Electric vehicle trip energy consumption probability prediction method and system based on traffic flow information

Abstract

The application discloses a kind of electric vehicle trip energy consumption probability prediction method and system based on traffic flow information, it is related to electric vehicle energy consumption prediction field, this method includes, obtaining actual road driving data;Calculate traffic flow characteristics and driving behavior characteristics;Build driving behavior characteristic edge probability distribution estimation model;Build driving behavior characteristic joint probability distribution estimation model;The traffic flow characteristics of given trip are input to feature joint probability distribution estimation model, and the joint probability distribution of driving behavior characteristics is obtained;Determine Nc trip characteristics samples that meet probability distribution and carry out feature clustering, obtain Mc clustering center characteristics;Build energy consumption model;The Mc typical driving behavior or determined driving behavior characteristic joint probability distribution is input to the energy consumption model, and the energy consumption probability of given trip is obtained.The application can greatly reduce the complexity of energy consumption related traffic flow characteristics, reduce the calculation requirement, improve model convergence and accuracy.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle energy consumption prediction, and in particular to a method and system for predicting the probability of electric vehicle trip energy consumption based on traffic flow information. Background Technology

[0002] In recent years, the electric vehicle industry has developed rapidly, with the global number of electric vehicles exceeding 16 million by 2021. Consumers are eager to obtain accurate predictions of future travel energy consumption to alleviate range anxiety caused by insufficient battery capacity in electric vehicles. In addition to the energy consumption characteristics of electric vehicles, the driving characteristics of future travel are also a major influencing factor on future travel energy consumption. However, existing methods for predicting future travel driving characteristics generally predict driving speed, which is a high-dimensional time series with complex influencing factors and strong randomness, making accurate prediction of this series extremely difficult.

[0003] This invention predicts the joint probability distribution or typical driving behavior distribution of energy consumption-related trip driving behavior features based on traffic flow information, thereby effectively supporting more accurate and comprehensive trip energy consumption prediction. The algorithm of this invention has low computational cost, can be deployed on actual vehicle systems, and provides accurate and comprehensive energy consumption predictions. Therefore, it is of great significance for reducing users' range anxiety and effectively promoting the further development of electric vehicles. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for predicting the probability of electric vehicle trip energy consumption based on traffic flow information, which can significantly reduce the complexity of energy consumption-related traffic flow features, reduce computational requirements, and improve model convergence and accuracy.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] In a first aspect, the present invention provides a method for predicting the probability of electric vehicle trip energy consumption based on traffic flow information, comprising:

[0007] Obtain actual road driving data;

[0008] Traffic flow characteristics are determined based on the actual road driving data.

[0009] Calculate driving behavior features; the traffic flow features and driving behavior features constitute a learning dataset;

[0010] A marginal probability distribution estimation model for driving behavior features is constructed based on the aforementioned learning dataset;

[0011] The marginal probability distribution of driving behavior characteristics is determined based on the marginal probability distribution estimation model.

[0012] A joint probability distribution estimation model for driving behavior features is constructed based on the marginal probability distribution and the learning dataset.

[0013] Given the traffic flow characteristics of the trip, and input them into the joint probability distribution estimation model of the features, the joint probability distribution of driving behavior characteristics is obtained;

[0014] Based on the joint probability distribution of the driving behavior characteristics, Nc trip feature samples conforming to the probability distribution are determined;

[0015] The Nc trip feature samples that conform to the probability distribution are clustered to obtain Mc cluster center features; the Mc cluster center features are Mc typical driving behaviors corresponding to the traffic flow features of the given trip;

[0016] Construct an energy consumption model;

[0017] The energy consumption model is input into the typical driving behaviors of Mc or the joint probability distribution of driving behavior characteristics to obtain the probability distribution of energy consumption for a given trip.

[0018] Optionally, determining traffic flow characteristics based on the actual road driving data specifically includes the following steps:

[0019] Extract N actual road trips from the actual road driving data;

[0020] Traffic flow characteristics are determined based on the N actual road journeys.

[0021] Optionally, the driving behavior characteristics include: relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity.

[0022] Optionally, the calculation of driving behavior characteristics specifically includes the following steps:

[0023] Calculate speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity;

[0024] The relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity are calculated based on the speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity.

