CN118494507A - Vehicle energy consumption prediction method and device - Google Patents

Abstract

The invention discloses a vehicle energy consumption prediction method and device, relates to the technical field of electric automobiles, and mainly aims to solve the problems that the conventional energy consumption prediction method based on a recurrent neural network is difficult to realize high-precision energy consumption prediction, and unreasonable measures are taken for drivers, so that uncertainty and anxiety in a journey are aggravated. The main technical scheme of the invention is as follows: acquiring historical energy consumption data of a vehicle, current external data of the vehicle on a current navigation path and predicted external data; according to the vehicle energy consumption prediction model, historical energy consumption data, current external data and predicted external data are processed to obtain a target energy consumption prediction result; and determining the driving measures of the driver according to the target energy consumption prediction result so that the driver can drive according to the driving measures. The method is used for energy consumption prediction of the vehicle.

Description

Vehicle energy consumption prediction method and device

Technical Field

The invention relates to the technical field of electric automobiles, in particular to a vehicle energy consumption prediction method and device.

Background

In the field of new energy automobiles, an energy consumption prediction technology is important. The technique aims to predict the energy consumption of a vehicle over a period of time in the future. Through the energy consumption condition, corresponding measures can be given to a driver, so that the driver can effectively relieve mileage anxiety in the driving process through the measures.

Currently, vehicle energy consumption prediction methods rely mostly on traditional recurrent neural network models. The method predicts future vehicle energy consumption by integrating vehicle driving data at different times as input into the model, i.e. the method fully considers the influence of vehicle history energy consumption data on future energy consumption, for example, when the vehicle is used in high intensity in peak time, the battery temperature may be obviously increased, even if the vehicle is driven in low peak time, the energy conversion efficiency of the battery is continuously influenced by the higher initial battery temperature, and the energy consumption is indirectly increased.

However, in practical applications, in addition to historical energy consumption data, vehicle energy consumption is subject to multiple external variable interventions, including air conditioning usage, weather conditions, and the like. This reveals that the existing energy consumption prediction method based on the traditional recurrent neural network is difficult to realize high-precision energy consumption prediction, so that measures for drivers are unreasonable, and uncertainty and anxiety in the journey are aggravated.

Disclosure of Invention

In view of the above problems, the invention provides a vehicle energy consumption prediction method and device, and mainly aims to solve the problems that the existing energy consumption prediction method based on the traditional recurrent neural network is difficult to realize high-precision energy consumption prediction, and the measures for drivers are unreasonable, so that uncertainty and anxiety in the journey are aggravated.

In order to solve the technical problems, the invention provides the following scheme:

In a first aspect, the present invention provides a vehicle energy consumption prediction method, the method comprising:

Acquiring historical energy consumption data of a vehicle, current external data of the vehicle on a current navigation path and predicted external data;

processing the historical energy consumption data, the current external data and the predicted external data according to a vehicle energy consumption prediction model to obtain a target energy consumption prediction result;

And determining a driving measure of the driver according to the target energy consumption prediction result so that the driver can drive according to the driving measure.

In a second aspect, the present invention provides a vehicle energy consumption prediction apparatus, the apparatus comprising:

the data acquisition unit is used for acquiring historical energy consumption data of the vehicle, current external data of the vehicle on a current navigation path and predicted external data;

the data processing unit is used for processing the historical energy consumption data, the current external data and the predicted external data acquired by the data acquisition unit according to a vehicle energy consumption prediction model to acquire a target energy consumption prediction result;

And the measure determining unit is used for determining the driving measure of the driver according to the target energy consumption prediction result obtained by the data processing unit so that the driver can drive according to the driving measure.

In order to achieve the above object, according to a third aspect of the present invention, there is provided a storage medium including a stored program, wherein a device in which the storage medium is controlled to execute the vehicle energy consumption prediction method of the first aspect is controlled when the program is run.

In order to achieve the above object, according to a fourth aspect of the present invention, there is provided a processor for running a program, wherein the program, when run, performs the vehicle energy consumption prediction method of the first aspect described above.

By means of the technical scheme, the vehicle energy consumption prediction method and device provided by the invention can be used for firstly acquiring historical energy consumption data of the vehicle, current external data of the vehicle on a current navigation path and predicted external data. And then, processing the data by using the vehicle energy consumption prediction model so as to obtain a target energy consumption prediction result. Finally, according to the target energy consumption prediction result, the driving measures of the driver can be determined, so that the driver drives according to the measures. Compared with the prior art, the input of the vehicle energy consumption prediction model not only has historical energy consumption data, but also comprises current external data and predicted external data. The model can consider the instant influence of the current external data on the energy consumption, and bridge the current driving environment and the future driving environment through the predicted external data, so that the large deviation between the predicted result and the actual situation caused by the abrupt change of the external conditions is effectively reduced. Therefore, the predicted target energy consumption prediction result is more accurate and is closer to the actual situation. Based on the energy consumption prediction result, more reasonable driving measures can be provided for a driver, mileage anxiety in the driving process is remarkably reduced, and uncertainty in the driving process is reduced.

The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.

Drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:

Fig. 1 shows a schematic structural diagram of a vehicle energy consumption prediction model according to an embodiment of the present invention;

FIG. 2 shows a flow chart of a vehicle energy consumption prediction method provided by an embodiment of the invention;

FIG. 3 shows a flowchart of another vehicle energy consumption prediction method provided by an embodiment of the present invention;

fig. 4 shows a block diagram of a vehicle energy consumption prediction apparatus according to an embodiment of the present invention;

fig. 5 shows a block diagram of another vehicle energy consumption prediction apparatus according to an embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Aiming at the problem that the current vehicle energy consumption prediction method only depends on historical energy consumption data, and the actual influence of external variables on energy consumption is not fully considered, so that the prediction result is not accurate enough, the inventor provides a vehicle energy consumption prediction method through intensive research. The method consumes time in prediction, not only deeply digs historical energy consumption data, but also takes real-time external data of the vehicle on the current navigation path and predicted external data into consideration. By fusing the current external data and the predicted external data, the accuracy of the target energy consumption prediction result can be improved, and meanwhile, the problem that large deviation exists between the target energy consumption prediction result and the actual situation due to abrupt change of external conditions is further avoided, so that the target energy consumption result output by the vehicle energy consumption prediction model is more attached to the actual driving scene. According to the target energy consumption prediction result, more reasonable and more practical running measure suggestions can be provided for a driver. The measures not only can help a driver to manage the energy consumption of the vehicle more effectively in the driving process, but also can obviously reduce mileage anxiety and driving pressure caused by the uncertainty of the energy consumption, thereby greatly improving driving experience.

Next, a vehicle energy consumption prediction method provided by the present invention will be described with reference to fig. 1, where specific execution steps of the method are shown in fig. 1, and the method includes:

101. Historical energy consumption data of the vehicle, current external data of the vehicle on a current navigation path and predicted external data are obtained.

The current external data comprise a vehicle speed characteristic parameter, a current navigation parameter, a current accessory energy consumption parameter, a current environment data parameter and a current vehicle equipment data parameter; the predicted external data includes a current vehicle speed characteristic parameter, a current navigation parameter, a current accessory energy consumption parameter, a predicted environment data parameter, a current vehicle device data parameter.

In view of the above external data, in addition to the environmental data parameters, the other parameters are mostly based on the individual driving habits of the driver, which are themselves unpredictable. Thus, in vehicle energy consumption prediction, the main difference between the predicted external data and the current external data may be focused on the environmental data parameters.

Of course, in addition to the environmental data parameters, the predicted vehicle speed characteristic parameter and the navigation parameter in the external data may be used. That is, in this case, the predicted external data may be selected according to actual situations, including a predicted vehicle speed feature parameter, a predicted navigation parameter, a current accessory energy consumption parameter, a predicted environment data parameter, and a current vehicle device data parameter, or the predicted external data may be selected to include a current vehicle speed feature parameter, a current navigation parameter, a current accessory energy consumption parameter, a predicted environment data parameter, and a current vehicle device data parameter, which are not specifically limited herein.

The vehicle speed characteristic parameters include: the running speed, the running state accelerator pedal opening, the running state brake pedal opening and the running state longitudinal acceleration; the navigation parameters comprise the vehicle speed, coordinates of a starting point position and an ending point position, distance and traffic flow state, road speed limit, traffic jam degree and charging pile information; the accessory energy consumption parameters include: energy consumption data of high-voltage accessories such as air conditioners, PTC, air compressors and the like; the environmental data parameters include: temperature, humidity, altitude, weather, etc.; the vehicle equipment data parameters include: vehicle weight, whether windows or sunroofs are open, tire pressure data, etc.

After comprehensively considering the content of the external data, the navigation path of the vehicle at present can be defined according to the vehicle navigation system, and then the acquisition mode of each parameter in the external data on the current navigation path is as follows: the vehicle speed characteristic parameter, the accessory energy consumption parameter and the equipment data parameter can be obtained through a vehicle end sensor equipment; the navigation parameters depend on real-time information provided by the vehicle navigation system; and environmental data parameters can be collected through an on-board weather forecast APP or related environmental sensors. In addition, the historical energy consumption data of the vehicle in the historical time period can be effectively extracted through a cloud big data platform or a vehicle management system or an on-board diagnostic system (OBD).

When the systems acquire corresponding data, the vehicle-mounted communication module is utilized to transmit the data to the cloud server. The cloud server is used as an execution main body of the invention and is responsible for receiving the data from different systems, including the historical energy consumption data of the vehicle, the current external data of the vehicle on the current navigation path and the predicted external data.

Wherein the historical energy consumption data may include: battery energy consumption data: including battery charge and discharge capacity (kWh), mileage (km), battery efficiency (e.g., hundred km power consumption), etc.; energy consumption efficiency: such as "power consumption per kilometer" or "hundred kilometers power consumption"; charging record: information including the time of starting and ending charging, the charging place, the charging mode (fast charging, slow charging, etc.), the charge amount, the battery temperature at the time of charging, etc.; travel data: including travel time, mileage, average speed, maximum speed, acceleration, deceleration, etc. of the vehicle.

102. And processing the historical energy consumption data, the current external data and the predicted external data according to the vehicle energy consumption prediction model to obtain a target energy consumption prediction result.

