Real-time vehicle speed recommendation method and device, electronic equipment and storage medium

Abstract

The invention relates to the technical field of vehicles, and discloses a real-time vehicle speed recommendation method and device, electronic equipment and a storage medium, and the method comprises the steps: constructing a current parameter vector according to current vehicle related parameters, and constructing a historical time sequence vector corresponding to a current moment according to historical parameters corresponding to a corresponding historical moment in a historical period; inputting the current parameter vector into a target first model to obtain a first vehicle speed combination, and inputting the historical time sequence vector into a target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is smaller than the number of structural layers of the target second model; and determining a current recommended vehicle speed interval according to the first vehicle speed combination and the second vehicle speed combination. According to the method, the accurate recommended vehicle speed interval can be provided for the driver to optimize the driving behavior of the driver, so that the vehicle driving energy consumption is reduced.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, and in particular to a method, device, electronic device and storage medium for recommending real-time vehicle speed. Background Art

[0002] Vehicle endurance has always been a hot topic in the field of electric and hybrid vehicles. Generally, the energy consumption of various components of the vehicle is reasonably adjusted to ensure the endurance of the vehicle, thereby improving the endurance of the vehicle. The Intelligent Speed ​​Adaptation (ISA) system is also mentioned in the relevant technology, which provides the driver with a recommended speed to save energy consumption, so as to improve the efficiency of vehicle energy use by optimizing the driver's driving behavior.

[0003] However, the intelligent speed recommendation system will be affected by factors such as vehicle conditions, resulting in low accuracy of the recommended speed it provides. It is unable to provide the driver with an accurate recommended speed, and thus cannot effectively reduce the vehicle's driving energy consumption. Summary of the invention

[0004] In view of the above problems, the present application provides a real-time vehicle speed recommendation method, device, electronic device and computer-readable storage medium for providing accurate recommended vehicle speed ranges in real time to optimize the driver's driving behavior and thus reduce vehicle driving energy consumption.

[0005] According to one aspect of the present application, a method for recommending real-time vehicle speed is provided, the recommendation method comprising: constructing a current parameter vector based on current vehicle-related parameters, and constructing a historical time series vector corresponding to the current moment based on historical parameters corresponding to corresponding historical moments in a historical period; inputting the current parameter vector into a target first model to obtain a first vehicle speed combination, and inputting the historical time series vector into a target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model; and determining a current recommended vehicle speed range based on the first vehicle speed combination and the second vehicle speed combination.

[0006] In an optional manner, the current vehicle-related parameters include current road parameters, and the current road parameters include vehicle detection road parameters; preset laser points are set around the vehicle; the current parameter vector is constructed based on the current vehicle-related parameters, further including: sending a detection laser to the surrounding area through the preset laser points to receive reflected light of the detection laser; determining the three-dimensional coordinate parameters of objects in the surrounding area based on the reflected light, and determining the current parameter vector corresponding to the vehicle detection road parameters based on the three-dimensional coordinate parameters.

[0007] In an optional manner, determining the current parameter vector corresponding to the vehicle detection road parameter based on the three-dimensional coordinate parameters further includes: multiplying each dimensional parameter in the three-dimensional coordinate parameters by their corresponding preset scaling factors, and summing the calculated products to obtain a normalized value; constructing the current parameter vector corresponding to the vehicle detection road parameter based on the normalized value and the number of preset laser points.

[0008] In an optional manner, the recommendation method also includes: matching the current user identity information with the preset user identity information, and using the preset first model and the preset second model corresponding to the successfully matched preset user identity information as the target first model and the target second model, respectively; wherein the structural layer of the preset first model includes a convolutional neural network layer and a fully connected neural network layer, and the structural layer of the preset second model includes a recursive neural network layer and a long short-term memory neural network layer.

[0009] In an optional manner, constructing a historical time series vector corresponding to the current moment based on historical parameters corresponding to corresponding historical moments in the historical cycle further includes: determining the historical parameter vectors corresponding to each corresponding historical moment based on the historical parameters corresponding to the corresponding historical moments in the historical cycle; inputting each historical parameter vector into a blank vector matrix according to a time series to obtain the historical time series vector corresponding to the current moment; wherein the number of vectors in the blank vector matrix is ​​the same as the number of historical parameter vectors.

[0010] In an optional manner, the first speed combination includes a first maximum speed and a first minimum speed, and the second speed combination includes a second maximum speed and a second minimum speed; determining the current recommended speed range based on the first speed combination and the second speed combination further includes: multiplying the first maximum speed and the second maximum speed by their respective corresponding preset weight coefficients, and summing the calculated products to obtain the recommended maximum speed; multiplying the first minimum speed and the second minimum speed by their respective corresponding preset weight coefficients, and summing the calculated products to obtain the recommended minimum speed; determining the current recommended speed range based on the recommended maximum speed and the recommended minimum speed.

[0011] In an optional manner, the current vehicle-related parameters represent a set of multiple different types of parameters; the current parameter vector constructed based on the current vehicle-related parameters further includes: numerically processing each type of parameter to obtain a vector element corresponding to each type of parameter; constructing a parameter vector corresponding to each type of parameter based on the vector elements corresponding to each type of parameter to obtain the current parameter vector corresponding to the current vehicle-related parameters; wherein the current parameter vector includes a current battery parameter vector, a current motor parameter vector, a current driving parameter vector, and a current road parameter vector.