[0025] Secondly, the present invention further provides an electric vehicle trip energy consumption probability prediction system based on traffic flow information, comprising:

[0026] The actual road driving data acquisition module is used to acquire actual road driving data;

[0027] A traffic flow feature determination module is used to determine traffic flow features based on the actual road driving data.

[0028] A driving behavior feature calculation module is used to calculate driving behavior features; the traffic flow features and driving behavior features constitute a learning dataset;

[0029] The marginal probability distribution estimation model building module is used to build a marginal probability distribution estimation model for driving behavior features based on the learning dataset.

[0030] The marginal probability distribution determination module for driving behavior features is used to determine the marginal probability distribution of driving behavior features based on the marginal probability distribution estimation model.

[0031] The feature joint probability distribution estimation model construction module is used to construct a driving behavior feature joint probability distribution estimation model based on the marginal probability distribution and the learning dataset.

[0032] The joint probability distribution determination module for driving behavior features is used to determine the joint probability distribution of driving behavior features given the traffic flow features of a trip, and input them into the joint probability distribution estimation model of the features to obtain the joint probability distribution of driving behavior features.

[0033] The trip feature sample determination module is used to determine Nc trip feature samples that conform to the probability distribution based on the joint probability distribution of the driving behavior features;

[0034] The clustering module is used to perform feature clustering on the Nc trip feature samples that conform to the probability distribution to obtain Mc cluster center features; the Mc cluster center features are Mc typical driving behaviors corresponding to the traffic flow features of the given trip;

[0035] The energy consumption model building module is used to build energy consumption models;

[0036] The energy consumption probability prediction module is used to input the joint probability distribution of the Mc typical driving behaviors or the driving behavior characteristics into the energy consumption model to obtain the probability distribution of energy consumption for a given trip.

[0037] Optionally, the traffic flow feature determination module specifically includes the following steps:

[0038] Extract N actual road trips from the actual road driving data;

[0039] Traffic flow characteristics are determined based on the N actual road journeys.

[0040] Optionally, the driving behavior characteristics include: relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity.

[0041] Optionally, the driving behavior feature calculation module specifically includes the following steps:

[0042] Calculate speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity;

[0043] The relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity are calculated based on the speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity.

[0044] Thirdly, the present invention also provides an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to perform the above-described electric vehicle trip energy consumption probability prediction method based on traffic flow information.

[0045] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the electric vehicle trip energy consumption probability prediction method based on traffic flow information as described above.

[0046] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0047] This invention, by introducing correlations between features, can effectively improve the accuracy and convergence of predicted feature probability distributions, significantly reduce the complexity of energy consumption-related traffic flow features, lower computational requirements, and improve model convergence and accuracy; at the same time, it provides more comprehensive probability estimation results or multiple typical driving behavior estimation results compared to deterministic estimation, alleviating users' "range anxiety". Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 The flowchart of the electric vehicle trip energy consumption probability prediction method based on traffic flow information provided by the present invention is shown below.

[0050] Figure 2 This is a schematic diagram of the electric vehicle trip energy consumption probability prediction system based on traffic flow information provided by the present invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] The purpose of this invention is to provide a method and system for predicting the probability of electric vehicle trip energy consumption based on traffic flow information, which can significantly reduce the complexity of energy consumption-related traffic flow features, reduce computational requirements, and improve model convergence and accuracy.

[0053] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0054] Figure 1 The flowchart of the electric vehicle trip energy consumption probability prediction method based on traffic flow information provided by the present invention is as follows: Figure 1 As shown, the method in this invention includes:

[0055] Step 1: Obtain actual road driving data.

[0056] The acquired actual road driving data is used to train the trip driving feature prediction model. This data includes, but is not limited to, time, vehicle speed, and GPS information of the driving trajectory. It is recommended that the sampling frequency for time and vehicle speed be no less than 1Hz. The greater the coverage of data samples across different time periods and roads, the more traffic flow features the subsequent training model can select for more accurate predictions of trip energy consumption. Finally, N actual road trips (i.e., trips where both the starting and ending speeds are 0, and the sample data is continuous from the start time to the end time) are extracted from the acquired actual road driving data. The number of speed sampling points within each trip is N. j .

[0057] Step 2: Determine traffic flow characteristics based on the actual road driving data.