The vehicle energy consumption prediction model mentioned in the step is trained in advance based on historical data, and is deployed to the cloud after the training is completed, so that the data is processed. The model may be a NARX neural network model, or other non-linear neural network model suitable for processing non-linear data having time dependence and multiple dimensions.

In this step, the vehicle history energy consumption data acquired in step 101, the current external data on the current navigation path, and the predicted external data may be transmitted as input data to a pre-trained vehicle energy consumption prediction model. After the model receives the data, the model calculates and outputs a target energy consumption prediction result.

In processing these data, the vehicle energy consumption prediction model may employ two different strategies. One strategy is to firstly process historical energy consumption data of a vehicle and current external data on a current navigation path respectively to obtain a first energy consumption prediction result; and then, processing the historical energy consumption data of the vehicle and the predicted external data to obtain a second energy consumption prediction result. Finally, the model integrates the two prediction results to obtain a more comprehensive and accurate target energy consumption prediction result, and the strategy enables the model to analyze the influence of different data sources on energy consumption more accurately by processing current external data and predicting external data respectively, so that the prediction result is more targeted.

Another strategy is that the vehicle energy consumption prediction model can process these three types of data simultaneously-the historical energy consumption data, the current external data, and the predicted external data of the vehicle. And directly outputting a target energy consumption prediction result fused with all information through complex calculation in the model. The strategy reduces the calculation steps and improves the prediction efficiency by processing all data simultaneously.

In this embodiment, the processing strategy of the vehicle energy consumption prediction model may be flexibly selected according to the actual application scenario and requirement, and is not limited to a specific use strategy.

The target energy consumption prediction result may include: the energy consumption is predicted: representing the amount of energy that the vehicle is expected to consume over a certain period of time or a certain travel path in the future or the amount of electricity that the vehicle is expected to need to consume when the vehicle travels from the current location to the end location, typically expressed in terms of electricity consumption (kWh) or electricity consumption per kilometer (kWh/km); the remaining range is predicted: based on the predicted energy consumption and the current vehicle residual quantity, estimating the residual mileage of the vehicle which can be driven in the future; energy consumption in different driving modes; energy consumption under different road conditions, energy consumption change trend and the like.

Of course, the above-mentioned target energy consumption prediction result refers specifically to the predicted energy consumption of the vehicle on the current navigation path. However, to provide a more comprehensive and personalized driving experience, the present invention is also able to predict the target energy consumption of all possible paths of the vehicle from the starting position to the end position and feed back to the driver. The design not only enables drivers to know the energy consumption condition of the current navigation path more accurately, but also helps the drivers to compare the energy consumption differences of different paths, so that the most energy-saving driving route is selected. Through the optimization, a driver can better manage the energy consumption of the vehicle in the driving process, and the driving experience is improved.

103. And determining the driving measures of the driver according to the target energy consumption prediction result so that the driver can drive according to the driving measures.

After the target energy consumption prediction result is obtained, appropriate driving measures can be formulated according to the target energy consumption prediction result. And then, the formulated driving measures and the target energy consumption prediction result are sent to a driver together, so that the driver is helped to more comprehensively know the energy consumption condition of the current driving route. Of course, according to the actual requirement, the target energy consumption prediction result can be selected to be sent to the driver independently, so that the driver can make a corresponding driving decision in advance.

Among other things, travel measures may include, but are not limited to: according to the current position, the terminal position and the charging pile information in the navigation parameters of the vehicle, an optimal charging path is planned; reminding a driver to slow down or use a brake at a proper time to maximize the utilization of the energy recovery system; based on the navigation parameters, the optimal time for pure electric driving and range-extending mode switching in the current navigation path is given; and the start and stop and the power distribution of the range extender are intelligently controlled, so that the use proportion of the whole vehicle battery and the fuel is coordinated and distributed, and the strategy considers the fuel economy to the greatest extent on the basis of meeting the balance of the battery SOC (state of charge) and the power performance of the whole vehicle, effectively reduces the energy consumption and improves the residual driving mileage.

Based on the implementation manner of fig. 1, it can be seen that the vehicle energy consumption prediction method provided by the invention can firstly obtain the historical energy consumption data of the vehicle, the current external data of the vehicle on the current navigation path and the predicted external data. And then, processing the data by using the vehicle energy consumption prediction model so as to obtain a target energy consumption prediction result. Finally, according to the target energy consumption prediction result, the driving measures of the driver can be determined, so that the driver drives according to the measures. Compared with the prior art, the input of the vehicle energy consumption prediction model not only has historical energy consumption data, but also comprises current external data and predicted external data. The model can consider the instant influence of the current external data on the energy consumption, and bridge the current driving environment and the future driving environment through the predicted external data, so that the large deviation between the predicted result and the actual situation caused by the abrupt change of the external conditions is effectively reduced. Therefore, the predicted target energy consumption prediction result is more accurate and is closer to the actual situation. Based on the energy consumption prediction result, more reasonable driving measures can be provided for a driver, mileage anxiety in the driving process is remarkably reduced, and uncertainty in the driving process is reduced.

