Vehicle control method and device, vehicle and storage medium

Abstract

The invention relates to a vehicle control method and device, a vehicle and a storage medium. The method comprises the steps that current environment data and user feature data are obtained; vehicle use scene features are constructed based on the current environment data and the user feature data; determining a correlation coefficient between the vehicle using scene feature and a vehicle control element of a vehicle; based on the correlation coefficient, determining a plurality of target vehicle control elements matched with the vehicle use scene features and corresponding control parameters; and in response to a confirmation operation for the plurality of target vehicle control elements, controlling the plurality of target vehicle control elements based on the control parameters. According to the method and the device, a plurality of target vehicle control elements can be accurately recommended, so that a user can simultaneously start a plurality of vehicle control elements at one time in different vehicle use scenes, and the vehicle use experience of the user is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle interactive control, and in particular to a vehicle control method, a vehicle control device, a vehicle and a computer-readable storage medium. Background Art

[0002] With the development of electronic information technology, vehicle body control is becoming more and more convenient. Current technology can control the entire vehicle through the screen of the vehicle driver's seat, including air conditioning temperature adjustment, window position adjustment, seat adjustment, ambient light adjustment, front and rear wiper adjustment, defogger and defrost adjustment, and rearview mirror adjustment, etc. It can even be controlled remotely and the vehicle status can be preset in advance. Due to the large number of functions, an initial scene needs to be provided for users to edit the scene on this basis. At present, there is technology that generates recommended vehicle control scenes based on user habits. After the vehicle control scene is generated, the preset settings can be awakened with one click. However, current technology usually only considers the relationship between a single scene element and a single vehicle control component; the recommended vehicle control component cannot meet user needs. Summary of the invention

[0003] One of the objects of the present invention is to provide a vehicle control method to solve the problem that the prior art only considers the relationship between a single scene element and a single vehicle control component; the recommended vehicle control component cannot meet user needs; the second object is to provide a vehicle control device; the third object is to provide a vehicle; and the fourth object is to provide a computer-readable storage medium.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] A vehicle control method, comprising:

[0006] Obtain current environment data and user feature data;

[0007] Constructing a vehicle use scenario feature based on the current environment data and the user feature data;

[0008] Determining a correlation coefficient between the vehicle use scenario characteristics and a vehicle control element of the vehicle;

[0009] Based on the correlation coefficient, determining a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics;

[0010] In response to the confirmation operation on the plurality of target vehicle control components, the plurality of target vehicle control components are controlled based on the control parameter.

[0011] Furthermore, the method further comprises:

[0012] Obtain historical environment data and historical user data;

[0013] A Pearson one-dimensional convolution model is constructed based on the historical environment data and the historical user data.

[0014] Furthermore, the step of determining the correlation coefficient between the vehicle usage scenario feature and the vehicle control element of the vehicle includes:

[0015] The vehicle usage scenario features are input into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output a correlation coefficient based on the vehicle usage scenario features.

[0016] Furthermore, the step of constructing the vehicle use scenario characteristics based on the current environment data and the user characteristic data includes:

[0017] Converting the current environment data into a first characteristic character;

[0018] Converting the user characteristic data into a second characteristic character;

[0019] The first characteristic character and the second characteristic character are combined to generate a vehicle use scenario feature.

[0020] Furthermore, the step of determining a plurality of target vehicle control components and corresponding control parameters matching the vehicle usage scenario characteristics based on the correlation coefficient includes:

[0021] Determining vehicle control components of multiple vehicles whose correlation coefficients are greater than a preset threshold as target vehicle control components;

[0022] Determine the corresponding control parameters of the target vehicle control component.

[0023] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the historical environment data and the historical user data includes:

[0024] Constructing training samples based on the historical environment data and the historical user data;

[0025] Constructing a Pearson coefficient matrix based on the training samples;

[0026] A Pearson one-dimensional convolution model is constructed based on the training samples and the Pearson coefficient matrix fitting.

[0027] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the training samples and the Pearson coefficient matrix fitting includes:

[0028] Multiplying the training sample and the Pearson coefficient matrix to obtain a feature vector;

[0029] A Pearson one-dimensional convolution model is constructed based on the feature vector and a preset one-dimensional residual convolution network.

[0030] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network includes:

[0031] Inputting the feature vector into a preset one-dimensional residual convolutional network for convolution processing to obtain an intermediate one-dimensional residual convolutional network;

[0032] The intermediate one-dimensional residual convolutional network is skipped based on a preset residual block to obtain a Pearson one-dimensional convolutional model.

[0033] Furthermore, the method further comprises:

[0034] The current environment data and the user characteristic data are preprocessed.

[0035] A vehicle control device, comprising:

[0036] The current acquisition module is used to obtain current environment data and user feature data;

[0037] A construction module, used to construct a vehicle use scenario feature based on the current environment data and the user feature data;

[0038] A first determination module is used to determine a correlation coefficient between the vehicle use scenario feature and a vehicle control element of the vehicle;

[0039] A second determination module, configured to determine, based on the correlation coefficient, a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics;

[0040] A response module is used to control the plurality of vehicle control components based on the control parameters in response to a confirmation operation on the plurality of vehicle control components.

[0041] A vehicle comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, wherein the computer program implements the steps of the vehicle control method as described above when executed by the processor.

[0042] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vehicle control method described above are implemented.

