Electric vehicle travel characteristic analysis method and device, medium and product

Abstract

The embodiment of the invention provides an electric vehicle travel feature analysis method and device, a medium and a product. The method comprises the following steps: acquiring first information of communication service of a vehicle-mounted terminal and second information of communication service of a user terminal; determining electric vehicle information corresponding to the first information according to the first information; determining electric vehicle owner information corresponding to the second information according to the second information; performing matching connection verification on the electric vehicle information and the electric vehicle owner information to obtain a vehicle identification result; and according to the vehicle identification result, analyzing travel behavior characteristics of the target electric vehicle. According to the scheme, the limitation of traditional single-source data is broken through, and the problem of inaccuracy in the aspect of feature analysis in the prior art is solved through collaborative analysis of the first information and the second information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle travel feature analysis, and in particular to an electric vehicle travel feature analysis method, device, medium and product. BACKGROUND

[0002] Under the background of global energy crisis and increasing environmental protection awareness, new energy vehicles, especially electric vehicles, are becoming the trend of future vehicle development. The rapid development of electric vehicles has brought new challenges to urban transportation systems and power infrastructure. In order to accurately perceive the urban traffic situation and optimize the layout of urban charging infrastructure, it is essential to analyze the travel characteristics of electric vehicles. At present, the technical solutions for analyzing the travel characteristics of electric vehicles can be roughly divided into three categories: the first category is to face the electric vehicle owners and directly obtain the travel behavior characteristics of some electric vehicle owners through questionnaire survey; the second category is to obtain the travel characteristics of electric vehicles through simulation by presetting the travel environment and overall travel parameters; the third category is to obtain the travel characteristics of electric vehicles by extracting and analyzing the historical driving and parking state information of electric vehicles based on the vehicle running state data reported by the vehicle Telematics BOX (T-BOX) on the Internet.

[0003] The existing technical solutions have the following problems: the questionnaire survey method has high cost, limited sample coverage, and subjective bias; the simulation deduction method relies on preset parameters and is difficult to dynamically reflect the real scene; the vehicle data analysis method is limited by the data of the vehicle enterprise itself, and lacks key dimension data of the behavior of the vehicle owner. In summary, the existing technology has certain limitations in data authenticity, dynamic modeling and data integrity, etc. SUMMARY

[0004] At least one embodiment of the present application provides an electric vehicle travel feature analysis method, device, medium and product, which is used to solve the problem of inaccuracy of the existing vehicle travel feature analysis scheme in feature analysis.

[0005] In order to solve the above technical problems, the present application is implemented as follows:

[0006] In a first aspect, the embodiments of the present application provide an electric vehicle travel feature analysis method, comprising:

[0007] obtaining first information of a vehicle terminal for communication service and second information of a user terminal for communication service;

[0008] determining electric vehicle information corresponding to the first information according to the first information;

[0009] determining electric vehicle owner information corresponding to the second information according to the second information;

[0010] matching connection verification of the electric vehicle information and the electric vehicle owner information, to obtain a vehicle identification result;

[0011] According to the vehicle identification result, the travel behavior characteristics of the target electric vehicle are analyzed.

[0012] Optionally, according to the first information, the electric vehicle information corresponding to the first information is determined, including:

[0013] Obtaining vehicle travel behavior characteristics and online behavior characteristics of a vehicle-mounted terminal;

[0014] According to the vehicle travel behavior characteristics, the online behavior characteristics, and a preset initial model using an unsupervised algorithm, an electric vehicle identification model for identifying electric vehicle information is constructed.

[0015] The first information is input into the electric vehicle identification model to determine the electric vehicle information corresponding to the first information.

[0016] Optionally, obtaining vehicle travel behavior characteristics and online behavior characteristics of a vehicle-mounted terminal includes:

[0017] Obtaining first sample data corresponding to the Internet of Vehicles card data;

[0018] According to a preset identifier in the first sample data, position signaling data and online log data of a vehicle-mounted terminal are obtained.

[0019] According to the position signaling data, the vehicle travel behavior characteristics are obtained. The vehicle travel behavior characteristics include at least one of the maximum distance of a single trip, the average distance of a single trip, the maximum speed of a single trip, the average speed of a single trip, the number of visits to a gas station within a preset time, and the number of visits to a charging station within a preset time.

[0020] According to the online log data, the online behavior characteristics are obtained. The online behavior characteristics include application class characteristics and traffic class characteristics.

[0021] Optionally, according to the vehicle travel behavior characteristics, the online behavior characteristics, and a preset initial model using an unsupervised algorithm, an electric vehicle identification model for identifying electric vehicle information is constructed, including:

[0022] The preprocessed vehicle travel behavior characteristics and online behavior characteristics are spliced to obtain a fusion feature vector.

