A method for controlling vehicle sideslip speed in curves

Abstract

The present invention belongs to the technical field of vehicle speed control. The present invention discloses a method for controlling the side slip speed of a vehicle on a curve, comprising: firstly collecting historical static data and a corresponding speed and adhesion coefficient reference table, then grouping the static data and establishing classification labels; for each data group, establishing a speed and adhesion coefficient relationship model, extracting static features, and then training a static classification model; obtaining current static data, extracting static features, obtaining classification labels through the static classification model, and then finding the corresponding speed and adhesion coefficient model, establishing a side slip speed model, and calculating the critical speed; finally, according to the comparison between the real-time vehicle speed and the critical speed, issuing a reminder to the driver. Calculating the critical speed by finding the corresponding speed and adhesion coefficient model based on the real-time static data can effectively improve the accuracy of critical speed calculation, and then accurately make warnings or interventions, thereby improving traffic safety.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle speed control, and more specifically, to a method for controlling the sideslip speed of a vehicle on a curve. Background Art

[0002] Vehicles are prone to skidding during driving, especially when turning or on slippery roads. Vehicle skidding refers to the phenomenon that the vehicle loses directional control due to the lateral force exceeding the adhesion between the tires and the road during driving. This situation can easily lead to traffic accidents and pose a threat to the safety of drivers and passengers. Therefore, how to effectively prevent and control vehicle skidding has always been a research focus in the field of automotive safety.

[0003] Existing vehicle side slip control technology usually calculates the critical speed for cornering first, and then controls the vehicle or reminds the driver based on the critical speed and the real-time vehicle speed. When calculating the critical speed, the adhesion coefficient is a key parameter. However, the existing technology usually uses the empirical value of the adhesion coefficient when calculating the critical speed. Since the adhesion coefficient varies with the vehicle condition and road conditions, different factors such as different vehicles and road conditions will affect the value of the adhesion coefficient. The empirical value alone cannot accurately reflect the adhesion coefficient under the actual vehicle condition and road conditions, which leads to large errors in the calculation of the critical speed, making it impossible for the side slip control system to issue warnings or take intervention measures in time at critical moments, increasing the risk of traffic accidents. Summary of the invention

[0004] In order to overcome the problem of large error in calculating critical speed in the prior art, the present invention proposes a method for controlling the sideslip speed of a vehicle on a curve to solve the above problem.

[0005] The present invention provides the following technical solutions:

[0006] A method for controlling the sideslip speed of a vehicle on a curve, comprising:

[0007] S1. Acquire a number of historical static data, the static data including vehicle data and road condition data, and simultaneously acquire a speed and adhesion coefficient reference table corresponding to the static data, the speed and adhesion coefficient reference table including a number of speed values ​​and corresponding adhesion coefficient values;

[0008] S2. According to the speed and adhesion coefficient reference table, all static data are divided into several static data groups, and a classification label is added to each static data group;

[0009] S3, establishing a speed and adhesion coefficient relationship model corresponding to each static data group according to a speed and adhesion coefficient reference table corresponding to the static data in each static data group;

[0010] S4, obtaining static features according to the static data, and marking the static features with corresponding classification labels according to the classification labels of the static data group where the static data is located; using a number of static features marked with classification labels to form a learning data set, and using the learning data set to train a static classification model;

[0011] S5. Obtain current static data, obtain static features according to the current static data, and input the static features into a static classification model to obtain a classification label;

[0012] S6. Acquire a corresponding static data group according to the classification label, and acquire a corresponding speed and adhesion coefficient relationship model according to the static data group;

[0013] S7, obtaining a curve radius, and establishing a sideslip speed model according to the curve radius, current static data, and a speed-adhesion coefficient relationship model; and obtaining a critical speed according to the sideslip speed model;

[0014] S8. Obtain the real-time vehicle speed, and issue a reminder to the driver according to the real-time vehicle speed and the critical speed.

[0015] Preferably, the vehicle data includes vehicle weight, tire pressure, tire width, tread depth, lift reference area and lift coefficient; the road condition data includes road surface type and road surface condition, the road surface type includes asphalt road surface, concrete road surface, gravel and dirt road surface, and the road surface condition includes dry and wet.

