Commercial vehicle transient fuel consumption accurate prediction method, device and equipment based on random hybrid model and storage medium

Abstract

The invention discloses a commercial vehicle transient fuel consumption accurate prediction method, device and equipment based on a random hybrid model and a storage medium, and relates to the technical field of commercial vehicle fuel consumption prediction.The commercial vehicle transient fuel consumption accurate prediction method comprises the steps that model input data and a transient fuel consumption model set are obtained, and the model input data are preprocessed to obtain preprocessed model input data; empirical distribution is constructed based on the model input data and the transient fuel consumption model set, a model empirical distribution set is determined, and the model empirical distribution set comprises a model prior distribution set and a model posterior distribution set; and based on the model input data and the model empirical distribution set, selecting a credible model to predict fuel consumption and update prior distribution, determining a target prediction transient fuel consumption set, and based on the target prediction transient fuel consumption set, completing transient fuel consumption accurate prediction. According to the method, the preprocessed data are input into various transient fuel consumption models for training, empirical distribution is constructed to select a credible model for prediction, the prediction precision is remarkably improved through dynamic updating, and fuel consumption is effectively reduced.

Description

Technical Field

[0001] The present application relates to the technical field of commercial vehicle fuel consumption prediction, and in particular to a method, device, equipment and storage medium for accurately predicting transient fuel consumption of commercial vehicles based on a random mixing model. Background Art

[0002] Commercial vehicles are an important carrier of transportation, and they undertake most of the logistics and personnel transportation work in global economic activities. In terms of economic costs, commercial vehicles consume a lot of fuel. The number of commercial vehicles only accounts for 12% of the total number of cars, but they contribute up to 55% of carbon emissions. Fuel consumption accounts for a considerable proportion of the operating costs of commercial vehicles. In addition to affecting the economic benefits of transportation business, fuel consumption also affects the transportation cost of goods, and ultimately affects the cost of the entire industry chain and even the commodity prices of the whole society. In terms of environmental impact, commercial vehicles are the main contributors to exhaust emissions in the transportation field due to their long mileage and high fuel consumption. The gases and particulate matter they emit have a significant impact on air quality and global climate. Therefore, reducing the fuel consumption of commercial vehicles is of great significance to the economy and the environment. The fuel consumption model is the core of the predictive cruise control system of commercial vehicles, and its accuracy directly affects the energy consumption performance of the predictive cruise system. Therefore, how to establish an accurate and fast fuel consumption model is the focus and difficulty of the current field.

[0003] The existing fuel consumption models in existing practices can be divided into two categories: steady-state fuel consumption models and transient fuel consumption models. The steady-state fuel consumption model is used to describe the fuel consumption of the vehicle under stable driving conditions, and is achieved by establishing a mapping function between the speed, torque, power and other parameters of the power system and the fuel consumption rate. The transient fuel consumption model is used to describe the fuel consumption of the vehicle under dynamically changing conditions, including two modeling methods: "steady-state initial value + transient correction" and direct modeling based on transient variables. The "steady-state initial value + transient correction" modeling method in the transient fuel consumption model is based on the steady-state fuel consumption model, and adds a transient correction amount or multiplies the predicted fuel consumption by a transient correction coefficient based on transient variables. The direct modeling method based on transient variables in the transient fuel consumption model directly uses transient variables to model the fuel consumption, and corrects the model based on historical information.

[0004] However, the steady-state fuel consumption model in existing practices requires that the vehicle power system is in a relatively stable working state, and different power systems need to be calibrated separately, which results in high test costs and is not conducive to large-scale deployment. The parameters of the transient correction value of the "steady-state initial value + transient correction" modeling method in the transient fuel consumption model are not easy to obtain, and the historical information and data set based on the transient variable direct modeling method in the transient fuel consumption model are strongly coupled, making it difficult to promote and apply to other working conditions. Therefore, how to more efficiently and accurately predict transient fuel consumption to adapt to complex and changeable driving conditions has become an urgent problem to be solved.

[0005] The above content is only used to assist in understanding the technical solution of this application, and does not represent an admission that the above content is prior art. Summary of the Invention

[0006] The main object of this application is to provide a method, device, equipment and storage medium for accurately predicting the transient fuel consumption of commercial vehicles based on a random hybrid model, aiming to solve the technical problem of how to more efficiently and accurately predict transient fuel consumption to adapt to complex and changeable driving conditions.

[0007] To achieve the above object, this application proposes a method for accurately predicting the transient fuel consumption of commercial vehicles based on a random hybrid model, and the method includes:

[0008] Obtain model input data and a transient fuel consumption model set, and preprocess the model input data to obtain preprocessed model input data;

[0009] Construct an empirical distribution based on the model input data and the transient fuel consumption model set, and determine a model empirical distribution set, where the model empirical distribution set includes a model prior distribution set and a model posterior distribution set;

[0010] Select a credible model to predict fuel consumption based on the model input data and the model empirical distribution set and update the prior distribution, determine a target predicted transient fuel consumption set, and complete the accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

[0011] In one embodiment, the step of obtaining model input data and a transient fuel consumption model set includes:

[0012] Obtain vehicle characteristic parameters, driver operation signals, road section map information, steady-state initial value transient correction model, polynomial transient fuel consumption prediction model and target fuel consumption prediction model. The vehicle characteristic parameters include vehicle speed information, vehicle acceleration information and vehicle mass information. The driver operation signals include accelerator pedal opening information, brake pedal opening information, steering wheel angle information, steering wheel torque information and vehicle gear information. The road section map information includes road surface adhesion information, slope information and speed limit information;

[0013] Determine model input data based on the vehicle characteristic parameters, the driver operation signals and the road section map information;

[0014] Determine a transient fuel consumption model set based on the steady-state initial value transient correction model, the polynomial transient fuel consumption prediction model and the target fuel consumption prediction model.

[0015] In one embodiment, the step of constructing an empirical distribution based on the model input data and the transient fuel consumption model set and determining a model empirical distribution set includes:

[0016] Input the model input data into the transient fuel consumption model set for training to determine the predicted transient fuel consumption set;

[0017] Screen the fuel consumption data based on the model input data to construct the prior distribution set of the model;

[0018] Calculate the posterior distribution set of the model based on the predicted transient fuel consumption set and the prior distribution set of the model;

[0019] Obtain the empirical distribution set of the model based on the prior distribution set of the model and the posterior distribution set of the model.

[0020] In one embodiment, the step of screening the fuel consumption data based on the model input data to construct the prior distribution set of the model includes:

[0021] Screen the fuel consumption data based on the model input data to determine the historical fuel consumption data and the operating condition similarity value;

[0022] Construct the prior distribution set of the model based on the historical fuel consumption data and the operating condition similarity value.

[0023] In one embodiment, the step of calculating the posterior distribution set of the model based on the predicted transient fuel consumption set and the prior distribution set of the model includes:

[0024] Obtain the posterior distribution calculation method;

[0025] Calculate the posterior distribution set of the model based on the posterior distribution calculation method, the predicted transient fuel consumption set and the prior distribution set of the model.

[0026] In one embodiment, the step of selecting a credible model to predict fuel consumption and updating the prior distribution based on the model input data and the empirical distribution set of the model to determine the target predicted transient fuel consumption set includes:

[0027] Obtain the position of the posterior distribution confidence interval;

[0028] Calculate the confidence level based on the posterior distribution set of the model in the empirical distribution set of the model and the position of the posterior distribution confidence interval;

[0029] Judge the credibility of the model based on the confidence level, determine the fuel consumption credibility model, and predict the transient fuel consumption prediction value based on the fuel consumption credibility model;

[0030] Adjust the historical fuel consumption data based on the model input data and the transient fuel consumption prediction value, reconstruct the prior distribution set of the model, and determine the updated prior distribution set of the model;

[0031] Train the fuel consumption model based on the updated prior distribution set of the model to determine the optimized fuel consumption prediction model set;

[0032] Output a set of target predicted transient fuel consumptions based on the set of fuel consumption prediction optimization models.

