle environment information. Prompt control information is information that needs to be reminded to the user based on the comprehensive feature evaluation result and the individual feature evaluation results.

[0081] In this embodiment, optionally, determining prompt control information based on the comprehensive feature evaluation result and the individual feature evaluation result, and displaying the prompt control information includes the following steps A1 to A3:

[0082] Step A1: Determine whether the comprehensive feature evaluation result meets the first preset condition, and determine whether the single feature evaluation result meets the second preset condition.

[0083] Among them, the first preset condition indicates that the current comprehensive feature evaluation result is greater than the preset comprehensive evaluation result, indicating that the user is relatively comfortable and does not need to be reminded. The second preset condition indicates that the current single feature evaluation result is greater than the preset single evaluation result, indicating that the user is relatively comfortable and does not need to be reminded. For example, when the user is riding or driving a vehicle, the comprehensive feature evaluation result is 3 points, which is greater than the preset comprehensive evaluation result (2 points). The feature evaluation result of each item in the single feature evaluation result is also greater than the preset single evaluation result. It can be determined that the comprehensive feature evaluation result meets the first preset condition and the single feature evaluation result meets the second preset condition.

[0084] Step A2: If the comprehensive feature evaluation result meets the first preset condition, determine the single feature evaluation result that does not meet the second preset condition from the single feature evaluation results, and determine the prompt control information based on the single feature evaluation result that does not meet the second preset condition.

[0085] When the comprehensive feature evaluation result meets the first preset condition, it means that the user is relatively comfortable and does not need to be reminded. At this time, it is necessary to determine whether the single evaluation result meets the second preset condition. If there is a single feature evaluation result in the single feature evaluation result that does not meet the second preset condition, it means that there is characteristic information in the single feature evaluation result that is not greater than the preset single evaluation result. It can be further determined that there is information in the driving feature information and the in-vehicle environment information within the preset time period that makes the user feel uncomfortable. Assuming that it is determined that the single feature evaluation result of the maximum longitudinal acceleration information is not greater than the preset single evaluation result characteristic, a prompt control information is issued to the user based on the maximum longitudinal acceleration information. For example, the user is shown through a mobile phone app and a car screen: the score is low at this moment, pay attention to controlling the accelerator pedal opening and reducing the vehicle acceleration. At the same time, the prompt control information is broadcast to the user through a voice announcer.

[0086] Step A3: If the comprehensive feature evaluation result does not meet the first preset condition, the target feature evaluation result is determined from the single feature evaluation results, and the prompt control information is determined based on the target feature evaluation result and the comprehensive feature evaluation result; the target feature evaluation result includes the single feature evaluation result that does not meet the second preset condition among the single feature evaluation results, or the target feature evaluation result includes the single feature evaluation result with the lowest score among the single feature evaluation results.

[0087] When the comprehensive feature evaluation result does not meet the first preset condition, it indicates that the user may be feeling uncomfortable and needs to be reminded. In this case, a target feature evaluation result needs to be determined from the individual feature evaluation results, and prompt control information is determined based on the target feature evaluation result and the comprehensive feature evaluation result. Specifically, it is determined whether the individual evaluation results meet the second preset condition. If any individual feature evaluation result among the individual feature evaluation results does not meet the second preset condition, the prompt control information is determined based on the individual feature evaluation result that does not meet the second preset condition. If no individual feature evaluation result among the individual feature evaluation results does not meet the second preset condition, the prompt control information is determined based on the lowest-scoring individual feature evaluation result among the individual feature evaluation results. For example, while the user is riding or driving a vehicle, the comprehensive feature evaluation result is 2 points, which is no greater than the preset comprehensive evaluation result (2 points), but the feature evaluation results of each individual feature evaluation result are greater than the preset individual evaluation results. Assume that the lowest-scoring individual feature evaluation result determined from all individual feature evaluation results is 3 points, and the corresponding feature information is the maximum longitudinal acceleration information. Then, a prompt control information is issued to the user based on the maximum longitudinal acceleration information.

[0088] Through the above steps, the prompt control information can be accurately determined based on the individual feature scoring results and the comprehensive feature scoring results. The user can make timely adjustments based on the content of the prompt control information, thereby improving the user's comfort when driving or riding in the vehicle and further enhancing the user experience.