[0016] Preferably, for each static data, obtain the corresponding speed and adhesion coefficient reference table

[0017] Preset n speed values, and obtain corresponding adhesion coefficient values ​​from a corresponding speed and adhesion coefficient reference table according to the preset n speed values ​​to form a reference array corresponding to the static data;

[0018] sorting the reference array according to its elements;

[0019] The static data corresponding to the reference arrays classified into the same category are classified into the same static data group.

[0020] Preferably, the classifying the reference array according to the elements of the reference array comprises:

[0021] S21, combining all reference arrays to obtain a reference array set;

[0022] S22, randomly selecting two reference arrays as center points in the reference array set, each center point represents a category, and marking all other reference arrays with labels to be assigned;

[0023] S23, calculating the comprehensive similarity between each array to be allocated and the two center points in the reference array set, using the larger median of the two calculated comprehensive similarities as the selected similarity, and comparing the selected similarity with a preset similarity threshold;

[0024] If the selected similarity is greater than or equal to the similarity threshold, the array to be assigned is assigned to the category represented by the center point corresponding to the selected similarity, and the label to be assigned of the reference array is removed;

[0025] If the selected similarity is less than the similarity threshold, the to-be-assigned label of the reference array is retained;

[0026] S24, determining whether there is a reference array with a tag to be assigned in the reference array set;

[0027] If yes, randomly select a reference array from all reference arrays with labels to be assigned, remove the labels to be assigned, and use it as a new center point to represent a new category, calculate the comprehensive similarity between all reference arrays with labels to be assigned and the new center point; compare the calculated comprehensive similarity with the similarity threshold;

[0028] If the comprehensive similarity is greater than or equal to the similarity threshold, the reference array is assigned to the category represented by the new center point, and the label to be assigned of the reference array is removed;

[0029] S25. Repeat step S4 until there is no reference array with a label to be assigned, and the classification of the reference array is completed.

[0030] Preferably, the step of calculating the comprehensive similarity includes:

[0031] Get the reference array corresponding to the center point and record it as an array , get the reference array with the label to be assigned and record it as array ;

[0032] The comprehensive similarity is calculated using the following formula:

[0033] ,

[0034] In the formula, represents the comprehensive similarity, Representing an array Middle elements, Representing an array Middle elements.

[0035] Preferably, the step of establishing a speed and adhesion coefficient relationship model corresponding to each static data group according to a speed and adhesion coefficient reference table corresponding to the static data in each static data group comprises:

[0036] Obtaining an adhesion coefficient reference table corresponding to all static data in the static data group;

[0037] For each static data, m fitting data points are obtained from the corresponding adhesion coefficient reference table to form a corresponding fitting data set, wherein the fitting data points include a speed value and a corresponding adhesion coefficient value;

[0038] According to the fitting data set, the speed value is used as the independent variable and the adhesion coefficient value is used as the dependent variable, and the fitting method is used to obtain the corresponding fitting equation;

[0039] For each obtained fitting equation corresponding to static data, the determination coefficient of the fitting equation is calculated using the fitting data sets corresponding to other static data in the static data group, and all the calculated determination coefficients are summed to obtain a total determination coefficient;

[0040] The fitting equation corresponding to the largest total determination coefficient is used as the speed-adhesion coefficient relationship model corresponding to the static data set.

[0041] Preferably, the step of acquiring the static features includes:

[0042] Set a corresponding unique hot code for each road type and a corresponding representative value for each road condition;

[0043] Obtain vehicle weight, tire pressure, tire width, and tread depth, perform normalization, obtain road condition data, and obtain the corresponding unique hot encoding and representative value based on the road condition data;

[0044] Use the obtained unique hot encoding, representative value, normalized vehicle weight, tire pressure, tire width and tread depth values ​​as static features;

[0045] The step of using the learning data set to train the static classification model includes:

[0046] Use the fully connected neural network as the basic framework to build a static classification model; obtain a learning data set, use the static features in the learning data set as the input of the model, and the classification label as the output of the model, divide the learning data set into a training data set and a validation data set according to a certain ratio, use the training data set and the validation data set to train the model, and obtain a trained static classification model.

[0047] Preferably, the step of obtaining the curve radius includes:

[0048] Use the camera to mark two points on the center line of the curve and record them as Point and Click to get Point and Distance of points ;

[0049] On the center line of the curve, Point and Draw a tangent to the center line at point 1, and obtain the angle between the two tangents toward the center of the bend. ;

[0050] Use the formula: , calculate the curve radius .