Further, as a refinement and extension to the embodiment shown in fig. 1, the embodiment of the present invention further provides another vehicle energy consumption prediction method, as shown in fig. 2, which specifically includes the following steps:

201. and judging whether the current navigation path of the vehicle is a preset typical path or not.

202. If yes, determining a target energy consumption prediction result of the vehicle on a preset typical path.

In steps 201 to 202, the cloud server first determines, through the vehicle navigation system and the vehicle communication module, a navigation path where the vehicle is currently located. Then, a matching value between the current navigation path and a preset typical path is calculated by using a path matching algorithm. The preset typical path is a navigation path which is planned in advance through a vehicle navigation system and is screened out from a starting position to an end position according to historical data, wherein the navigation path has the largest number of times of being selected.

After the calculation of the matching value is completed, it may be evaluated whether the matching value exceeds a preset threshold. If the matching value exceeds the preset threshold value, the system determines that the current navigation path is a preset typical path, and calculates a target energy consumption prediction result of the vehicle running on the typical path according to the current navigation path. However, if the match value fails to reach the preset threshold, indicating that the current navigation path is an atypical path, the system proceeds to step 203 to execute subsequent atypical path processing logic. Such a procedure ensures the accuracy of the energy consumption prediction and the flexibility of the system.

The preset threshold value can be 95 percent or can be adjusted according to actual conditions.

The target energy consumption prediction result of the vehicle running on the preset typical path may be aggregated vehicle historical energy consumption data, where the aggregated vehicle historical energy consumption data is data obtained by comprehensively analyzing energy consumption data generated by a plurality of vehicles in a plurality of historical time periods on the preset typical path. The method not only can remarkably improve the processing efficiency when the navigation path is matched with the preset typical path, but also can ensure that the target energy consumption prediction result under the condition has higher precision.

203. Historical energy consumption data of the vehicle, current external data of the vehicle on a current navigation path and predicted external data are obtained.

The implementation of step 203 is the same as that of step 101, and the same technical effects can be achieved, so that the same technical problems are solved, and the detailed description is not repeated here.

204. And processing the historical energy consumption data, the current external data and the predicted external data according to the vehicle energy consumption prediction model to obtain a target energy consumption prediction result.

In the present embodiment, it is preferable to process the current external data and the predicted external data separately. Specifically, historical energy consumption data and current external data are input into a vehicle energy consumption prediction model to obtain a first energy consumption prediction result of a vehicle; and simultaneously, inputting the historical energy consumption data and the predicted external data into a vehicle energy consumption prediction model to obtain a second energy consumption prediction result of the vehicle. And then obtaining a target energy consumption prediction result according to the first energy consumption prediction result and the second energy consumption prediction result.

The vehicle energy consumption prediction model is an NARX neural network model.

When the first energy consumption prediction result and the second energy consumption prediction result are combined to obtain the target energy consumption prediction result, first, a first weight corresponding to the first energy consumption prediction result and a second weight corresponding to the second energy consumption prediction result are calculated according to a preset weight calculation method. To ensure the rationality of the weights, the sum of the two weights, i.e., the weight sum, can be calculated. Then, it may be determined whether this weight sum value exceeds a predetermined specified value.

If the weight sum exceeds a specified value, this means that a weight overflow situation has occurred. In order to correct this situation, the first weight and the second weight need to be adjusted accordingly according to the relation between the weight sum value and the specified value, so as to obtain a first target weight corresponding to the first weight and a second target weight corresponding to the second weight.

Finally, a more accurate target energy consumption prediction result can be calculated by combining the first energy consumption prediction result and the second energy consumption prediction result by using the adjusted first target weight and the second target weight. The processing flow ensures the rationality of the weight and improves the accuracy of the energy consumption prediction result.

Of course, if the weights do not overflow, a more accurate target energy consumption prediction result is calculated according to the first weight and the second weight calculated for the first time by combining the first energy consumption prediction result and the second energy consumption prediction result. Weight overflow means that the sum of the first weight and the second weight is greater than one hundred percent (i.e., the specified value).

According to a preset weight calculation method, the steps of calculating a first weight corresponding to a first energy consumption prediction result and a second weight corresponding to a second energy consumption prediction result are as follows:

First, a first confidence coefficient corresponding to current external data corresponding to the first energy consumption prediction result and a second confidence coefficient corresponding to predicted external data corresponding to the second energy consumption prediction result can be confirmed.

Second, a sum between the first confidence and the second confidence is calculated.

Third, a first ratio between the first confidence coefficient and the sum value is calculated, and the first ratio is determined as a first weight corresponding to the first energy consumption prediction result.

Fourth, a second ratio between the second confidence and the sum is calculated, and the second ratio is determined as a second weight corresponding to the second energy consumption prediction result.

When determining the confidence of the current external data and the predicted external data, the confidence of the environment data parameters corresponding to the current external data and the predicted external data is mainly determined. In order to obtain the confidence level of the environmental data parameters, a professional weather service or API can be used, which can provide detailed data and corresponding confidence indexes.

Specifically, whether the current external data or the predicted external data is aimed at, specific values of various environmental data parameters such as temperature, humidity, altitude, weather and the like and corresponding confidence levels thereof can be obtained through the professional weather service or the API. The confidence levels of these different environmental parameters are then integrated using a weighted average or other suitable statistical method to derive a confidence estimate of the current external data or the predicted external data overall.