[0023] Obtaining second sample data of vehicle travel of a vehicle without labeled vehicle type;

[0024] inputting the second sample data and the fusion feature vector into a preset initial model using an unsupervised algorithm, training the initial model, and obtaining an electric vehicle identification model for identifying electric vehicle information after training.

[0025] Optionally, according to the second information, the electric vehicle owner information corresponding to the second information is determined, including:

[0026] Obtaining third sample data corresponding to the mobile phone card data;

[0027] Determining the application association rule and the time duration rule of the user terminal for communication service;

[0028] Based on the third sample data, the application association rule and the time duration rule, an electric vehicle owner identification model for identifying electric vehicle owner information is constructed;

[0029] Inputting the second information into the electric vehicle owner identification model to determine the electric vehicle owner information corresponding to the second information.

[0030] Optionally, according to the vehicle identification result, the travel behavior characteristics of the target electric vehicle are analyzed, including:

[0031] According to the vehicle identification result, the target owner associated with the vehicle identification result and the target electric vehicle are determined;

[0032] Obtaining mobile phone location signaling data of the target owner and vehicle networking card signaling data corresponding to the target electric vehicle;

[0033] Based on the mobile phone location signaling data and the vehicle networking card signaling data, a first traffic travel amount sequence in the mobile phone location signaling data and a second traffic travel amount sequence in the vehicle networking card signaling data are extracted respectively;

[0034] In the case that the starting point spatial position and the ending point spatial position of the first traffic travel amount sequence and the second traffic travel amount sequence are the same, and the starting point departure time difference and the ending point arrival time difference are less than a preset time length, the target travel data of the target owner driving the target electric vehicle is determined;

[0035] Statistics of the total mileage, total time length of the owner driving the electric vehicle in all target travel data in a preset period, and the proportion of the total mileage and the total time length in all travel mileage and time length are generated to generate the travel behavior preference of the target owner driving the electric vehicle.

[0036] Optionally, according to the vehicle identification result, the travel behavior characteristics of the target electric vehicle are analyzed, including:

[0037] According to the vehicle identification result, a micro-analysis result and a macro-analysis result of a target area passed by the target electric vehicle are determined; the micro-analysis result is used to represent a result of counting vehicle flow and parking behavior in a grid after the target area is divided into the grid; and the macro-analysis result is used to represent a result of overall travel regularity in the target area.

[0038] The micro-analysis result and the macro-analysis result are taken as a comprehensive analysis result of the target electric vehicle.

[0039] In a second aspect, an embodiment of the present application provides an electric vehicle travel feature analysis device, comprising:

[0040] An acquisition module is configured to acquire first information of a vehicle terminal for communication service and second information of a user terminal for communication service.

[0041] A first determination module is configured to determine electric vehicle information corresponding to the first information according to the first information.

[0042] A second determination module is configured to determine electric vehicle owner information corresponding to the second information according to the second information.

[0043] A first processing module is configured to match, connect and verify the electric vehicle information and the electric vehicle owner information to obtain a vehicle identification result.

[0044] A second processing module is configured to analyze travel behavior features of a target electric vehicle according to the vehicle identification result.

[0045] In a third aspect, an embodiment of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement steps of the method in any one of the first aspect.

[0046] In a fourth aspect, an embodiment of the present application provides a computer program product, comprising computer instructions, and the computer instructions are executed by a processor to implement steps of the method in any one of the first aspect.

[0047] Compared with the prior art, the electric vehicle travel feature analysis method, device, medium and product provided by the embodiment of the application can identify fake or abnormal data through cross verification of the first information of the vehicle terminal and the second information of the user terminal, and can also match the vehicle owner identity information with the vehicle features to avoid feature misjudgment caused by vehicle transfer in the traditional scheme; the electric vehicle information determined based on the first information and the electric vehicle owner information determined based on the second information are matched and connected for verification to obtain a vehicle identification result, and dynamic verification is realized; the vehicle identification result is used to analyze the travel behavior features of the target electric vehicle, overcoming the problem of single data in the traditional analysis method and effectively improving the accuracy of travel feature analysis. The scheme provided by the application determines the vehicle identification result through cooperative analysis of the first information and the second information, and analyzes the travel behavior features of the target electric vehicle by using the vehicle identification result, solving the problem of inaccuracy of the existing vehicle travel feature analysis scheme in feature analysis. BRIEF DESCRIPTION OF DRAWINGS

[0048] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included to provide a description of the preferred embodiments and are not intended to limit the scope of the application. Moreover, the same reference numerals are used throughout the same figures. In the drawings:

[0049] Figure 1 One of the schematic diagrams of the electric vehicle travel feature analysis method provided by the embodiment of the application;

[0050] Figure 2 The second schematic diagram of the electric vehicle travel feature analysis method provided by the embodiment of the application;

[0051] Figure 3 The structural diagram of the electric vehicle travel feature analysis method provided by the embodiment of the application. DETAILED DESCRIPTION

[0052] The terms "first", "second", and the like in the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged under appropriate circumstances, so that the embodiments of the application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second" are usually a class, and are not limited to the number of objects, for example, the first object can be one or more. In addition, "or" in the present application means at least one of the connected objects. For example, "A or B" covers three schemes, namely, scheme one: including A and not including B; scheme two: including B and not including A; scheme three: including A and including B. The character " / " generally represents that the objects before and after are in an "or" relationship.

[0053] The term "indication" in the present application can be a direct indication (or explicit indication) or an indirect indication (or implicit indication). The direct indication can be understood as that the sender explicitly informs the receiver of specific information, operation to be performed or requested result, etc. in the sent indication. The indirect indication can be understood as that the receiver determines the corresponding information according to the indication sent by the sender, or judges and determines the operation to be performed or the requested result according to the judgment result.

[0054] As the technical problem described in the background, the scheme of the present application fully utilizes the communication operator data from the two dimensions of the vehicle owner and the vehicle, and more comprehensively realizes the analysis of the electric vehicle travel characteristics.

[0055] Please refer to Figure 1 The electric vehicle travel characteristic analysis method provided by the embodiment of the present application comprises:

[0056] Step 11, obtaining first information of the communication service of the vehicle terminal and second information of the communication service of the user terminal.

[0057] In the embodiment of the present application, the first information is used to indicate all the information of the communication of the vehicle terminal. The first information herein can include real-time collection of the vehicle driving track, charging record, energy consumption state, etc. The second information is used to represent the online communication information of the user terminal. The second information herein can include user positioning information, mobile network connection log, etc. Specifically, the vehicle terminal and the mobile communication network can be used to collect the vehicle network card data and the mobile phone card data in real time. The vehicle network card data includes the vehicle driving track, the charging record, the energy consumption state, etc. The mobile phone card data includes the user positioning information and the mobile network connection log.

[0058] It should be noted that the communication operator provides communication services for terminal users through a SIM (Subscriber Identity Module) card. According to different use scenarios and functions, it can be divided into two types of mobile phone cards and Internet of Things cards. The mobile phone card is used in the mobile phone terminal device and can provide voice, short message and data transmission functions. The Internet of Things card is used in the Internet of Things device, so that the device can be connected to the network and realize data exchange. For the Internet of Things card, according to the subscription scenario information, the vehicle network card can be further identified, that is, the Internet of Things card loaded in the vehicle T-BOX terminal for realizing the intelligent networking service of the vehicle. The subscription scenario information can be as shown in Table 1.

[0059] Table 1: Internet of Things user subscription scenario information.

[0060] Encoding Encoding name Sign-up scenario 11001 Vehicle front loading Vehicle-to-everything service 11002 Vehicle rear loading Vehicle-to-everything service 12001 Public safety Public service 12002 Mobile office Public service 13001 Financial payment Retail service 14001 Personal wearable device Smart furniture ...... ...... ......

[0061] The application breaks through the limitation of traditional single data source, and provides vehicle running state parameters by Internet of Vehicles data and user behavior characteristics by mobile phone data, to form multi-dimensional data complementation.

[0062] In step 12, the electric vehicle information corresponding to the first information is determined according to the first information.

[0063] In the embodiment of the application, the first information collected above is real-time data, and the potential electric vehicle identification model needs to be constructed by using historical Internet of Vehicles data. At present, electric vehicles and some intelligent fuel vehicles are equipped with Internet of Vehicles cards as standard, and the information of the vehicle enterprises using the Internet of Vehicles cards is registered when the cards are handled. Therefore, for some vehicle enterprises that only produce electric vehicles, such as T manufacturers, X manufacturers and L manufacturers, all the registered Internet of Vehicles cards can be directly extracted, and the vehicles equipped with these Internet of Vehicles cards can be directly determined as electric vehicles. For other Internet of Vehicles cards, they may be equipped on electric vehicles or intelligent fuel vehicles, so it is necessary to further construct an identification model. The identification model is constructed by an unsupervised clustering method, and the constructed identification model and the first information are used to distinguish the vehicle types and determine the electric vehicle information after the distinction. In this way, the dependence on artificial labels or fixed rules can be avoided, the vehicle type adaptive division can be realized through the inherent law of data, and the classification generalization ability can be improved.

[0064] In step 13, the electric vehicle owner information corresponding to the second information is determined according to the second information.