[0089] Step 405: Obtain feedback information from the user on the vehicle regarding the prompt control information.

[0090] The feedback information includes information indicating whether the user accepts the prompt control information and whether the user's physiological state is normal at the time. Specifically, while the user is riding or driving the vehicle, the OMS or DMS collects the user's heart rate variability information in real time. When the user's heart rate variability decreases significantly, it can be inferred that the user is experiencing discomfort. If the vehicle does not prompt the user at this time, it indicates that the comprehensive feature evaluation results and the individual feature evaluation results are incorrect. Feedback information is then determined based on the driving characteristics information, user characteristics information, driving characteristics information, in-vehicle environment information, and the user's heart rate variability at the time. While riding or driving the vehicle, after receiving the prompt control information, the user can decide whether to accept the prompt control information based on their physiological state and feelings. If the user does not accept the prompt control information, feedback information is determined based on the driving characteristics information, user characteristics information, driving characteristics information, in-vehicle environment information corresponding to the prompt control information, as well as the prompt control information that the user did not accept.

[0091] Step 406: Update the multi-task evaluation model based on the feedback information.

[0092] After the feedback information is determined, the feedback information is recorded and stored, and the feedback information is used for the next training of the multi-task evaluation model to update the multi-task evaluation model. Figure 7 This is the overall flow chart of the vehicle comfort evaluation method provided by the embodiment of the present invention. Figure 7 As shown, the individual feature evaluation results of driving characteristics and in-vehicle environment information, as well as the driving characteristics and user characteristics, are input into the multi-task evaluation model to obtain a comprehensive feature evaluation result. Based on the individual and comprehensive feature evaluation results, prompt control information is displayed to the user, and feedback information about the prompt control information is obtained. The multi-task evaluation model is further updated based on the feedback information.

[0093] The solution of an embodiment of the present invention obtains the vehicle's current driving characteristic information and user characteristic information, as well as the vehicle's driving characteristic information and in-vehicle environment information within a preset time period. The driving characteristic information and in-vehicle environment information within the preset time period are input into a time-series feature evaluation model of corresponding categories to obtain individual feature evaluation results corresponding to the driving characteristic information and in-vehicle environment information, respectively. The individual feature evaluation results are used to represent the user comfort evaluation results for the corresponding features. The individual feature evaluation results, driving characteristic information, and user characteristic information corresponding to the driving characteristic information and in-vehicle environment information are input into a multi-task evaluation model to obtain a comprehensive feature evaluation result for the vehicle, which is used to represent the user comfort evaluation results for the vehicle. Based on the comprehensive and individual feature evaluation results, prompt control information is determined and displayed. Feedback from users on the vehicle regarding the prompt control information is obtained, and the multi-task evaluation model is updated based on the feedback. The solution of this embodiment accurately determines prompt control information based on the individual feature scoring results and the comprehensive feature scoring results. Users can make timely adjustments based on the content of the prompt control information, thereby improving the user experience. Furthermore, in this solution, the comprehensive feature evaluation model can be updated based on user feedback, thereby improving the accuracy of the comprehensive feature evaluation results.

[0094] Figure 8 : is a schematic diagram of the structure of a vehicle comfort evaluation device provided by an embodiment of the present invention. The embodiment of the present invention provides a vehicle comfort evaluation device, the device comprising:

[0095] The information acquisition module 801 is used to obtain the current driving characteristic information and user characteristic information of the vehicle, and obtain the driving characteristic information and in-vehicle environment information of the vehicle within a preset time period;

[0096] A single result determination module 802 is configured to input the driving characteristic information and the in-vehicle environment information within the preset time period into a time series feature evaluation model of a corresponding category, and obtain single feature evaluation results corresponding to the driving characteristic information and the in-vehicle environment information, respectively, wherein the single feature evaluation results are used to represent user comfort evaluation results of the corresponding features;

[0097] The comprehensive result determination module 803 is used to input the single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the driving feature information and the user feature information into a multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle, and the comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle.

[0098] Optionally, the time series feature evaluation model is trained as follows:

[0099] Obtaining a first training data set, wherein each sample in the first training data set includes sample data and a sample label; wherein the sample data includes historical driving characteristic information and each type of characteristic information in the in-vehicle environment information, and the sample label includes a comfort evaluation result fed back by an experimenter for the sample data;

[0100] Inputting the sample data into an initial time series feature model for training to obtain output data of the initial time series feature model;

[0101] Determine a first loss function according to the output data and the corresponding sample label;

[0102] The model parameters of the initial time series feature model are reversely optimized based on the first loss function to obtain the time series feature evaluation model.