[0051] Preferably, the process of establishing the sideslip velocity model is as follows:

[0052] Establish the theoretical formula of centrifugal force: ,in represents the centrifugal force, Indicates vehicle weight. Indicates vehicle speed;

[0053] Establish the theoretical formula of gravity: ,in Indicates gravity;

[0054] Establish lift theory formula: ,in represents the air density, represents the lift reference area, represents the lift coefficient;

[0055] The relationship model between acquisition speed and adhesion coefficient is recorded as , the following sideslip velocity model is established:

[0056] ;

[0057] The obtaining of the critical speed according to the sideslip speed model comprises: using a numerical method to solve a solution set of the sideslip speed model, and using a maximum value in the solution set as the critical speed.

[0058] Preferably, the reminding the driver according to the real-time vehicle speed and the critical speed includes:

[0059] Calculate the ratio of critical speed to real-time vehicle speed ;

[0060] like , no reminder will be issued or a normal reminder will be issued; if , a normal reminder is issued; if , an alarm will be issued; if , an emergency reminder is issued, , It is the preset reminder ratio threshold.

[0061] The present invention provides a method for controlling the sideslip speed of a vehicle on a curve, which has the following beneficial effects:

[0062] By collecting historical static data and the corresponding speed and adhesion coefficient reference table, the static data is grouped, and a speed and adhesion coefficient model is established for each group. Each static group is regarded as a category, and the static features of each group are extracted. The relationship between the static features and the classification is learned using machine learning methods. Then, in subsequent use, the static features can be obtained based on real-time static data to find the corresponding speed and adhesion coefficient model. Based on the corresponding speed and adhesion coefficient model, the accuracy of critical speed calculation can be effectively improved, and warnings or interventions can be made accurately to improve traffic safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 A schematic flow chart of a method for controlling the sideslip speed of a vehicle on a curve according to the present invention;

[0064] Figure 2 A schematic diagram of a reference array classification process of a vehicle curve sideslip speed according to the present invention;

[0065] Figure 3 Schematic diagram of the geometric relationship at the bend. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] Example 1, please refer to Figure 1 In this embodiment, a method for controlling the sideslip speed of a vehicle on a curve includes:

[0068] S1. Acquire a number of historical static data, the static data including vehicle data and road condition data, and simultaneously acquire a speed and adhesion coefficient reference table corresponding to the static data, the speed and adhesion coefficient reference table including a number of speed values ​​and corresponding adhesion coefficient values;

[0069] The vehicle data includes vehicle weight, tire pressure, tire width, tread depth, lift reference area and lift coefficient; the road condition data includes road surface type and road surface condition, the road surface type includes asphalt road surface, concrete road surface, gravel and dirt road surface, and the road surface condition includes dry and wet.

[0070] In this embodiment, historical static data is obtained in an experimental or real environment, wherein vehicle data can be obtained by the manufacturer or actual measurement, and road condition data can be obtained by personnel input or image recognition. The speed and adhesion coefficient reference table is constructed by measuring and counting the adhesion coefficient values ​​generated by vehicles with different vehicle data and driving at different speeds on roads with different road conditions. This reference table provides data support for understanding and predicting the adhesion coefficients of different vehicles under different road conditions and speed conditions.

[0071] S2. According to the speed and adhesion coefficient reference table, all static data are divided into several static data groups, and a classification label is added to each static data group;

[0072] According to the speed and adhesion coefficient reference table, the steps of dividing all static data into several static data groups include:

[0073] For each static data, obtain the corresponding speed and adhesion coefficient reference table

[0074] Preset n speed values, and obtain corresponding adhesion coefficient values ​​from a corresponding speed and adhesion coefficient reference table according to the preset n speed values ​​to form a reference array corresponding to the static data;

[0075] sorting the reference array according to its elements;

[0076] The static data corresponding to the reference arrays classified into the same category are classified into the same static data group.