[0065] Based on the online log data in the second information and in combination with the full list of electric vehicle brand application programs (APPs) and charging APPs, the electric vehicle owner identification, i.e., the identification of potential vehicle owners, is performed by using the electric vehicle owner identification model constructed in advance. In the electric vehicle owner identification model, at least an application program association rule indicating the association between vehicle use and charging behavior and a time persistence rule for excluding accidental behavior interference are required. The second information is input into the electric vehicle owner identification model, and the high-confidence potential electric vehicle owner can be determined, and the identity information and behavior characteristic data of the potential electric vehicle owner can be output. The second information and the electric vehicle owner identification model verification rule are used to improve the identification accuracy by combining the behavior association and long-term stability.

[0066] In step 14, the electric vehicle information and the electric vehicle owner information are matched, connected and verified to obtain a vehicle identification result.

[0067] An example can be constructed to build a vehicle matching model to realize matching and verification of potential electric vehicle and vehicle owner identification results. According to the real-name authentication information of the Internet of Vehicles card corresponding to the first information and the mobile card corresponding to the second information, the association analysis of the two can be performed, so as to realize the matching and verification of the potential electric vehicle and the potential electric vehicle owner identification result. Through the Internet of Vehicles card and the mobile card of the potential electric vehicle owner, the corresponding real-name authentication information, that is, the user ID information, is extracted, and then the table connection operation is performed by taking the information as the connection field. The matching successful Internet of Vehicles card and mobile card are extracted. Then the vehicle loaded with the Internet of Vehicles card can be determined as an electric vehicle, and the corresponding mobile card is the mobile card used by the vehicle owner.

[0068] Step 15, according to the vehicle identification result, analyzing the travel behavior characteristics of the target electric vehicle.

[0069] The travel behavior characteristics of the present application include but are not limited to: time dimension: high-frequency travel period, single travel duration distribution; spatial dimension: resident area heat map, charging station preference distribution; energy consumption characteristics: correlation analysis of charging efficiency and travel energy consumption. The travel behavior characteristics obtained by analysis can support charging pile layout optimization, user trip planning recommendation and carbon emission reduction accounting.

[0070] The scheme of the present application can identify fake or abnormal data through cross verification of the first information of the vehicle terminal and the second information of the user terminal, and can also match the vehicle owner's identity information with the vehicle characteristics to avoid feature misjudgment caused by vehicle transfer in traditional schemes. The electric vehicle information determined based on the first information and the electric vehicle owner information determined based on the second information are matched and connected to verify the vehicle identification result, which realizes dynamic verification. Using the vehicle identification result, the travel behavior characteristics of the target electric vehicle are analyzed. From the multi-dimensional data obtained to the analysis process of the result, the problem of single data of traditional analysis methods is overcome, and the accuracy of travel characteristic analysis is effectively improved. Based on the travel behavior characteristics, the spatio-temporal characteristics and behavior correlation can be determined to construct an interpretable travel portrait to support accurate decision-making.

[0071] Optionally, the step 12 described above comprises:

[0072] Obtaining vehicle travel behavior characteristics and online behavior characteristics of the vehicle terminal;

[0073] According to the vehicle travel behavior characteristics, the online behavior characteristics and the initial model of the preset unsupervised algorithm, an electric vehicle identification model for identifying electric vehicle information is constructed;

[0074] The first information is input into the electric vehicle identification model to determine the electric vehicle information corresponding to the first information.

[0075] It should be noted that the Internet of Vehicles card and the mobile phone card will generate location signaling and online log data in the use process. The location signaling records the connection interaction relationship between the terminal device and the communication base station and the spatial position information of the base station, and can reflect the space-time trajectory of the terminal. The specific data form is shown in Table 2.

[0076] Table 2: Location signaling data field.

[0077]

[0078] It should be noted that the online log records the information of the terminal device accessing the Internet application and using the traffic, and can reflect the online behavior of the terminal. The specific data form is shown in Table 3. Both types of data are collected and provided by the communication operator in the process of providing communication services for terminal users.

[0079] Table 3: Online log data field.

[0080]

[0081] Specifically, since only part of the Internet of Vehicles cards loaded on the electric vehicles can be determined as positive samples through the vehicle enterprise information, and the negative samples, i.e. the Internet of Vehicles cards loaded on the intelligent fuel vehicles, cannot be obtained, the vehicle travel behavior characteristics and the online behavior characteristics of the vehicle-mounted terminal are used as training data, and an initial model using a preset unsupervised algorithm is used to construct an electric vehicle identification model for identifying electric vehicle information. The electric vehicle identification model is an identification model based on unsupervised clustering, and can be used to distinguish and identify electric vehicles and intelligent fuel vehicles.