[0012] According to another aspect of the present application, a real-time vehicle speed recommendation device is provided, the recommendation device comprising: a construction module, which constructs a current parameter vector according to current vehicle-related parameters, and constructs a historical time series vector corresponding to the current moment according to historical parameters corresponding to corresponding historical moments in historical cycles; a vehicle speed combination module, which inputs the current parameter vector into a target first model to obtain a first vehicle speed combination, and inputs the historical time series vector into a target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model; and a recommendation module, which determines a current recommended vehicle speed range according to the first vehicle speed combination and the second vehicle speed combination.

[0013] According to one aspect of the present application, an electronic device is provided, including: a controller; and a memory for storing one or more programs, which, when the one or more programs are executed by the controller, perform the above-mentioned recommendation method.

[0014] According to one aspect of the present application, a computer-readable storage medium is also provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes the above-mentioned recommendation method.

[0015] According to one aspect of the present application, a computer program product or a computer program is also provided, the computer program product or the computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above-mentioned recommendation method.

[0016] This application constructs the current parameter vector in real time according to the vehicle-related parameters, that is, the point value vector at the real time, and constructs the historical time series vector corresponding to the current moment according to the historical parameters, that is, the segment value vector of the historical time period; the current parameter vector and the historical time series vector are respectively input into models with different structural layers to obtain two vehicle speed combinations, and the recommended vehicle speed range corresponding to the current moment is determined according to the two vehicle speed combinations, which can provide the driver with an accurate recommended vehicle speed range to optimize the driver's driving behavior, thereby reducing the vehicle's driving energy consumption. Among them, the historical time period vector is a segment value vector with a time series composed of point value vectors corresponding to multiple historical moments. Inputting it into a model with a higher number of structural layers can better extract features from the sequence data, making the obtained vehicle speed combination more accurate.

[0017] The above description is only an overview of the technical solution of the embodiment of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 It is a flowchart of a method for recommending real-time vehicle speed shown in an exemplary embodiment of the present application.

[0020] Figure 2 is based on Figure 1 The exemplary embodiment shown is a flowchart of another method for recommending real-time vehicle speed.

[0021] Figure 3 is based on Figure 1 The exemplary embodiment shown is a flowchart of another method for recommending real-time vehicle speed.

[0022] Figure 4 It is a schematic diagram of an application scenario of the real-time vehicle speed recommendation method of the present application.

[0023] Figure 5 It is a structural schematic diagram of a device for recommending real-time vehicle speed shown in an exemplary embodiment of the present application.

[0024] Figure 6 It is a structural diagram of a computer system of an electronic device shown in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0025] Here, exemplary embodiments will be described in detail, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0026] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0027] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0028] The term "multiple" as used in this application refers to two or more than two. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0029] The intelligent vehicle speed recommendation system in the related technology can provide the driver with a recommended vehicle speed that saves energy consumption, so as to improve the energy efficiency of the vehicle by optimizing the driver's driving behavior.

[0030] However, the intelligent speed recommendation system will be affected by factors such as vehicle conditions, resulting in low accuracy of the recommended speed it provides. It is unable to provide the driver with an accurate recommended speed, and thus cannot effectively reduce the vehicle's driving energy consumption.

[0031] To this end, one aspect of the present application provides a method for recommending real-time vehicle speed. Figure 1 , Figure 1 1 is a flowchart of a method for recommending real-time vehicle speed according to an exemplary embodiment of the present application. The recommendation method at least includes S110 to S130, which are described in detail as follows:

[0032] S110: constructing a current parameter vector according to current vehicle-related parameters, and constructing a historical time series vector corresponding to the current moment according to historical parameters corresponding to corresponding historical moments in the historical cycle.

[0033] The current vehicle-related parameters are parameters that characterize the vehicle during its driving process and are related to the vehicle, including but not limited to status parameters of various vehicle components, traffic flow parameters of the road on which the vehicle is driving, parameters of the vehicle's surrounding environment, etc.

[0034] The current parameter vector is a one-dimensional vector composed of different vehicle-related parameters during the vehicle's driving process at the current moment. It is a point value vector that represents a point in time, specifically including various vehicle-related parameters at the current moment.

[0035] The historical cycle is a relative cycle determined by taking the current time as the time scale. The historical cycle is the previous cycle immediately before the current time. For example, if the current time is 10:00, the cycle before 10:00 is determined as the historical cycle.

[0036] The historical time series vector is a vector matrix constructed from the historical vectors of multiple historical moments, that is, each element in the historical time series vector is a historical vector of a historical moment. The elements in the historical time series vector are arranged according to the time series, which can represent the historical vector set of multiple historical moments, similar to a segment value vector. The number of vector elements in the historical time series vector is related to the number of historical moments in the historical period.

[0037] Example of constructing historical time series vectors: According to the historical parameters corresponding to the corresponding historical moments in the historical period, determine the historical parameter vectors corresponding to each corresponding historical moment; input each historical parameter vector into the blank vector matrix according to the time series to obtain the historical time series vector corresponding to the current moment; wherein, the number of vectors in the blank vector matrix is ​​the same as the number of historical parameter vectors.

[0038] The historical parameter vector corresponding to the historical moment in the historical cycle that is longest from the current moment is input as the first vector in the blank vector matrix. Similarly, each historical parameter vector is input in chronological order to obtain a historical time series vector.