[0033] In one embodiment, the steps of completing the accurate prediction of transient fuel consumption based on the set of target predicted transient fuel consumptions include:

[0034] Integrate the set of target predicted transient fuel consumptions to obtain a target accuracy transient fuel consumption prediction value;

[0035] Complete the accurate prediction of transient fuel consumption based on the target accuracy transient fuel consumption prediction value.

[0036] In addition, to achieve the above object, the present application also proposes a device for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model. The device for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model includes:

[0037] An acquisition module, configured to acquire model input data and a set of transient fuel consumption models, and preprocess the model input data to obtain preprocessed model input data;

[0038] A processing module, configured to construct an empirical distribution based on the model input data and the set of transient fuel consumption models, and determine a set of model empirical distributions, where the set of model empirical distributions includes a set of model prior distributions and a set of model posterior distributions;

[0039] An execution module, configured to select a credible model to predict fuel consumption and update the prior distribution based on the model input data and the set of model empirical distributions, determine a set of target predicted transient fuel consumptions, and complete the accurate prediction of transient fuel consumption based on the set of target predicted transient fuel consumptions.

[0040] In addition, to achieve the above object, the present application also proposes a device for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model. The device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor. The computer program is configured to implement the steps of the method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model as described above.

[0041] In addition, to achieve the above object, the present application also proposes a storage medium. The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model as described above.

[0042] One or more technical solutions proposed by the present application have at least the following technical effects:

[0043] A precise transient fuel consumption prediction method for commercial vehicles based on a random hybrid model proposed in this embodiment obtains model input data and a transient fuel consumption model set, preprocesses the model input data to obtain preprocessed model input data; constructs an empirical distribution based on the model input data and the transient fuel consumption model set to determine a model empirical distribution set, where the model empirical distribution set includes a model prior distribution set and a model posterior distribution set; selects a credible model to predict fuel consumption and updates the prior distribution based on the model input data and the model empirical distribution set to determine a target predicted transient fuel consumption set, and completes precise transient fuel consumption prediction based on the target predicted transient fuel consumption set. This application preprocesses the model input data to improve the signal-to-noise ratio, inputs multiple transient fuel consumption models for training, constructs a model prior distribution and a posterior distribution, and by evaluating the credibility of the prediction results of each model, selects a credible model for transient fuel consumption prediction and updates the prior distribution according to the prediction results. Finally, it integrates the prediction results of each model to output a high-precision transient fuel consumption prediction value, significantly improving the prediction accuracy. Applied to different driving conditions, it enhances the generalization ability of the model, facilitates large-scale deployment under different vehicles and different driving conditions, effectively reduces the fuel consumption of commercial vehicles, and reduces carbon emissions. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with the present application and, together with the specification, are used to explain the principles of the present application.

[0045] To more clearly illustrate the technical solutions in the embodiments of the present application or in the prior art, the following will briefly introduce the accompanying drawings required for use in the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0046] Figure 1 It is a schematic flowchart provided for Embodiment 1 of the precise transient fuel consumption prediction method for commercial vehicles based on a random hybrid model of the present application;

[0047] Figure 2 It is a schematic flowchart provided for Embodiment 2 of the precise transient fuel consumption prediction method for commercial vehicles based on a random hybrid model of the present application;

[0048] Figure 3 It is a schematic flowchart provided for the precise transient fuel consumption prediction method for commercial vehicles based on a random hybrid model of the embodiments of the present application;

[0049] Figure 4 It is a schematic module structure diagram of the precise transient fuel consumption prediction device for commercial vehicles based on a random hybrid model of the embodiments of the present application;

[0050] Figure 5Schematic diagram of the device structure of the hardware operating environment involved in the method for accurately predicting the transient fuel consumption of commercial vehicles based on a random hybrid model in the embodiments of the present application.

[0051] The implementation, functional features and advantages of the present application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. Detailed implementation manners

[0052] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0053] To better understand the technical solutions of the present application, the following will be described in detail in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0054] The main solution of the embodiments of the present application is: obtaining model input data and a transient fuel consumption model set, preprocessing the model input data to obtain preprocessed model input data; constructing an empirical distribution based on the model input data and the transient fuel consumption model set, determining a model empirical distribution set, the model empirical distribution set including a model prior distribution set and a model posterior distribution set; selecting a credible model to predict fuel consumption and updating the prior distribution based on the model input data and the model empirical distribution set, determining a target predicted transient fuel consumption set, and completing the accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

[0055] In this embodiment, for the convenience of description, the following will be described with the device for accurately predicting the transient fuel consumption of commercial vehicles based on a random hybrid model as the execution subject.

[0056] Since the steady-state fuel consumption model in the prior art requires the vehicle power system to be in a relatively stable working state and needs to be separately calibrated for different power systems, the test cost is high and it is not conducive to large-scale deployment. Moreover, the parameters of the transient correction amount in the "steady-state initial value + transient correction" modeling method in the transient fuel consumption model are not easy to obtain, and the historical information and data set of the direct modeling method based on transient variables in the transient fuel consumption model are strongly coupled, making it difficult to be popularized and applied to other working conditions.

[0057] The present application provides a solution, obtaining model input data and a transient fuel consumption model set, preprocessing the model input data to obtain preprocessed model input data; constructing an empirical distribution based on the model input data and the transient fuel consumption model set, determining a model empirical distribution set, the model empirical distribution set including a model prior distribution set and a model posterior distribution set; selecting a credible model to predict fuel consumption and updating the prior distribution based on the model input data and the model empirical distribution set, determining a target predicted transient fuel consumption set, and completing the accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

[0058] As can be seen from the above embodiments, the present application preprocesses the model input data to improve the signal-to-noise ratio, inputs multiple transient fuel consumption models for training, constructs the prior distribution and posterior distribution of the model, and by evaluating the credibility of the prediction results of each model, selects a credible model for transient fuel consumption prediction, and updates the prior distribution according to the prediction results. Finally, the prediction results of each model are integrated to output a high-precision transient fuel consumption prediction value, significantly improving the prediction accuracy. Applied to different driving conditions, the generalization ability of the model is enhanced, facilitating large-scale deployment under different vehicles and different driving conditions, effectively reducing the fuel consumption of commercial vehicles and reducing carbon emissions.

[0059] Based on this, an embodiment of the present application provides a method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model. Referring to Figure 1 , Figure 1 is a schematic flowchart of the first embodiment of the method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model of the present application.

[0060] In this embodiment, the method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model includes steps S10 to S30:

[0061] Step S10, obtain model input data and a set of transient fuel consumption models, and preprocess the model input data to obtain preprocessed model input data;

[0062] It should be noted that the model input data reflects the characteristics of various states and environmental factors of the vehicle during actual driving, and the set of transient fuel consumption models reflects the characteristics of a set of models that model and predict transient fuel consumption from different perspectives.

[0063] It can be understood that the model input data can characterize the fuel consumption changes of the vehicle under different working conditions, and the set of transient fuel consumption models can adapt to various complex driving conditions through different modeling methods and data processing methods, more accurately predict the vehicle fuel consumption. By obtaining and preprocessing the model input data and using multiple transient fuel consumption models for prediction, the accuracy of transient fuel consumption prediction can be significantly improved, and the accuracy of fuel consumption prediction can be improved. At the same time, the applicability of the model under different vehicles and different driving conditions is enhanced, effectively reducing the fuel consumption of commercial vehicles.