[0082] Further, the vehicle travel behavior characteristics and the online behavior characteristics of the vehicle-mounted terminal are obtained, including:

[0083] Obtain the first sample data corresponding to the Internet of Vehicles card data;

[0084] According to the preset identifier in the first sample data, obtain the location signaling data and the online log data of the vehicle-mounted terminal;

[0085] According to the location signaling data, obtain the vehicle travel behavior characteristics; the vehicle travel behavior characteristics include at least one of the maximum distance of single trip, the average distance of single trip, the maximum speed of single trip, the average speed of single trip, the number of visits to gas stations within a preset time, and the number of visits to charging stations within a preset time.

[0086] According to the online log data, obtain the online behavior characteristics; the online behavior characteristics include application type characteristics and traffic type characteristics.

[0087] In the embodiments of the present application, the first sample data is sample data of a vehicle networking card, which can be historical vehicle networking card data, the preset identifier can be "uid" in Table 2, "uid" is taken as a unique identifier, the location signaling data and the online log data of the vehicle terminal in the sample data of the vehicle networking card are associated and integrated, and duplicate records are removed to ensure that each record is unique. For the location signaling and online log data with the same "uid", the time stamps are sorted from early to late, and aligned to ensure that the vehicle travel behavior characteristics and online behavior characteristics in a given time period can be analyzed. Based on the location signaling data, the travel behavior characteristics X mobility of the vehicle are extracted, specifically including the maximum distance D max of a single trip, the average distance D mean of a single trip, the maximum speed V max of a single trip, the average speed V mean of a single trip, the number of visits to a gas station in a month M, the number of visits to a charging station in a month N, and the like.

[0088] Specifically, according to the location signaling data, the vehicle travel behavior characteristics are obtained, including:

[0089] Based on the location signaling data, the parking points of the vehicle are identified. For example, if the position of the trajectory point of the vehicle does not change significantly in a period of time (i.e., the distance d between adjacent trajectory points is less than d0, and the duration t is greater than t0, where d0 and t0 are preset threshold values), it is considered that the vehicle is parked at the position, and in the present application, d0 and t0 are set to 500 meters and 30 minutes, respectively.

[0090] All parking points of a single vehicle are sorted from early to late, and a trip OD is formed between adjacent parking points, where O represents the previous parking point and D represents the next parking point. The distance between OD is calculated, and the distance of a single trip of the vehicle is obtained. The distance is divided by the time difference between OD, and the speed of a single trip of the vehicle is obtained.

[0091] The distance and speed of all single trips of a single vehicle in a given data period are calculated, and the maximum value and the average value are obtained, respectively, to obtain the characteristics D max , D mean , V max , and V mean .

[0092] For each trajectory point, the distance from the known gas station position is calculated. If the distance from the gas station position of a plurality of consecutive trajectory points is less than 500 meters and the duration exceeds 3 minutes, it is considered that the vehicle visits the gas station and the refueling behavior may occur. The number of visits to the gas station by a single vehicle in a month is counted, and if the data covers multiple months, the average value of multiple months is calculated to obtain the number of visits to the gas station in a month M.

[0093] For each stop point, calculate its distance from the known charging station location. If the distance is less than 500 meters, it is considered that the vehicle has visited the charging station and charging behavior may have occurred. Count the number of times a single vehicle visits a charging station in a month, and if the data covers multiple months, calculate the average number of times a charging station is visited in a month, and obtain the number of times N in a month.

[0094] It should be noted that all the above distance calculations are obtained by converting latitude and longitude and surface distance. The distance between positions i and j is denoted as d j,j , and its calculation formula is (unit: meter):

[0095]

[0096] Where lat_i, lat_j: latitude of position i and j, representing the angle between the point and the equatorial plane (north latitude is positive, south latitude is negative). lon_i, lon_j: longitude of position i and j, representing the angle between the point and the prime meridian (east longitude is positive, west longitude is negative). Composition of intermediate variable a(i,j): sin2((lat_i-lat_j) / 2): calculate the half-angle sine square of the difference between the two latitudes, reflecting the influence of the latitude difference on the distance. cos(lat_i)×cos(lat_j): the product of the two latitude cosines, which corrects the distance contraction effect in the longitude direction caused by the difference in latitude. sin2((lon_i-lon_j) / 2): calculate the half-angle sine square of the difference between the two longitudes, representing the contribution of the longitude difference to the distance. Derivation of the final distance d(i,j): sin -1 a i,j ×6378173: take the square root of the intermediate variable and take the inverse sine to get the central angle (radian) between the two points on the earth's surface. 6378173 (meters) represents the equatorial radius in the earth ellipsoid model, which is used to convert radians to actual distance (this value corresponds to the international geodetic standard ellipsoid parameter). The formula in this application is suitable for approximate calculation of short distances (<200 kilometers) on the earth's surface, ignoring the influence of terrain undulations. Here, if the input latitude and longitude are angles (such as decimal degrees), they need to be converted to radians (angle x π / 180).