[0039] In some embodiments, the current vehicle-related parameters represent a collection of multiple different types of parameters, and the different types of parameters need to be processed to unify the different parameter units, for example: each type of parameter is digitized separately to obtain the vector elements corresponding to each type of parameter; based on the vector elements corresponding to each type of parameter, a parameter vector corresponding to each type of parameter is constructed to obtain the current parameter vector corresponding to the current vehicle-related parameters; wherein the current parameter vector includes the current battery parameter vector, the current motor parameter vector, the current driving parameter vector, and the current road parameter vector.

[0040] Exemplarily, the current battery parameters include battery power, battery voltage, battery temperature, and battery pack balancing parameters, which are numerically processed to obtain the corresponding vector elements a1, a2, a3, and a4, and the current battery parameter vector A is constructed. n =(a1, a2, a3, a4), the battery parameter vector of the corresponding historical moment in the historical period is also constructed in this way, and the historical battery parameter time series vector B corresponding to the current moment is constructed n =(A n-4 , A n-3 , A n-2 , A n-1 ), wherein the number of vector elements in the historical battery parameter time series vector is related to the corresponding historical moments in the historical period, and the number of historical moments here is 4.

[0041] In another exemplary embodiment, the current motor parameter vector includes motor temperature, engine speed, and engine torque, which are numerically processed to obtain respective corresponding vector elements c1, c2, and c3, and the current battery parameter vector C is constructed. n =(c1, c2, c3), the battery parameter vector of the corresponding historical moment in the historical period is also constructed in this way, and the historical motor parameter time series vector D corresponding to the current moment is constructed n =(C n-3 , C n-2 , C n-1 ). Here, the number of historical moments is 3, which unifies the number of vector elements in the point value vector and the segment value vector, in order to balance the characteristic elements of the two vectors and facilitate the fair calculation of subsequent vehicle speeds.

[0042] In another exemplary embodiment, the current driving parameters include vehicle speed, wind resistance, tire pressure, and weather, which are numerically processed to obtain the corresponding vector elements e1, e2, e3, and e4, and the current driving parameter vector E is constructed. n =(e1, e2, e3, e4), the driving parameter vector of the corresponding historical moment in the historical period is also constructed in this way, and the historical driving parameter time series vector F corresponding to the current moment is constructed n =(E n-4 , E n-3 , E n-2 , E n-1 ).

[0043] In another exemplary embodiment, the current road parameters include vehicle detection road parameters, satellite positioning road parameters, traffic flow parameters, and road condition parameters, which are numerically processed to obtain the corresponding vector elements g1, g2, g3, and g4. g1 is a two-dimensional parameter of the vehicle laser detection of surrounding objects; g2 = (whether it is congested, the distance of the congested road section, the total length of the road section, and the time to the destination), which represents the road conditions determined by satellite positioning; g3 = (the number of vehicles passing through the intersection per minute, the number of people passing through the intersection per minute, and the speed of other vehicles passing in front), which represents the road traffic flow; g4 = (whether there is construction ahead, the duration of vehicle congestion, and the distance of the vehicle congestion section), which represents the road maintenance situation. The current road parameter vector G is constructed based on g1, g2, g3, and g4. n =(g1, g2, g3, g4), the road parameter vector of the corresponding historical moment in the historical period is also constructed in this way, and the historical road parameter time series vector H corresponding to the current moment is constructed n =(G n-4 , G n-3 , G n-2 , G n-1 ).

[0044] S120: Input the current parameter vector into the target first model to obtain a first vehicle speed combination, and input the historical time series vector into the target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model.

[0045] The target first model and the target second model are vehicle speed recommendation models adapted to the current user (ie, the current driver).

[0046] The more structural layers a model has, the deeper its feature extraction is. For point-like structural data such as the current parameter vector, there is no need to perform deep feature extraction, that is, the first model with a relatively large number of structural layers can quickly extract the corresponding features, so that the first model can quickly output the first vehicle speed combination. For time series data such as historical time series vectors, deep feature extraction is required to accurately extract the corresponding features, so that the second vehicle speed combination output by the second model is more accurate.

[0047] In some embodiments, it is necessary to clean the current parameter vector and the historical time series vector to remove invalid or erroneous records, and to standardize or normalize the cleaned data so that the parameter values ​​are between [0, 1], which is conducive to the learning of the corresponding model. If there are missing parameter values, the data processing can be completed by interpolation or deletion. The processed current parameter vector and historical time series vector can be better applied to the corresponding model so that the corresponding model can output an accurate vehicle speed combination.

[0048] S130: Determine a current recommended vehicle speed range according to the first vehicle speed combination and the second vehicle speed combination.

[0049] Exemplarily, the lowest vehicle speed and the highest vehicle speed in the first vehicle speed combination and the second vehicle speed combination are respectively used as the lowest vehicle speed and the highest vehicle speed of the current recommended vehicle speed range to obtain the recommended vehicle speed range.