In addition, when the weight sum exceeds the specified value, the steps when the first weight and the second weight are correspondingly adjusted are as follows:

Firstly, calculating the ratio between a specified numerical value and the weight sum value, and determining the ratio as a proportionality coefficient;

Second, calculating a first product value between the proportionality coefficient and the first weight, and determining the first product value as a first target weight;

third, a second product value between the scaling factor and the second weight is calculated, and the second product value is determined as the second target weight.

Finally, a formula for obtaining the target energy consumption prediction result according to the first energy consumption prediction result, the first target weight, the second energy consumption prediction result and the second target weight is as follows:

target energy consumption prediction result=first energy consumption prediction result×first target weight (first weight) +second energy consumption prediction result×second target weight (second weight) (formula one).

Next, the training steps of the vehicle energy consumption prediction model proposed in the present embodiment will be described in detail:

first: defining a network structure (as shown in fig. 5, fig. 5 is a schematic diagram of a vehicle energy consumption prediction model according to the present embodiment)

(1) Defining the number of neurons for each layer

Input layer: the number of neurons of the input layer is set to 10 (the number of delay steps of the historical energy consumption data is 5 steps+5 dimensions of the external data (the characteristic parameter of the vehicle speed, the navigation parameter, the energy consumption parameter of the accessories, the environmental data parameter and the data parameter of the vehicle equipment).

Hidden layer: 2 hidden layers were set, with 20 neurons per hidden layer.

Output layer: the number of output layer neurons is 1.

(2) Initializing weights

All weights and offsets are initialized with random numbers, ensuring that they are within reasonable limits.

(3) Data preparation

The data package is divided into a training set, a verification set and a test set.

(4) Determining delay steps

Input delay 1:5, output delay 1: and 5, namely the delay steps of the historical energy consumption data and the delay steps of the external data are 5 steps.

(5) Selecting a nonlinear activation function

The nonlinear activation function selects a sigmoid function.

(6) Selecting a loss function

And (3) selecting a mean square error (Mean Square Error, MSE) as an error evaluation index, evaluating the predicted data, and verifying the effectiveness of the constructed network, wherein the smaller the mean value is, the better the predicted result is. The calculation formula is as follows:

wherein the method comprises the steps of As predicted values, y _t is a true value; the MSE is the degree of matching of the average predicted value to the true value, representing the expected value of the square of the error.

(7) Determination optimizer

An Adam optimization algorithm (Adaptive Moment Estimation, adam) is chosen as the optimizer of the network model, responsible for adjusting the weights in the network to minimize the loss function.

(8) Selecting a machine learning framework

Constructing a network using TensorFlow machine learning framework

And determining a network structure through the parameter setting, and constructing and completing the NARX neural network model.

Second,: training model

(1) Obtaining sample data

Training samples can be obtained through cloud big data platforms or vehicle ends (including vehicle end sensor devices, vehicle-mounted navigation systems, vehicle-mounted weather forecast APP or related environment sensors and the like), and the method comprises the following steps: historical energy consumption data and current external data of the vehicle end and predicted external data (the current external data at this time refers to the current time when the model is trained)

(2) Data processing

A: and performing data exploration, data cleaning and feature processing on the training samples.

First, the requirements are explicitly obtained, including the characteristic requirements of the data, the data format, or other requirements. And then probing from the sample data according to the required characteristic requirement and a given data format to obtain the required specified sample data. And secondly, cleaning, normalizing, missing value and abnormal value processing are carried out on the appointed sample data, so that target sample data is obtained. And finally, carrying out unified feature processing on the target sample data, wherein the method comprises the following steps: and (3) carrying out standardized normalization, coding, data conversion, feature mapping and feature reconstruction to obtain processed target sample data, so that the target sample data has accuracy, integrity, consistency and usability. And storing the processed target sample data for training and testing the neural network model.

After obtaining the target sample data, the sample data may be divided into a training set, a validation set, and a test set.

B: the step of the characteristic processing mainly comprises specific signals (namely speed characteristic parameters) representing driving styles, wherein the specific signals comprise a running state accelerator pedal opening, a running state brake pedal opening, a running speed and a running state longitudinal acceleration, and the in-stroke average value and standard deviation of the parameters are calculated to obtain a running state accelerator pedal opening average value, a running state accelerator pedal opening standard deviation, a running state vehicle speed average value, a running state vehicle speed standard deviation, a running state acceleration average value, a running state acceleration standard deviation, a running state deceleration average value and a running state deceleration standard deviation. The above 8 parameters and the driving energy consumption have a certain relation, but in actual calculation, the driving style cannot be represented by the related analysis results of all the parameters at the same time, and the data needs to be subjected to dimension reduction processing. And performing dimension reduction treatment on the high-dimension data sample by using a PCA principal component analysis method, wherein the purpose is to reduce the dimension of the sample on the basis of reserving the element characteristics of the sample to the greatest extent. One can synthesize 1 parameter, i.e. the first principal component parameter, from 8 parameters that characterize the driving style. The final target training data comprises a first principal component parameter, a processed navigation parameter, an accessory energy consumption parameter, an environment data parameter, equipment parameter data and historical energy consumption data.