[0064] For ease of understanding, taking the acquisition of model input data and a set of transient fuel consumption models as an example for illustration, where the information acquisition device is an information acquisition module and the storage device is a memory.

[0065] The information acquisition module obtains vehicle characteristic parameters, driver operation signals, road section map information, steady-state initial value transient correction model, polynomial transient fuel consumption prediction model, and target fuel consumption prediction model. The vehicle characteristic parameters include vehicle speed information, vehicle acceleration information, and vehicle mass information. The driver operation signals include accelerator pedal opening information, brake pedal opening information, steering wheel angle information, steering wheel torque information, and vehicle gear information. The road section map information includes road surface adhesion information, slope information, and speed limit information. Based on the vehicle characteristic parameters, the driver operation signals, and the road section map information, model input data is determined, that is, the input required for the model such as vehicle parameters, driver operation signals, and road section map information where the vehicle is located is collected to obtain the model input data. Among them, vehicle parameters include dynamic and static parameters of the vehicle such as speed, acceleration, and mass. Driver operation signals include inputs to the vehicle by the driver during driving such as accelerator pedal opening, brake pedal opening, steering wheel angle, steering wheel torque, and vehicle gear. Map information includes road surface adhesion, slope, speed limit, etc. of the road section where the vehicle is located and will pass through in the future. The vehicle parameters and driver operation signals can be directly measured by on-vehicle sensors or indirectly measured through algorithms. The map information is extracted from the high-precision map in combination with vehicle GPS coordinates, vehicle path planning, etc., so as to obtain the model input data.

[0066] The model input data is preprocessed to obtain preprocessed model input data, that is, the input data required for the model collected is preprocessed. Among them, the purpose of preprocessing is to reduce noise in the data and exclude outliers. The noise reduction and outliers include, but are not limited to, noise generated by sensors or data acquisition modules, etc., and data with large errors. After excluding abnormal data, the data is denoised by symmetric exponential moving filtering to improve the signal-to-noise ratio of the input data. The expression of symmetric exponential moving filtering is:

[0067]

[0068] D = min{3Δ, i - 1, N α - 1}

[0069] where, x α (t k ) is the original data of vehicle α at the sampling moment k, is the smoothed data, N α is the total number of α sample points, T represents the time span of the smoothing operation, and dt represents the time interval between each frame.

[0070] Determine a transient fuel consumption model set based on the steady-state initial value transient correction model, the polynomial transient fuel consumption prediction model, and the target fuel consumption prediction model. That is, the selected transient fuel consumption models include, but are not limited to, two types of transient fuel consumption models: "steady-state initial value + transient correction" or direct modeling based on transient variables.

[0071] The model based on "steady-state initial value + transient correction" can be represented by multiple equations.

[0072] Perform steady-state initial value fuel consumption estimation, expressed as:

[0073]

[0074] Among them, is the steady-state fuel consumption rate, and the logarithmic steady-state fuel consumption rate m is obtained by logarithmic transformation s , and the engine torque T e and the rotational speed ω e are used as inputs for polynomial fitting. Since the vehicle fuel consumption rate is generally a binary function of the engine rotational speed or torque, and the order is mostly 2 or 3, three polynomial structures are respectively used for estimation.

[0075] Use the transient correction module to perform fitting using the vehicle speed v and acceleration a, expressed as:

[0076]

[0077] Among them, β is the parameter corresponding to the power.

[0078] Then combine the steady-state estimated value with the transient correction coefficient to predict the transient fuel consumption rate, expressed as:

[0079]

[0080] Thus, the "steady-state initial value + transient correction" model is obtained, that is, the steady-state initial value transient correction model is obtained.

[0081] The polynomial transient fuel consumption estimation model based on the least squares method is expressed as:

[0082]

[0083] Among them, α 0 、α 1 、α 2 、α 3 are the coefficients obtained by fitting using the least squares method, β 0 is the fuel consumption rate at vehicle idling, a is the vehicle acceleration, and v is the vehicle speed.

[0084] Thus, a polynomial transient fuel consumption estimation model based on the least squares method is obtained, that is, a polynomial transient fuel consumption prediction model is obtained.

[0085] The CNN-LSTM fuel consumption prediction model is expressed as:

[0086]

[0087] where f is a function representing the transient fuel consumption prediction neural network, x is the network input vector, and Φ is the network parameter.

[0088] The input vector x can be represented by relevant variables as:

[0089] x(t) = {v(t), a(t), P x (t), P y (t), y(t)}

[0090] where v(t), a(t), P x (t), P y (t), and y(t) are the vehicle speed, acceleration, longitude, latitude, and instantaneous fuel consumption at time t, respectively.

[0091] Thus, the CNN-LSTM fuel consumption prediction model is obtained, that is, the target fuel consumption prediction model is obtained.

[0092] Determine a transient fuel consumption model set based on the steady-state initial value transient correction model, the polynomial transient fuel consumption prediction model, and the target fuel consumption prediction model.

[0093] Perform subsequent processing based on the model input data and the transient fuel consumption model set.

[0094] In a feasible implementation manner, step S10 may include steps A11 to A13:

[0095] Step A11, obtain vehicle characteristic parameters, driver operation signals, road section map information, a steady-state initial value transient correction model, a polynomial transient fuel consumption prediction model, and a target fuel consumption prediction model. The vehicle characteristic parameters include vehicle speed information, vehicle acceleration information, and vehicle mass information. The driver operation signals include accelerator pedal opening information, brake pedal opening information, steering wheel angle information, steering wheel torque information, and vehicle gear information. The road section map information includes road surface adhesion information, slope information, and speed limit information;

[0096] It should be noted that the vehicle characteristic parameters reflect the characteristics of the vehicle's dynamic driving state, the driver's operation signal reflects the characteristics of the driver's operation intention and behavior, the road section map information reflects the characteristics of the detailed information of the vehicle driving environment, the steady-state initial value transient correction model reflects the characteristic of correcting the predicted fuel consumption by introducing transient variables on the basis of the steady-state fuel consumption model, the polynomial transient fuel consumption prediction model reflects the characteristic of directly modeling and predicting the transient fuel consumption by fitting variables such as the vehicle speed and acceleration using the least squares method, and the target fuel consumption prediction model reflects the characteristic of integrating multi-source information for high-precision fuel consumption prediction.

[0097] Step A12: Determine the model input data based on the vehicle characteristic parameters, the driver's operation signal, and the road section map information.

[0098] It can be understood that the vehicle characteristic parameters may include vehicle speed information, vehicle acceleration information, and vehicle mass information, which are collected in real time by on-vehicle sensors to more accurately predict fuel consumption. The driver's operation signal may include accelerator pedal opening information, brake pedal opening information, steering wheel angle information, steering wheel torque information, and vehicle gear information, which directly reflect the driver's operation intention and behavior and thus directly affect fuel consumption prediction. The road section map information may include road surface adhesion information, slope information, and speed limit information, which are extracted from the high-precision map through the vehicle's GPS coordinates and path planning and are used to predict fuel consumption changes under different road conditions.

[0099] Step A13: Determine the transient fuel consumption model set based on the steady-state initial value transient correction model, the polynomial transient fuel consumption prediction model, and the target fuel consumption prediction model.

[0100] It can be understood that the steady-state initial value transient correction model is a combination of steady-state initial value fuel consumption estimation, transient correction module, and transient fuel consumption rate, which can more accurately reflect the fuel consumption of the vehicle under dynamic changing working conditions. The polynomial transient fuel consumption prediction model can estimate the fuel consumption through a polynomial structure and accurately capture the non-linear relationship between fuel consumption and vehicle dynamic parameters. The target fuel consumption prediction model can provide a more comprehensive and accurate fuel consumption prediction result by integrating multi-source information.