[0077] The classifying the reference array according to the elements of the reference array comprises:

[0078] S21, combining all reference arrays to obtain a reference array set;

[0079] S22, randomly selecting two reference arrays as center points in the reference array set, each center point represents a category, and marking all other reference arrays with labels to be assigned;

[0080] S23, calculating the comprehensive similarity between each array to be allocated and the two center points in the reference array set, using the larger median of the two calculated comprehensive similarities as the selected similarity, and comparing the selected similarity with a preset similarity threshold;

[0081] If the selected similarity is greater than or equal to the similarity threshold, the array to be assigned is assigned to the category represented by the center point corresponding to the selected similarity, and the label to be assigned of the reference array is removed;

[0082] If the selected similarity is less than the similarity threshold, the to-be-assigned label of the reference array is retained;

[0083] S24, determining whether there is a reference array with a tag to be assigned in the reference array set;

[0084] If yes, randomly select a reference array from all reference arrays with labels to be assigned, remove the labels to be assigned, and use it as a new center point to represent a new category, calculate the comprehensive similarity between all reference arrays with labels to be assigned and the new center point; compare the calculated comprehensive similarity with the similarity threshold;

[0085] If the comprehensive similarity is greater than or equal to the similarity threshold, the reference array is assigned to the category represented by the new center point, and the label to be assigned of the reference array is removed;

[0086] S25. Repeat step S4 until there is no reference array with a label to be assigned, and the classification of the reference data is completed.

[0087] The steps of calculating the comprehensive similarity include:

[0088] Get the reference array corresponding to the center point and record it as an array , get the reference array with the label to be assigned and record it as array ;

[0089] The comprehensive similarity is calculated using the following formula:

[0090] ,

[0091] In the formula, represents the comprehensive similarity, Representing an array Middle elements, Representing an array Middle elements.

[0092] See also Figure 2In this embodiment, after completing the construction of the speed and adhesion coefficient reference table, this embodiment further groups all the collected static data and labels each group of data with a classification label for subsequent analysis and application. The specific steps are: for each static data record, first extract information from its corresponding speed and adhesion coefficient reference table. Preset n representative speed values, such as 0km / h, 10km / h, 20km / h, etc., with an interval of 10km / h, and then look up the adhesion coefficient values ​​corresponding to these speed values ​​in the reference table to form a reference array corresponding to the static data. All generated reference arrays are combined into a reference array set. Subsequently, two reference arrays are randomly selected in the reference array set as the initial center points, and the two center points represent two different categories respectively. At this time, all the remaining reference arrays in the set are labeled "to be assigned", indicating that they have not yet been assigned to a specific category.

[0093] For each reference array with a "to be assigned" label, calculate the comprehensive similarity between it and the two center points. For each reference array to be assigned, select the one with the higher comprehensive similarity between the two center points as the basis for assignment. If the calculated selection similarity is greater than or equal to the preset similarity threshold, such as 0.7, the reference array is assigned to the corresponding category and its "to be assigned" label is removed. If the selection similarity is less than the threshold, keep its to be assigned status.

[0094] When there are reference arrays that have not been successfully assigned, a reference array with a "to be assigned" label is randomly selected as the new center point, and the above assignment process is repeated. The addition of new center points is intended to process data points that are significantly different from the initial center points, thereby increasing the accuracy and coverage of classification. Repeat until there are no reference arrays with the "to be assigned" label in the reference array set. At this point, all reference arrays are assigned to the corresponding categories, forming several static data groups.

[0095] Finally, a corresponding classification label is assigned to each static data group, where the classification label can be a randomly generated code or a specific value, and each static data group corresponds to a different classification label.

[0096] S3, establishing a speed and adhesion coefficient relationship model corresponding to each static data group according to a speed and adhesion coefficient reference table corresponding to the static data in each static data group;

[0097] The step of establishing a speed and adhesion coefficient relationship model corresponding to each static data group according to a speed and adhesion coefficient reference table corresponding to the static data in each static data group comprises:

[0098] Obtaining an adhesion coefficient reference table corresponding to all static data in the static data group;

[0099] For each static data, m fitting data points are obtained from the corresponding adhesion coefficient reference table to form a corresponding fitting data set, wherein the fitting data points include a speed value and a corresponding adhesion coefficient value;

[0100] According to the fitting data set, the speed value is used as the independent variable and the adhesion coefficient value is used as the dependent variable, and the fitting method is used to obtain the corresponding fitting equation;

[0101] For each obtained fitting equation corresponding to static data, the determination coefficient of the fitting equation is calculated using the fitting data sets corresponding to other static data in the static data group, and all the calculated determination coefficients are summed to obtain a total determination coefficient;

[0102] The fitting equation corresponding to the largest total determination coefficient is used as the speed-adhesion coefficient relationship model corresponding to the static data set.