[0103] Optionally, the multi-task evaluation model is trained according to the following method:

[0104] Obtaining a second training data set, wherein each sample in the second training data set includes sample data and a sample label; wherein the sample data includes: comfort evaluation results corresponding to each type of feature information in historical driving feature information and in-vehicle environment information, and each item of driving feature information and user feature information within a historical period; and the sample label includes: comfort evaluation results fed back by an experimenter for the sample data;

[0105] Inputting the sample data into the initial task model for training to obtain output data of the initial task model;

[0106] Determine a second loss function based on the output data and the corresponding sample label;

[0107] The model parameters of the initial task model are reversely optimized based on the second loss function to obtain the multi-task evaluation model.

[0108] Optionally, the initial task model includes a shared layer and a personalized layer, and the reverse optimization of model parameters of the initial task model based on the second loss function includes:

[0109] The model parameters of the shared layer and the model parameters of the personalized layer of the initial task model are reversely optimized based on the second loss function.

[0110] Optionally, the device further includes:

[0111] An information display unit is used to determine prompt control information based on the comprehensive feature evaluation result and the single feature evaluation result, and to display the prompt control information.

[0112] Optionally, the information display unit is specifically used to:

[0113] Determining whether the comprehensive feature evaluation result meets a first preset condition, and determining whether the individual feature evaluation result meets a second preset condition;

[0114] If the comprehensive feature evaluation result satisfies the first preset condition, determining the individual feature evaluation results that do not satisfy the second preset condition from the individual feature evaluation results, and determining the prompt control information based on the individual feature evaluation results that do not satisfy the second preset condition;

[0115] If the comprehensive feature evaluation result does not meet the first preset condition, the target feature evaluation result is determined from the single feature evaluation results, and the prompt control information is determined based on the target feature evaluation result and the comprehensive feature evaluation result; the target feature evaluation result includes the single feature evaluation result among the single feature evaluation results that does not meet the second preset condition, or the target feature evaluation result includes the single feature evaluation result with the lowest score among the single feature evaluation results.

[0116] Optionally, the information display unit is further used to:

[0117] Obtaining feedback information from a user on the vehicle regarding the prompt control information;

[0118] The multi-task evaluation model is updated based on the feedback information.

[0119] The vehicle comfort evaluation device provided in the embodiment of the present invention can execute the vehicle comfort evaluation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0120] Figure 9 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, with reference to Figure 9 , which shows a structural diagram of a computer system 12 of an electronic device suitable for implementing an embodiment of the present invention. Figure 9 The electronic device shown is merely an example and should not limit the functionality and scope of use of the embodiments of the present invention. Components of the electronic device 12 may include, but are not limited to, one or more processors or processing units 16, a system memory 28, and a bus 18 that connects various system components (including the system memory 28 and the processing unit 16).

[0121] Bus 18 represents one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or a local bus using any of a variety of bus architectures. Examples of these architectures include, but are not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MAC) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnect (PCI) bus.

[0122] The electronic device 12 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the electronic device 12, including volatile and non-volatile media, removable and non-removable media.

[0123] The system memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. The electronic device 12 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the storage system 34 may be configured to read and write non-removable, non-volatile magnetic media ( Figure 9 Not shown, often called a "hard drive"). Although Figure 9 Not shown, a magnetic disk drive for reading and writing to a removable non-volatile magnetic disk (e.g., a "floppy disk"), and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0124] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which, or some combination thereof, may include an implementation of a network environment. Program modules 42 generally perform the functions and / or methodologies of the embodiments described herein.