[0097] Further, according to the electric vehicle identification model and the vehicle networking card data, the electric vehicle information after being distinguished by the vehicle type is identified, including: based on the online log data, extracting the online behavior features X internetincluding access application class features and traffic class features. The extraction step includes: according to the application type classification corresponding to "app_type" as shown in Table 3, such as navigation class, entertainment class, charging class (charging pile query, payment), etc., the access frequency of each class of application is counted respectively as the access application type feature. Compared with fuel vehicles, electric vehicles access entertainment class and charging class applications more frequently. The sum of uplink traffic and downlink traffic and the proportion of uplink traffic of each vehicle in a given data period are counted as traffic class features. The battery capacity of fuel vehicles is very small and cannot continuously meet the power supply demand of intelligent vehicle-machine interaction systems, and the use of network traffic of the vehicle networking card is much smaller than that of electric vehicles, and electric vehicles need to continuously report vehicle operating state information, so the proportion of uplink traffic is higher.

[0098] Optionally, according to the vehicle travel behavior features and the online behavior features, an electric vehicle identification model of an unsupervised clustering method is constructed, including:

[0099] The preprocessed vehicle travel behavior features and online behavior features are spliced to obtain a fusion feature vector;

[0100] Second sample data of vehicle travel of unmarked vehicle types is obtained;

[0101] The second sample data and the fusion feature vector are input into a preset initial model using an unsupervised algorithm, the initial model is trained, and an electric vehicle identification model for identifying electric vehicle information after training is obtained.

[0102] In the embodiments of the application, the preprocessing can be standardization data processing of the vehicle travel behavior features and the online behavior features, for example, Z-score standardization or Min-Max standardization method can be used to standardize the data, to solve the problem of large dimension difference of different features. The standardized travel behavior feature vector X nobility and the online behavior feature vector X internet are spliced to obtain a fusion feature vector, and the calculation process is as follows: X mix =[X mobility ,X internet ].

[0103] Further, the second sample data and the fusion feature vector are input into an initial model using a preset unsupervised algorithm, the initial model is trained, and an electric vehicle identification model for identifying electric vehicle information after training is obtained, including: setting a clustering coefficient, for example, setting the clustering coefficient k = 2, and the cluster center of one of the clusters is calculated by averaging the feature vectors corresponding to the part of the vehicle cards determined as electric vehicles according to the vehicle enterprise information, denoted as C0, where C0 represents the initial cluster center of the electric vehicle, that is, the mean of the fusion feature vector is determined as the fixed cluster center representing the electric vehicle; the cluster center of the other cluster is randomly selected, that is, a feature vector is randomly selected from the second sample data; then the preset clustering algorithm in the initial model is trained by an unsupervised algorithm, and an electric vehicle identification model for identifying electric vehicle information after training is obtained. The preset clustering algorithm is, for example, a k-means clustering algorithm, and training the initial model can be understood as iterative training using the k-means clustering algorithm, which divides the unknown type of Internet of Vehicles cards into two categories of electric vehicles and fuel vehicles, and calculates the distance between the cluster center of each cluster and C0, and the category with the smallest distance belongs to the electric vehicle, while the other category belongs to the fuel vehicle. The electric vehicle identification model obtained after training has the ability to identify the type of vehicle (such as electric vehicle or fuel vehicle).

[0104] Specifically, C0 is the mean calculated according to the known electric vehicle samples provided by the vehicle enterprise, that is, the fusion feature vector described above, and C0 represents the typical behavior mode of the electric vehicle, such as high-frequency short-distance travel and peak online traffic associated with night charging. The C0 cluster center remains fixed during the clustering process and serves as an "anchor point" for classification. For example, if the vehicle enterprise has labeled part of the vehicles as electric vehicles, the mean of their fusion feature vectors (traveling behavior and online behavior) is defined as C0, which solidifies the cluster center of the electric vehicle.

[0105] In k-means clustering (k = 2), one cluster center is directly set to C0 at initialization, and the other cluster center is randomly selected (representing a potential fuel vehicle mode). Random selection is represented as: a feature vector is randomly selected from the second sample data in all unmarked vehicle types and is denoted as initial C1 (not dependent on vehicle enterprise data). The algorithm updates the cluster center position and data point assignment through iteration: (1) assignment step: the feature vector (i.e. second sample data) of each unmarked Internet of Vehicles card is assigned to the nearest cluster center, such as using Euclidean distance, etc. (2) update step: only the cluster center position other than C0 is dynamically updated, denoted as updated C1, and C0 remains unchanged; repeat until the cluster center converges, with a change less than a threshold value. Finally, two clusters are formed: a first cluster A (cluster center C0) and a second cluster B (cluster center C1 after convergence).