[0103] In this embodiment, first, the adhesion coefficient reference table corresponding to all static data is extracted from each static data group. These reference tables contain the adhesion coefficient values ​​corresponding to different speeds and are the basic data for building the relationship model. For each static data in the static data group, 10 fitting data points can be selected from the corresponding adhesion coefficient reference table. These fitting data points include specific speed values ​​and corresponding adhesion coefficient values, which together constitute the fitting data set of the static data.

[0104] With the speed value as the independent variable and the adhesion coefficient value as the dependent variable, the binomial fitting method is used, for example, the quadratic polynomial y=ax^2+bx+c is used to fit the fitting data set. Through the fitting functions in mathematical software such as MATLAB, Python's NumPy and SciPy libraries, the fitting equation coefficients a, b, and c corresponding to each static data can be solved, and then the fitting equation can be obtained. For the fitting equation obtained for each static data, the determination coefficient R-squared of the equation is calculated using the fitting data sets corresponding to all other static data in the static data group. The determination coefficient measures the goodness of fit of the model to the data. The closer its value is to 1, the better the fitting effect of the model. The determination coefficients of the fitting equations corresponding to all static data are summed to obtain the total determination coefficient. The fitting equation with the largest total determination coefficient is selected as the final speed and adhesion coefficient relationship model corresponding to the static data group. This model is a function with speed as the independent variable and adhesion coefficient as the dependent variable, which can accurately reflect the dynamic relationship between speed and adhesion coefficient.

[0105] S4, obtaining static features according to the static data, and marking the static features with corresponding classification labels according to the classification labels of the static data group where the static data is located; using a number of static features marked with classification labels to form a learning data set, and using the learning data set to train a static classification model;

[0106] The step of acquiring the static features comprises:

[0107] Set a corresponding unique hot code for each road type and a corresponding representative value for each road condition;

[0108] Obtain vehicle weight, tire pressure, tire width, and tread depth, perform normalization, obtain road condition data, and obtain the corresponding unique hot encoding and representative value based on the road condition data;

[0109] Use the obtained unique hot encoding, representative value, normalized vehicle weight, tire pressure, tire width and tread depth values ​​as static features;

[0110] The step of using the learning data set to train the static classification model includes:

[0111] Use the fully connected neural network as the basic framework to build a static classification model; obtain a learning data set, use the static features in the learning data set as the input of the model, and the classification label as the output of the model, divide the learning data set into a training data set and a validation data set according to a certain ratio, use the training data set and the validation data set to train the model, and obtain a trained static classification model.

[0112] In this embodiment, for a given set of road surface types, including asphalt road surface, concrete road surface, gravel road surface and soil road surface, we use one-hot encoding to convert them into vector form. Each road surface type can be mapped to a four-dimensional vector, in which only the index position corresponding to the road surface type is 1, and the rest of the positions are 0. For example, asphalt road surface is encoded as [1,0,0,0], concrete road surface is encoded as [0,1,0,0], gravel road surface is encoded as [0,0,1,0], and soil road surface is encoded as [0,0,0,1]. At the same time, in this embodiment, the representative value of the dry road surface condition can be set to 1, and the representative value of the wet road surface condition can be set to 0.

[0113] In this embodiment, a fully connected neural network (FCNN) is used as the basic framework to build a static classification model. The model structure may include the following levels: The input layer is used to receive the above-mentioned combined static feature vector as input. Hidden layer: Multiple hidden layers can be set as needed, each layer contains a certain number of neurons, and activation functions such as ReLU and Sigmoid are used to enhance the nonlinear ability of the model. For example, two hidden layers can be set, each with 64 neurons. Output layer: The number of neurons in the output layer corresponds to the number of classification labels, and the Softmax activation function is used to output the predicted probability of each category.

[0114] This embodiment uses a fully connected neural network (FCNN) as a framework to build a static classification model. The model is designed as follows: an input layer for directly receiving the combined static feature vector, two hidden layers, each with 64 neurons, and activation functions such as ReLU or Sigmoid are applied to enhance the model's ability to handle complex data relationships and capture nonlinear features in the data. An output layer, the number of neurons in the output layer corresponds one-to-one to the classification label, and the Softmax activation function is used to ensure that the output is the predicted probability distribution of each category, which is convenient for subsequent decision-making or further analysis.

[0115] After obtaining a learning dataset containing static features and corresponding classification labels, divide it into a training dataset and a validation dataset according to a certain ratio, such as 70% training set and 30% validation set. Use the training dataset to train the model, and adjust the model parameters through the back propagation algorithm to minimize the error between the predicted label and the actual label, such as cross entropy loss. At the same time, use the validation dataset to regularly evaluate the model to avoid overfitting, and adjust the model structure and training parameters in a timely manner. Finally, we can train a static classification model that obtains classification labels based on static data.

