A method, apparatus and electronic device for predicting a range

Abstract

The application belongs to the field of electric vehicles, and discloses a method and device for predicting the cruising range and electronic equipment, the method comprising: dividing micro segments according to historical vehicle driving data; predicting according to the micro segments by using a Markov Monte Carlo model; determining the corresponding energy consumption prediction value according to the obtained driving state prediction result, and determining the corresponding cruising range prediction value according to the energy consumption prediction value. The application divides the historical vehicle driving data into micro segments to capture the subtle changes in driving characteristics. Then, the random simulation capability of the Markov Monte Carlo model is used to predict the driving state of the vehicle according to these micro segments, and finally the cruising range prediction value is calculated by predicting the driving state. The application can adapt to the changes in driving conditions in real time, provide a cruising range prediction that is more in line with the actual driving situation, make the prediction result more accurate, help optimize the energy management of electric vehicles and improve the overall performance, and at the same time improve the user satisfaction.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric vehicles, and in particular to a method and device for predicting the range of an electric vehicle, an electronic device and a computer readable storage medium. BACKGROUND

[0002] With the popularity of electric vehicles, range prediction has become a key factor in evaluating the performance of electric vehicles. The range not only affects consumers' purchasing decisions, but also is an important consideration in the daily use of electric vehicles. Currently, the range prediction methods for electric vehicles on the market are mainly based on historical vehicle driving data, which estimates the energy consumption and remaining range of the vehicle under different conditions by analyzing these data.

[0003] However, the existing range prediction methods have some limitations. First, these methods often rely on simplified energy consumption models, which cannot fully consider the complex factors in the actual driving process. At the same time, traditional prediction methods cannot be updated in real time and adapt to new driving conditions, resulting in a large deviation between the predicted results and the actual range.

[0004] Therefore, how to improve the accuracy of electric vehicle range prediction is a technical problem to be solved by those skilled in the art. SUMMARY

[0005] In order to solve the problem of low accuracy of electric vehicle range prediction in the prior art, the present application provides a method and device for predicting the range of an electric vehicle, an electronic device and a computer readable storage medium.

[0006] A method for predicting the range of an electric vehicle, comprising:

[0007] Obtaining historical vehicle driving data and dividing micro segments according to the historical vehicle driving data;

[0008] Using a Markov Monte Carlo model to predict the micro segments and obtain a driving state prediction result;

[0009] Determining a corresponding energy consumption prediction value according to the driving state prediction result, and determining a corresponding range prediction value according to the energy consumption prediction value.

[0010] Optionally, the method for predicting the range of an electric vehicle using a Markov Monte Carlo model to predict the micro segments and obtain a driving state prediction result, comprising:

[0011] Calculating a first characteristic parameter of each micro segment; the first characteristic parameter includes at least one of the average speed, the difference between the terminal starting speed, the average acceleration, the maximum acceleration and the minimum acceleration;

[0012] cluster the first characteristic parameters of each of the micro-segments by using a fuzzy C-means clustering algorithm to obtain a first correspondence relationship between driving state categories and the first characteristic parameters;

[0013] determine a driving state category of each of the micro-segments according to the first correspondence relationship to obtain a first driving state sequence;

[0014] calculate a driving state transition probability according to the first driving state sequence, and predict a second driving state sequence in a preset time period by using the Markov Monte Carlo model according to the driving state transition probability.

[0015] Optionally, the determining of the corresponding energy consumption prediction value according to the driving state prediction result comprises:

[0016] obtaining a second correspondence relationship between vehicle speed states and vehicle speeds;

[0017] determining a first vehicle speed state sequence corresponding to different categories of driving states according to the second correspondence relationship;

[0018] calculating a vehicle speed state transition probability corresponding to each category of driving states according to the first vehicle speed state sequence;

[0019] predicting a second vehicle speed state sequence corresponding to the second driving state sequence by using the Markov Monte Carlo model according to the vehicle speed state transition probability corresponding to each category of driving states;

[0020] generating a corresponding vehicle speed-time curve according to the second vehicle speed state sequence;

[0021] determining a corresponding energy consumption prediction value according to the vehicle speed-time curve.

[0022] Optionally, the predicting of the second vehicle speed state sequence corresponding to the second driving state sequence by using the Markov Monte Carlo model according to the vehicle speed state transition probability corresponding to each category of driving states comprises:

[0023] repeatedly predicting a plurality of second vehicle speed state sequences corresponding to the second driving state sequence by using the Markov Monte Carlo model according to the vehicle speed state transition probability corresponding to each category of driving states;

[0024] the generating of the corresponding vehicle speed-time curve according to the second vehicle speed state sequence comprises:

[0025] generating a plurality of vehicle speed-time curves corresponding to the plurality of second vehicle speed state sequences;

[0026] the determining of the corresponding energy consumption prediction value according to the vehicle speed-time curve comprises:

[0027] determining a corresponding energy consumption prediction value according to each of the vehicle speed-time curves, and determining an average value of a plurality of the energy consumption prediction values as a final energy consumption prediction value.

[0028] Optionally, the determining a corresponding energy consumption prediction value according to each of the vehicle speed-time curves comprises:

[0029] dividing the vehicle speed-time curve according to a preset short trip to obtain a plurality of first short trip segments;

[0030] calculating a second characteristic parameter of each of the first short trip segments; the second characteristic parameter comprises at least one of an average ambient temperature, an average vehicle speed, an average acceleration, a speed 5% quantile point, a speed 95% quantile point, an acceleration 5% quantile point, and an acceleration 95% quantile point;

[0031] inputting each of the second characteristic parameters into a pre-trained energy consumption prediction model to obtain an energy consumption prediction value of each of the first short trip segments output by the energy consumption prediction model;

[0032] calculating a corresponding energy consumption prediction value of the vehicle speed-time curve according to the energy consumption prediction value of each of the first short trip segments.

[0033] Optionally, the training process of the energy consumption prediction model comprises:

[0034] dividing the historical vehicle driving data according to the preset short trip to obtain a plurality of second short trip segments;

[0035] calculating a second characteristic parameter and an energy consumption of each of the second short trip segments;

[0036] screening the second short trip segments according to the second characteristic parameters according to a preset screening rule;

[0037] performing normalization processing on the second characteristic parameters of the screened second short trip segments to obtain training data;

[0038] establishing a relationship between the training data and the energy consumption based on a meta-heuristic algorithm to obtain the energy consumption prediction model.

[0039] Optionally, the determining a corresponding energy consumption prediction value according to each of the vehicle speed-time curves comprises:

[0040] obtaining a battery apparent charge of the vehicle, and calculating a battery residual energy according to the battery apparent charge;

[0041] calculating a residual driving range according to the battery residual energy in combination with the corresponding energy consumption prediction value of the vehicle speed-time curve.