(3) Model training

A: forward propagation: sample data (including historical energy consumption data and external data) in a given training set is calculated from an input layer to a hidden layer and then from the hidden layer to an output layer through a network. At each hidden layer and output layer, the weighted inputs of neurons are processed using an activation function.

B: calculating a prediction error: the output result is compared with a target value (actual energy consumption data), and a prediction error is calculated using an loss function (MSE).

C: back propagation: the gradient of each weight in the network is calculated using a back propagation algorithm based on the error value calculated by the loss function.

D: and (5) weight updating: weights and bias values in the network are updated using Adam optimization algorithms.

E: iterative training: repeating the steps 1 to 4, and training by using all samples in the training set. The weights of the network are updated each iteration.

F: verification and super-parameter adjustment: during the training process, the performance of the network is periodically assessed using the validation set. Parameters of the network are adjusted according to the performance of the validation set. Training is stopped when a predetermined number of iterations is reached or performance on the validation set can no longer be improved or when validation performance begins to drop.

G: model evaluation: and evaluating the performance of the finally trained model by using the test set, and calculating the prediction error on the test set to evaluate the generalization capability of the model.

H: and the real vehicle uses the trained neural network model to conduct energy consumption prediction, and the prediction accuracy is verified.

And if the verification accuracy does not reach the standard, readjusting the model until the prediction accuracy is reached, and obtaining the trained vehicle energy consumption prediction model.

It should be noted that, in the model training stage, since the predicted external data does not have an actual energy consumption value corresponding to the predicted external data as a label, model training is usually performed using only the external data and the historical energy consumption data acquired at that time.

205. And determining the driving measures of the driver according to the target energy consumption prediction result so that the driver can drive according to the driving measures.

In this step, a predicted consumption value in the target energy consumption prediction result may be obtained first, the predicted consumption value is an amount of electricity that is predicted to be consumed when the vehicle travels from the current position to the end position, then it may be determined whether the predicted consumption value is greater than the current amount of electricity of the vehicle, and if so, a charging path between the current position and the end position of the vehicle may be determined according to charging efficiencies of a plurality of charging stations between the current position and the end position of the vehicle and the predicted waiting time.

The information of the plurality of charging stations can be obtained according to navigation parameters.

Of course, if the predicted consumption value is not greater than the current electric quantity of the vehicle, other driving measures may be formulated for the vehicle, including: firstly, reminding a driver to slow down or use a brake at a proper time so as to maximize the utilization of the energy recovery system; secondly, based on navigation parameters, giving the best time for switching the pure electric driving and range-extending modes in the current navigation path; thirdly, starting and stopping of the range extender and power distribution are intelligently controlled, so that the use proportion of the whole vehicle battery and the fuel is coordinated and distributed.

Of course, the above other driving measures may also be formulated for the driver without distinguishing whether the predicted consumption value is greater than the current electric quantity of the vehicle, that is, whether the predicted consumption value is greater than the current electric quantity of the vehicle.

It should be noted that all the driving measures and the target energy consumption prediction results in the invention can be displayed on the large-screen at the vehicle end in a visual manner so as to be convenient for a driver to check.

In addition, the invention can also intuitively display the predicted energy consumption distribution diagram at the vehicle end by adopting a fusion augmented reality technology (Augmented Reality, AR) no matter aiming at the energy consumption prediction results of all navigation paths or the current navigation path, and a driver can observe the real-time change of the energy consumption on different driving paths through an AR interface and simulate the effect after taking specific energy-saving measures so as to enhance the intuitiveness of the participation feeling and the energy consumption management of the driver.

Further, as an implementation of the method shown in fig. 1, the embodiment of the invention further provides a vehicle energy consumption prediction device, which is used for implementing the method shown in fig. 1. The embodiment of the device corresponds to the embodiment of the method, and for convenience of reading, details of the embodiment of the method are not repeated one by one, but it should be clear that the device in the embodiment can correspondingly realize all the details of the embodiment of the method. As shown in fig. 3, the apparatus includes:

A data acquisition unit 301, configured to acquire historical energy consumption data of a vehicle, current external data of the vehicle on a current navigation path, and predicted external data;

A data processing unit 302, configured to process the historical energy consumption data, the current external data and the predicted external data obtained by the data obtaining unit 301 according to a vehicle energy consumption prediction model, so as to obtain a target energy consumption prediction result;

A measure determining unit 303 for determining a driving measure of the driver according to the target energy consumption prediction result obtained by the data processing unit 302, so that the driver can drive according to the driving measure.

Further, as an implementation of the method shown in fig. 2, another vehicle energy consumption prediction apparatus is provided in the embodiment of the present invention, which is configured to implement the method shown in fig. 2. The embodiment of the device corresponds to the embodiment of the method, and for convenience of reading, details of the embodiment of the method are not repeated one by one, but it should be clear that the device in the embodiment can correspondingly realize all the details of the embodiment of the method. As shown in fig. 4, the apparatus includes:

A data acquisition unit 301, configured to acquire historical energy consumption data of a vehicle, current external data of the vehicle on a current navigation path, and predicted external data;

A data processing unit 302, configured to process the historical energy consumption data, the current external data and the predicted external data obtained by the data obtaining unit 301 according to a vehicle energy consumption prediction model, so as to obtain a target energy consumption prediction result;

A measure determining unit 303 for determining a driving measure of the driver according to the target energy consumption prediction result obtained by the data processing unit 302, so that the driver can drive according to the driving measure.