[0106] It should be noted that the initial C1 has no clear vehicle direction, which can be a fuel vehicle or a hybrid electric vehicle, etc.; the converged C1 represents the typical behavior characteristics of the fuel vehicle.

[0107] The Euclidean distance from the cluster center of the two clusters to C0 is calculated respectively: the cluster center of the first cluster A is C0, so the distance is always 0 (dist_A = 0). The distance from the cluster center C1 of the second cluster B to C0 is dist_B (the absolute value of C1-C0), and the cluster with the smallest distance (i.e. dist_A = 0) is determined as the electric vehicle class because its cluster center coincides with C0, indicating that the behavior pattern of this cluster is closest to the known electric vehicle sample. The cluster with a larger distance (dist_B > 0) is determined as the fuel vehicle class because its cluster center deviates from C0, representing different behavior characteristics such as long-distance driving and stable flow pattern.

[0108] The present application fixes C0 to integrate a small amount of labeled information of vehicle enterprises into the initial model of unsupervised learning for model training, avoiding the bias of pure random initialization and improving classification accuracy, such as distinguishing the charging / driver pattern difference between electric vehicles and fuel vehicles; distance calculation quantifies the similarity between the cluster and the reference prototype, and the smaller the distance, the closer the behavior characteristics to the typical mode of electric vehicles, ensuring that the classification result is interpretable; even if the data distribution is complex, this rule can be compared globally through the cluster center rather than single-point distribution, reducing the noise effect and being suitable for vehicle networking big data scenarios.

[0109] For example, assuming that a certain feature dimension (such as the average daily charging frequency) in the vehicle networking card data is as follows: known electric vehicle samples: vehicle A (1 time), vehicle B (0.8 times), vehicle C (1.2 times), C0 mean = (1 + 0.8 + 1.2) / 3 = 1 time. Unknown type samples: vehicle D (0.3 times), vehicle E (0.5 times), vehicle F (0.2 times). Initialize C0 = 1, and randomly select C1 = 0.3 (the feature value of vehicle D). Assign samples: vehicle D to C1 (|0.3-0.3| = 0 < |0.3-1| = 0.7); vehicle E to C1 (|0.5-0.3| = 0.2 < |0.5-1| = 0.5); vehicle F to C1 (|0.2-0.3| = 0.1 < |0.2-1| = 0.8); update C1: new mean = (0.3 + 0.5 + 0.2) / 3 = 0.33. Iterate until convergence (assuming that the final C1 = 0.35). Determine the classification result: calculate the cluster center distance: |1-0.35| = 0.65. All vehicles (D, E, F) assigned to C1 (0.35) are determined as fuel vehicles (far from C0), and potential new samples close to C0 are determined as electric vehicles.

[0110] Optionally, the step 13 described above comprises:

[0111] Obtain third sample data corresponding to the mobile phone card data;

[0112] determine application program association rules and time duration rules of the user terminal for communication services;

[0113] based on the third sample data, the application program association rules and the time duration rules, build an electric vehicle owner identification model for identifying electric vehicle owner information;

[0114] input the second information into the electric vehicle owner identification model to determine the electric vehicle owner information corresponding to the second information.

[0115] It should be noted that for electric vehicle owners, they usually need to install and register the corresponding vehicle machine application program (APP) on the mobile phone to realize communication with the intelligent vehicle networking system, so as to realize remote operation, system management and the like of the vehicle. At the same time, due to the limitation of battery capacity, electric vehicles need to be charged frequently to maintain endurance, and when charging, the owner also needs to install and register the related APP and perform corresponding operations on the mobile phone, such as charging pile code scanning, account login, starting / stopping charging, settlement payment and the like. Therefore, by analyzing the mobile phone card online log data, the identification of potential electric vehicle owners can be realized.

[0116] In the embodiments of the present application, third sample data corresponding to the mobile phone card data is obtained, and the third sample data usually refers to user behavior data collected through a mobile terminal, mainly including APP installation list, usage frequency, geographic location and the like. Based on such data, an electric vehicle related APP full list can be constructed, including two categories: one is the APP corresponding to each electric vehicle brand, such as T manufacturer APP, X manufacturer APP, L manufacturer APP and the like; the other is electric vehicle charging related APP, such as "e charging, special electricity, fast electricity, star charging" and the like. For each mobile phone user, the online log data and the electric vehicle related APP list constructed in the previous step are associated and analyzed to determine the application program association rules and time duration rules of the user terminal for communication services. For example, whether the user's mobile phone simultaneously installs electric vehicle brand class and electric vehicle charging class two types of APPs, that is, there are access records of the two types of APPs in the online log data; select at least 3 months of data to determine whether the user has access records of electric vehicle related APPs every month for 3 consecutive months.