[0042] A device for predicting a range, comprising:

[0043] An acquisition module configured to acquire historical vehicle driving data and divide micro segments according to the historical vehicle driving data;

[0044] A driving state prediction module configured to predict, by using a Markov Monte Carlo model, according to the micro segments, to obtain a driving state prediction result;

[0045] A range prediction module configured to determine a corresponding energy consumption prediction value according to the driving state prediction result, and determine a corresponding range prediction value according to the energy consumption prediction value.

[0046] An electronic device, comprising:

[0047] A processor and a memory configured to store at least one instruction, which is loaded and executed by the processor to implement the method for predicting a range according to any one of the above.

[0048] A computer readable storage medium having a computer program stored thereon, which is executed by a processor to implement the method for predicting a range according to any one of the above.

[0049] The method for predicting a range provided by the embodiments of the present application, by dividing micro segments according to historical vehicle driving data, and predicting according to the micro segments by using a Markov Monte Carlo model, then determining a corresponding energy consumption prediction value according to the obtained driving state prediction result, and finally determining a corresponding range prediction value according to the energy consumption prediction value. The present application divides the historical vehicle driving data into micro segments to capture the subtle changes in driving characteristics by acquiring and analyzing the historical vehicle driving data. Then, by using the random simulation capability of the Markov Monte Carlo model, the driving state of the vehicle is predicted according to these micro segments, and finally the range prediction value is calculated by predicting the driving state. The present application can adapt to the changes of driving conditions in real time, provide a range prediction that is more in line with the actual driving situation, make the prediction result more accurate, help to optimize the energy management of electric vehicles and improve the overall performance, and at the same time improve the user satisfaction. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0051] Figure 1 A flowchart of the method for predicting a range provided by the embodiments of the present application;

[0052] Figure 2 A method for predicting the range of a vehicle provided by the present application, wherein S02 is a flowchart of an actual embodiment of the method; Figure 1

[0053] Figure 3 An effect diagram of the present application, wherein the number of clusters is 3;

[0054] Figure 4 A method for predicting the range of a vehicle provided by the present application, wherein S03 is a flowchart of an actual embodiment of the method; Figure 1

[0055] An effect diagram of the present application, wherein the number of clusters is 3; Figure 5

[0056] A method for predicting the range of a vehicle provided by the present application, wherein S03 is a flowchart of an actual embodiment of the method; Figure 6 Figure 4 A flowchart of a training process of an energy consumption prediction model provided by the present application;

[0057] Figure 7 A flowchart of a training process of an energy consumption prediction model provided by the present application;

[0058] Figure 8 A structure diagram of a range prediction device provided by the present application. DETAILED DESCRIPTION

[0059] In order to better understand the technical solutions of the present application, the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0060] It should be clear that the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0061] The terms used in the embodiments of the present application are only for the purpose of describing the specific embodiments, and are not intended to limit the present application. The singular forms "a", "an" and "the" used in the embodiments of the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0062] It should be understood that the term "and / or" used herein is only to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " in this paper generally represents that the front and rear associated objects are a "or" relationship.

[0063] ​​With the popularity of electric vehicles, range prediction has become a key factor in evaluating the performance of electric vehicles. Range not only affects consumers' purchasing decisions, but also is an important consideration in the daily use of electric vehicles. Currently, the range prediction methods for electric vehicles on the market are mainly based on historical vehicle driving data, which estimates the energy consumption and remaining range of the vehicle under different conditions by analyzing these data.

[0064] However, the existing range prediction methods have some limitations. First, these methods often rely on simplified energy consumption models, which cannot fully consider the complex factors in actual driving. At the same time, traditional prediction methods cannot update and adapt to new driving conditions in real time, resulting in a large deviation between the predicted results and the actual range.

[0065] Therefore, the present application provides a range prediction method to solve the above problems.

[0066] Please refer to Figure 1 The flow chart of the range prediction method provided by the embodiment of the present application includes the following steps:

[0067] Step S01, obtaining historical vehicle driving data, and dividing micro segments according to the historical vehicle driving data.

[0068] In this embodiment, the historical vehicle driving data is the data generated by the vehicle during actual driving. The historical vehicle driving data can be collected by using a data reading module, which may include driving time stamp, vehicle speed, battery voltage, battery current, battery SOC, ambient temperature, etc.

[0069] These continuous driving data are divided into smaller time periods, called "micro segments". Each micro segment contains the driving characteristics within a specific time. Such division helps to identify and analyze the behavior patterns of the vehicle under different driving conditions, providing detailed and representative data basis for subsequent energy consumption and range prediction.

[0070] In this embodiment, the purpose of obtaining historical vehicle driving data and dividing micro segments is to build a detailed database for analyzing and predicting the energy consumption and range of electric vehicles under different driving conditions. By dividing the driving data into micro segments, rapid changes and details in the driving process can be captured, such as acceleration, deceleration, constant speed driving, and idling. These micro segments contain rich information such as speed change, acceleration, ambient temperature, etc. Through the analysis of these micro segments, the energy consumption performance of the vehicle under similar conditions can be more accurately simulated and predicted, because short-term acceleration or deceleration events have a significant impact on energy consumption. This method enables the prediction model to be trained and optimized based on these detailed driving patterns, thereby improving the accuracy and reliability of range prediction.

[0071] In some embodiments, a specific time interval or event-driven approach can be employed to segment the data into micro-segments. For example, micro-segments can be divided based on the vehicle's start-stop, significant changes in speed, or preset time windows such as 10 seconds or 1 minute. Each micro-segment represents a specific phase of driving, such as an acceleration phase, a constant-speed cruising phase, or a deceleration phase.

[0072] On this basis, in order to improve the accuracy of the prediction, machine learning techniques can also be used to further classify and feature extract the micro-segments, so as to identify the key factors affecting energy consumption and range. This method enables the prediction model to be trained and optimized based on these detailed driving patterns, thereby improving the accuracy and reliability of the range prediction.

[0073] Step S02, using a Markov Monte Carlo model to predict according to the micro-segments, obtaining a driving state prediction result.

[0074] In this embodiment, the Markov Monte Carlo (MCMC) model combines the Markov chain's assumption of no memory and the randomness of Monte Carlo simulation, so that it can predict the possible future driving state of the vehicle based on the current driving state, without considering all previous driving history. By analyzing the micro-segments in the historical vehicle driving data, the MCMC model can simulate the transition probability of the vehicle in different driving states.