[0116] S5. Obtain current static data, obtain static features according to the current static data, and input the static features into a static classification model to obtain a classification label;

[0117] In this embodiment, the vehicle data in the current static data can be the original data provided by the manufacturer, or it can be modified by the driver based on the original data. The road condition data can be obtained by combining navigation data with weather data, or it can be obtained through driver input and image recognition.

[0118] S6. Acquire a corresponding static data group according to the classification label, and acquire a corresponding speed and adhesion coefficient relationship model according to the static data group;

[0119] In this embodiment, the input data is first classified using a static classification model to obtain classification labels corresponding to the current road surface and vehicle status, and then a static data group is obtained based on the classification labels. Finally, the corresponding speed and adhesion coefficient relationship model is obtained based on the static data group, ensuring that the selected model is closely matched with the real-time conditions, thereby significantly improving the accuracy and reliability of subsequent critical speed calculations.

[0120] S7, obtaining a curve radius, and establishing a sideslip speed model according to the curve radius, current static data, and a speed-adhesion coefficient relationship model; and obtaining a critical speed according to the sideslip speed model;

[0121] The process of establishing the sideslip velocity model is as follows:

[0122] Establish the theoretical formula of centrifugal force: ,in represents the centrifugal force, Indicates vehicle weight. Indicates vehicle speed;

[0123] Establish the theoretical formula of gravity: ,in Indicates gravity;

[0124] Establish lift theory formula: ,in represents the air density, represents the lift reference area, represents the lift coefficient;

[0125] The relationship model between acquisition speed and adhesion coefficient is recorded as , the following sideslip velocity model is established:

[0126] ;

[0127] The obtaining of the critical speed according to the sideslip speed model comprises: using a numerical method to solve a solution set of the sideslip speed model, and using a maximum value in the solution set as the critical speed.

[0128] In this embodiment, in order to ensure the safety and stability of the vehicle when driving on a curve, we first obtain the radius of the current curve, and build a side slip speed model by combining the current static data and the pre-established speed and adhesion coefficient relationship model. Then, the critical speed of the vehicle when driving on the curve is calculated through the model, providing important reference information for the driver.

[0129] First, the curvature information of the curve ahead is detected in real time using on-board sensors such as GPS, radar or visual sensors, and the radius of the curve is calculated. When driving on a curve, the vehicle is subjected to centrifugal force due to inertia, and its magnitude depends on the vehicle mass, driving speed and curve radius. At the same time, the stability of the vehicle is affected by sliding friction, which is jointly determined by the vehicle's gravity, lift and adhesion coefficient. When the centrifugal force exceeds the sliding friction, the vehicle is at risk of skidding. Based on the above physical principles, a skidding speed model is constructed, which comprehensively considers various factors that affect the vehicle's skidding. In order to solve this model, numerical methods such as dichotomy and Newton iteration are used to obtain a series of speed solution sets. Among these solutions, the maximum speed value is defined as the critical speed under the curve condition, which represents the maximum speed limit at which the vehicle can safely pass the curve without skidding.

[0130] S8. Obtain the real-time vehicle speed, and issue a reminder to the driver according to the real-time vehicle speed and the critical speed.

[0131] The reminding the driver according to the real-time vehicle speed and the critical speed includes:

[0132] Calculate the ratio of critical speed to real-time vehicle speed ;

[0133] like , no reminder will be issued or a normal reminder will be issued; if , a normal reminder is issued; if , an alarm will be issued; if , an emergency reminder is issued, , It is the preset reminder ratio threshold.

[0134] In this embodiment, It can be set to 0.5. Can be set to 0.8; It means that the current vehicle speed is far below the critical speed and the vehicle is driving safely. The system can choose not to issue a reminder or issue a normal reminder to remind the driver that the current driving status is safe. It means that the current vehicle speed is close to but has not yet reached the critical speed. The system should issue a normal reminder to remind the driver to control the speed. Indicates that the current vehicle speed has approached or reached the critical speed, and there is a high risk of skidding. The system should issue an alarm and strongly recommend the driver to slow down. It means that the current vehicle speed has exceeded the critical speed and the vehicle is in a highly dangerous state. The system should immediately issue an emergency reminder to remind the driver that he must slow down immediately to ensure driving safety.