[0101] Step S20: Construct an empirical distribution based on the model input data and the transient fuel consumption model set, and determine the model empirical distribution set, which includes the model prior distribution set and the model posterior distribution set.

[0102] It should be noted that the model empirical distribution set reflects the characteristic of more accurately estimating the fuel consumption distribution through different distribution calculation methods.

[0103] It can be understood that the model empirical distribution set only represents the inclusion relationship between the model prior distribution set and the model posterior distribution set, that is, it cannot be obtained by jointly calculating the model prior distribution set and the model posterior distribution set. Among them, the model prior distribution set can provide statistical information of historical data, while the model posterior distribution set is dynamically adjusted according to new data, thereby significantly improving the accuracy of transient fuel consumption prediction.

[0104] For the convenience of understanding, taking the determination of the model empirical distribution set as an example for illustration, where the information acquisition device is the information acquisition module, the storage device is the memory, and the processing device is the processing module.

[0105] The information acquisition module obtains the model input data and inputs it into the transient fuel consumption model set, trains the model input data by inputting it into the transient fuel consumption model set to determine the predicted transient fuel consumption set, that is, pre-trains the model input data using the fuel consumption data set, and then inputs the pre-trained data into selected multiple transient fuel consumption models respectively to predict the transient fuel consumption at the next moment. Based on the model input data, fuel consumption data is screened to determine historical fuel consumption data and the working condition similarity value, and a model prior distribution set is constructed based on the historical fuel consumption data and the working condition similarity value, that is, sampling the historical fuel consumption data to construct the prior distribution of each model respectively. The historical fuel consumption data includes the existing data set and the fuel consumption data continuously collected during the vehicle operation process. According to the model input data, similar working conditions are extracted from the fuel consumption data set. The fuel consumption data on the same section or under the same road condition with similar vehicle loads is extracted, and the data is weighted based on the degree of working condition similarity to obtain the model prior distribution set, and the Gaussian mixture model is used to fit the prior distribution.

[0106] Among them, the Gaussian mixture model is a linear combination of multiple normal distribution functions and can fit different types of distributions. A single multi-dimensional normal distribution can be expressed as:

[0107]

[0108] The Gaussian mixture model is obtained by using a weighted combination of multiple multi-dimensional normal distributions, which is expressed as:

[0109]

[0110] Among them, X is the sample, μ is the sample mean, Σ is the variance matrix, and α j represents the weight of the j-th multi-dimensional normal distribution.

[0111] Obtain the method for calculating the posterior distribution. Based on the method for calculating the posterior distribution, the predicted transient fuel consumption set, and the model prior distribution set, calculate the model posterior distribution set, that is, use approximate Bayesian calculation to estimate the posterior distribution according to the predicted values of each model respectively. According to the prediction results of each model and the constructed prior distribution, use the approximate Bayesian calculation method to estimate the posterior distribution. The basic idea of approximate Bayesian calculation is to estimate the posterior distribution of parameters by simulating the data and the observed data. Estimate the posterior distribution through the prediction results of each model and the fuel consumption data under similar working conditions to obtain the model posterior distribution set. Among them, the posterior distribution given by Bayes' theorem is expressed as:

[0112] P(θ|D) ∝ P(D|θ)P(θ)

[0113] Among them, P(θ|D) is the posterior probability, P(D|θ) is the likelihood function representing the probability of observing data D under parameter θ, and P(θ) is the prior distribution of the parameter.

[0114] Since the likelihood function is difficult to calculate, the posterior distribution can be approximated by comparing the prediction results of each model and the fuel consumption data under similar working conditions.

[0115] The distance metric methods for comparison include Euclidean distance, Wasserstein distance, etc.

[0116] For two vectors x and y, the Euclidean distance can be expressed as:

[0117]

[0118] The Wasserstein distance is also called the Earth-Mover Distance and is often used to compare the distribution similarity between simulated data and observed data. The Wasserstein distance measures the minimum value of the average distance required to move the data from distribution P to distribution Q.

[0119] The Wasserstein distance can be expressed as:

[0120]

[0121] Among them, in the formula, P and Q are two probability distributions, and Π(P,Q) is the set of joint distributions γ.

[0122] Perform subsequent processing based on the model prior distribution set and the model posterior distribution set.

[0123] Step S30, select the credible model predicted fuel consumption based on the model input data and the model empirical distribution set, update the prior distribution, determine the target predicted transient fuel consumption set, and complete the accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

[0124] It should be noted that the target predicted transient fuel consumption set reflects the characteristics of the set of model predicted values of the transient fuel consumption predicted values finally determined for achieving accurate prediction of transient fuel consumption.

[0125] It can be understood that the target predicted transient fuel consumption set is a set of model predicted values with high confidence and strong credibility obtained by screening the model posterior distribution sets of each model using the position of the posterior distribution confidence interval from multiple model prediction results, and organizing the model predicted values.

[0126] For the convenience of understanding, taking the determination of the target predicted transient fuel consumption set as an example, the information collection device is the information collection module, the storage device is the memory, and the execution device is the execution module.

[0127] The information collection module obtains model input data, the model empirical distribution set, and the position of the posterior distribution confidence interval, calculates the confidence based on the model posterior distribution set in the model empirical distribution set and the position of the posterior distribution confidence interval, judges the model credibility based on the confidence, determines the fuel consumption credibility model, and predicts the transient fuel consumption predicted value based on the fuel consumption credibility model, that is, evaluates the credibility of the transient fuel consumption prediction results of each model for real-time input. According to the position of the model prediction result in the posterior distribution confidence interval, the credibility of the model prediction result is evaluated. If the confidence is lower than the set threshold, the model is directly rejected. The Metropolis-Hasting method can be used to evaluate the acceptance degree of the sample. The Metropolis-Hasting method is expressed as:

[0128]

[0129] Among them, the Metropolis-Hasting method determines the acceptance degree of the sample with the acceptance rate ρ and assigns corresponding weights.

[0130] Judge whether the prediction credibility is lower than the set threshold. If the prediction credibility is lower than the set threshold, retrain each model in the model library using all historical fuel consumption data to obtain new model parameters.

[0131] If the prediction credibility is not lower than the set threshold, adjust the historical fuel consumption data based on the model input data and the transient fuel consumption predicted value, reconstruct the model prior distribution set, determine the updated model prior distribution set, and train the fuel consumption model based on the updated model prior distribution set to determine the fuel consumption prediction optimization model set, that is, store the model input data and the output transient fuel consumption predicted value as historical fuel consumption data, and update the prior distribution. The historical fuel consumption data is stored in the form of the input matrix X and the output value Y, where the input vector X is composed of the preprocessed map information vector, vehicle information vector, and driver driving behavior vector spliced together. The input matrix X can be expressed as:

[0132]

[0133] Through threshold judgment, the dynamic update of the model library model can be achieved, significantly improving the accuracy of model prediction.

[0134] Based on the output of the fuel consumption prediction optimization model set, the target predicted transient fuel consumption set is obtained. The target predicted transient fuel consumption set is integrated to obtain the target accuracy transient fuel consumption prediction value. Based on the target accuracy transient fuel consumption prediction value, the accurate prediction of transient fuel consumption is completed, that is, the prediction results of each model are integrated, and the transient fuel consumption prediction value is output. The predictions of all accepted models for the transient fuel consumption at the next moment are output, and normalized weighted summation is performed according to the calculated corresponding weights. The prediction results of multiple models are integrated, and the historical fuel consumption data is used to integrate the predictions of each model. Finally, a high-precision transient fuel consumption prediction value is obtained, thus completing the accurate prediction of transient fuel consumption.