[0075] During the prediction process, the MCMC model uses random sampling to combine the randomness of Monte Carlo simulation to predict the various states that the vehicle may encounter during future driving. This prediction takes into account the uncertainty and complexity of vehicle driving, such as traffic conditions, driving behavior, and environmental changes. Through multiple iterations and simulations, the MCMC model can provide a series of possible driving state sequences, thereby obtaining the prediction result of the driving state, which provides an important basis for determining the energy consumption and range of the vehicle, making the range prediction more accurate and reliable.

[0076] Step S03, determining the corresponding energy consumption prediction value according to the driving state prediction result, and determining the corresponding range prediction value according to the energy consumption prediction value.

[0077] Because traditional energy consumption prediction may rely too much on theoretical models or simplified assumptions without fully considering dynamic changes in actual driving, it results in low accuracy of electric vehicle range prediction.

[0078] In this embodiment, it is determined that the vehicle's energy consumption is closely related to its driving state. Different driving states, such as acceleration, deceleration, constant speed driving, or idling, will significantly affect the battery consumption of the electric vehicle. Therefore, by accurately predicting the vehicle's energy consumption under various driving states, this embodiment can more accurately estimate the remaining battery energy and the vehicle's remaining driving distance.

[0079] By combining driving status predictions, energy consumption forecasts can reflect the vehicle's actual energy consumption under specific driving conditions. These energy consumption forecasts can then be used to calculate the predicted driving range, providing electric vehicle users with more reliable mileage information to help them better plan their trips and avoid inconvenience caused by insufficient battery power. Furthermore, this forecasting method can also provide data support for the vehicle's energy management system, optimizing battery usage strategies, extending battery life, and improving overall energy efficiency.

[0080] Based on the above technical solution, the driving range prediction method provided in this invention divides historical vehicle driving data into micro-segments, uses a Markov Monte Carlo model to predict the driving range based on these micro-segments, determines the corresponding energy consumption prediction value based on the obtained driving state prediction results, and finally determines the corresponding driving range prediction value based on the energy consumption prediction value. This invention acquires and analyzes historical vehicle driving data, dividing it into micro-segments to capture subtle changes in driving characteristics. Then, utilizing the stochastic simulation capability of the Markov Monte Carlo model, it predicts the vehicle's driving state based on these micro-segments, and finally calculates the driving range prediction value based on the predicted driving state. This invention can adapt to changes in driving conditions in real time, providing a driving range prediction that is more consistent with actual driving scenarios, making the prediction results more accurate, helping to optimize the energy management of electric vehicles and improve overall performance, while also enhancing user satisfaction.

[0081] Please refer to Figure 2 ,for Figure 1 A flowchart illustrating an actual implementation of step S02 in the provided method for predicting driving range. In some embodiments, step S02, which involves using a Markov Monte Carlo model to predict driving status based on micro-segments, may specifically include the following steps:

[0082] Step S11: Calculate the first characteristic parameter of each micro-segment.

[0083] The first characteristic parameter includes at least one of the following: average vehicle speed, difference between starting and ending speeds, average acceleration, maximum acceleration, and minimum acceleration.

[0084] In this embodiment, the first characteristic parameter of each micro-segment is extracted by analyzing the vehicle driving data, which quantitatively describes the driving characteristics of the vehicle during the micro-segment.

[0085] By processing a large amount of collected vehicle driving data, variables that significantly affect energy consumption and driving state can be identified. For example, the average vehicle speed reflects the average driving speed of the vehicle over a period of time, while the maximum and minimum values of acceleration reveal the intensity of acceleration and deceleration of the vehicle during that period of time. Through these parameters, the driving behavior of the vehicle can be understood more meticulously, and specific patterns and trends of vehicle driving can be identified, such as frequent acceleration and deceleration which may increase energy consumption, and smooth driving which helps to reduce energy consumption. These parameters provide basic data for subsequent driving state clustering and energy consumption prediction, enabling the prediction model to more accurately simulate the behavior of the vehicle under different driving conditions.

[0086] In some embodiments, the vehicle acceleration can be calculated by formula (1):

[0087]

[0088] where V t+1 is the vehicle speed at time t+1, V t-1 is the vehicle speed at time t-1, the vehicle speed unit is km / h, a t is the acceleration at time t, the acceleration unit is m / s 2 .

[0089] Step S12, using the fuzzy C-means clustering algorithm to cluster the first characteristic parameters of each micro-segment, obtaining the first corresponding relationship between the driving state category and the first characteristic parameter.

[0090] In this embodiment, the first characteristic parameters of the micro-segments are taken as input variables, and the fuzzy C-means clustering algorithm (FCM) algorithm is used to assign them to different driving state categories, forming the corresponding relationship between the driving state category and the first characteristic parameter.

[0091] The FCM algorithm is a clustering method based on fuzzy logic, which optimizes the objective function to obtain the membership degree of each sample point to all cluster centers, thereby automatically classifying the samples. FCM algorithm allows a data point to belong to multiple clusters with different membership degrees, which is particularly useful in handling driving states with fuzzy boundaries. For example, a micro-segment may be similar to both acceleration state and constant speed state in characteristics. In this way, different driving states such as smooth driving, sudden acceleration or sudden deceleration can be more accurately identified and distinguished, thereby providing a more detailed classification basis for energy consumption prediction and range prediction.

[0092] In some embodiments, after clustering the first feature parameters of each micro-segment using the fuzzy C-means clustering algorithm, the effectiveness of the classification can also be checked using the silhouette coefficient.

[0093] The silhouette coefficient is an index for evaluating the effectiveness of clustering, which measures the tightness and separation of data points in clusters. The value of the silhouette coefficient ranges from -1 to 1, where a value close to 1 indicates that the point fits well into its own cluster. Please refer to Figure 3 An effect diagram with a cluster number of 3 is provided in an embodiment of the present application. Based on this, the mean of the first feature parameters of each driving state after clustering is shown in Table 1:

[0094]

[0095] Table 1. Mean of first feature parameters of different driving states

[0096] Step S13, determining the driving state category of each micro-segment according to the first correspondence relationship, to obtain a first driving state sequence.

[0097] In this embodiment, the classification result obtained by processing using the FCM algorithm is used to assign a driving state category to each micro-segment, i.e., matching the first feature parameters of the micro-segment with the driving state categories determined by the FCM algorithm, thereby identifying the driving state to which each micro-segment most likely belongs, such as driving state 1, driving state 2, or driving state 3, etc.

[0098] Through this matching, a first driving state sequence can be constructed, which records in detail the driving state changes of the vehicle in a series of micro-segments. The first driving state sequence provides a detailed description of the driving behavior of the vehicle, enabling the model to predict the subsequent driving state of the vehicle based on these state changes. This process is the basis for calculating the driving state transition probability, and after the driving state of each micro-segment is determined, the transition probability between these driving states can be further analyzed and calculated.