[0125] The electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboards, pointing devices, displays 24, etc.), and may also communicate with one or more devices that enable a user to interact with the electronic device 12, and / or any device that enables the electronic device 12 to communicate with one or more other computing devices (e.g., a network card, a modem, etc.). Such communication may be performed through an input / output (I / O) interface 22. In addition, the electronic device 12 in this embodiment, the display 24 does not exist as an independent individual, but is embedded in the mirror surface. When the display surface of the display 24 is not displayed, the display surface of the display 24 and the mirror surface are visually integrated. Furthermore, the electronic device 12 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 through the bus 18. It should be understood that although Figure 9 Not shown, other hardware and / or software modules may be used in conjunction with the electronic device 12, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0126] The processing unit 16 executes various functional applications and vehicle comfort assessments by running programs stored in the system memory 28, such as implementing a vehicle comfort assessment method provided in an embodiment of the present invention: obtaining the vehicle's current driving characteristic information and user characteristic information, and obtaining the vehicle's driving characteristic information and in-vehicle environment information within a preset time period; inputting the driving characteristic information and in-vehicle environment information within the preset time period into a time series feature assessment model of a corresponding category to obtain single feature assessment results corresponding to the driving characteristic information and the in-vehicle environment information, respectively, and the single feature assessment results are used to represent the user comfort assessment results of the corresponding features; inputting the single feature assessment results corresponding to the driving characteristic information and the in-vehicle environment information, the driving characteristic information, and the user characteristic information into a multi-task assessment model to obtain a comprehensive feature assessment result of the vehicle, and the comprehensive feature assessment result is used to represent the user comfort assessment result of the vehicle.

[0127] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon. When executed by a processor, the program implements a vehicle comfort assessment method as provided in all embodiments of the present invention: obtaining current driving characteristic information and user characteristic information of a vehicle, and obtaining driving characteristic information and in-vehicle environment information of the vehicle within a preset time period; inputting the driving characteristic information and in-vehicle environment information within the preset time period into a time series feature assessment model of a corresponding category to obtain individual feature assessment results corresponding to the driving characteristic information and the in-vehicle environment information, respectively, wherein the individual feature assessment results are used to represent a user comfort assessment result for the corresponding feature; inputting the individual feature assessment results corresponding to the driving characteristic information and the in-vehicle environment information, the driving characteristic information, and the user characteristic information into a multi-task assessment model to obtain a comprehensive feature assessment result of the vehicle, wherein the comprehensive feature assessment result is used to represent a user comfort assessment result of the vehicle. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: an electrical connection having one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0128] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0129] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0130] Computer program code for performing the operations of the present invention can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0131] Note that the above are only preferred embodiments of the present invention and the technical principles employed. Those skilled in the art will appreciate that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments, and substitutions are possible for those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments and may include many other equivalent embodiments without departing from the scope of the present invention. The scope of the present invention is determined by the scope of the appended claims.

Claims

1. A vehicle comfort evaluation method, characterized in that: The method comprises: Obtaining the vehicle's current driving characteristic information and user characteristic information, and obtaining the vehicle's driving characteristic information and in-vehicle environment information within a preset time period; the vehicle's current driving characteristic information is information generated by the user while driving or riding the vehicle, including the current time, driving duration, and driving history information; the vehicle's driving characteristic information includes maximum longitudinal acceleration, maximum longitudinal deceleration, maximum longitudinal impact degree, maximum lateral acceleration, and acceleration and deceleration frequency information within the preset time period; The driving feature information and the in-vehicle environment information within the preset time period are input into the time series feature evaluation model of the corresponding category to obtain the single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information respectively, and the single feature evaluation results are used to represent the user comfort evaluation results of the corresponding features; the time series feature evaluation model is trained in the following manner: a first training data set is obtained, and each sample in the first training data set includes sample data and a sample label; wherein the sample data includes each category of feature information in the historical driving feature information and the in-vehicle environment information, and the sample label includes the comfort evaluation results fed back by the experimenter for the sample data; the sample data is input into the initial time series feature model for training to obtain the output data of the initial time series feature model; a first loss function is determined according to the output data and the corresponding sample label; and the model parameters of the initial time series feature model are reversely optimized based on the first loss function to obtain the time series feature evaluation model; The single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the driving feature information and the user feature information are input into a multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle, and the comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle; the multi-task evaluation model is trained according to the following method: a second training data set is obtained, and each sample in the second training data set includes sample data and sample labels; wherein the sample data includes: the comfort evaluation results corresponding to each type of feature information in the historical driving feature information and the in-vehicle environment information, and each item of information in the driving feature information and user feature information within a historical period; the sample labels include: the comfort evaluation results fed back by the experimenter for the sample data; the sample data is input into the initial task model for training to obtain the output data of the initial task model; a second loss function is determined according to the output data and the corresponding sample labels; the model parameters of the initial task model are reversely optimized based on the second loss function to obtain the multi-task evaluation model.