[0050] In another exemplary embodiment, the first speed combination includes a first maximum speed and a first minimum speed, and the second speed combination includes a second maximum speed and a second minimum speed. The first maximum speed and the second maximum speed are respectively multiplied by their corresponding preset weight coefficients, and the calculated products are summed to obtain the recommended maximum speed; the first minimum speed and the second minimum speed are respectively multiplied by their corresponding preset weight coefficients, and the calculated products are summed to obtain the recommended minimum speed; according to the recommended maximum speed and the recommended minimum speed, the current recommended speed range is determined. The corresponding calculation formula is as follows:

[0051] final_min_speed=weight_structured*V1+weight_sequence*V3;

[0052] final_max_speed=weight_structured*V2+weight_sequence*V4;

[0053] Among them, final_min_speed represents the minimum speed of the recommended speed interval, final_max_speed represents the maximum speed of the recommended speed interval, weight_structured represents the first preset weight coefficient corresponding to each speed in the first speed combination, weight_sequence represents the second preset weight coefficient corresponding to each speed in the second speed combination, V1 represents the minimum speed in the first speed combination (i.e., the first minimum speed), V2 represents the minimum speed in the second speed combination (i.e., the second minimum speed), V3 represents the maximum speed in the first speed combination (i.e., the first maximum speed), and V4 represents the maximum speed in the second speed combination (i.e., the second maximum speed). The first preset weight coefficient + the second preset weight coefficient = 1. This application does not limit the preset weight coefficients corresponding to the two speed combinations. They can be the same or different.

[0054] This embodiment constructs the current parameter vector, i.e., the point value vector at the real time, based on the vehicle-related parameters in real time, and constructs the historical time series vector corresponding to the current moment, i.e., the segment value vector of the historical period, based on the historical parameters; the current parameter vector and the historical time series vector are respectively input into models with different structural layers to obtain two vehicle speed combinations, and the recommended vehicle speed range corresponding to the current moment is determined based on the two vehicle speed combinations, which can provide the driver with an accurate recommended vehicle speed range to optimize the driver's driving behavior, thereby reducing the vehicle's driving energy consumption. Among them, the historical period vector is a segment value vector with a time series composed of point value vectors corresponding to multiple historical moments. Inputting it into a model with a higher number of structural layers can better extract features from the sequence data, making the obtained vehicle speed combination more accurate.

[0055] The process of determining the relevant recommended vehicle speed only takes into account the vehicle's own condition, and does not take into account other parameters that affect the vehicle's driving, resulting in the determined recommended speed being unsuitable for the current vehicle. The driver adjusts the speed based on the recommended speed with low accuracy, which cannot optimize the vehicle's energy consumption.

[0056] To this end, in another exemplary embodiment of the present application, by introducing road parameters, a detailed description is given of how to construct a current parameter vector according to the current vehicle-related parameters. Figure 2 , Figure 2 is based on Figure 1 The exemplary embodiment shown is a flow chart of another method for recommending real-time vehicle speed. Figure 1 The S110 shown further includes S210 to S220; wherein preset laser points are set around the vehicle, which are described in detail as follows:

[0057] S210: Sending a detection laser to the surrounding area through a preset laser point to receive reflected light of the detection laser.

[0058] The number and position of preset laser points can be adjusted according to different vehicles, and there is no fixed number and position limit. For example, a vehicle is preset with 192 laser points, and the pulse frequency is 10HZ, that is, theoretically the vehicle can send out a maximum of 1920 detection laser points per second, and obtain 1920 three-dimensional coordinates based on the reflected light of the received detection laser.

[0059] S220: Determine the three-dimensional coordinate parameters of the object in the surrounding area according to the reflected light, and determine the current parameter vector corresponding to the vehicle detection road parameter according to the three-dimensional coordinate parameters.

[0060] Exemplarily, each dimension parameter in a certain three-dimensional coordinate parameter is summed up, and the sum is used as a vector element, and so on, and a current parameter vector corresponding to the vehicle detection road parameter is constructed based on multiple vector elements.

[0061] In another exemplary embodiment, each dimension parameter in the three-dimensional coordinate parameter is multiplied by its corresponding preset scaling factor, and the calculated products are summed to obtain a normalized value; based on the normalized value and the number of preset laser points, a current parameter vector corresponding to the vehicle detection road parameter is constructed.

[0062] For example, the three-dimensional coordinate parameters (2.00, 0.00, 0.50) indicate that the object is 2 meters in front of the corresponding preset laser point, with a horizontal offset of 0 meters and a height offset of 0.5 meters. Each coordinate value is normalized to the interval [0, 1] to obtain (x1, x2, x3), and then the three values ​​are combined into one by multiplying by an appropriate scaling factor.

[0063] Y = round down (a1*x1) + round down (a2*x2) + round down (a3*x3);

[0064] Among them, a1, a2, and a3 are preset scaling factors corresponding to different dimensional parameters. In order to convert the normalized value into an integer, a1=a2=a3=10 5 , at this time, we get a two-dimensional vector [1920, Y], and flatten the data to get the current parameter vector corresponding to the vehicle detection road parameters.

[0065] The first model and the second model in the related art are relatively fixed, and their learning data are derived from the parameters generated during the vehicle driving process. In the scenario where the first model and the second model are applied, the lack of personalized adjustment for the driver will affect the output results of the model and the daily training of the model.

[0066] To this end, in another exemplary embodiment of the present application, the model is linked to the preset user identity information to adapt the personalized model for different drivers. For details, please refer to Figure 3 , Figure 3 is based on Figure 1 The exemplary embodiment shown is a flow chart of another method for recommending real-time vehicle speed. Figure 1 Based on the S110 to S130 shown, at least S310 is also included, which is described in detail as follows:

[0067] S310: Match the current user identity information with the preset user identity information, and use the preset first model and the preset second model corresponding to the successfully matched preset user identity information as the target first model and the target second model, respectively; wherein the structural layer of the preset first model includes a convolutional neural network layer and a fully connected neural network layer, and the structural layer of the preset second model includes a recursive neural network layer and a long short-term memory neural network layer.