In an alternative embodiment, before the data acquisition unit 301, the apparatus further comprises a path determination unit 304, the path determination unit 304 comprising:

the path judging module 3041 is configured to judge whether the current navigation path is a preset typical path, where the preset typical path is a path with the largest number of selected navigation paths from a starting position to an end position, which is planned in advance;

The data acquisition module 3042 is configured to acquire, if the path determination module 3041 determines that the path determination module is yes, aggregate vehicle historical energy consumption data on the preset typical path, and take the aggregate vehicle historical energy consumption data as a target energy consumption prediction result of the vehicle, where the aggregate vehicle historical energy consumption data is data obtained by comprehensively analyzing energy consumption data generated by a plurality of vehicles in a plurality of historical time periods on the preset typical path.

In an alternative embodiment, the path determining module 3041 is specifically configured to:

calculating a matching value between the current navigation path and the preset typical path by using a path matching algorithm;

Judging whether the matching value exceeds a preset threshold value or not;

if yes, determining the current navigation path as a preset typical path.

In an alternative embodiment, the data processing unit 302 includes:

The first input module 3021 is configured to input the historical energy consumption data and the current external data into the vehicle energy consumption prediction model to obtain a first energy consumption prediction result of the vehicle, where the current external data includes a current vehicle speed feature parameter, a current navigation parameter, a current accessory energy consumption parameter, a current environment data parameter, and a current vehicle equipment data parameter;

The second input module 3022 is configured to input the historical energy consumption data and the predicted external data into the vehicle energy consumption prediction model, to obtain a second energy consumption prediction result of the vehicle, where the vehicle energy consumption prediction model is a NARX neural network model, and the predicted external data includes a current vehicle speed feature parameter, a current navigation parameter, a current accessory energy consumption parameter, a predicted environment data parameter, and a current vehicle equipment data parameter;

The result determining module 3023 is configured to obtain a target energy consumption prediction result according to the first energy consumption prediction result obtained by the first input module 3021 and the second energy consumption prediction result obtained by the second input module 3022.

In an alternative embodiment, the result determining module 3023 is specifically configured to:

Calculating a first weight corresponding to the first energy consumption prediction result and a second weight corresponding to the second energy consumption prediction result according to a preset weight calculation method;

calculating a weight sum value between the first weight and the second weight;

judging whether the weight sum value exceeds a specified numerical value;

if yes, the first weight and the second weight are adjusted according to the weight sum value and the appointed numerical value, and a first target weight corresponding to the first weight and a second target weight corresponding to the second weight are obtained;

And obtaining a target energy consumption prediction result according to the first energy consumption prediction result, the first target weight, the second energy consumption prediction result and the second target weight.

In an alternative embodiment, the result determining module 3023 is specifically configured to, when adjusting the first weight and the second weight according to the weight sum value and the specified numerical value:

calculating the ratio between the appointed numerical value and the weight sum value, and determining the ratio as a proportionality coefficient;

Calculating a first product value between the proportionality coefficient and the first weight, and determining the first product value as the first target weight;

a second product value between the scaling factor and the second weight is calculated and determined as the second target weight.

In an alternative embodiment, the measure determining unit 303 includes:

The consumption value obtaining module 3031 is configured to obtain a predicted consumption value in the target energy consumption prediction result, where the predicted consumption value is an amount of electricity that needs to be consumed when the predicted vehicle travels from the current position to the end position;

the electric quantity judging module 3032 is configured to judge whether the predicted consumption value acquired by the consumption value acquiring module 3031 is greater than the current electric quantity of the vehicle;

the path planning module 3033 is configured to determine a charging path between the current position and the destination position according to charging efficiencies of a plurality of charging stations between the current position and the destination position and the estimated waiting time if the determination result of the electric quantity determining module 3032 is greater than the predetermined threshold value.

Further, an embodiment of the present invention further provides a storage medium, where the storage medium is configured to store a computer program, where the computer program controls, when running, a device where the storage medium is located to execute the vehicle energy consumption prediction method described in fig. 1-2.

Further, an embodiment of the present invention further provides a processor, where the processor is configured to execute a program, where the program executes the method for predicting vehicle energy consumption described in fig. 1-2.

In the foregoing embodiments, the descriptions of the embodiments are emphasized, and for parts of one embodiment that are not described in detail, reference may be made to related descriptions of other embodiments.

It will be appreciated that the relevant features of the methods and apparatus described above may be referenced to one another. In addition, the "first", "second", and the like in the above embodiments are for distinguishing the embodiments, and do not represent the merits and merits of the embodiments.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.

The algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general-purpose systems may also be used with the teachings herein. The required structure for a construction of such a system is apparent from the description above. In addition, the present invention is not directed to any particular programming language. It will be appreciated that the teachings of the present invention described herein may be implemented in a variety of programming languages, and the above description of specific languages is provided for disclosure of enablement and best mode of the present invention.