[0058] Traffic flow characteristics F1, F2...F are constructed based on sample data combined with map, navigation, and other information. P This includes, but is not limited to: trip length (total distance traveled between the starting and ending points), number of intersections with traffic lights (number of intersections with traffic lights along the route between the starting and ending points), number of intersections without traffic lights (number of intersections without traffic lights along the route between the starting and ending points that may affect vehicle speed), trip time, and road type.

[0059] Step 3: Calculate driving behavior features; the traffic flow features and driving behavior features constitute a learning dataset. The learning dataset includes a training set and a validation set.

[0060] Construct driving behavior features e1, e2, ... e q It includes, but is not limited to, four characteristics: relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity. The specific calculation method is as follows:

[0061] First, the speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-down intensity are calculated using equations (1)-(4).

[0062] Speed ​​intensity:

[0063]

[0064] Braking strength:

[0065] First, define braking energy:

[0066]

[0067] E brk，j (i)=max(v j (i) 2 -v j (i+1) 2 ,0) (3)

[0068] Among them, v j (i) represents the speed at the i-th sampling point during the j-th actual road journey, t s Let m be the sampling interval, and m be the mass of the vehicle. The average speed of the vehicle from sampling point i to sampling point i+1 can be expressed as:

[0069]

[0070] A, B, and C are target coefficients related to the calculation of load force.

[0071] In calculating braking energy E brk，j (i) If the vehicle's coasting braking characteristics are known, then formula (2) is applied; otherwise, formula (3) is applied.

[0072] The braking intensity can be determined based on a certain threshold velocity v. th Divide it into two parts:

[0073] Low-speed braking intensity:

[0074]

[0075] High-speed braking intensity:

[0076]

[0077] Slow traffic intensity:

[0078]

[0079] Calculate the relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity by selecting a baseline operating condition (such as CLTC-P (China Driving Test Cycle)).

[0080]

[0081]

[0082]

[0083]

[0084] Among them, the baseline speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity are... Calculate according to equations (1)-(7).

[0085] For the p traffic flow features F of the j-th trip 1,j ,F 2,j ...F p,j and q driving behavior characteristics e′ 1,j ,e′ 2,j ...e′ q,j A trip sample is formed, and N trip samples form the model learning dataset.

[0086] Step 4: Construct a marginal probability distribution estimation model for driving behavior features based on the learning dataset.

[0087] Step 5: Determine the marginal probability distribution of driving behavior features based on the marginal probability distribution estimation model.

[0088] Constructing a marginal probability distribution estimation model for driving behavior involves applying probabilistic data-driven methods (such as Bayesian methods, quantile models, kernel density estimation, etc.) based on historical datasets, and incorporating traffic flow characteristics F... 1，j F 2，j ...F p，j As input, driving behavior characteristics e 1，j e 2，j …e q，j As the output training data model, the marginal probability distribution estimation model for each feature is trained separately.

[0089] Step 6: Construct a joint probability distribution estimation model for driving behavior features based on the marginal probability distribution and the learning dataset.

[0090] Step 7: Given the traffic flow characteristics of the trip, input them into the joint probability distribution estimation model of the features to obtain the joint probability distribution of driving behavior characteristics.

[0091] Based on historical datasets and acquired marginal probability distributions, a model is constructed to estimate the joint probability distribution (cumulative probability distribution form) of driving features, P(e′1≤e). 1k , e′2≤e 2k ,..…e′ q ≤e qk Specifically, methods such as copula and correlation matrix can be used. The following section uses the copula method as an example to introduce the basic steps for obtaining the joint probability distribution. Different copula functions C(·) are compared using model performance evaluation metrics (such as the Akaike Information Criterion, AIC), and the best copula function suitable for the correlation of features in the historical dataset is selected. At this point, the joint probability distribution of the driving features can be expressed as:

[0092] P(e′1≤e 1k , e′2≤e 2k ,…...e′ q ≤e qk ,)=C(CDF1(e′1),CDF2(e′2),…,CDF q (e′ q ))(12)

[0093] Among them, CDF i (e′ i ) represents the marginal probability distribution (cumulative probability distribution form) of the i-th driving feature.

[0094] Step 8: Determine Nc trip feature samples that conform to the probability distribution based on the joint probability distribution of the driving behavior features.