2. The method according to claim 1, characterized in that The initial task model includes a shared layer and a personalized layer, and the reverse optimization of model parameters of the initial task model based on the second loss function includes: The model parameters of the shared layer and the model parameters of the personalized layer of the initial task model are reversely optimized based on the second loss function.

3. The method according to claim 1, characterized in that The method further comprises: Prompt control information is determined based on the comprehensive feature evaluation result and the individual feature evaluation result, and the prompt control information is displayed.

4. The method according to claim 3, wherein: The determining of prompt control information based on the comprehensive feature evaluation result and the individual feature evaluation result includes: Determining whether the comprehensive feature evaluation result meets a first preset condition, and determining whether the individual feature evaluation result meets a second preset condition; If the comprehensive feature evaluation result satisfies the first preset condition, determining the individual feature evaluation results that do not satisfy the second preset condition from the individual feature evaluation results, and determining the prompt control information based on the individual feature evaluation results that do not satisfy the second preset condition; If the comprehensive feature evaluation result does not meet the first preset condition, the target feature evaluation result is determined from the single feature evaluation results, and the prompt control information is determined based on the target feature evaluation result and the comprehensive feature evaluation result; the target feature evaluation result includes the single feature evaluation result among the single feature evaluation results that does not meet the second preset condition, or the target feature evaluation result includes the single feature evaluation result with the lowest score among the single feature evaluation results.

5. The method according to claim 3, characterized in that The method further comprises: Obtaining feedback information from a user on the vehicle regarding the prompt control information; The multi-task evaluation model is updated based on the feedback information.

6. A vehicle comfort evaluation device, characterized in that: The device comprises: an information acquisition module for acquiring the vehicle's current driving characteristic information and user characteristic information, and acquiring the vehicle's driving characteristic information and in-vehicle environment information over a preset time period; the vehicle's current driving characteristic information is information generated by the user while driving or riding the vehicle, including the current time, driving duration, and driving history; the vehicle's driving characteristic information includes maximum longitudinal acceleration, maximum longitudinal deceleration, maximum longitudinal impact, maximum lateral acceleration, and acceleration / deceleration frequency information over the preset time period; A single result determination module is used to input the driving feature information and the in-vehicle environment information within the preset time length into the time series feature evaluation model of the corresponding category, and obtain the single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information respectively, and the single feature evaluation results are used to represent the user comfort evaluation results of the corresponding features, and the single feature evaluation results are used to represent the user comfort evaluation results of the corresponding features; the time series feature evaluation model is trained in the following manner: a first training data set is obtained, and each sample in the first training data set includes sample data and sample labels; wherein the sample data includes historical driving feature information and each category of feature information in the in-vehicle environment information, and the sample labels include the comfort evaluation results fed back by the experimenter for the sample data; the sample data is input into the initial time series feature model for training to obtain the output data of the initial time series feature model; a first loss function is determined according to the output data and the corresponding sample labels; the model parameters of the initial time series feature model are reversely optimized based on the first loss function to obtain the time series feature evaluation model; A comprehensive result determination module is used to input the single feature evaluation results corresponding to the driving feature information and the in-vehicle environment information, the driving feature information and the user feature information into a multi-task evaluation model to obtain a comprehensive feature evaluation result of the vehicle, and the comprehensive feature evaluation result is used to represent the user comfort evaluation result of the vehicle; the multi-task evaluation model is trained according to the following method: a second training data set is obtained, and each sample in the second training data set includes sample data and sample labels; wherein the sample data includes: the comfort evaluation results corresponding to each type of feature information in the historical driving feature information and the in-vehicle environment information, and each item of information in the driving feature information and user feature information within a historical period; the sample labels include: the comfort evaluation results fed back by the experimenter for the sample data; the sample data is input into the initial task model for training to obtain the output data of the initial task model; a second loss function is determined based on the output data and the corresponding sample labels; the model parameters of the initial task model are reversely optimized based on the second loss function to obtain the multi-task evaluation model.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the vehicle comfort evaluation method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the vehicle comfort evaluation method according to any one of claims 1 to 5 is implemented.

Images