[0135] In a feasible implementation manner, step S30 may include steps B11 to B16:

[0136] Step B11, obtaining the position of the posterior distribution confidence interval;

[0137] It should be noted that the position of the posterior distribution confidence interval reflects the characteristic of the specific position of the confidence interval in which the set model prediction results are located.

[0138] It can be understood that the position of the posterior distribution confidence interval can characterize the reliability of the model prediction result. The closer the position is to the center of the posterior distribution, the higher the credibility of the prediction result.

[0139] Step B12, calculating the confidence level based on the model posterior distribution set in the model empirical distribution set and the position of the posterior distribution confidence interval;

[0140] It should be noted that the confidence level reflects the characteristic of the credibility degree of the model prediction result in the posterior distribution.

[0141] It can be understood that the confidence level can characterize the relative position of the model prediction result in the posterior distribution, usually represented by a value between 0 and 1. The higher the confidence level, the more reliable the model prediction result and the closer it is to the true value.

[0142] Step B13, judging the credibility of the model based on the confidence level, determining the fuel consumption credibility model, and predicting the transient fuel consumption prediction value based on the fuel consumption credibility model;

[0143] It should be noted that the transient fuel consumption prediction value reflects the characteristic of the value obtained by predicting the fuel consumption of a commercial vehicle at a specific moment based on the selected credible model.

[0144] It is understandable that the transient fuel consumption prediction value can characterize the fuel consumption of the vehicle under dynamically changing working conditions. By integrating the prediction results of multiple models and performing weighted summation according to the confidence level, a high-precision prediction value is obtained, so as to perform energy-saving control and route planning for commercial vehicles.

[0145] Step B14: Based on the model input data and the transient fuel consumption prediction value, adjust the historical fuel consumption data, reconstruct the prior distribution set of the model, and determine the updated prior distribution set of the model.

[0146] It should be noted that the updated prior distribution set of the model reflects the characteristics of the new prior distribution set obtained by dynamically updating the prior distribution of the model according to the new prediction results and historical data during the model prediction process.

[0147] It is understandable that the updated prior distribution set of the model can enable the model to continuously learn and adapt to new data, thereby improving the prediction accuracy and generalization ability of the model. By continuously adjusting the prior distribution, the model can better reflect the current driving conditions and vehicle status.

[0148] Step B15: Train the fuel consumption model based on the updated prior distribution set of the model, and determine the optimized fuel consumption prediction model set.

[0149] It should be noted that the optimized fuel consumption prediction model set reflects the characteristics of the model set with higher prediction accuracy and adaptability obtained by retraining after dynamically updating the prior distribution set of the model during the model prediction process.

[0150] It is understandable that the optimized fuel consumption prediction model set can more accurately predict the transient fuel consumption of commercial vehicles under different driving conditions.

[0151] Step B16: Output the target predicted transient fuel consumption set based on the optimized fuel consumption prediction model set.

[0152] It is understandable that the target predicted transient fuel consumption set can more accurately characterize the transient fuel consumption of commercial vehicles under different driving conditions, so as to perform energy-saving control and route planning for commercial vehicles.

[0153] In another feasible implementation manner, step S30 may include steps C11 to C12:

[0154] Step C11: Integrate the target predicted transient fuel consumption set to obtain the target accuracy transient fuel consumption prediction value.

[0155] It should be noted that the target accuracy transient fuel consumption prediction value reflects the characteristics of the final transient fuel consumption prediction value obtained by integrating the prediction results of multiple credible models and performing weighted summation according to the confidence level of each model.

[0156] It can be understood that the target-precision transient fuel consumption prediction value can characterize the fuel consumption of a commercial vehicle at a specific moment, and has high precision and high reliability.

[0157] Step C12, complete the accurate prediction of transient fuel consumption based on the target-precision transient fuel consumption prediction value.

[0158] It can be understood that by integrating the prediction results of multiple credible models and performing weighted summation according to the confidence levels of each model, the precision of the transient fuel consumption prediction value is significantly improved, ensuring the high precision and high reliability of the prediction result, and continuously learning and adapting to new data, significantly reducing the fuel consumption of commercial vehicles.

[0159] A method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model proposed in this embodiment obtains model input data and a transient fuel consumption model set, and preprocesses the model input data to obtain preprocessed model input data; constructs an empirical distribution based on the model input data and the transient fuel consumption model set, determines a model empirical distribution set, and the model empirical distribution set includes a model prior distribution set and a model posterior distribution set; selects a credible model to predict fuel consumption and updates the prior distribution based on the model input data and the model empirical distribution set, determines a target predicted transient fuel consumption set, and completes the accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set. This solves the technical problem of how to more efficiently predict transient fuel consumption to adapt to complex and changeable driving conditions. Compared with the prior art, in this application, prior and posterior distributions are constructed by combining multiple transient fuel consumption models, credible models are screened for fuel consumption prediction, and at the same time, the prior distribution is dynamically updated, and the prediction results are integrated to complete the high-precision prediction of transient fuel consumption, significantly improving the prediction precision, enhancing the model adaptability, applying to different driving conditions, and effectively reducing the fuel consumption of commercial vehicles.

[0160] Based on the first embodiment of this application, in the second embodiment of this application, for the same or similar content as in the above-mentioned first embodiment, reference can be made to the above introduction and will not be elaborated hereinafter.

[0161] In this embodiment, referring to Figure 2 , Figure 2 is a schematic flowchart provided for the second embodiment of the method for accurately predicting the transient fuel consumption of a commercial vehicle based on a stochastic hybrid model in this application. Step S20 specifically includes steps S21 to S24:

[0162] Step S21, input the model input data into the transient fuel consumption model set for training to determine a predicted transient fuel consumption set;

[0163] It should be noted that the predicted transient fuel consumption set reflects the characteristics of a set of fuel consumption prediction values obtained by using multiple different transient fuel consumption models for fuel consumption prediction.

[0164] It is understandable that each model has its own unique modeling method and data processing method, which can predict the transient fuel consumption from different angles. By inputting the model input data into multiple transient fuel consumption models, a combination of predicted values ​​can be obtained, that is, the predicted transient fuel consumption set, which can characterize the fuel consumption of commercial vehicles under different driving conditions.

[0165] For ease of understanding, an example of determining a predicted transient fuel consumption set is used for explanation, wherein the information collection device is an information collection module, the storage device is a memory, and the processing device is a processing module.

[0166] The information acquisition module obtains model input data and inputs it into the transient fuel consumption model set, inputs the model input data into the transient fuel consumption model set for training, and determines a predicted transient fuel consumption set, that is, the model input data is pre-trained using the fuel consumption data set, and then the pre-trained data is respectively input into a plurality of selected transient fuel consumption models to respectively predict the transient fuel consumption at the next moment, to obtain a predicted transient fuel consumption set, and subsequent processing is performed based on the predicted transient fuel consumption set.

[0167] Step S22, filtering fuel consumption data based on the model input data and constructing a model prior distribution set;

[0168] It should be noted that the model prior distribution set reflects the characteristics of the prior probability distribution set obtained by fitting the Gaussian mixture model based on historical fuel consumption data and operating condition similarity values.

[0169] For ease of understanding, the construction of a model prior distribution set is taken as an example for explanation, wherein the information collection device is an information collection module, the storage device is a memory, and the processing device is a processing module.

[0170] The information acquisition module obtains the model input data and the predicted transient fuel consumption set, filters the fuel consumption data based on the model input data, determines the historical fuel consumption data and the similarity value of the working condition, and constructs the model prior distribution set based on the historical fuel consumption data and the similarity value of the working condition, that is, samples the historical fuel consumption data and constructs the prior distribution of each model respectively. The historical fuel consumption data includes the existing data set and the fuel consumption data continuously collected during the operation of the vehicle. According to the model input data, similar working conditions are extracted from the fuel consumption data set. The fuel consumption data on the same road section or the same road conditions that are similar to the vehicle load are extracted, and the data is weighted based on the similarity of the working conditions to obtain the model prior distribution set, and the Gaussian mixture model is used to fit the prior distribution.