[0095] Step 9: Perform feature clustering on the Nc trip feature samples that conform to the probability distribution to obtain Mc cluster center features; the Mc cluster center features are Mc typical driving behaviors corresponding to the traffic flow features of the given trip.

[0096] Specifically, for a given trip's traffic flow information (such as geographic and traffic information in navigation), the traffic flow feature F is obtained. 1，j F 2，j ...F p，j Substitute the values ​​from steps 4 and 6 to obtain the joint probability distribution of driving behavior features. Based on the joint probability distribution, apply the Monte Carlo method to generate Nc trip feature samples that conform to the probability distribution. Perform feature clustering on the Nc samples (using methods such as KNN) to obtain Mc clusters. The features of the Mc cluster centers can be used as the Mc typical driving behaviors that are possible for given traffic flow information.

[0097] Step 10: Construct an energy consumption model.

[0098] Step 11: Input the Mc typical driving behaviors into the energy consumption model to obtain the energy consumption probability for a given trip.

[0099] Specifically, the joint probability distribution of driving behavior features for a given trip, or typical driving behavior, obtained in step 9, is used as input to the energy consumption model. The results can be used for energy consumption probability estimation for a given trip or energy consumption value for a typical trip.

[0100] In order to implement the method corresponding to Embodiment 1 above and achieve the corresponding functions and technical effects, an electric vehicle trip energy consumption probability prediction system based on traffic flow information is provided below.

[0101] Alternatively, the aforementioned electronic device may be a server.

[0102] See Figure 2 , Figure 2 This is a schematic diagram of the electric vehicle trip energy consumption probability prediction system based on traffic flow information provided by the present invention. The system in the present invention includes:

[0103] The actual road driving data acquisition module 201 is used to acquire actual road driving data.

[0104] Traffic flow feature determination module 202 is used to determine traffic flow features based on the actual road driving data.

[0105] The driving behavior feature calculation module 203 is used to calculate driving behavior features; the traffic flow features and driving behavior features constitute a learning dataset.

[0106] Marginal probability distribution estimation model building module 204 is used to build a marginal probability distribution estimation model of driving behavior features based on the learning dataset.

[0107] The marginal probability distribution determination module 205 for driving behavior features is used to determine the marginal probability distribution of driving behavior features based on the marginal probability distribution estimation model.

[0108] The feature joint probability distribution estimation model construction module 206 is used to construct a driving behavior feature joint probability distribution estimation model based on the marginal probability distribution and the learning dataset.

[0109] The joint probability distribution determination module 207 for driving behavior features is used to determine the joint probability distribution of driving behavior features given the traffic flow features of the trip and input them into the joint probability distribution estimation model of the features to obtain the joint probability distribution of driving behavior features.

[0110] The trip feature sample determination module 208 is used to determine Nc trip feature samples that conform to the probability distribution based on the joint probability distribution of the driving behavior features.

[0111] Clustering module 209 is used to perform feature clustering on the Nc trip feature samples that conform to the probability distribution to obtain Mc cluster center features; the Mc cluster center features are Mc typical driving behaviors corresponding to the traffic flow features of the given trip.

[0112] Energy consumption model building module 210 is used to build energy consumption models.

[0113] The energy consumption probability prediction module 211 is used to input the joint probability distribution of the Mc typical driving behaviors or the driving behavior characteristics into the energy consumption model to obtain the energy consumption probability for a given trip.

[0114] In addition, embodiments of the present invention also provide a computer scale storage medium storing a computer program, which, when executed by a processor, implements a probability prediction of electric vehicle trip energy consumption based on traffic flow information as described in the above embodiments.

[0115] Based on the above description, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned computer storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory, random access memory, magnetic disks, or optical disks.