Furthermore, the memory may include volatile memory, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM), in a computer readable medium, the memory including at least one memory chip.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises an element.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and variations of the present application will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the application are to be included in the scope of the claims of the present application.

Claims (10)

1. A method for predicting vehicle energy consumption, the method comprising:

Acquiring historical energy consumption data of a vehicle, current external data of the vehicle on a current navigation path and predicted external data;

processing the historical energy consumption data, the current external data and the predicted external data according to a vehicle energy consumption prediction model to obtain a target energy consumption prediction result;

And determining a driving measure of the driver according to the target energy consumption prediction result so that the driver can drive according to the driving measure.

2. The method of claim 1, wherein prior to obtaining historical energy consumption data of the vehicle, current external data of the vehicle on the current navigation path, and predicted external data, the method further comprises:

Judging whether the current navigation path is a preset typical path or not, wherein the preset typical path is a path with the largest selected times in a plurality of navigation paths from a starting position to an end position, which are planned in advance;

if so, acquiring the historical energy consumption data of the aggregate vehicle on the preset typical path, and taking the historical energy consumption data of the aggregate vehicle as a target energy consumption prediction result of the vehicle, wherein the historical energy consumption data of the aggregate vehicle is obtained by comprehensively analyzing the energy consumption data generated by a plurality of vehicles in a plurality of historical time periods on the preset typical path.

3. The method of claim 2, wherein determining whether the current navigation path is a preset typical path comprises:

calculating a matching value between the current navigation path and the preset typical path by using a path matching algorithm;

Judging whether the matching value exceeds a preset threshold value or not;

if yes, determining the current navigation path as a preset typical path.

4. The method of claim 1, wherein processing the historical energy consumption data, the current external data, and the predicted external data according to a vehicle energy consumption prediction model to obtain a target energy consumption prediction result comprises:

Inputting the historical energy consumption data and the current external data into the vehicle energy consumption prediction model to obtain a first energy consumption prediction result of the vehicle, wherein the current external data comprises a current vehicle speed characteristic parameter, a current navigation parameter, a current accessory energy consumption parameter, a current environment data parameter and a current vehicle equipment data parameter;

Inputting the historical energy consumption data and the predicted external data into the vehicle energy consumption prediction model to obtain a second energy consumption prediction result of the vehicle, wherein the vehicle energy consumption prediction model is an NARX neural network model, and the predicted external data comprises a current vehicle speed characteristic parameter, a current navigation parameter, a current accessory energy consumption parameter, a predicted environment data parameter and a current vehicle equipment data parameter;

and obtaining a target energy consumption prediction result according to the first energy consumption prediction result and the second energy consumption prediction result.

5. The method of claim 4, wherein obtaining a target energy consumption prediction result from the first energy consumption prediction result and the second energy consumption prediction result comprises:

Calculating a first weight corresponding to the first energy consumption prediction result and a second weight corresponding to the second energy consumption prediction result according to a preset weight calculation method;

calculating a weight sum value between the first weight and the second weight;

judging whether the weight sum value exceeds a specified numerical value;

if yes, the first weight and the second weight are adjusted according to the weight sum value and the appointed numerical value, and a first target weight corresponding to the first weight and a second target weight corresponding to the second weight are obtained;

And obtaining a target energy consumption prediction result according to the first energy consumption prediction result, the first target weight, the second energy consumption prediction result and the second target weight.

6. The method of claim 5, wherein adjusting the first weight and the second weight according to the weights and values and the specified values comprises:

calculating the ratio between the appointed numerical value and the weight sum value, and determining the ratio as a proportionality coefficient;

Calculating a first product value between the proportionality coefficient and the first weight, and determining the first product value as the first target weight;

a second product value between the scaling factor and the second weight is calculated and determined as the second target weight.

7. The method of claim 1, wherein determining a driver's travel measure based on the target energy consumption prediction result comprises:

obtaining a predicted consumption value in the target energy consumption prediction result, wherein the predicted consumption value is the predicted electric quantity required to be consumed when the vehicle runs to the end position at the current position;

Judging whether the predicted consumption value is larger than the current electric quantity of the vehicle or not;

If the charging time is greater than the preset charging time, determining a charging path between the current position and the end position according to the charging efficiency of a plurality of charging stations between the current position and the end position and the expected waiting time.

8. A vehicle energy consumption prediction apparatus, characterized by comprising:

the data acquisition unit is used for acquiring historical energy consumption data of the vehicle, current external data of the vehicle on a current navigation path and predicted external data;

the data processing unit is used for processing the historical energy consumption data, the current external data and the predicted external data acquired by the data acquisition unit according to a vehicle energy consumption prediction model to acquire a target energy consumption prediction result;

And the measure determining unit is used for determining the driving measure of the driver according to the target energy consumption prediction result obtained by the data processing unit so that the driver can drive according to the driving measure.

9. A storage medium comprising a stored program, wherein the program, when run, controls a device in which the storage medium is located to execute the vehicle energy consumption prediction method according to any one of claims 1 to 7.

10. A processor for running a program, wherein the program when run performs the vehicle energy consumption prediction method according to any one of claims 1 to 7.