[0171] Among them, the Gaussian mixture model is a linear combination of multiple normal distribution functions, which can fit different types of distributions. A single multidimensional normal distribution can be expressed as:

[0172]

[0173] The Gaussian mixture model, on the other hand, is obtained by a weighted combination of multiple multi-dimensional normal distributions, which is expressed as:

[0174]

[0175] where X is the sample, μ is the sample mean, Σ is the variance matrix, and α j represents the weight of the j-th multi-dimensional normal distribution.

[0176] In a feasible implementation, step S22 may include steps D11 to D12:

[0177] Step D11, screening fuel consumption data based on the model input data to determine historical fuel consumption data and operating condition similarity values;

[0178] It should be noted that the historical fuel consumption data reflects the fuel consumption data continuously collected and accumulated during vehicle operation, including the characteristics of fuel consumption records of the vehicle under different road sections and different driving conditions, and the operating condition similarity value reflects the characteristics of the similarity index calculated by comparing multiple parameters of the current operating condition and the historical operating condition.

[0179] It can be understood that the historical fuel consumption data characterizes the fuel consumption of the vehicle under various operating conditions and is used to construct the prior distribution, and the operating condition similarity value is used to screen the historical fuel consumption data most similar to the current operating condition to ensure that the prior distribution can accurately reflect the fuel consumption under the current operating condition.

[0180] Step D12, constructing a set of model prior distributions based on the historical fuel consumption data and the operating condition similarity values.

[0181] It can be understood that the set of model prior distributions can characterize the prior distributions constructed based on the historical statistical information of commercial vehicle fuel consumption under different driving conditions.

[0182] Step S23, calculating a set of model posterior distributions based on the predicted transient fuel consumption set and the set of model prior distributions;

[0183] It should be noted that the set of model posterior distributions reflects the characteristics of the set of posterior probability distributions obtained by an approximate Bayesian calculation method, combining the model prediction results and historical data, given the model input data and the prior distribution.

[0184] For ease of understanding, an example of calculating the set of model posterior distributions is given, where the information collection device is the information collection module, the storage device is the memory, and the processing device is the processing module.

[0185] The information acquisition module obtains the model prior distribution set, the predicted transient fuel consumption set, and the posterior distribution calculation method, and calculates the model posterior distribution set based on the posterior distribution calculation method, the predicted transient fuel consumption set, and the model prior distribution set, that is, uses approximate Bayesian calculation to estimate the posterior distribution according to the predicted values of each model respectively. According to the prediction results of each model and the constructed prior distribution, the approximate Bayesian calculation method is used to estimate the posterior distribution, and the basic idea of approximate Bayesian calculation is to estimate the posterior distribution of parameters by simulating data and observed data, estimate the posterior distribution by the prediction results of each model and the fuel consumption data under similar working conditions, and obtain the model posterior distribution set. Among them, the posterior distribution given by Bayes' theorem is expressed as:

[0186] P(θ|D) ∝ P(D|θ)P(θ)

[0187] Among them, P(θ|D) is the posterior probability, P(D|θ) is the likelihood function representing the probability of observing data D under parameter θ, and P(θ) is the prior distribution of the parameter.

[0188] Since the likelihood function is difficult to calculate, the posterior distribution can be approximated by comparing the prediction results of each model and the fuel consumption data under similar working conditions.

[0189] The distance metric methods for comparison include Euclidean distance, Wasserstein distance, etc.

[0190] For two vectors x and y, the Euclidean distance can be expressed as:

[0191]

[0192] The Wasserstein distance, also known as the Earth-Mover Distance, is often used to compare the distribution similarity of simulated data and observed data. The Wasserstein distance measures the minimum of the average distance required to move data from distribution P to distribution Q.

[0193] The Wasserstein distance can be expressed as:

[0194]

[0195] Among them, in the formula, P and Q are two probability distributions, and Π(P, Q) is the set of joint distributions γ.

[0196] In a feasible implementation manner, step S23 may include steps E11 to E12:

[0197] Step E11, obtain the posterior distribution calculation method;

[0198] It should be noted that the posterior distribution calculation method reflects the characteristics of estimating the posterior distribution of parameters by combining the model prediction results and the prior distribution through an approximate Bayesian calculation method.

[0199] It can be understood that the posterior distribution calculation method estimates the posterior distribution by comparing the simulated data with the observed data and using a distance metric to evaluate the similarity between the model prediction results and the historical data, so as to ensure that the model can better adapt to new data and working conditions.

[0200] Step E12, calculate the model posterior distribution set based on the posterior distribution calculation method, the predicted transient fuel consumption set, and the model prior distribution set.

[0201] It can be understood that by using an approximate Bayesian calculation method combined with a distance metric method to estimate the posterior distribution, the prediction accuracy of the model is significantly improved, and the dynamic update mechanism of the posterior distribution can enable the model to continuously learn and adapt to new data and working conditions. By continuously adjusting the posterior distribution, the model can better adapt to different driving conditions and vehicle states, enhancing the adaptability of the model. At the same time, calculating a high-precision posterior distribution can more efficiently perform energy-saving control and path planning for commercial vehicles, effectively reducing the fuel consumption of commercial vehicles.

[0202] Step S24, obtain the model empirical distribution set based on the model prior distribution set and the model posterior distribution set.

[0203] It should be noted that the model prior distribution set is obtained by screening historical fuel consumption data similar to the current working condition and fitting it with a Gaussian mixture model to provide a prior probability distribution of fuel consumption. The model posterior distribution set uses an approximate Bayesian calculation method, combines the predicted transient fuel consumption set and the model prior distribution set, and updates the posterior distribution by comparing the similarity between the prediction results and the historical data, so as to more accurately characterize the fuel consumption distribution under the current data conditions.

[0204] For the convenience of understanding, taking the obtaining of the model empirical distribution set as an example for illustration, where the information collection device is the information collection module, the storage device is the memory, and the processing device is the processing module.

[0205] The model empirical distribution set only represents the inclusion relationship between the model prior distribution set and the model posterior distribution set, that is, it cannot be obtained by jointly calculating the model prior distribution set and the model posterior distribution set, and subsequent processing is carried out based on the model prior distribution set and the model posterior distribution set.

[0206] A precise transient fuel consumption prediction method for commercial vehicles based on a random hybrid model proposed in this embodiment inputs the model input data into the transient fuel consumption model set for training to determine the predicted transient fuel consumption set; screens the fuel consumption data based on the model input data to construct a model prior distribution set; calculates a model posterior distribution set based on the predicted transient fuel consumption set and the model prior distribution set; and obtains a model empirical distribution set based on the model prior distribution set and the model posterior distribution set. This solves the technical problem of how to more accurately predict transient fuel consumption to adapt to complex and variable driving conditions. Compared with the prior art, this application inputs the preprocessed model input data into multiple transient fuel consumption models for training, establishes a fuel consumption empirical distribution as the prior distribution based on the existing data set and continuously accumulated fuel consumption data, then establishes a transient fuel consumption model library composed of multiple transient models such as "steady-state initial value + transient correction" or directly modeling based on transient variables, uses approximate Bayesian calculation methods to adjust the weights of each model, obtains a hybrid model that can accurately predict transient fuel consumption, is more conveniently deployed on different vehicles, and is applied to different driving conditions, significantly improving the accuracy and generalization ability of the model.