[0043] Beneficial effects of the present invention:

[0044] (1) The vehicle usage scenario characteristics are constructed by identifying the current environmental data and user characteristic data, and multiple related target vehicle control components and corresponding control parameters are determined based on the correlation coefficient. The determination process does not involve human experience knowledge that may have subjective factors, and recommends specific parameters for each vehicle control component, so that users can enable multiple vehicle control components at the same time in different vehicle usage scenarios, thereby improving the user's vehicle experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart of steps of an embodiment of a vehicle control method of the present invention;

[0046] Figure 2 A flowchart of another vehicle control method embodiment of the present invention;

[0047] Figure 3 It is a schematic diagram of the characteristic modules of the model in the present invention;

[0048] Figure 4 It is a schematic diagram of the prediction module of the model in the present invention;

[0049] Figure 5 A flowchart of an exemplary vehicle control method of the present invention;

[0050] Figure 6 is a structural block diagram of an embodiment of a vehicle control device of the present invention;

[0051] Figure 7 A schematic diagram of a processor and a storage medium of a vehicle embodiment of the present invention;

[0052] Figure 8 A schematic diagram of an embodiment of a computer-readable storage medium of the present invention. DETAILED DESCRIPTION

[0053] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0054] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0055] The current vehicle control scene recommendation method often does not take current environmental factors into consideration, or involves very few controls. It usually artificially classifies users in advance according to preset fixed simple scenes, determines the user category by user characteristics, and recommends corresponding vehicle control parameters to users when the corresponding scene is triggered. This method cannot be applied to situations where the external environment is more complex, and can often only make independent recommendations for fewer vehicle control parameters based on fixed scenes. In reality, the external environment is often more complex, and there is correlation between various controls and various scene elements. For example, on a rainy day, the wipers and air conditioning dehumidifier may be turned on at the same time, and the windows will be closed. The settings that users will make will be very different in the hot summer afternoon at 29 degrees and in the cool autumn sky at 29 degrees. Previous inventions did not consider the relationship between vehicle control components and scene elements. In order to solve the above problems, the present invention provides a vehicle control method, a vehicle control device, a vehicle and a computer-readable storage medium. The external complex environment and user habits are jointly constructed and modeled. The scene features and user usage data are analyzed for correlation. The analysis results are involved in model construction. The Pearson one-dimensional convolution model is used to learn environmental features, user habit features, and further learn the relationship between the two. The final result is predicted using a fully connected network. The user's historical data and environmental data are input into the model, and the vehicle control scene data considering the current comprehensive environment will be pushed to the user. The present invention does not preset scene information and user information. By learning and modeling scene features and user features, it does not involve human experience knowledge that may have subjective factors. Finally, based on the historical usage data and scenes of the target user, the specific parameters of each vehicle control are recommended.

[0056] Reference Figure 1 , shows a flow chart of steps of an embodiment of a vehicle control method of the present invention. The vehicle control method may specifically include the following steps:

[0057] Step 101, obtaining current environment data and user characteristic data;

[0058] In the embodiment of the present invention, the current environment data and user characteristic data may be first obtained. The user characteristic data may include various identity characteristics of the user and characteristic data such as vehicle use habits. The current environment data is the current environment data inside and outside the vehicle.

[0059] Step 102, constructing a vehicle use scenario feature based on the current environment data and the user feature data;

[0060] The current environment data is combined with the user feature data to construct the user's current vehicle usage scenario feature. The vehicle usage scenario feature is the data that characterizes the scenario in which the vehicle is currently used.

[0061] Step 103, determining a correlation coefficient between the vehicle use scenario feature and a vehicle control element of the vehicle;

[0062] The correlation coefficient between the vehicle scene feature and each vehicle control element of the vehicle is calculated. The correlation coefficient may be a Pearson correlation coefficient, a Spearman rank correlation coefficient, a Kendall rank correlation coefficient, or the like.

[0063] Among them, the Pearson correlation coefficient characterizes the degree of linear correlation between the vehicle-use scenario characteristics and the vehicle's vehicle control components. The Pearson correlation coefficient can be obtained by calculating the deviation of the vehicle-use scenario characteristics and the vehicle's vehicle control components from their respective average values, then multiplying and summing these deviations, and finally dividing by the number of samples. When the value of the Pearson correlation coefficient is 1, it means that the vehicle-use scenario characteristics and the vehicle's vehicle control components are completely positively correlated; when the value is -1, it means that it is completely negatively correlated; when the value is 0, it means that there is no linear correlation.

[0064] The Spearman rank correlation coefficient measures the monotonic relationship between the vehicle usage scenario characteristics and the vehicle's vehicle control components. Sort the original data from small to large, and assign a rank to each data. If there are multiple data that are equal, the corresponding rank is the average of these data. Then, calculate the Pearson correlation coefficient between the ranks, that is, the Spearman correlation coefficient. The value range of the Spearman correlation coefficient is between -1 and 1. When the Spearman rank correlation coefficient is equal to 1, it means that there is a completely monotonic positive correlation between the vehicle usage scenario characteristics and the vehicle's vehicle control components; when the Spearman rank correlation coefficient is equal to -1, it means that there is a completely monotonic negative correlation between the vehicle usage scenario characteristics and the vehicle's vehicle control components; when the Spearman rank correlation coefficient is equal to 0, it means that there is no monotonic relationship between the vehicle usage scenario characteristics and the vehicle control components.

[0065] Kendall's rank correlation coefficient is a statistic used to characterize the ordinal association between vehicle usage scenario features and vehicle control components. The vehicle usage scenario features and the levels of the vehicle control components are compared according to their magnitude, and then the number of logarithms of the vehicle usage scenario features and the vehicle control components with the same level is counted. The Kendall rank correlation coefficient is then obtained by dividing the number of logarithms of the same level by the total number of logarithms. Vehicle usage scenario features and vehicle control components.