[0068] Exemplarily, the current user identity information is matched with the preset user identity information in Table 1 to quickly determine the target first model and the target second model suitable for the current user.

[0069] Table 1: Relationship between preset user identity information and corresponding preset models

[0070]

[0071]

[0072] The establishment of the preset first model is described below by way of example, and the relevant code examples are as follows:

[0073] model = models.Sequential()

[0074] #One-dimensional convolution layer, 64 filters, kernelsize is 6, do not use the activation function model.add(layers.Conv1D(64,6,activation=None,input_shape=input_shape))

[0075] #Add BatchNormalization layer

[0076] model.add(BatchNormalization())

[0077] #Add activation function

[0078] model.add(ReLU())

[0079] #Pooling layer

[0080] model.add(layers.MaxPooling1D(3))

[0081] #The second one-dimensional convolution layer, 128 filters

[0082] model.add(layers.Conv1D(128,3,activation='relu'))

[0083] #Add BatchNormalization layer

[0084] model.add(BatchNormalization())

[0085] #Add activation function

[0086] model.add(ReLU())

[0087] #Pooling layer

[0088] model.add(layers.MaxPooling1D(3))

[0089] # Flatten the layers

[0090] model.add(layers.Flatten())

[0091] #Fully connected layer, 64 neurons

[0092] model.add(layers.Dense(64,activation='relu'))

[0093] #Add a Dropout layer with a dropout rate of 0.2, which means that about 20% of the neurons will be randomly dropped

[0094] model.add(Dropout(0.2))

[0095] #Add a hidden layer with 128 neurons

[0096] model.add(layers.Dense(128,activation='relu'))

[0097] #Add a Dropout layer with a dropout rate of 0.5, which means that about 50% of the neurons will be randomly dropped

[0098] model.add(Dropout(0.5))

[0099] #Second hidden layer, 64 neurons

[0100] model.add(layers.Dense(64,activation='relu'))

[0101] #Add a Dropout layer with a dropout rate of 0.2, which means that about 20% of the neurons will be randomly dropped

[0102] model.add(Dropout(0.2))

[0103] #The third hidden layer, 32 neurons

[0104] model.add(layers.Dense(32,activation='relu'))

[0105] #Add a Dropout layer with a dropout rate of 0.1, which means that about 10% of the neurons will be randomly dropped

[0106] model.add(Dropout(0.1))

[0107] #Branch 1: Predict the minimum speed value (i.e. the lowest speed output by the preset first model)

[0108] min_speed_branch=layers.Dense(1,name='min_speed_output')(inputs)

[0109] #Branch 2: Predict the maximum speed value (i.e. the maximum speed output by the preset first model)

[0110] max_speed_branch=layers.Dense(num_classes,activation='softmax',name='max_speed_output')(inputs)

[0111] #Merge the two branches into one model

[0112] model=Model(inputs=inputs,outputs=[min_speed_branch,max_speed_branch])

[0113] # Compile the model

[0114] #Use mean square error as the loss function for both outputs

[0115] model.compile(optimizer='adam',loss=['mean_squared_error','mean_squared_error'])

[0116] Among them, the use of layers.Conv1D() is mainly to extract features from the input vector values ​​and learn more local features. At the same time, by stacking multiple convolutional layers, the network can learn feature representations of different scales, while reducing the amount of calculation and memory usage compared to the fully connected layer.

[0117] layers.MaxPooling1D() mainly performs dimensionality reduction operations on one-dimensional data, retaining the most important features in the sequence, ignoring insignificant changes, reducing the amount of calculation in subsequent layers, and improving the speed of model training and reasoning.

[0118] The purpose of layers.Flatten() is to better input data to the next layer.

[0119] layers.Dense() can integrate the features from the previous layer and extract more global features based on the previous local features. This time, three Dense() layers are designed for feature extraction.

[0120] Dropout layer: Appropriate regularization techniques are used to help the model better generalize to unseen data (for example, there are many parameters that characterize road conditions, and the number of parameters can be reduced in this way). It can also prevent network overfitting and improve generalization ability.

[0121] Batch normalization before the activation function can help stabilize and speed up the training process because it normalizes the output of the convolutional layer so that the activation function can work more effectively.

[0122] Max pooling can reduce the spatial size of the data after the activation function while retaining the important features processed by the activation function.

[0123] Activation function (ReLU): A non-linear activation function is applied after batch normalization to introduce non-linear characteristics.

[0124] This model combines 2 layers of layers.Conv1D() and 4 layers of layers.Dense() to obtain more local features and reduce model parameters. At the same time, more global features can be extracted based on the previous local features. At the same time, 2 layers of layers.Conv1D() will not lead to a large number of model parameters. Adding the corresponding Dropout() in the middle of 4 layers of layers.Dense() can effectively reduce the number of parameters and prevent overfitting. The dropout rates of 0.2, 0.5, 0.2, and 0.1 correspond to 64, 128, 64, and 2 layers of layers.Dense(), which can effectively reduce the computing power requirements.