[0099] Step S14, calculating the driving state transition probability according to the first driving state sequence, and predicting a second driving state sequence within a predetermined time period using the Markov Monte Carlo model based on the driving state transition probability.

[0100] In this embodiment, the driving state transition probability is a key parameter in the prediction of the range, which describes the likelihood of the vehicle transitioning from one driving state to another. By analyzing the first driving state sequence, the regularity of driving state transition can be identified, thereby predicting the state changes of the vehicle in future driving.

[0101] The driving state transition probability is then used in the MCMC model to simulate and predict the driving state sequence of the vehicle within a preset time period. The MCMC model, by combining the memoryless property of Markov chain and the randomness of Monte Carlo simulation, can generate the possible driving state sequence of the vehicle within a future preset time period, i.e., a second driving state sequence. The second driving state sequence can provide a more accurate basis for the subsequent prediction of energy consumption and range.

[0102] In some embodiments, the driving state transition probability can be calculated according to formula (2):

[0103]

[0104] where p ij is the probability of driving state i transitioning to driving state j, M ij is the number of times driving state i transitions to driving state j, M i is the total number of times driving state i transitions.

[0105] The calculated driving state transition probability is shown in Table 2:

[0106] Driving state 1 Driving state 2 Driving state 3 Driving state 1 0.681 0.010 0.309 Driving state 2 0.188 0.692 0.120 Driving state 3 0.203 0.489 0.308

[0107] Table 2. Driving state transition probability

[0108] In some embodiments, the next driving state can be predicted using a Markov model according to formula (3):

[0109]

[0110] where M is the total number of driving states, n is a random number between 0 and 1 generated by Monte Carlo, p ah is the probability of transitioning from driving state a to driving state h; k represents the sequence number of the next driving state.

[0111] Based on the above technical solutions, the present embodiment provides an accurate and efficient means for predicting the driving state of an electric vehicle by combining the Markov Monte Carlo model and the fuzzy C-means clustering algorithm. The first corresponding relationship between the driving state categories and the first characteristic parameters is established by analyzing the first characteristic parameters (such as average speed, speed difference, acceleration, etc.) of the micro-fragments to identify and classify the driving states. The first corresponding relationship enables the model to more accurately understand and predict the driving behavior of the vehicle. In addition, by calculating the driving state transition probability and using the MCMC model to predict the second driving state sequence within a future time period, the present method can provide a more accurate data basis for energy consumption and range prediction. This not only improves the accuracy of the prediction, but also enhances the adaptability of the model to different driving conditions.

[0112] Please refer to Figure 4 ,for Figure 1 A flowchart illustrating an actual implementation of step S03 in a provided method for predicting driving range. In some embodiments, step S03, which involves determining the corresponding energy consumption prediction value based on the driving state prediction result, may specifically include the following steps:

[0113] Step S21: Obtain the second correspondence between vehicle speed status and vehicle speed.

[0114] Vehicle speed is one of the key factors affecting the energy consumption and driving range of electric vehicles; different vehicle speeds lead to different energy consumption levels. Therefore, this embodiment obtains a second correspondence between vehicle speed state and vehicle speed, the purpose of which is to predict the specific vehicle speed at each moment in each micro-segment based on historical vehicle speed data. The second correspondence enables the model to determine the vehicle's speed state based on historical vehicle driving data.

[0115] In some embodiments, considering the maximum speed limit, the daily driving speed does not exceed 130 km / h. Therefore, a 1 km / h interval can be used as a vehicle speed state, and the vehicle speed state can be divided into 130 vehicle speed state intervals. This second correspondence can be shown in Table 3:

[0116] Table 3. Second Correspondence

[0117] Step S22: Determine the first vehicle speed state sequence corresponding to different categories of driving states based on the second correspondence.

[0118] In this embodiment, the vehicle speed in micro-segments corresponding to different driving states is classified using a second correspondence. This process involves matching the vehicle speed data in the micro-segment with predefined vehicle speed states, thereby determining a corresponding vehicle speed state for each vehicle speed data point in the micro-segment. This allows the construction of a first vehicle speed state sequence for the entire vehicle speed dataset in the micro-segment. The first vehicle speed state sequence records in detail the vehicle speed state at different time points within the micro-segment. The first vehicle speed state sequences corresponding to different driving states provide a detailed description of the vehicle speed state changes in different driving states, enabling the model to predict the changes in vehicle speed state in micro-segments corresponding to different driving states.

[0119] Step S23: Calculate the vehicle speed state transition probability corresponding to each type of driving state based on the first vehicle speed state sequence.

[0120] In this embodiment, the vehicle speed state transition probability is a key parameter in the range prediction, which describes the possibility of the vehicle transitioning from one speed state to another. By analyzing the first speed state sequence, the regularity of speed state transition corresponding to each driving state can be identified, thereby predicting the speed change of the vehicle in each driving state.

[0121] In some embodiments, the transition probability between speed states can also be calculated according to the above formula (2).

[0122] Step S24, using the Markov Monte Carlo model to predict the second speed state sequence corresponding to the second driving state sequence according to the speed state transition probability corresponding to each driving state.

[0123] In this embodiment, the speed state change of the vehicle in future driving is simulated and predicted by using the MCMC model. This process is based on the speed state transition probability corresponding to each driving state calculated in step S23, which describes the possibility of the vehicle transitioning between different speed states under different driving states.

[0124] By sequentially taking the speed state transition probability corresponding to each micro-segment in the second driving state sequence as the input parameter of the model, the MCMC model can generate a series of possible speed state sequences, which reflect the speed change that the vehicle may experience in future driving.

[0125] For example, assuming that the predicted second driving state sequence is

driving state 1, driving state 3, driving state 2

speed state 10, speed state 15, speed state 20, speed state 33

speed state 35, speed state 40, speed state 44, speed state 45

speed state 33, speed state 32, speed state 30, speed state 20

speed state 10, speed state 15, speed state 20, speed state 33, speed state 35, speed state 40, speed state 44, speed state 45, speed state 33, speed state 32, speed state 30, speed state 20

[0126] Since energy consumption is closely related to speed state, predicting speed state helps to build a more accurate energy consumption model, thereby making the range prediction more accurate and reliable.

[0127] Step S25, generating a corresponding speed-time curve according to the second speed state sequence.