[0066] Step 104, determining a plurality of target vehicle control components and corresponding control parameters matching the vehicle usage scenario characteristics based on the correlation coefficient;

[0067] According to the correlation coefficient, the vehicle control components associated with the vehicle use scenario characteristics are determined from all the vehicle control components, that is, multiple target vehicle control components and corresponding control parameters. The control parameters are the vehicle control component control parameters that control the corresponding target vehicle control components to meet the vehicle use scenario characteristics. The vehicle control components are controls that control a certain function of the vehicle, such as air conditioning controls that control the temperature, door controls that control the opening and closing of doors, and so on.

[0068] Step 105 : In response to the confirmation operation on the plurality of target vehicle control components, controlling the plurality of target vehicle control components based on the control parameters.

[0069] After obtaining multiple target vehicle control components and corresponding control parameters, the relevant multiple target vehicle control components can be recommended to the user, and the user can confirm, cancel or change the multiple target vehicle control components through various operations, such as clicking to confirm. Generate confirmation operations for multiple target vehicle control components. In response to the confirmation operations for multiple target vehicle control components, the multiple target vehicle control components can be controlled to reach corresponding states based on the control parameters, so that multiple vehicle control components can be turned on synchronously when the user uses the vehicle, which is convenient for the user's operation and avoids the user from controlling the vehicle control components one by one.

[0070] The embodiment of the present invention obtains current environment data and user characteristic data; constructs vehicle scene characteristics based on the current environment data and the user characteristic data; determines the correlation coefficient between the vehicle scene characteristics and the vehicle control components of the vehicle; determines multiple target vehicle control components and corresponding control parameters that match the vehicle scene characteristics based on the correlation coefficient; and controls the multiple target vehicle control components based on the control parameters in response to confirmation operations on the multiple target vehicle control components. The vehicle scene characteristics are constructed by identifying the current environment data and user characteristic data, and multiple related target vehicle control components and corresponding control parameters are determined based on the correlation coefficient. The determination process does not involve human experience knowledge that may have subjective factors, and recommends specific parameters for each vehicle control component, so that users can enable multiple vehicle control components at the same time in different vehicle use scenarios, thereby improving the user's vehicle experience.

[0071] Reference Figure 2 , shows a flow chart of steps of another vehicle control method embodiment of the present invention. The vehicle control method may specifically include the following steps:

[0072] Step 201, obtaining historical environment data and historical user data;

[0073] First, based on the data recorded in the vehicle storage, historical environment data and historical user data are obtained. The historical environment data refers to the environment inside and outside the vehicle when the user used the vehicle in the past. The historical user data refers to the user feature data corresponding to the user's use of the vehicle in the past.

[0074] Step 202, constructing a Pearson one-dimensional convolution model based on the historical environment data and the historical user data;

[0075] The historical environment data and the historical user data can be used together for model training, verification, and testing to obtain a Pearson one-dimensional convolution model. The Pearson one-dimensional convolution model is a one-dimensional convolution model based on the Pearson coefficient matrix as the input layer.

[0076] In an optional embodiment of the present invention, the step of constructing a Pearson one-dimensional convolution model based on the historical environment data and the historical user data includes: constructing training samples based on the historical environment data and the historical user data; constructing a Pearson coefficient matrix based on the training samples; and constructing a Pearson one-dimensional convolution model based on fitting the training samples and the Pearson coefficient matrix.

[0077] For the construction of the Pearson one-dimensional convolution model, the historical environment data and historical user data can be first converted into corresponding feature vectors to construct training samples, and then the Pearson coefficient matrix is ​​constructed based on these training samples as matrix elements. The training samples and the Pearson coefficient matrix are then fitted together to generate the Pearson one-dimensional convolution model.

[0078] Specifically, the step of constructing a Pearson one-dimensional convolution model based on the training sample and the Pearson coefficient matrix fitting includes: multiplying the training sample and the Pearson coefficient matrix to obtain a feature vector; and constructing a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network.

[0079] The training sample and the Pearson coefficient matrix can be multiplied to obtain a feature vector. Then the feature vector is used as input to the preset one-dimensional residual convolution network, and the preset one-dimensional residual convolution network is trained until the convergence requirements are met to obtain the Pearson one-dimensional convolution model. Among them, the preset one-dimensional residual convolution network is a pre-acquired initial one-dimensional residual convolution network, which is a special convolutional neural network structure that combines the characteristics of one-dimensional convolution and residual connection. The one-dimensional residual convolution network may include one-dimensional convolution and residual connection. One-dimensional convolution means that the convolution kernel (vector) slides only in one dimension, and its main purpose is to capture features within a local range (convolution kernel size). The larger the convolution kernel, the more contextual information can be captured. The calculation of one-dimensional convolution is simpler than that of two-dimensional convolution, which can reduce the amount of calculation of the entire network. At the same time, the receptive field of the one-dimensional convolution kernel can cover the entire frequency range, so it can recognize richer patterns within the frequency range. Residual connection is used to solve the gradient vanishing and gradient exploding problems in the deep network training process, thereby improving the performance of the model. In the residual network, the input data can directly skip one or more layers and connect to the following layers. The one-dimensional residual convolutional network can not only capture the features within the local range, but also solve the gradient problem in the deep network training process. And compared with the two-dimensional convolution, the one-dimensional convolution has less computational complexity, so the one-dimensional residual convolutional network has higher computational efficiency.