[0116] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0117] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for predicting the probability of electric vehicle trip energy consumption based on traffic flow information, characterized in that, include: Obtain actual road driving data; Traffic flow characteristics are determined based on the actual road driving data. Calculate driving behavior characteristics; The traffic flow features and driving behavior features constitute the learning dataset; A marginal probability distribution estimation model for driving behavior features is constructed based on the aforementioned learning dataset; The marginal probability distribution of driving behavior characteristics is determined based on the marginal probability distribution estimation model. A joint probability distribution estimation model for driving behavior features is constructed based on the marginal probability distribution and the learning dataset. Given the traffic flow characteristics of the trip, and input them into the joint probability distribution estimation model of the features, the joint probability distribution of driving behavior characteristics is obtained; Based on the joint probability distribution of the driving behavior characteristics, Nc trip feature samples conforming to the probability distribution are determined; The Nc trip feature samples that conform to the probability distribution are clustered to obtain Mc cluster center features; the Mc cluster center features are Mc typical driving behaviors corresponding to the traffic flow features of the given trip; Construct an energy consumption model; The probability distribution of the Mc typical driving behaviors or the joint probability distribution of driving behavior characteristics is input into the energy consumption model to obtain the probability distribution of energy consumption for a given trip.

2. The method for predicting the probability of electric vehicle trip energy consumption based on traffic flow information according to claim 1, characterized in that, Determining traffic flow characteristics based on the actual road driving data specifically includes the following steps: Extract N actual road trips from the actual road driving data; Traffic flow characteristics are determined based on the N actual road journeys.

3. The method for predicting the probability of electric vehicle trip energy consumption based on traffic flow information according to claim 1, characterized in that, The driving behavior characteristics include: relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity.

4. The method for predicting the probability of electric vehicle trip energy consumption based on traffic flow information according to claim 1, characterized in that, The calculation of driving behavior characteristics specifically Includes the following steps Calculate speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity; The relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity are calculated based on the speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity.

5. A probability prediction system for electric vehicle trip energy consumption based on traffic flow information, characterized in that, include: The actual road driving data acquisition module is used to acquire actual road driving data; A traffic flow feature determination module is used to determine traffic flow features based on the actual road driving data. The driving behavior characteristic calculation module is used to calculate driving behavior characteristics; The traffic flow features and driving behavior features constitute the learning dataset; The marginal probability distribution estimation model building module is used to build a marginal probability distribution estimation model for driving behavior features based on the learning dataset. The marginal probability distribution determination module for driving behavior features is used to determine the marginal probability distribution of driving behavior features based on the marginal probability distribution estimation model. The feature joint probability distribution estimation model construction module is used to construct a driving behavior feature joint probability distribution estimation model based on the marginal probability distribution and the learning dataset. The joint probability distribution determination module for driving behavior features is used to determine the joint probability distribution of driving behavior features given the traffic flow features of a trip, and input them into the joint probability distribution estimation model of the features to obtain the joint probability distribution of driving behavior features. The trip feature sample determination module is used to determine Nc trip feature samples that conform to the probability distribution based on the joint probability distribution of the driving behavior features; The clustering module is used to perform feature clustering on the Nc trip feature samples that conform to the probability distribution to obtain Mc cluster center features; the Mc cluster center features are Mc typical driving behaviors corresponding to the traffic flow features of the given trip; The energy consumption model building module is used to build energy consumption models; The energy consumption probability prediction module is used to input the joint probability distribution of the Mc typical driving behaviors or the driving behavior characteristics into the energy consumption model to obtain the probability distribution of energy consumption for a given trip.

6. The electric vehicle trip energy consumption probability prediction system based on traffic flow information according to claim 5, characterized in that, The traffic flow feature determination module specifically includes the following steps: Extract N actual road trips from the actual road driving data; Traffic flow characteristics are determined based on the N actual road journeys.

7. The electric vehicle trip energy consumption probability prediction system based on traffic flow information according to claim 5, characterized in that, The driving behavior characteristics include: relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity.

8. The electric vehicle trip energy consumption probability prediction system based on traffic flow information according to claim 5, characterized in that, The driving behavior characteristic calculation module specifically includes the following steps: Calculate speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity; The relative speed intensity, relative low-speed braking intensity, relative high-speed braking intensity, and relative slow-moving intensity are calculated based on the speed intensity, low-speed braking intensity, high-speed braking intensity, and slow-moving intensity.

9. An electronic device, characterized in that, The device includes a memory and a processor, the memory being used to store a computer program, and the processor running the computer program to cause the electronic device to perform the electric vehicle trip energy consumption probability prediction method based on traffic flow information according to any one of claims 1 to 4.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the electric vehicle trip energy consumption probability prediction method based on traffic flow information as described in any one of claims 1 to 4.

Images