[0125] The establishment of the preset second model is described below by way of example, and the relevant code examples are as follows:

[0126] #Assume that the input data is time series data with time_steps time steps, and the feature dimension of each time step is features

[0127] time_steps=100

[0128] features=16

[0129] #Create a sequential model

[0130] model=Sequential()

[0131] #Add a RNN layer

[0132] model.add(SimpleRNN(32,return_sequences=True,input_shape=(time_steps,features)))

[0133] #Add an LSTM layer, you can stack more RNN or LSTM layers

[0134] model.add(LSTM(64,return_sequences=True))

[0135] model.add(SimpleRNN(64,return_sequences=True))

[0136] model.add(LSTM(128,return_sequences=True))

[0137] #Add a fully connected layer, the TimeDistributed layer can apply a Dense layer to the output of each time step

[0138] model.add(TimeDistributed(Dense(32,activation='relu')))

[0139] #Add additional fully connected layers to extract features

[0140] model.add(Dense(64,activation='relu'))

[0141] #Branch 1: Predict the minimum speed value (i.e. the lowest speed output by the preset second model)

[0142] min_speed_branch=Dense(1,name='min_speed_output')(inputs)

[0143] #Branch 2: Predict the maximum speed value (i.e. the maximum speed output by the preset second model)

[0144] max_speed_branch=Dense(num_classes,activation='softmax',name='max_spe ed_output')(inputs)

[0145] #Merge the two branches into one model

[0146] model=Model(inputs=inputs,outputs=[min_speed_branch,max_speed_branch])

[0147] # Compile the model

[0148] #Because there are two outputs, we need to define two loss functions

[0149] model.compile(optimizer='adam',loss={'min_speed_output':'mean_squared_er ror','max_speed_output':'categorical_crossentropy'})

[0150] Among them, the RNN layer is particularly suitable for processing sequence data because it can capture the dynamic characteristics in the time series. Time series data mainly refers to historical records such as vehicle speed, acceleration, and battery charge status.

[0151] The LSTM layer is a special type of RNN that can learn long-term dependencies and is very suitable for processing and predicting time series data. It is mainly used to capture dynamic features and long-term dependencies in time series. It uses the fast processing capabilities of the RNN layer and the long-term dependency learning capabilities of the LSTM layer.

[0152] The RNN layer is used to better capture the dynamic features in the time series, and the LSTM layer can learn long-term dependencies. Adding an RNN layer between two LSTM layers can capture dynamic features again after learning long-term dependencies and analyze features in more depth. The number of layers of 32, 64, 64, and 128 is also set to fit the data size. At the same time, additional fully connected layers are added later to extract features again, so as to better integrate dynamic features, dependencies, and global and local relationships for subsequent output calculations.

[0153] In another exemplary embodiment of the present application, the application scenarios of the above multiple recommendation methods are exemplarily described. Figure 4 , Figure 4 Schematic diagram of the application scenario of the real-time vehicle speed recommendation method of the present application, wherein the vehicle 100, the server 200 and the display screen 300 are included, and the three ends can be connected by wireless communication, and the present application does not limit the connection method between them.

[0154] The display screen 300 is a component placed in the vehicle 100 and can display the recommended speed range determined by the server 200. The server 200 can Figure 4 The device is shown as being placed in the vehicle 100, or may be independent of the vehicle 100, and serves as an execution subject of the recommended method shown in any of the above exemplary embodiments to execute any of the above recommended methods, as exemplarily described as follows:

[0155] The server 200 constructs a current parameter vector based on the relevant parameters of the current vehicle 100, and constructs a historical time series vector corresponding to the current moment based on the historical parameters corresponding to the corresponding historical moment in the historical period; the server 200 inputs the current parameter vector into the target first model to obtain a first vehicle speed combination, and inputs the historical time series vector into the target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model; the server 200 determines the current recommended vehicle speed range based on the first vehicle speed combination and the second vehicle speed combination.

[0156] If the server 300 is independent of the vehicle 100, it can be a server cluster or a distributed system composed of multiple physical servers, wherein multiple servers can form a blockchain, and the server is a node on the blockchain. The server 300 can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network) and big data and artificial intelligence platforms, and this is not restricted here.

[0157] Another aspect of the present application also provides a real-time vehicle speed recommendation device, such as Figure 5 As shown, Figure 5 : is a schematic diagram of the structure of a real-time vehicle speed recommendation device shown in an exemplary embodiment of the present application. The recommendation device 500 includes:

[0158] The construction module 510 constructs a current parameter vector according to the current vehicle-related parameters, and constructs a historical time series vector corresponding to the current moment according to the historical parameters corresponding to the corresponding historical moments in the historical cycle.

[0159] The vehicle speed combination module 530 inputs the current parameter vector into the target first model to obtain the first vehicle speed combination, and inputs the historical time series vector into the target second model to obtain the second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model.

[0160] The recommendation module 550 determines a current recommended vehicle speed range according to the first vehicle speed combination and the second vehicle speed combination.

[0161] In another exemplary embodiment, the current vehicle-related parameters include current road parameters, which include vehicle-detected road parameters; preset laser points are set around the vehicle; and the construction module 510 further includes:

[0162] The detection unit sends a detection laser to the surrounding area through a preset laser point position to receive the reflected light of the detection laser.

[0163] The determination unit determines the three-dimensional coordinate parameters of the object in the surrounding area according to the reflected light, and determines the current parameter vector corresponding to the vehicle detection road parameter according to the three-dimensional coordinate parameters.