[0080] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network includes: inputting the feature vector into a preset one-dimensional residual convolution network for convolution processing to obtain an intermediate one-dimensional residual convolution network; and performing jump processing on the intermediate one-dimensional residual convolution network based on a preset residual block to obtain a Pearson one-dimensional convolution model.

[0081] Correspondingly, the specific training process of the Pearson one-dimensional convolution model can input the feature vector into a preset one-dimensional residual convolution network for convolution processing to obtain an intermediate one-dimensional residual convolution network; based on the preset residual block, the intermediate one-dimensional residual convolution network is jump-connected to obtain the Pearson one-dimensional convolution model.

[0082] In summary, for example, refer to Figure 3 and Figure 4 The model is divided into two modules: feature module and prediction module. In the feature module, the Pearson coefficient matrix is ​​used as the input layer. The original feature vector x is input into the model and first enters the input layer Pearson matrix P, and then multiplied by the Pearson matrix y = P·x,y i =p i1 x1+p i2 ·x2+…p ii ·x i+…, introduce Pearson correlation knowledge for features, and then enter the one-dimensional residual convolutional network, convolve with multiple convolutional kernels, learn environmental features and user features, and the relationship between them. The one-dimensional residual convolutional network consists of multiple one-dimensional convolutional layers, pooling layers, and activation function layers, and introduces residual blocks for jump processing to prevent gradient disappearance and explosion problems, and effectively extract features from one-dimensional data. The classification module inputs the processed feature vector into the fully connected network for control parameter prediction.

[0083] The following are the basic components of the model:

[0084] Pearson input layer: also known as the Pearson coefficient matrix, which linearly changes the eigenvector to supplement the relationship between each environmental element and control element.

[0085] Convolutional layer: The convolutional layer slides the convolution filter on the input feature vector, moving one step at a time. For each sub-region that slides over, the filter is multiplied by the one-dimensional signal in the region, and then the sum is obtained to obtain an output value. In this way, a new feature vector can be obtained, which is shorter than the original sequence, but the width depends on the number of convolution kernels.

[0086] Activation function: used to introduce nonlinearity to help the network learn more complex patterns. The output of a neuron is usually linear, that is, the weight multiplied by the input plus the bias. By applying an activation function to the output of each neuron, we can transform these linear combinations into nonlinear mappings, enabling the network to learn complex decision boundaries.

[0087] Residual block: Contains one or more convolutional layers and corresponding activation functions. The goal of each residual block is to learn the "residual" mapping from input to output, that is, the difference between the target signal and the input signal.

[0088] Skip connection: The input data is directly added to the result processed by the residual block and then passed to the next layer. Skip connection allows the network to directly propagate gradients, helping to solve the gradient disappearance problem in deep neural network training.

[0089] Batch Normalization: is an auxiliary layer for neural networks that normalizes the input in each training batch, improving the training speed and stability of deep neural networks and enhancing the generalization ability of the model.

[0090] Pooling layer: samples the input data, changes the feature dimension, and improves the generalization ability of the model by extracting key features.

[0091] Fully connected layer: Each node in each layer is connected to all nodes in the previous layer and linearly transformed through the weight matrix. The fully connected layer contains a set of weights and a set of biases. For each node in the input layer, the fully connected layer multiplies the input value of each node by the corresponding weight and adds a bias.

[0092] Step 203, obtaining current environment data and user characteristic data;

[0093] After the model training is completed, when the user is using the car, the current environment data and user feature data can be obtained.

[0094] Step 204, preprocessing the current environment data and the user characteristic data;

[0095] Preprocess the current environment data and user feature data and delete abnormal data to make the recommendation results more accurate.

[0096] Step 205, constructing a vehicle use scenario feature based on the current environment data and the user feature data;

[0097] The current environment data can be integrated with the user characteristic data to determine the characteristics of the vehicle usage scenario.

[0098] Specifically, the step of constructing a vehicle usage scenario feature based on the current environment data and the user characteristic data includes: converting the current environment data into a first characteristic character; converting the user characteristic data into a second characteristic character; and combining the first characteristic character and the second characteristic character to generate a vehicle usage scenario feature.

[0099] The first characteristic character may be a character pre-set to characterize the current environmental data, and may be represented by a number. Accordingly, the second characteristic character may be a character pre-set to characterize the user characteristic data, and may also be represented by a number. In the environmental data, weather, season and time are divided into several fixed states, and each state is represented by a specific number. For example, sunny, cloudy, cloudy, light rain, moderate rain, heavy rain, light snow, heavy snow, hail, and heavy fog are represented by 0, 1, 2, 3, 4, 5, 6, 7, 8, and 9 respectively; early spring, spring, late spring, early summer, summer, late summer, early autumn, autumn, late autumn, early winter, winter, and late winter are represented by 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11 respectively; 0-6 o'clock, 7-11 o'clock, 12-16 o'clock, 17-20 o'clock, and 20-23 o'clock are represented by 0, 1, 2, 3, and 4 respectively; temperature, humidity, and visibility are represented by specific numbers. In the user characteristic data, according to the use of the vehicle, the front and rear wiper switches, defogger and defrost switches, ambient light status, etc. are divided into several fixed states, each of which is represented by a specific number, for example, on and off are represented by 0 and 1, while the air conditioning temperature, window opening degree, seat folding degree, rearview mirror folding degree, etc. are represented by actual values.