[0135] Example 2, please refer to Figure 3Based on Example 1, the step of obtaining the curve radius includes:

[0136] Use the camera to mark two points on the center line of the curve and record them as Point and Click to get Point and Distance of points ;

[0137] On the center line of the curve, Point and Draw a tangent to the center line at point 1, and obtain the angle between the two tangents toward the center of the bend. ;

[0138] Use the formula: , calculate the curve radius .

[0139] In this embodiment, two points are marked on the center line of the curve and are respectively recorded as Point and Click to get Point and Distance of points ; On the center line of the curve, respectively Point and Draw a tangent to the center line at point 1, and obtain the angle between the two tangents toward the center of the bend. ; The distance The angle and the angle can be measured by the on-board computer. The line segments perpendicular to the tangent line through points A and B will intersect at one point, which is the center of the circle O. From the geometric relationship, it can be concluded that ω=180-β, α=ω / 2=90-β / 2. At the same time, from the trigonometric function relationship, sinα=L / 2R, that is, sin(90-β / 2)=L / 2R, and thus, R=L / 2sin(90-β / 2).

[0140] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the units is only one, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0141] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0142] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for controlling the sideslip speed of a vehicle on a curve, characterized in that: include: S1. Acquire a number of historical static data, the static data including vehicle data and road condition data, and simultaneously acquire a speed and adhesion coefficient reference table corresponding to the static data, the speed and adhesion coefficient reference table including a number of speed values ​​and corresponding adhesion coefficient values; S2. According to the speed and adhesion coefficient reference table, all static data are divided into several static data groups, and a classification label is added to each static data group; S3, establishing a speed and adhesion coefficient relationship model corresponding to each static data group according to a speed and adhesion coefficient reference table corresponding to the static data in each static data group; S4, obtaining static features according to the static data, and marking the static features with corresponding classification labels according to the classification labels of the static data group where the static data is located; using a number of static features marked with classification labels to form a learning data set, and using the learning data set to train a static classification model; S5. Obtain current static data, obtain static features according to the current static data, and input the static features into a static classification model to obtain a classification label; S6. Acquire a corresponding static data group according to the classification label, and acquire a corresponding speed and adhesion coefficient relationship model according to the static data group; S7, obtaining a curve radius, and establishing a sideslip speed model according to the curve radius, current static data, and a speed-adhesion coefficient relationship model; and obtaining a critical speed according to the sideslip speed model; S8. Obtain the real-time vehicle speed, and issue a reminder to the driver according to the real-time vehicle speed and the critical speed.

2. The method for controlling the vehicle's sideslip speed on a curve according to claim 1, characterized in that: The vehicle data includes vehicle weight, tire pressure, tire width, tread depth, lift reference area and lift coefficient; the road condition data includes road surface type and road surface condition, the road surface type includes asphalt road surface, concrete road surface, gravel and dirt road surface, and the road surface condition includes dry and wet.

3. The method for controlling the vehicle's sideslip speed on a curve according to claim 2, characterized in that: According to the speed and adhesion coefficient reference table, the steps of dividing all static data into several static data groups include: For each static data, obtain the corresponding speed and adhesion coefficient reference table Preset n speed values, and obtain corresponding adhesion coefficient values ​​from a corresponding speed and adhesion coefficient reference table according to the preset n speed values ​​to form a reference array corresponding to the static data; sorting the reference array according to its elements; The static data corresponding to the reference arrays classified into the same category are classified into the same static data group.

4. The method for controlling the vehicle's sideslip speed on a curve according to claim 3, characterized in that: The classifying the reference array according to the elements of the reference array comprises: S21, combining all reference arrays to obtain a reference array set; S22, randomly selecting two reference arrays as center points in the reference array set, each center point represents a category, and marking all other reference arrays with labels to be assigned; S23, calculating the comprehensive similarity between each array to be allocated and the two center points in the reference array set, using the larger median of the two calculated comprehensive similarities as the selected similarity, and comparing the selected similarity with a preset similarity threshold; If the selected similarity is greater than or equal to the similarity threshold, the array to be assigned is assigned to the category represented by the center point corresponding to the selected similarity, and the label to be assigned of the reference array is removed; If the selected similarity is less than the similarity threshold, the to-be-assigned label of the reference array is retained; S24, determining whether there is a reference array with a tag to be assigned in the reference array set; If yes, randomly select a reference array from all reference arrays with labels to be assigned, remove the labels to be assigned, and use it as a new center point to represent a new category, calculate the comprehensive similarity between all reference arrays with labels to be assigned and the new center point; compare the calculated comprehensive similarity with the similarity threshold; If the comprehensive similarity is greater than or equal to the similarity threshold, the reference array is assigned to the category represented by the new center point, and the label to be assigned of the reference array is removed; S25. Repeat step S4 until there is no reference array with a label to be assigned, and the classification of the reference array is completed.