[0164] In another exemplary embodiment, the determining unit further comprises:

[0165] Normalization section, multiply each dimension parameter in the three-dimensional coordinate parameter by its corresponding preset scaling factor, and sum the calculated products to obtain a normalized value.

[0166] The construction section constructs the current parameter vector corresponding to the vehicle detection road parameters based on the normalized value and the number of preset laser points.

[0167] In another exemplary embodiment, the recommendation device 500 further includes:

[0168] The identity matching module matches the current user identity information with the preset user identity information, and uses the preset first model and the preset second model corresponding to the successfully matched preset user identity information as the target first model and the target second model, respectively; wherein the structural layer of the preset first model includes a convolutional neural network layer and a fully connected neural network layer, and the structural layer of the preset second model includes a recursive neural network layer and a long short-term memory neural network layer.

[0169] In another exemplary embodiment, the building block 510 further includes:

[0170] The historical parameter vector determination unit determines the historical parameter vector corresponding to each corresponding historical moment according to the historical parameters corresponding to the corresponding historical moment in the historical cycle.

[0171] The historical time series vector unit inputs each historical parameter vector into the blank vector matrix according to the time series to obtain the historical time series vector corresponding to the current moment; wherein the number of vectors in the blank vector matrix is ​​the same as the number of historical parameter vectors.

[0172] In another exemplary embodiment, the first vehicle speed combination includes a first maximum vehicle speed and a first minimum vehicle speed, and the second vehicle speed combination includes a second maximum vehicle speed and a second minimum vehicle speed; the recommendation module 550 further includes:

[0173] The first calculation unit multiplies the first maximum vehicle speed and the second maximum vehicle speed by their corresponding preset weight coefficients, and sums the calculated products to obtain a recommended maximum vehicle speed.

[0174] The second calculation unit multiplies the first minimum vehicle speed and the second minimum vehicle speed by their corresponding preset weight coefficients, and sums the calculated products to obtain a recommended minimum vehicle speed.

[0175] The vehicle speed recommendation unit determines a current recommended vehicle speed range according to the recommended maximum vehicle speed and the recommended minimum vehicle speed.

[0176] In another exemplary embodiment, the current vehicle-related parameters represent a collection of multiple different types of parameters; the building module 510 further includes:

[0177] The numerical processing unit performs numerical processing on each type parameter to obtain a vector element corresponding to each type parameter.

[0178] The current parameter vector construction unit constructs the parameter vector corresponding to each type of parameter according to the vector elements corresponding to each type of parameter, so as to obtain the current parameter vector corresponding to the current vehicle-related parameters; wherein the current parameter vector includes the current battery parameter vector, the current motor parameter vector, the current driving parameter vector, and the current road parameter vector.

[0179] The recommendation device of the present application constructs the current parameter vector, i.e., the point value vector at the real time, in real time according to the vehicle-related parameters, and constructs the historical time series vector corresponding to the current moment, i.e., the segment value vector of the historical period, according to the historical parameters; the current parameter vector and the historical time series vector are respectively input into models with different structural layers to obtain two vehicle speed combinations, and the recommended vehicle speed range corresponding to the current moment is determined according to the two vehicle speed combinations, which can provide the driver with an accurate recommended vehicle speed range to optimize the driver's driving behavior, thereby reducing the vehicle's driving energy consumption. Among them, the historical period vector is a segment value vector with a time series composed of point value vectors corresponding to multiple historical moments. Inputting it into a model with a higher number of structural layers can better extract features from the sequence data, making the obtained vehicle speed combination more accurate.

[0180] It should be noted that the recommendation device provided in the above embodiment and the recommendation method provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs the operation has been described in detail in the method embodiment and will not be repeated here.

[0181] Another aspect of the present application provides an electronic device, including: a controller; a memory for storing one or more programs, when the one or more programs are executed by the controller, the above-mentioned recommendation method is executed.

[0182] See also Figure 6 , Figure 6 1 is a schematic diagram of the structure of a computer system of an electronic device shown in an exemplary embodiment of the present application, which shows a schematic diagram of the structure of a computer system of an electronic device suitable for implementing an embodiment of the present application.

[0183] It should be noted that Figure 6 The computer system 600 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0184] like Figure 6 As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 602 or the program loaded from the storage part 608 to the random access memory (RAM) 603, such as executing the method in the above embodiment. In the RAM 603, various programs and data required for system operation are also stored. The CPU 601, ROM 602 and RAM 603 are connected to each other through a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0185] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed so that a computer program read therefrom is installed into the storage section 608 as needed.

[0186] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication section 609, and / or installed from a removable medium 611. When the computer program is executed by a central processing unit (CPU) 601, various functions defined in the system of the present application are executed.

[0187] It should be noted that the computer-readable medium shown in the embodiment of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as a part of a carrier wave, wherein a computer-readable computer program is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0188] The flowchart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operations of the system, method and computer program product according to various embodiments of the present application. Wherein, each box in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0189] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. The names of these units do not, in some cases, constitute limitations on the units themselves.

[0190] Another aspect of the present application further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the aforementioned recommended method is implemented. The computer-readable storage medium may be included in the electronic device described in the above embodiment, or may exist independently without being assembled into the electronic device.

[0191] Another aspect of the present application also provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the recommendation method provided in each of the above embodiments.