[0100] The current environment data can be converted into the first characteristic character, and the user characteristic data can be converted into the second characteristic character; the environment data and the user characteristic data are quantified. Then the first characteristic character and the second characteristic character are merged to form a set, which is the vehicle scene feature. For example, in summer, the temperature is 38 degrees, the humidity is 50%, the visibility is 24 kilometers, it is sunny, at 13:10, the vehicle's windows are closed, the air conditioner is turned on and set to 26 degrees, the defogger and defrost are turned off, the front and rear wipers are turned off, the ambient light is turned off, the rearview mirror is fully opened, and the seat is tilted back 15 degrees. The vehicle scene feature is [0,4,38,50,2,24,0,0,0,0,26,0,15,0].

[0101] Step 206, determining a correlation coefficient between the vehicle usage scenario feature and a vehicle control element of the vehicle;

[0102] The vehicle usage scenario feature is correlated with each vehicle control element to determine the correlation coefficient between each vehicle control element and the vehicle usage scenario feature. The correlation coefficient may be a Pearson correlation coefficient.

[0103] Specifically, the step of determining the correlation coefficient between the vehicle usage scenario characteristics and the vehicle control components of the vehicle includes: inputting the vehicle usage scenario characteristics into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output the correlation coefficient based on the vehicle usage scenario characteristics.

[0104] For the calculation of the correlation coefficient, the vehicle usage scenario characteristics may be input into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output the Pearson correlation coefficient as the correlation coefficient based on the vehicle usage scenario characteristics.

[0105] Step 207, based on the correlation coefficient, determining a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics;

[0106] Based on the correlation coefficient of each vehicle control component, multiple target vehicle control components and corresponding control parameters matching the vehicle usage scenario characteristics are determined from all the vehicle control components.

[0107] Specifically, the step of determining multiple target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics based on the correlation coefficient includes: determining the vehicle control components of multiple vehicles whose correlation coefficients are greater than a preset threshold as target vehicle control components; and determining the corresponding control parameters of the target vehicle control components.

[0108] For the screening of target vehicle control components and corresponding control parameters, the correlation coefficient of each vehicle control component can be first compared with a preset threshold based on the correlation coefficient of each vehicle control component, and the vehicle control components with correlation coefficients greater than the preset threshold are determined as target vehicle control components, and the corresponding control parameters of these target vehicle control components are determined at the same time.

[0109] Step 208 : In response to the confirmation operation on the plurality of target vehicle control components, control the plurality of target vehicle control components based on the control parameters.

[0110] When the user sees the target vehicle control element and the corresponding control parameter, he can operate to make the vehicle end perform the corresponding control. In response to the confirmation operation for multiple target vehicle control elements, multiple target vehicle control elements can be controlled based on the control parameter.

[0111] The embodiment of the present invention collects the user's current scene data and user feature data and performs preprocessing and feature construction, inputs historical environment data and historical user data features into the model, and the model predicts the parameters of all vehicle control components that match the comprehensive environment, calculates the recommended data and pushes it to the user terminal. After receiving the recommended scene arrangement, the user terminal can wake up with one click and quickly set the relevant controls to improve the user experience.

[0112] In order to make the implementation process of the embodiments of the present invention clear to those skilled in the art, refer to Figure 5 , the following is an example:

[0113] Data collection: In order to make the prediction results more in line with the current environment and meet user expectations, it is necessary to collect a large amount of user data and environmental data to build a data set. User data includes air conditioning temperature, window position, seat position, ambient light, front and rear wipers, defogger, defrost and rearview mirror position, etc. Environmental data includes the current season, time, temperature, humidity, weather, etc. The seasons are divided into spring, summer, autumn and winter, and the time, temperature and humidity are divided into multiple intervals. The weather includes rain, strong wind, sunny, snowy, cloudy, etc. Data collection begins after the vehicle is started, and data is recorded when the environmental state changes. A sample is the longest-lasting user operation on the above controls within a period of time under a certain environmental state.

[0114] Data preprocessing: Delete abnormal data. Data in extreme weather conditions are considered abnormal. Values ​​outside the normal range are considered abnormal. The window position, seat position, ambient light, front and rear wipers, defogger, defrost or rearview mirror position are considered abnormal if they are maintained for too short a time. This is considered an emergency and is considered abnormal. Data from a vehicle that has not been driven for a long time is also considered abnormal.

[0115] Feature construction: The original data is processed by features, and the environment status, control switch and opening degree are described by numbers. The features are standardized, and the environment features are spliced ​​with the user features to obtain the final features, which are divided into training sets, validation sets and test sets. The features are standardized and the training sets, validation sets and test sets are divided in a ratio of 7:2:1.

[0116] Feature correlation analysis: In order to fully explore the relationship between environmental factors and user behavior, the association between various vehicle control components, and improve the accuracy of the model, we introduced feature correlation knowledge into the model. We conducted correlation analysis on environmental features and user features, analyzed the correlation between various environmental elements and various vehicle control components, and calculated the Pearson correlation matrix of environmental features and user features. The Pearson correlation matrix is ​​a square matrix used to show the degree of linear correlation between a set of variables. Each element in the matrix represents the Pearson correlation coefficient between the corresponding two variables. The rows and columns of the matrix correspond to different variables in the data set, and the cells inside the matrix represent the correlation between each pair of variables. The Pearson correlation coefficient is a statistic used to measure the degree of linear correlation between two continuous variables, with a range of values ​​between -1 and 1. When the value is 1, it means that the two variables are completely positively correlated; when the value is -1, it means that the correlation is completely negative; and a value close to 0 means that there is no obvious linear relationship. Its correlation degree is used as a weight to participate in model construction, providing environmental element-related information and vehicle control component-related information for the prediction of the user's vehicle control setting behavior.