[0128] In this embodiment, by converting the predicted vehicle speed state into a specific vehicle speed value and plotting it on the time axis, a curve reflecting the change of vehicle speed over time, i.e., a speed-time curve, is formed. The generated speed-time curve can show the speed change of the vehicle in the preset time period in detail. By analyzing the speed-time curve, the energy consumption in different driving states can be calculated more accurately, thereby providing data support for energy management and driving strategy of electric vehicles.

[0129] In some embodiments, the predicted vehicle speed state can be converted into a vehicle speed according to formula (4):

[0130] v s = [(s-1)+n] x Ad formula (4)

[0131] where v s is the speed value corresponding to the vehicle speed state s; n is a random number between 0 and 1 generated by Monte Carlo; and Ad is the interval length corresponding to the vehicle speed state s.

[0132] Step S26, determining a corresponding energy consumption prediction value according to the speed-time curve.

[0133] In this embodiment, the energy consumption of the electric vehicle under specific driving conditions is estimated by using the speed-time curve. This process is based on the speed change information provided by the speed-time curve and combines the energy consumption model of the vehicle to calculate the energy consumption in different speed and time intervals.

[0134] Specifically, the determination of the energy consumption prediction value takes into account the energy consumption characteristics of the vehicle at each speed state, such as higher energy consumption during acceleration and lower energy consumption during constant speed or deceleration. By matching the speed at each time point on the speed-time curve with the energy consumption characteristics of the vehicle, the energy consumption in each time period can be calculated. Adding up the energy consumption in these time periods, the total energy consumption prediction value in the entire driving process can be obtained.

[0135] Based on the above technical solution, this embodiment can more carefully understand the speed change pattern under different driving states by obtaining the second correspondence between the vehicle speed state and the vehicle speed. This method not only considers the transition of driving states, but also goes deeper into the level of vehicle speed state, providing more abundant data support for energy consumption prediction.

[0136] In some embodiments, as mentioned in step S24, the second speed state sequence corresponding to the second driving state sequence is predicted according to the speed state transition probability corresponding to each type of driving state by using the Markov Monte Carlo model, which can be specifically:

[0137] By repeatedly using the Markov Monte Carlo model, multiple second vehicle speed state sequences corresponding to the second driving state sequence are predicted based on the vehicle speed state transition probability corresponding to each type of driving state.

[0138] Based on this, as mentioned in step S25, the corresponding vehicle speed-time curve is generated according to the second vehicle speed state sequence, which can be:

[0139] Multiple speed-time curves are generated based on multiple second speed state sequences.

[0140] Please refer to Figure 5 This is a schematic diagram of multiple vehicle speed-time curves provided in an embodiment of the present invention.

[0141] Based on this, as mentioned in step S26, determining the corresponding energy consumption prediction value according to the vehicle speed-time curve can be as follows:

[0142] For each vehicle speed-time curve, determine the corresponding energy consumption prediction value, and then use the average of multiple energy consumption prediction values ​​as the final energy consumption prediction value.

[0143] In this embodiment, by repeatedly predicting multiple second vehicle speed state sequences, the behavior of the vehicle under different driving conditions can be simulated more comprehensively. By generating multiple speed-time curves, vehicle speed changes during driving can be captured, providing richer data for energy consumption prediction. By determining the energy consumption prediction value corresponding to each speed-time curve and using the average of these energy consumption prediction values ​​as the final energy consumption prediction value, the randomness of a single prediction result can be reduced, improving the accuracy and reliability of energy consumption prediction. This method considers the uncertainties and variability during driving and provides a more robust energy consumption prediction by integrating information from multiple prediction curves.

[0144] Please refer to Figure 6 ,for Figure 4 The flowchart illustrates one practical manifestation of step S26 in the method shown. Based on the above embodiments, in some embodiments, step S26, which mentions determining the corresponding energy consumption prediction value based on the vehicle speed-time curve, may specifically include the following steps:

[0145] Step S31: Divide the vehicle speed-time curve according to a preset short stroke to obtain multiple first short stroke segments.

[0146] In this embodiment, the continuous vehicle speed-time curve is divided into a series of smaller time intervals of equal or unequal length, each interval representing a short travel segment. These short travel segments are designed to capture the vehicle's driving characteristics over a specific time period, such as acceleration, deceleration, or constant speed.

[0147] With this division, the driving behavior and energy consumption patterns of the vehicle can be analyzed in more detail. Each short trip segment can be independently feature extracted and energy analyzed, providing more accurate data input for the energy consumption prediction model. This embodiment takes into account the dynamic changes during driving, allowing for a more refined assessment of the vehicle's energy consumption rather than simply averaging the data for the entire trip. Such subdivision helps to identify and quantify key driving stages that may affect energy consumption, supporting the optimization of energy management for electric vehicles and improving the accuracy of range prediction.

[0148] Step S32, calculate the second feature parameter of each first short trip segment.

[0149] Among them, the second feature parameter includes at least one of the average ambient temperature, the average vehicle speed, the average acceleration, the speed 5% quantile, the speed 95% quantile, the acceleration 5% quantile, and the acceleration 95% quantile.

[0150] In this embodiment, the second feature parameter is a series of key indicators that can describe the driving characteristics of the vehicle within the short trip segment. The calculation of the second feature parameter helps to deeply understand the performance of the vehicle under different driving conditions, for example, the average vehicle speed reflects the average driving speed of the vehicle in that period, and the quantile of acceleration reveals the intensity of vehicle acceleration or deceleration. By analyzing the second feature parameter, the energy consumption characteristics of the vehicle can be more accurately evaluated, providing necessary input for the energy consumption prediction model, so that the model can predict energy consumption based on the actual driving behavior of the vehicle, thereby improving the accuracy of range prediction.

[0151] Step S33, input each second feature parameter into a pre-trained energy consumption prediction model to obtain an energy consumption prediction value of each first short trip segment output by the energy consumption prediction model.

[0152] In this embodiment, the second feature parameter is provided as an input variable to an energy consumption prediction model that has been trained with historical data. The energy consumption prediction model can predict the energy consumption of the vehicle in the corresponding short trip segment. The model's prediction is based on the patterns and relationships it has learned from the training data, which reveal the relationship between the second feature parameter and energy consumption. In this way, the energy consumption prediction model can provide an energy consumption prediction value for each short trip segment, which takes into account the driving characteristics and environmental conditions of the vehicle in that period of time and their impact on energy consumption.

[0153] Step S34, calculate the energy consumption prediction value corresponding to the vehicle speed-time curve according to the energy consumption prediction value of each first short trip segment.

[0154] In this embodiment, the total energy consumption prediction for the entire driving cycle is obtained by integrating the energy consumption prediction values of the vehicle within each short trip segment. This process first relies on the energy consumption data provided by the energy consumption prediction model for each short trip segment, which reflects the energy consumption of the vehicle under specific time periods and driving conditions.