[0207] Exemplarily, to facilitate understanding of the implementation process of the precise transient fuel consumption prediction method for commercial vehicles based on the random hybrid model obtained by combining the above-mentioned Embodiment 1, please refer to Figure 3 , Figure 3 A brief process schematic diagram of a precise transient fuel consumption prediction method for commercial vehicles based on a random hybrid model is provided. Specifically:

[0208] Referring to Embodiment 1, obtain the model input data and the transient fuel consumption model set, preprocess the model input data to obtain the preprocessed model input data; construct an empirical distribution based on the model input data and the transient fuel consumption model set, determine the model empirical distribution set, and the model empirical distribution set includes a model prior distribution set and a model posterior distribution set; select a credible model to predict the fuel consumption and update the prior distribution based on the model input data and the model empirical distribution set, determine the target predicted transient fuel consumption set, and complete the accurate prediction of the transient fuel consumption based on the target predicted transient fuel consumption set. Referring to Embodiment 2, input the model input data into the transient fuel consumption model set for training to determine the predicted transient fuel consumption set; screen the fuel consumption data based on the model input data to construct a model prior distribution set; calculate a model posterior distribution set based on the predicted transient fuel consumption set and the model prior distribution set; obtain a model empirical distribution set based on the model prior distribution set and the model posterior distribution set. Obtain map information, including vehicle route planning, vehicle GPS coordinates, road surface adhesion and gradient, and other map information, obtain vehicle information, including vehicle speed, vehicle acceleration, vehicle load, and other vehicle information, obtain the driver's driving behavior, including pedal information, steering wheel signal, gear signal, and other driving behaviors, so as to obtain the model input data, and perform preprocessing, construct a prior distribution using historical fuel consumption data, and input the preprocessed data into a multi - variable non - linear fuel consumption model, a least - squares fitting linear fuel consumption model, and a neural network model, construct a posterior distribution, evaluate the prediction credibility, determine whether it is lower than the credibility threshold, and retrain or update the historical fuel consumption data, so as to output the final predicted value of the transient fuel consumption.

[0209] It should be noted that the above examples are only for understanding the present application and do not constitute a limitation to the method for accurately predicting the transient fuel consumption of commercial vehicles based on a stochastic hybrid model. Based on this technical concept, more forms of simple transformations are within the protection scope of the present application.

[0210] The present application also provides a device for accurately predicting the transient fuel consumption of commercial vehicles based on a stochastic hybrid model. Please refer to Figure 4 , the device for accurately predicting the transient fuel consumption of commercial vehicles based on a stochastic hybrid model includes:

[0211] An acquisition module 10, configured to acquire model input data and a transient fuel consumption model set, and preprocess the model input data to obtain preprocessed model input data;

[0212] A processing module 20, configured to construct an empirical distribution based on the model input data and the transient fuel consumption model set, determine a model empirical distribution set, and the model empirical distribution set includes a model prior distribution set and a model posterior distribution set;

[0213] The execution module 30 is configured to select a credible model to predict fuel consumption based on the model input data and the model empirical distribution set, update the prior distribution, determine a target predicted transient fuel consumption set, and complete accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

[0214] The acquisition module 10 is further configured to acquire vehicle characteristic parameters, driver operation signals, road segment map information, a steady-state initial value transient correction model, a polynomial transient fuel consumption prediction model, and a target fuel consumption prediction model. The vehicle characteristic parameters include vehicle speed information, vehicle acceleration information, and vehicle mass information. The driver operation signals include accelerator pedal opening information, brake pedal opening information, steering wheel angle information, steering wheel torque information, and vehicle gear information. The road segment map information includes road surface adhesion information, slope information, and speed limit information.

[0215] Determine model input data based on the vehicle characteristic parameters, the driver operation signals, and the road segment map information.

[0216] Determine a transient fuel consumption model set based on the steady-state initial value transient correction model, the polynomial transient fuel consumption prediction model, and the target fuel consumption prediction model.

[0217] The processing module 20 is further configured to input the model input data into the transient fuel consumption model set for training to determine a predicted transient fuel consumption set.

[0218] Screen fuel consumption data based on the model input data to construct a model prior distribution set.

[0219] Calculate a model posterior distribution set based on the predicted transient fuel consumption set and the model prior distribution set.

[0220] Obtain a model empirical distribution set based on the model prior distribution set and the model posterior distribution set.

[0221] The processing module 20 is further configured to screen fuel consumption data based on the model input data to determine historical fuel consumption data and a working condition similarity value.

[0222] Construct a model prior distribution set based on the historical fuel consumption data and the working condition similarity value.

[0223] The processing module 20 is further configured to obtain a posterior distribution calculation method.

[0224] Calculate a model posterior distribution set based on the posterior distribution calculation method, the predicted transient fuel consumption set, and the model prior distribution set.

[0225] The execution module 30 is further configured to obtain the position of the posterior distribution confidence interval.

[0226] Calculate the confidence level based on the concentrated model posterior distribution set of the model empirical distribution and the position of the posterior distribution confidence interval;

[0227] Judge the credibility of the model based on the confidence level, determine the fuel consumption credibility model, and predict the transient fuel consumption prediction value based on the fuel consumption credibility model;

[0228] Adjust the historical fuel consumption data based on the model input data and the transient fuel consumption prediction value, reconstruct the model prior distribution set, and determine the updated model prior distribution set;

[0229] Train the fuel consumption model based on the updated model prior distribution set, and determine the fuel consumption prediction optimization model set;

[0230] Output the target predicted transient fuel consumption set based on the fuel consumption prediction optimization model set.

[0231] The execution module 30 is further configured to integrate the target predicted transient fuel consumption set to obtain the target accuracy transient fuel consumption prediction value;

[0232] Complete the accurate prediction of transient fuel consumption based on the target accuracy transient fuel consumption prediction value.

[0233] The commercial vehicle transient fuel consumption accurate prediction device based on the stochastic hybrid model provided by the present application adopts the commercial vehicle transient fuel consumption accurate prediction method in the above embodiment, and can solve the technical problem of how to more efficiently and accurately predict the transient fuel consumption to adapt to complex and changeable driving conditions. Compared with the prior art, the beneficial effects of the commercial vehicle transient fuel consumption accurate prediction device based on the stochastic hybrid model provided by the present application are the same as those of the commercial vehicle transient fuel consumption accurate prediction method provided by the above embodiment, and other technical features in the commercial vehicle transient fuel consumption accurate prediction device based on the stochastic hybrid model are the same as the features disclosed in the above embodiment method, and will not be elaborated here.

[0234] The present application provides a commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model. The commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the commercial vehicle transient fuel consumption accurate prediction method in Embodiment 1 above.

[0235] Next, refer to Figure 5, which shows a schematic structural diagram of a commercial vehicle transient fuel consumption accurate prediction device suitable for implementing the embodiments of the present application. The commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The shown commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present application.

[0236] As Figure 5 shown, the commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 into the random access memory (RAM: Random Access Memory) 1004. In the RAM 1004, various programs and data required for the operation of the commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model are also stored. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. The input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems may be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the commercial vehicle transient fuel consumption accurate prediction device based on a stochastic hybrid model to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows a commercial vehicle transient fuel consumption accurate prediction device with various systems, it should be understood that it is not required to implement or have all the shown systems. More or fewer systems may be alternatively implemented or had.

[0237] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by a processing device 1001, the above functions defined in the methods of the embodiments disclosed in the present application are executed.