[0192] According to one aspect of an embodiment of the present application, a computer system is also provided, including a central processing unit (CPU), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage portion into a random access memory (RAM), such as executing the method in the above embodiment. In RAM, various programs and data required for system operation are also stored. CPU, ROM and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0193] The following components are connected to the I / O interface: an input part including a keyboard, a mouse, etc.; an output part including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker; a storage part including a hard disk, etc.; and a communication part including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication part performs communication processing via a network such as the Internet. A drive is also connected to the I / O interface as needed. Removable media, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., are installed on the drive as needed so that the computer program read therefrom is installed into the storage part as needed.

[0194] The above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. A person skilled in the art can easily make corresponding changes or modifications based on the main concept and spirit of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection required by the claims.

Claims

1. A method for recommending real-time vehicle speed, characterized in that: The recommended methods include: The current parameter vector is constructed based on the current vehicle-related parameters, and the historical time series vector corresponding to the current moment is constructed based on the historical parameters corresponding to the corresponding historical moment in the historical cycle; Inputting the current parameter vector into the target first model to obtain a first vehicle speed combination, and inputting the historical time series vector into the target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model; A current recommended vehicle speed range is determined according to the first vehicle speed combination and the second vehicle speed combination.

2. The recommendation method according to claim 1, characterized in that: The current vehicle related parameters include current road parameters, and the current road parameters include vehicle detected road parameters; Preset laser points are set around the vehicle; The step of constructing a current parameter vector according to current vehicle related parameters further includes: Sending a detection laser to a surrounding area through the preset laser point to receive reflected light of the detection laser; The three-dimensional coordinate parameters of the object in the surrounding area are determined according to the reflected light, and the current parameter vector corresponding to the vehicle detection road parameter is determined according to the three-dimensional coordinate parameters.

3. The recommendation method according to claim 2, characterized in that: The step of determining a current parameter vector corresponding to the vehicle detected road parameter according to the three-dimensional coordinate parameter further includes: Multiplying each dimension parameter in the three-dimensional coordinate parameter by the corresponding preset scaling factor, and summing the calculated products to obtain a normalized value; According to the normalized value and the number of the preset laser points, a current parameter vector corresponding to the vehicle detection road parameter is constructed.

4. The recommendation method according to claim 1, characterized in that: The recommended method also includes: The current user identity information is matched with the preset user identity information, and the preset first model and the preset second model corresponding to the successfully matched preset user identity information are used as the target first model and the target second model, respectively; wherein the structural layer of the preset first model includes a convolutional neural network layer and a fully connected neural network layer, and the structural layer of the preset second model includes a recursive neural network layer and a long short-term memory neural network layer.

5. The recommendation method according to claim 1, characterized in that: The constructing a historical time series vector corresponding to the current moment according to the historical parameters corresponding to the corresponding historical moment in the historical cycle further includes: According to the historical parameters corresponding to the corresponding historical moments in the historical cycle, the historical parameter vectors corresponding to the corresponding historical moments are determined; Input each historical parameter vector into a blank vector matrix according to the time series to obtain the historical time series vector corresponding to the current moment; wherein the number of vectors in the blank vector matrix is ​​the same as the number of the historical parameter vectors.

6. The recommendation method according to any one of claims 1 to 5, characterized in that: The first vehicle speed combination includes a first maximum vehicle speed and a first minimum vehicle speed, and the second vehicle speed combination includes a second maximum vehicle speed and a second minimum vehicle speed; The determining of the current recommended vehicle speed range according to the first vehicle speed combination and the second vehicle speed combination further includes: Multiplying the first maximum vehicle speed and the second maximum vehicle speed by their respective corresponding preset weight coefficients, and summing the calculated products to obtain a recommended maximum vehicle speed; Multiplying the first minimum vehicle speed and the second minimum vehicle speed by their respective corresponding preset weight coefficients, and summing the calculated products to obtain a recommended minimum vehicle speed; A current recommended vehicle speed range is determined according to the recommended maximum vehicle speed and the recommended minimum vehicle speed.

7. The recommendation method according to any one of claims 1 to 5, characterized in that: The current vehicle-related parameters represent a collection of multiple different types of parameters; The step of constructing a current parameter vector according to current vehicle related parameters further includes: Numerically process each type parameter to obtain the vector element corresponding to each type parameter; According to the vector elements corresponding to each type of parameter, a parameter vector corresponding to each type of parameter is constructed to obtain a current parameter vector corresponding to the current vehicle-related parameters; wherein the current parameter vector includes a current battery parameter vector, a current motor parameter vector, a current driving parameter vector, and a current road parameter vector.

8. A device for recommending real-time vehicle speed, characterized in that: The recommended devices include: A construction module is used to construct a current parameter vector based on the current vehicle-related parameters, and to construct a historical time series vector corresponding to the current moment based on the historical parameters corresponding to the corresponding historical moment in the historical cycle; A vehicle speed combination module, inputting the current parameter vector into a target first model to obtain a first vehicle speed combination, and inputting the historical time series vector into a target second model to obtain a second vehicle speed combination; wherein the number of structural layers of the target first model is less than the number of structural layers of the target second model; The recommendation module determines a current recommended vehicle speed range according to the first vehicle speed combination and the second vehicle speed combination.

9. An electronic device, characterized in that: include: Controller; A memory for storing one or more programs, which, when executed by a controller, enables the controller to implement the recommended method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the recommendation method according to any one of claims 1 to 7.

Images