[0155] By accumulating or weightedly averaging the energy consumption prediction values of all short trip segments according to time periods, the energy consumption prediction value for the entire speed-time curve can be obtained. This total energy consumption prediction value provides a comprehensive estimate of the energy consumption of the vehicle under specific driving conditions and is a key indicator for evaluating the range of electric vehicles.

[0156] Based on the above technical solutions, this embodiment can more accurately capture the details of energy consumption changes during driving by subdividing the speed-time curve into multiple first short trip segments. By inputting the second characteristic parameter into the pre-trained energy consumption prediction model, an energy consumption prediction value can be generated for each first short trip segment. Finally, by aggregating the energy consumption prediction values of each short trip segment, the energy consumption prediction value corresponding to the entire speed-time curve can be obtained. This embodiment takes into account the subtle changes during driving and utilizes machine learning models to improve the accuracy of the prediction, providing more accurate energy consumption predictions.

[0157] On this basis, according to the energy consumption prediction value, the corresponding range prediction value is determined, which can specifically include the following steps:

[0158] Step S41, the battery apparent capacity of the vehicle is obtained, and the battery remaining energy is calculated according to the battery apparent capacity.

[0159] In some embodiments, the battery remaining energy can be calculated according to formula (5) based on the current battery apparent capacity of the vehicle:

[0160] E r = E T · SOC Formula (5)

[0161] where E r is the battery remaining energy, E T is the total battery energy (kWh), and SOC is the current apparent capacity (%).

[0162] Step S42, the remaining driving range is calculated according to the battery remaining energy combined with the energy consumption prediction value corresponding to the speed-time curve.

[0163] In some embodiments, the remaining driving range can be calculated according to formula (6) combined with the energy consumption prediction value corresponding to the speed-time curve:

[0164] S r = E T / E predict Formula (5)

[0165] wherein S r is the remaining range, E predict is the energy consumption prediction value corresponding to the speed-time curve.

[0166] In this embodiment, by obtaining the apparent battery capacity of the vehicle, the remaining energy of the battery can be accurately calculated. By combining the energy consumption prediction value corresponding to the speed-time curve, the energy consumption of the vehicle under different driving conditions under the current remaining energy of the battery can be calculated. This combination takes into account the actual driving behavior and energy consumption characteristics of the vehicle, so that the calculation of the remaining range is more close to the actual driving situation.

[0167] Please refer to Figure 7 , a flowchart of the training process of an energy consumption prediction model provided by an embodiment of the present application. Based on the above embodiment, in some embodiments, the training process of the energy consumption prediction model mentioned in step S33 can include the following steps:

[0168] Step S51, dividing the historical vehicle driving data according to a preset short trip to obtain a plurality of second short trip segments.

[0169] In this embodiment, the collected historical vehicle driving data is divided into a series of fixed length time periods. These time periods are called second short trip segments, and each segment contains the driving information of the vehicle in that time period.

[0170] Through this division, the microscopic driving patterns can be extracted from the macroscopic driving data, so that each second short trip segment can independently reflect the driving characteristics of the vehicle in a specific time period. This method focuses on the behavior of the vehicle under specific driving conditions, rather than the average or overall characteristics of the entire trip, and can more accurately identify and quantify the factors affecting energy consumption in subsequent analysis, providing a more detailed and richer data basis for the energy consumption prediction model, thereby improving the accuracy and reliability of the model prediction.

[0171] Step S52, calculating the second characteristic parameters and energy consumption of each second short trip segment.

[0172] In this embodiment, each second short trip segment is analyzed in detail to extract key driving features (i.e. second characteristic parameters) and energy consumption data. This process first determines the second characteristic parameters of each short trip segment, which can include average ambient temperature, average speed, average acceleration, 5% and 95% quantile points of speed, 5% and 95% quantile points of acceleration, etc., which can fully reflect the driving conditions of the vehicle in that time period.

[0173] In this embodiment, the energy consumption of each second short trip segment is calculated, which can be achieved by analyzing the charge and discharge state of the battery, the kinetic energy change of the vehicle, and the possible energy recovery. This calculation process needs to accurately measure or estimate the energy consumption of the vehicle during each short trip segment in order to provide actual energy consumption data for model training.

[0174] In some embodiments, the energy consumption data of the second short trip segment can be calculated by formula (7):

[0175]

[0176] where E is the energy consumption data, U t is the battery voltage corresponding to time t; I t is the battery current corresponding to time t; and D is the mileage of the second short trip segment.

[0177] By calculating the second characteristic parameter and the energy consumption, rich training data can be provided for the energy consumption prediction model, which will be used to establish the relationship between the vehicle driving characteristics and the energy consumption. This embodiment is based on actual driving data and detailed energy consumption analysis, so that the energy consumption prediction model can more truly reflect the energy consumption behavior of the vehicle under different driving conditions.

[0178] Step S53, according to the preset screening rule, the second short trip segment is screened according to the second characteristic parameter.

[0179] In this embodiment, a series of pre-defined standards or rules are applied to select those most representative and relevant short trip segments, in order to exclude data that distorts the model training results due to abnormalities, errors or non-representative driving conditions.

[0180] In some embodiments, the preset screening rule can be as shown in Table 4:

[0181]

[0182] Table 4. Preset screening rule

[0183] Wherein the second short trip segment is the combination of idle segment and motion segment, and the car from the start to the next start is considered as an idle segment, and the car from the start to the next stop is defined as a motion segment.

[0184] In some embodiments, the preset screening rules can also include conditions such as excluding segments recorded in extreme environmental temperatures, or based on the reasonableness of energy consumption, such as excluding segments with abnormally high or low energy consumption, which may be caused by sensor failure or data entry errors. In this way, the quality of the data used for model training can be ensured, thereby improving the accuracy and reliability of the energy consumption prediction model.

[0185] The screening process helps to concentrate the information in the training data, making it more focused on short trip segments that reflect the vehicle's energy consumption characteristics under normal driving conditions. This not only improves the efficiency of model training, but also helps to build a more robust and accurate energy consumption prediction model, providing more accurate range predictions for electric vehicle users.

[0186] Step S54, the second feature parameters of the second short trip segments after screening are normalized to obtain training data.

[0187] In this embodiment, normalization is a data preprocessing technique that scales the values of feature parameters to a specific range, usually 0 to 1. The purpose of this step is to ensure that all feature parameters have the same importance in model training, avoiding the situation where some features occupy too much weight in the model due to large differences in parameter value ranges. Normalization allows the model to more fairly evaluate the contribution of each feature to energy consumption prediction, helping to improve the efficiency and accuracy of model training. At the same time, normalized data can speed up the convergence speed of optimization algorithms and improve the generalization ability of the model.