[0117] Model construction: Construct a Pearson one-dimensional convolution model, use the Pearson matrix as the input layer, and then build multiple one-dimensional convolution layers, pooling layers, activation function layers, and fully connected layers. Introduce residual blocks and jump connections to solve the gradient vanishing and gradient exploding problems in deep networks. The model is divided into two modules: feature module and prediction module. In the feature module, use the correlation matrix to construct the input layer, introduce environmental correlation knowledge and control correlation knowledge, and in the classification module, use the fully connected layer network for feature mapping and output the predicted vehicle control component parameters. When building the model, we use the Pearson correlation coefficient matrix as the input layer, change the feature vector space, and introduce correlation knowledge into the model.

[0118] Model selection: Set different numbers of convolution layers, different sizes of convolution kernels, different activation functions, different numbers of fully connected layers, several sets of learning rates, and different gradient descent functions. Train the models on the training set, then verify the model effects on the validation set, calculate the accuracy on the validation set, and the model with the highest accuracy is the final model.

[0119] Recommended vehicle control data: In the same way, the user's historical data and scene data are collected and preprocessed and feature constructed. The historical environment data and user data features are input into the model. The model predicts the parameters of all vehicle control components that match the comprehensive environment, calculates the recommended data and pushes it to the user terminal. After receiving the recommended scene arrangement, the user terminal can wake up with one click and quickly set the relevant controls.

[0120] It should be noted that, for the sake of simplicity, the method embodiments are described as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0121] Reference Figure 6 , shows a structural block diagram of an embodiment of a vehicle control device of the present invention, and the vehicle control device may specifically include the following sub-steps:

[0122] The current acquisition module 601 is used to acquire current environment data and user feature data;

[0123] A construction module 602 is used to construct a vehicle use scenario feature based on the current environment data and the user feature data;

[0124] A first determination module 603 is used to determine a correlation coefficient between the vehicle use scenario feature and a vehicle control element of the vehicle;

[0125] A second determination module 604 is used to determine a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics based on the correlation coefficient;

[0126] The response module 605 is configured to control the plurality of vehicle control components based on the control parameters in response to the confirmation operation on the plurality of vehicle control components.

[0127] In an optional embodiment of the present invention, the device further comprises:

[0128] A historical acquisition module is used to acquire historical environment data and historical user data;

[0129] A training module is used to build a Pearson one-dimensional convolution model based on the historical environment data and the historical user data.

[0130] In an optional embodiment of the present invention, the first determining module 603 includes:

[0131] The model using module is used to input the vehicle usage scenario features into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output a correlation coefficient based on the vehicle usage scenario features.

[0132] In an optional embodiment of the present invention, the construction module 602 includes:

[0133] A first conversion submodule, used for converting the current environment data into a first characteristic character;

[0134] A second conversion submodule, used for converting the user characteristic data into a second characteristic character;

[0135] The combination submodule is used to combine the first characteristic character and the second characteristic character to generate a vehicle use scenario feature.

[0136] In an optional embodiment of the present invention, the second determining module 604 includes:

[0137] A first determination submodule is used to determine that the vehicle control components of multiple vehicles whose correlation coefficients are greater than a preset threshold are target vehicle control components;

[0138] The second determination submodule is used to determine the corresponding control parameters of the target vehicle control component.

[0139] In an optional embodiment of the present invention, the training module includes:

[0140] A sample construction submodule, used to construct training samples based on the historical environment data and historical user data;

[0141] A matrix construction submodule, used for constructing a Pearson coefficient matrix based on the training samples;

[0142] The model building submodule is used to build a Pearson one-dimensional convolution model based on the training samples and the Pearson coefficient matrix fitting.

[0143] In an optional embodiment of the present invention, the model building submodule includes:

[0144] A first processing unit, configured to multiply the training sample and the Pearson coefficient matrix to obtain a feature vector;

[0145] The second processing unit is used to construct a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network.

[0146] In an optional embodiment of the present invention, the second processing unit includes:

[0147] A first training subunit is used to input the feature vector into a preset one-dimensional residual convolutional network for convolution processing to obtain an intermediate one-dimensional residual convolutional network;

[0148] The second training subunit is used to perform jump connection processing on the intermediate one-dimensional residual convolution network based on a preset residual block to obtain a Pearson one-dimensional convolution model.

[0149] In an optional embodiment of the present invention, the device further comprises:

[0150] The preprocessing module is used to preprocess the current environment data and the user characteristic data.

[0151] The embodiment of the present invention obtains current environment data and user characteristic data; constructs vehicle scene characteristics based on the current environment data and the user characteristic data; determines the correlation coefficient between the vehicle scene characteristics and the vehicle control components of the vehicle; determines multiple target vehicle control components and corresponding control parameters that match the vehicle scene characteristics based on the correlation coefficient; and controls the multiple target vehicle control components based on the control parameters in response to confirmation operations on the multiple target vehicle control components. The vehicle scene characteristics are constructed by identifying the current environment data and user characteristic data, and multiple related target vehicle control components and corresponding control parameters are determined based on the correlation coefficient. The determination process does not involve human experience knowledge that may have subjective factors, and recommends specific parameters for each vehicle control component, so that users can enable multiple vehicle control components at the same time in different vehicle use scenarios, thereby improving the user's vehicle experience.

[0152] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0153] Reference Figure 7 , an embodiment of the present invention further provides a vehicle, comprising:

[0154] A processor 701 and a storage medium 702, wherein the storage medium 702 stores a computer program executable by the processor 701, and when the vehicle is running, the processor 701 executes the computer program to perform the vehicle control method as described in any one of the embodiments of the present invention.