[0238] The accurate prediction device for transient fuel consumption of commercial vehicles based on a random hybrid model provided by the present application adopts the method for accurate prediction of transient fuel consumption of commercial vehicles based on a random hybrid model in the above embodiments, and can solve the technical problem of how to more efficiently and accurately predict transient fuel consumption to adapt to complex and changeable driving conditions. Compared with the prior art, the beneficial effects of the accurate prediction device for transient fuel consumption of commercial vehicles based on a random hybrid model provided by the present application are the same as those of the method for accurate prediction of transient fuel consumption of commercial vehicles based on a random hybrid model provided in the above embodiments, and other technical features in the accurate prediction device for transient fuel consumption of commercial vehicles based on a random hybrid model are the same as the features disclosed in the method of the previous embodiment, and will not be elaborated here.

[0239] It should be understood that the various parts disclosed in the present application can be implemented by hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0240] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all of them should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0241] The present application provides a computer-readable storage medium, on which there are computer-readable program instructions (i.e., computer programs), and the computer-readable program instructions are used to execute the method for accurate prediction of transient fuel consumption of commercial vehicles based on a random hybrid model in the above embodiments.

[0242] The computer-readable storage medium provided by this application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component. The program code contained on the computer-readable storage medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0243] The above computer-readable storage medium can be included in the commercial vehicle transient fuel consumption accurate prediction device based on a random hybrid model; it can also exist independently without being assembled into the commercial vehicle transient fuel consumption accurate prediction device based on a random hybrid model.

[0244] The above computer-readable storage medium carries one or more programs. When the above one or more programs are executed by the commercial vehicle transient fuel consumption accurate prediction device based on a random hybrid model, the commercial vehicle transient fuel consumption accurate prediction device based on a random hybrid model is enabled to: obtain model input data and a transient fuel consumption model set, preprocess the model input data to obtain preprocessed model input data; construct an empirical distribution based on the model input data and the transient fuel consumption model set, determine a model empirical distribution set, and the model empirical distribution set includes a model prior distribution set and a model posterior distribution set; select a credible model to predict fuel consumption and update the prior distribution based on the model input data and the model empirical distribution set, determine a target predicted transient fuel consumption set, and complete the accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

[0245] Computer program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The above-mentioned programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, by connecting through the Internet using an Internet service provider).

[0246] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of the code, and this module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks shown can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0247] The modules described in the embodiments of this application can be implemented in software or in hardware. Among them, the name of the module does not constitute a limitation on the unit itself in some cases.

[0248] The readable storage medium provided by this application is a computer-readable storage medium. The computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for performing the above-mentioned accurate prediction method of commercial vehicle transient fuel consumption based on a random hybrid model, and can solve the technical problem of how to more efficiently and accurately predict transient fuel consumption to adapt to complex and variable driving conditions. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by this application are the same as those of the accurate prediction method of commercial vehicle transient fuel consumption based on a random hybrid model provided in the above embodiments, and will not be elaborated here.

[0249] The above are only some embodiments of the present application, and thus do not limit the patent scope of the present application. Any equivalent structural transformation made under the technical concept of the present application by using the content of the specification and drawings of the present application, or any direct / indirect application in other related technical fields, is included in the patent protection scope of the present application.

Claims

1. A method for accurately predicting transient fuel consumption of commercial vehicles based on a random mixing model, characterized in that: The method includes: Acquire model input data and a transient fuel consumption model set, and preprocess the model input data to obtain preprocessed model input data; Constructing an empirical distribution based on the model input data and the transient fuel consumption model set, and determining a model empirical distribution set, wherein the model empirical distribution set includes a model prior distribution set and a model posterior distribution set; Based on the model input data and the model experience distribution set, a credible model is selected to predict fuel consumption and the prior distribution is updated, a target predicted transient fuel consumption set is determined, and accurate prediction of transient fuel consumption is completed based on the target predicted transient fuel consumption set.

2. The method according to claim 1, characterized in that The step of obtaining model input data and transient fuel consumption model set comprises: Acquire vehicle characteristic parameters, driver manipulation signals, road section map information, a steady-state initial value transient correction model, a polynomial transient fuel consumption prediction model, and a target fuel consumption prediction model, wherein the vehicle characteristic parameters include vehicle speed information, vehicle acceleration information, and vehicle mass information, the driver manipulation signals include accelerator pedal opening information, brake pedal opening information, steering wheel angle information, steering wheel torque information, and vehicle gear information, and the road section map information includes road adhesion information, slope information, and speed limit information; Determining model input data based on the vehicle characteristic parameters, the driver manipulation signal and the road section map information; A transient fuel consumption model set is determined based on the steady-state initial value transient correction model, the polynomial transient fuel consumption prediction model and the target fuel consumption prediction model.

3. The method according to claim 1, characterized in that The step of constructing an empirical distribution based on the model input data and the transient fuel consumption model set and determining the model empirical distribution set comprises: Inputting the model input data into the transient fuel consumption model set for training to determine a predicted transient fuel consumption set; Filtering fuel consumption data based on the model input data to construct a model prior distribution set; Calculating a model posterior distribution set based on the predicted transient fuel consumption set and the model prior distribution set; A model empirical distribution set is obtained based on the model prior distribution set and the model posterior distribution set.

4. The method according to claim 3, characterized in that The step of filtering the fuel consumption data based on the model input data and constructing a model prior distribution set comprises: Filtering fuel consumption data based on the model input data to determine historical fuel consumption data and operating condition similarity values; A model prior distribution set is constructed based on the historical fuel consumption data and the operating condition similarity value.

5. The method according to claim 3, characterized in that The step of calculating the model posterior distribution set based on the predicted transient fuel consumption set and the model prior distribution set comprises: Get the posterior distribution calculation method; A model posterior distribution set is calculated based on the posterior distribution calculation method, the predicted transient fuel consumption set and the model prior distribution set.

6. The method according to claim 1, characterized in that The step of selecting a credible model to predict fuel consumption based on the model input data and the model empirical distribution set and updating the prior distribution to determine a target predicted transient fuel consumption set comprises: Get the position of the confidence interval of the posterior distribution; Calculate confidence based on the model posterior distribution set in the model empirical distribution set and the posterior distribution confidence interval position; Determining the credibility of the model based on the confidence level, determining a fuel consumption confidence model, and predicting a transient fuel consumption prediction value based on the fuel consumption confidence model; Adjusting historical fuel consumption data based on the model input data and the transient fuel consumption prediction value, reconstructing a model prior distribution set, and determining to update the model prior distribution set; Training a fuel consumption model based on the updated model prior distribution set to determine a fuel consumption prediction optimization model set; A target predicted transient fuel consumption set is output based on the fuel consumption prediction optimization model set.

7. The method according to claim 1, characterized in that The step of completing accurate prediction of transient fuel consumption based on the target predicted transient fuel consumption set comprises: Integrating the target predicted transient fuel consumption set to obtain a target accuracy transient fuel consumption prediction value; Based on the target accuracy transient fuel consumption prediction value, accurate transient fuel consumption prediction is completed.

8. A commercial vehicle transient fuel consumption accurate prediction device based on a random mixing model, characterized in that: The device comprises: An acquisition module, used for acquiring model input data and a transient fuel consumption model set, and preprocessing the model input data to obtain preprocessed model input data; A processing module, configured to construct an empirical distribution based on the model input data and the transient fuel consumption model set, and determine a model empirical distribution set, wherein the model empirical distribution set includes a model prior distribution set and a model posterior distribution set; The execution module is used to select a credible model to predict fuel consumption based on the model input data and the model experience distribution set and update the prior distribution, determine a target predicted transient fuel consumption set, and complete the precise prediction of transient fuel consumption based on the target predicted transient fuel consumption set.

9. A commercial vehicle transient fuel consumption accurate prediction device based on a random mixing model, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the commercial vehicle transient fuel consumption accurate prediction method based on a random mixing model as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the commercial vehicle transient fuel consumption accurate prediction method based on a random mixing model as described in any one of claims 1 to 7 are implemented.

Images