[0188] After normalization, the second feature parameters together with the corresponding energy consumption data form the training data set of the model. This data set will be used to train the energy consumption prediction model, so that it can learn the relationship between feature parameters and energy consumption, and thus accurately predict unknown data in actual application. The benefit of this method is that it can provide more accurate and reliable energy consumption prediction, helping to optimize energy management of electric vehicles and improve the accuracy of range prediction.

[0189] Step S55, based on the meta-heuristic algorithm, the relationship between the training data and the energy consumption is established to obtain the energy consumption prediction model.

[0190] In this embodiment, meta-heuristic algorithms are a class of algorithms used to solve optimization problems, which simulate natural phenomena or social behavior to find the optimal or satisfactory solution to the problem. Common meta-heuristic algorithms include genetic algorithms, particle swarm optimization, simulated annealing, etc.

[0191] During the training process of the energy consumption prediction model, a meta-heuristic algorithm is used to explore the mapping relationship between the second feature parameters and the energy consumption. Through an iterative process, the algorithm continuously adjusts the model parameters to minimize the difference between the predicted energy consumption and the actual energy consumption. In this way, the model can learn the patterns in the data and establish a mathematical model that can accurately predict energy consumption.

[0192] Based on the above technical solution, the short trip segment is filtered by the preset filtering rule, which can exclude abnormal or non-representative data to ensure the quality of the training data. The filtered data is normalized, which can eliminate the dimensional influence between different feature parameters and make the data more standardized, which helps to improve the efficiency and accuracy of model training. The relationship between the training data and the energy consumption is established by using the meta-heuristic algorithm, which can discover the complex relationship between the feature parameters and the energy consumption and optimize the model parameters, so as to obtain a model that can accurately predict the energy consumption. Through this method, the accuracy of energy consumption prediction can be significantly improved, which provides strong support for energy management and user decision-making of electric vehicles.

[0193] Please refer to Figure 8 The structure of the range prediction device provided by the embodiment of the application can include:

[0194] The acquisition module 100 is configured to acquire historical vehicle driving data and divide micro segments according to the historical vehicle driving data.

[0195] The driving state prediction module 200 is configured to use a Markov Monte Carlo model to predict the micro segments and obtain a driving state prediction result.

[0196] The range prediction module 300 is configured to determine a corresponding energy consumption prediction value according to the driving state prediction result, and determine a corresponding range prediction value according to the energy consumption prediction value.

[0197] Based on the above embodiment, in a specific embodiment, the driving state prediction module 200 can be specifically configured to:

[0198] Calculate the first feature parameters of each micro segment, wherein the first feature parameters include at least one of the average speed, the terminal start speed difference, the average acceleration, the maximum acceleration, and the minimum acceleration.

[0199] Cluster the first feature parameters of each micro segment using a fuzzy C-means clustering algorithm to obtain a first correspondence relationship between the driving state categories and the first feature parameters.

[0200] Determine the driving state category of each micro segment according to the first correspondence relationship to obtain a first driving state sequence.

[0201] The driving state transition probability is calculated according to the first driving state sequence, and the second driving state sequence in a preset time period is predicted according to the driving state transition probability by using a Markov Monte Carlo model.

[0202] On the basis of the above-mentioned embodiments, in one specific embodiment, the endurance mileage prediction module 300 can be specifically used for:

[0203] obtaining a second correspondence relationship between the speed state and the speed;

[0204] determining a first speed state sequence corresponding to each category of driving state according to the second correspondence relationship;

[0205] calculating a speed state transition probability corresponding to each category of driving state according to the first speed state sequence;

[0206] predicting a second speed state sequence corresponding to the second driving state sequence according to the speed state transition probability corresponding to each category of driving state by using a Markov Monte Carlo model;

[0207] generating a speed-time curve corresponding to the second speed state sequence;

[0208] determining a corresponding energy consumption prediction value according to the speed-time curve.

[0209] On the basis of the above-mentioned embodiments, in one specific embodiment, the endurance mileage prediction module 300 can be specifically used for:

[0210] repeatedly predicting a plurality of second speed state sequences corresponding to the second driving state sequence according to the speed state transition probability corresponding to each category of driving state by using the Markov Monte Carlo model;

[0211] generating a plurality of speed-time curves corresponding to the plurality of second speed state sequences;

[0212] determining a corresponding energy consumption prediction value for each speed-time curve, and determining an average value of the plurality of energy consumption prediction values as a final energy consumption prediction value.

[0213] On the basis of the above-mentioned embodiments, in one specific embodiment, the endurance mileage prediction module 300 can be specifically used for:

[0214] dividing the speed-time curve into a plurality of first short-trip segments according to a preset short-trip;

[0215] calculating a second feature parameter of each first short-trip segment; the second feature parameter includes at least one of the average ambient temperature, the average speed, the average acceleration, the speed 5% quantile, the speed 95% quantile, the acceleration 5% quantile, and the acceleration 95% quantile;

[0216] input each second characteristic parameter into a pre-trained energy consumption prediction model to obtain an energy consumption prediction value of each first short trip segment output by the energy consumption prediction model;

[0217] According to the energy consumption prediction value of each first short trip segment, the energy consumption prediction value corresponding to the vehicle speed-time curve is calculated.

[0218] On the basis of the above-mentioned embodiments, in a specific embodiment, the endurance mileage prediction module 300 can be specifically used for:

[0219] The historical vehicle driving data is divided according to the preset short trip to obtain a plurality of second short trip segments;

[0220] The second characteristic parameters and energy consumptions of each second short trip segment are calculated;

[0221] The second short trip segments are screened according to the second characteristic parameters according to a preset screening rule;

[0222] The second characteristic parameters of the screened second short trip segments are normalized to obtain training data;

[0223] The relationship between the training data and the energy consumptions is established based on a meta-heuristic algorithm to obtain an energy consumption prediction model.

[0224] On the basis of the above-mentioned embodiments, in a specific embodiment, the endurance mileage prediction module 300 can be specifically used for:

[0225] The battery apparent charge of the vehicle is obtained, and the battery remaining energy is calculated according to the battery apparent charge;

[0226] The remaining driving range is calculated according to the battery remaining energy combined with the energy consumption prediction value corresponding to the vehicle speed-time curve.

[0227] The embodiment provides an electronic device, which comprises a processor and a memory, the memory is used for storing at least one instruction, the instruction is loaded and executed by the processor to realize the method for predicting the endurance mileage, and the execution mode and beneficial effects are similar, and details are not repeated here.