[0155] The vehicle control method comprises:

[0156] Obtain current environment data and user feature data;

[0157] Constructing a vehicle use scenario feature based on the current environment data and the user feature data;

[0158] Determining a correlation coefficient between the vehicle use scenario characteristics and a vehicle control element of the vehicle;

[0159] Based on the correlation coefficient, determining a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics;

[0160] In response to the confirmation operation on the plurality of target vehicle control components, the plurality of target vehicle control components are controlled based on the control parameter.

[0161] Furthermore, the method further comprises:

[0162] Obtain historical environment data and historical user data;

[0163] A Pearson one-dimensional convolution model is constructed based on the historical environment data and the historical user data.

[0164] Furthermore, the step of determining the correlation coefficient between the vehicle usage scenario feature and the vehicle control element of the vehicle includes:

[0165] The vehicle usage scenario features are input into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output a correlation coefficient based on the vehicle usage scenario features.

[0166] Furthermore, the step of constructing the vehicle use scenario characteristics based on the current environment data and the user characteristic data includes:

[0167] Converting the current environment data into a first characteristic character;

[0168] Converting the user characteristic data into a second characteristic character;

[0169] The first characteristic character and the second characteristic character are combined to generate a vehicle use scenario feature.

[0170] Furthermore, the step of determining a plurality of target vehicle control components and corresponding control parameters matching the vehicle usage scenario characteristics based on the correlation coefficient includes:

[0171] Determining vehicle control components of multiple vehicles whose correlation coefficients are greater than a preset threshold as target vehicle control components;

[0172] Determine the corresponding control parameters of the target vehicle control component.

[0173] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the historical environment data and the historical user data includes:

[0174] Constructing training samples based on the historical environment data and the historical user data;

[0175] Constructing a Pearson coefficient matrix based on the training samples;

[0176] A Pearson one-dimensional convolution model is constructed based on the training samples and the Pearson coefficient matrix fitting.

[0177] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the training samples and the Pearson coefficient matrix fitting includes:

[0178] Multiplying the training sample and the Pearson coefficient matrix to obtain a feature vector;

[0179] A Pearson one-dimensional convolution model is constructed based on the feature vector and a preset one-dimensional residual convolution network.

[0180] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network includes:

[0181] Inputting the feature vector into a preset one-dimensional residual convolutional network for convolution processing to obtain an intermediate one-dimensional residual convolutional network;

[0182] The intermediate one-dimensional residual convolutional network is skipped based on a preset residual block to obtain a Pearson one-dimensional convolutional model.

[0183] Furthermore, the method further comprises:

[0184] The current environment data and the user characteristic data are preprocessed.

[0185] The embodiment of the present invention obtains current environment data and user characteristic data; constructs vehicle scene characteristics based on the current environment data and the user characteristic data; determines the correlation coefficient between the vehicle scene characteristics and the vehicle control components of the vehicle; determines multiple target vehicle control components and corresponding control parameters that match the vehicle scene characteristics based on the correlation coefficient; and controls the multiple target vehicle control components based on the control parameters in response to confirmation operations on the multiple target vehicle control components. The vehicle scene characteristics are constructed by identifying the current environment data and user characteristic data, and multiple related target vehicle control components and corresponding control parameters are determined based on the correlation coefficient. The determination process does not involve human experience knowledge that may have subjective factors, and recommends specific parameters for each vehicle control component, so that users can enable multiple vehicle control components at the same time in different vehicle use scenarios, thereby improving the user's vehicle experience.

[0186] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the processor.

[0187] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0188] Reference Figure 8 The embodiment of the present invention further provides a computer-readable storage medium 801, on which a computer program is stored. When the computer program is executed by a processor, the vehicle control method as described in any one of the embodiments of the present invention is executed.

[0189] The vehicle control method comprises:

[0190] Obtain current environment data and user feature data;

[0191] Constructing a vehicle use scenario feature based on the current environment data and the user feature data;

[0192] Determining a correlation coefficient between the vehicle use scenario characteristics and a vehicle control element of the vehicle;

[0193] Based on the correlation coefficient, determining a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics;

[0194] In response to the confirmation operation on the plurality of target vehicle control components, the plurality of target vehicle control components are controlled based on the control parameter.

[0195] Furthermore, the method further comprises:

[0196] Obtain historical environment data and historical user data;

[0197] A Pearson one-dimensional convolution model is constructed based on the historical environment data and the historical user data.

[0198] Furthermore, the step of determining the correlation coefficient between the vehicle usage scenario feature and the vehicle control element of the vehicle includes:

[0199] The vehicle usage scenario features are input into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output a correlation coefficient based on the vehicle usage scenario features.

[0200] Furthermore, the step of constructing the vehicle use scenario characteristics based on the current environment data and the user characteristic data includes:

[0201] Converting the current environment data into a first characteristic character;

[0202] Converting the user characteristic data into a second characteristic character;

[0203] The first characteristic character and the second characteristic character are combined to generate a vehicle use scenario feature.

[0204] Furthermore, the step of determining a plurality of target vehicle control components and corresponding control parameters matching the vehicle usage scenario characteristics based on the correlation coefficient includes:

[0205] Determining vehicle control components of multiple vehicles whose correlation coefficients are greater than a preset threshold as target vehicle control components;

[0206] Determine the corresponding control parameters of the target vehicle control component.

[0207] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the historical environment data and the historical user data includes:

[0208] Constructing training samples based on the historical environment data and the historical user data;

[0209] Constructing a Pearson coefficient matrix based on the training samples;

[0210] A Pearson one-dimensional convolution model is constructed based on the training samples and the Pearson coefficient matrix fitting.