5. The method for controlling the vehicle's sideslip speed on a curve according to claim 4, characterized in that: The steps of calculating the comprehensive similarity include: Get the reference array corresponding to the center point and record it as an array , get the reference array with the label to be assigned and record it as array ; The comprehensive similarity is calculated using the following formula: ， In the formula, represents the comprehensive similarity, Representing an array Middle elements, Representing an array Middle elements.

6. The method for controlling the vehicle's sideslip speed on a curve according to claim 5, characterized in that: The step of establishing a speed and adhesion coefficient relationship model corresponding to each static data group according to a speed and adhesion coefficient reference table corresponding to the static data in each static data group comprises: Obtaining an adhesion coefficient reference table corresponding to all static data in the static data group; For each static data, m fitting data points are obtained from the corresponding adhesion coefficient reference table to form a corresponding fitting data set, wherein the fitting data points include a speed value and a corresponding adhesion coefficient value; According to the fitting data set, the speed value is used as the independent variable and the adhesion coefficient value is used as the dependent variable, and the fitting method is used to obtain the corresponding fitting equation; For each obtained fitting equation corresponding to static data, the determination coefficient of the fitting equation is calculated using the fitting data sets corresponding to other static data in the static data group, and all the calculated determination coefficients are summed to obtain a total determination coefficient; The fitting equation corresponding to the largest total determination coefficient is used as the speed-adhesion coefficient relationship model corresponding to the static data set.

7. The method for controlling the vehicle's sideslip speed on a curve according to claim 6, characterized in that: The step of acquiring the static features comprises: Set a corresponding unique hot code for each road type and a corresponding representative value for each road condition; Obtain vehicle weight, tire pressure, tire width, and tread depth, perform normalization, obtain road condition data, and obtain the corresponding unique hot encoding and representative value based on the road condition data; Use the obtained unique hot encoding, representative value, normalized vehicle weight, tire pressure, tire width and tread depth values ​​as static features; The step of using the learning data set to train the static classification model includes: Use the fully connected neural network as the basic framework to build a static classification model; obtain a learning data set, use the static features in the learning data set as the input of the model, and the classification label as the output of the model, divide the learning data set into a training data set and a validation data set according to a certain ratio, use the training data set and the validation data set to train the model, and obtain a trained static classification model.

8. The method for controlling the vehicle's sideslip speed on a curve according to claim 7, characterized in that: The step of obtaining the curve radius comprises: Use the camera to mark two points on the center line of the curve and record them as Point and Click to get Point and Distance of points ; On the center line of the curve, Point and Draw a tangent to the center line at point 1, and obtain the angle between the two tangents toward the center of the bend. ; Use the formula: , calculate the curve radius .

9. The method for controlling the vehicle's sideslip speed on a curve according to claim 8, characterized in that: The process of establishing the sideslip velocity model is as follows: Establish the theoretical formula of centrifugal force: ,in represents the centrifugal force, Indicates vehicle weight. Indicates vehicle speed; Establish the theoretical formula of gravity: ,in Indicates gravity; Establish lift theory formula: ,in represents the air density, represents the lift reference area, represents the lift coefficient; The relationship model between acquisition speed and adhesion coefficient is recorded as , establish the following sideslip velocity model: ； The obtaining of the critical speed according to the sideslip speed model comprises: using a numerical method to solve a solution set of the sideslip speed model, and using a maximum value in the solution set as the critical speed.

10. The method for controlling the vehicle's sideslip speed on a curve according to claim 9, characterized in that: The reminding the driver according to the real-time vehicle speed and the critical speed includes: Calculate the ratio of critical speed to real-time vehicle speed ; like , no reminder will be issued or a normal reminder will be issued; if , a normal reminder is issued; if , an alarm will be issued; if , an emergency reminder is issued, , It is the preset reminder ratio threshold.

Images