[0228] The embodiment of the application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the method for predicting the endurance mileage, and the execution mode and beneficial effects are similar, and details are not repeated here.

[0229] It should be noted that although the above describes each step in a specific order, it does not mean that each step must be executed in the above specific order, in fact, some of the steps can be executed concurrently, or even the order is changed, as long as the required function can be realized.

[0230] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for predicting driving range, characterized in that, include: Acquire historical vehicle driving data and divide the data into micro-segments. The driving state prediction result is obtained by using the Markov Monte Carlo model based on the micro-segment; Based on the driving status prediction results, the corresponding energy consumption prediction value is determined, and based on the energy consumption prediction value, the corresponding driving range prediction value is determined. The step of using a Markov Monte Carlo model to predict the driving state based on the micro-segments to obtain the driving state prediction result includes: Calculate the first characteristic parameter of each micro-segment; the first characteristic parameter includes at least one of the following: average vehicle speed, difference between the initial and final speeds, average acceleration, maximum acceleration, and minimum acceleration. The first feature parameter of each micro-segment is clustered using the fuzzy C-means clustering algorithm to obtain the first correspondence between the driving state category and the first feature parameter; The driving state category of each micro-segment is determined based on the first correspondence, resulting in a first driving state sequence; The driving state transition probability is calculated based on the first driving state sequence, and the second driving state sequence within a preset time period is predicted using the Markov Monte Carlo model based on the driving state transition probability. The step of determining the corresponding energy consumption prediction value based on the driving state prediction result includes: Obtain the second correspondence between vehicle speed status and vehicle speed; Based on the second correspondence, determine the first vehicle speed state sequence corresponding to different categories of driving states; Calculate the vehicle speed state transition probability corresponding to each type of driving state based on the first vehicle speed state sequence. The Markov Monte Carlo model is used to predict the second vehicle speed state sequence corresponding to the second driving state sequence based on the vehicle speed state transition probability corresponding to each type of driving state. Generate the corresponding vehicle speed-time curve based on the second vehicle speed state sequence; The corresponding energy consumption prediction value is determined based on the vehicle speed-time curve.

2. The method according to claim 1, characterized in that, The step of predicting the second vehicle speed state sequence corresponding to the second driving state sequence using the Markov Monte Carlo model based on the vehicle speed state transition probability corresponding to each driving state includes: The Markov Monte Carlo model is reused to predict multiple second vehicle speed state sequences corresponding to the second driving state sequence based on the vehicle speed state transition probability corresponding to each type of driving state. The step of generating a corresponding vehicle speed-time curve based on the second vehicle speed state sequence includes: Multiple vehicle speed-time curves are generated based on multiple second vehicle speed state sequences; The step of determining the corresponding energy consumption prediction value based on the vehicle speed-time curve includes: Each speed-time curve is used to determine the corresponding energy consumption prediction value, and the average of multiple energy consumption prediction values ​​is used to determine the final energy consumption prediction value.

3. The method according to claim 1, characterized in that, The step of determining the corresponding energy consumption prediction value based on the vehicle speed-time curve includes: The vehicle speed-time curve is divided according to a preset short stroke to obtain multiple first short stroke segments; Calculate a second characteristic parameter for each of the first short-distance segments; the second characteristic parameter includes at least one of the following: average ambient temperature, average vehicle speed, average acceleration, 5th percentile of speed, 95th percentile of speed, 5th percentile of acceleration, and 95th percentile of acceleration. Each of the second feature parameters is input into a pre-trained energy consumption prediction model to obtain the energy consumption prediction value of each of the first short-distance segments output by the energy consumption prediction model. The predicted energy consumption value corresponding to the speed-time curve is calculated based on the predicted energy consumption value for each of the first short-distance segments.

4. The method according to claim 3, characterized in that, The training process of the energy consumption prediction model includes: The historical vehicle driving data is divided according to the preset short distance to obtain multiple second short distance segments; Calculate the second characteristic parameter and energy consumption for each of the second short-stroke segments; The second short-stroke segment is filtered according to the second feature parameter based on the preset filtering rules; The second feature parameters of the selected second short-stroke segment are normalized to obtain training data; The relationship between the training data and the energy consumption is established based on a metaheuristic algorithm to obtain the energy consumption prediction model.

5. The method according to claim 3, characterized in that, The step of determining the corresponding predicted driving range based on the predicted energy consumption value includes: Obtain the battery level displayed on the vehicle's display and calculate the remaining battery energy based on the displayed battery level. The remaining driving range is calculated based on the remaining battery energy and the energy consumption prediction value corresponding to the vehicle speed-time curve.

6. A device for predicting driving range, characterized in that, include: The acquisition module is used to acquire historical vehicle driving data and divide the historical vehicle driving data into micro-segments. The driving state prediction module is used to predict the driving state based on the micro-segment using a Markov Monte Carlo model to obtain the driving state prediction result. The driving range prediction module is used to determine the corresponding energy consumption prediction value based on the driving status prediction result, and to determine the corresponding driving range prediction value based on the energy consumption prediction value. The step of using a Markov Monte Carlo model to predict the driving state based on the micro-segments to obtain the driving state prediction result includes: Calculate the first characteristic parameter of each micro-segment; the first characteristic parameter includes at least one of the following: average vehicle speed, difference between the initial and final speeds, average acceleration, maximum acceleration, and minimum acceleration. The first feature parameter of each micro-segment is clustered using the fuzzy C-means clustering algorithm to obtain the first correspondence between the driving state category and the first feature parameter; The driving state category of each micro-segment is determined based on the first correspondence, resulting in a first driving state sequence; The driving state transition probability is calculated based on the first driving state sequence, and the second driving state sequence within a preset time period is predicted using the Markov Monte Carlo model based on the driving state transition probability. The step of determining the corresponding energy consumption prediction value based on the driving state prediction result includes: Obtain the second correspondence between vehicle speed status and vehicle speed; Based on the second correspondence, determine the first vehicle speed state sequence corresponding to different categories of driving states; Calculate the vehicle speed state transition probability corresponding to each type of driving state based on the first vehicle speed state sequence. The Markov Monte Carlo model is used to predict the second vehicle speed state sequence corresponding to the second driving state sequence based on the vehicle speed state transition probability corresponding to each type of driving state. Generate the corresponding vehicle speed-time curve based on the second vehicle speed state sequence; The corresponding energy consumption prediction value is determined based on the vehicle speed-time curve.

7. An electronic device, characterized in that, include: A processor and a memory, the memory being used to store at least one instruction, which, when loaded and executed by the processor, implements the method for predicting driving range as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for predicting driving range as described in any one of claims 1-5.

Images