[0211] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the training samples and the Pearson coefficient matrix fitting includes:

[0212] Multiplying the training sample and the Pearson coefficient matrix to obtain a feature vector;

[0213] A Pearson one-dimensional convolution model is constructed based on the feature vector and a preset one-dimensional residual convolution network.

[0214] Furthermore, the step of constructing a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network includes:

[0215] Inputting the feature vector into a preset one-dimensional residual convolutional network for convolution processing to obtain an intermediate one-dimensional residual convolutional network;

[0216] The intermediate one-dimensional residual convolutional network is skipped based on a preset residual block to obtain a Pearson one-dimensional convolutional model.

[0217] Furthermore, the method further comprises:

[0218] The current environment data and the user characteristic data are preprocessed.

[0219] The embodiment of the present invention obtains current environment data and user characteristic data; constructs vehicle scene characteristics based on the current environment data and the user characteristic data; determines the correlation coefficient between the vehicle scene characteristics and the vehicle control components of the vehicle; determines multiple target vehicle control components and corresponding control parameters that match the vehicle scene characteristics based on the correlation coefficient; and controls the multiple target vehicle control components based on the control parameters in response to confirmation operations on the multiple target vehicle control components. The vehicle scene characteristics are constructed by identifying the current environment data and user characteristic data, and multiple related target vehicle control components and corresponding control parameters are determined based on the correlation coefficient. The determination process does not involve human experience knowledge that may have subjective factors, and recommends specific parameters for each vehicle control component, so that users can enable multiple vehicle control components at the same time in different vehicle use scenarios, thereby improving the user's vehicle experience.

[0220] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0221] It will be appreciated by those skilled in the art that the embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the embodiments of the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0222] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0223] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0224] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0225] The above embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Any equivalent substitution or change made by a person skilled in the art based on the present invention is within the protection scope of the present invention.

Claims

1. A vehicle control method, characterized in that: include: Obtain current environment data and user feature data; Constructing a vehicle use scenario feature based on the current environment data and the user feature data; Determining a correlation coefficient between the vehicle use scenario characteristics and a vehicle control element of the vehicle; Based on the correlation coefficient, determining a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics; In response to the confirmation operation on the plurality of target vehicle control components, the plurality of target vehicle control components are controlled based on the control parameter.

2. The method according to claim 1, characterized in that The method further comprises: Obtain historical environment data and historical user data; A Pearson one-dimensional convolution model is constructed based on the historical environment data and the historical user data.

3. The method according to claim 2, characterized in that The step of determining the correlation coefficient between the vehicle use scenario feature and the vehicle control element of the vehicle comprises: The vehicle usage scenario features are input into the Pearson one-dimensional convolution model, and the Pearson one-dimensional convolution model is used to output a correlation coefficient based on the vehicle usage scenario features.

4. The method according to claim 1, characterized in that: The step of constructing the vehicle use scenario feature based on the current environment data and the user feature data comprises: Converting the current environment data into a first characteristic character; Converting the user characteristic data into a second characteristic character; The first characteristic character and the second characteristic character are combined to generate a vehicle use scenario feature.

5. The method according to claim 1, characterized in that The step of determining a plurality of target vehicle control components and corresponding control parameters matching the vehicle usage scenario characteristics based on the correlation coefficient comprises: Determining vehicle control components of multiple vehicles whose correlation coefficients are greater than a preset threshold as target vehicle control components; Determine the corresponding control parameters of the target vehicle control component.

6. The method according to claim 2, characterized in that The step of constructing a Pearson one-dimensional convolution model based on the historical environment data and the historical user data comprises: Constructing training samples based on the historical environment data and the historical user data; Constructing a Pearson coefficient matrix based on the training samples; A Pearson one-dimensional convolution model is constructed based on the training samples and the Pearson coefficient matrix fitting.

7. The method according to claim 6, characterized in that The step of constructing a Pearson one-dimensional convolution model based on the training samples and the Pearson coefficient matrix fitting comprises: Multiplying the training sample and the Pearson coefficient matrix to obtain a feature vector; A Pearson one-dimensional convolution model is constructed based on the feature vector and a preset one-dimensional residual convolution network.

8. The method according to claim 7, characterized in that The step of constructing a Pearson one-dimensional convolution model based on the feature vector and a preset one-dimensional residual convolution network includes: Inputting the feature vector into a preset one-dimensional residual convolutional network for convolution processing to obtain an intermediate one-dimensional residual convolutional network; The intermediate one-dimensional residual convolutional network is skipped based on a preset residual block to obtain a Pearson one-dimensional convolutional model.

9. The method according to claim 1, characterized in that: The method further comprises: The current environment data and the user characteristic data are preprocessed.

10. A vehicle control device, characterized in that: include: The current acquisition module is used to obtain current environment data and user feature data; A construction module, used to construct a vehicle use scenario feature based on the current environment data and the user feature data; A first determination module is used to determine a correlation coefficient between the vehicle use scenario feature and a vehicle control element of the vehicle; A second determination module, configured to determine, based on the correlation coefficient, a plurality of target vehicle control components and corresponding control parameters that match the vehicle usage scenario characteristics; A response module is used to control the plurality of vehicle control components based on the control parameters in response to a confirmation operation on the plurality of vehicle control components.

11. A vehicle, characterized in that: The method comprises a processor, a memory and a computer program stored in the memory and capable of running on the processor, wherein the computer program implements the steps of the vehicle control method according to any one of claims 1 to 9 when executed by the processor.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vehicle control method according to any one of claims 1 to 9 are implemented.

Images