[0117] Based on the third sample data, the application program relevance rule and the time persistence rule, a new electric vehicle owner identification model is constructed for identifying the electric vehicle owner information. If the user needs to meet the application program relevance rule and the time persistence rule at the same time in the electric vehicle owner identification model, the user is identified as a potential electric vehicle owner. The application program relevance rule indicates that the user not only uses the electric vehicle telematics APP, but also may have charging behavior. The time persistence rule indicates that the behavior is long-term stable, not accidental or temporary. Therefore, the above two rules are combined, the second information is input into the electric vehicle owner identification model, and the electric vehicle owner information corresponding to the second information is determined, so that the identification of the potential electric vehicle owner can be realized more accurately.

[0118] Optionally, the step 15 comprises:

[0119] According to the vehicle identification result, the target owner associated with the vehicle identification result and the target electric vehicle are determined.

[0120] The mobile phone location signaling data of the target owner and the Internet of Vehicles card signaling data corresponding to the target electric vehicle are obtained.

[0121] Based on the mobile phone location signaling data and the Internet of Vehicles card signaling data, a first traffic travel amount sequence in the mobile phone location signaling data and a second traffic travel amount sequence in the Internet of Vehicles card signaling data are extracted respectively.

[0122] In the case that the starting point spatial position and the ending point spatial position of the first traffic travel amount sequence and the second traffic travel amount sequence are the same, and the starting point departure time difference and the ending point arrival time difference are less than a preset time length, the target travel data of the target owner driving the target electric vehicle is determined.

[0123] The total mileage and the total time length of the owner driving the electric vehicle in all target travel data in a preset period are counted, and the proportion of the total mileage and the total time length in all travel mileage and time length is calculated, so as to generate the travel behavior preference of the target owner driving the electric vehicle.

[0124] It should be noted that the electric vehicle travel behavior characteristics include two parts, one is whether the owner chooses to drive the electric vehicle to travel, and the other is the behavior characteristics when driving the electric vehicle to travel. Since the target electric vehicle and the target owner can be identified at the same time, and the corresponding signaling trajectories are extracted respectively, the analysis of the two types of characteristics can be realized at the same time.

[0125] In the embodiments of the present application, the target data source is: mobile phone location signaling data, i.e., user mobile phone positioning information (timestamp, longitude and latitude, moving speed, etc.) recorded by a mobile base station, reflecting the user's travel trajectory. Vehicle networking card signaling data, i.e., real-time position and driving state (such as ignition / off time, mileage increment, energy consumption value, etc.) uploaded by the built-in communication module of the vehicle. Further, invalid positioning points (such as signal drift, stationary state noise points) are filtered, timestamps are aligned, and the timestamps of the mobile phone and the vehicle data are synchronized.

[0126] Based on the continuous positioning points, a travel segment is generated, and the start time, start position (longitude and latitude), end time, and end position of each record are marked to realize feature extraction of the first traffic travel quantity sequence (mobile phone data). Example: the user's mobile phone moves from position A (longitude X1, latitude Y1) to position B (longitude X2, latitude Y2) at 08:00, and the end time is 08:30, forming a travel record. The travel segment is divided by the vehicle ignition / off signal, and the departure time, starting position, arrival time, and ending position are recorded, and the feature extraction of the second traffic travel quantity sequence (vehicle networking data) is realized. Example: the vehicle starts from position A' (longitude X1', latitude Y1') at 08:02, and arrives at position B' (longitude X2', latitude Y2') at 08:32, with a cumulative mileage of 15 kilometers; end point matching: similarly, the distance between the mobile phone end position B and the vehicle end position B' is ≤ threshold; time matching condition: departure time difference: the difference between the mobile phone departure time (08:00) and the vehicle departure time (08:02) is ≤ preset time length (such as 5 minutes). Arrival time difference: the difference between the mobile phone arrival time (08:30) and the vehicle arrival time (08:32) is ≤ preset time length. Matching judgment: the travel record that meets the spatial and temporal conditions at the same time is determined as the target travel data, i.e., the travel of the target electric vehicle driven by the vehicle owner.

[0127] Travel behavior statistics and preference generation: total mileage: statistics of the cumulative value of the mileage of all matching travel records (such as 100 kilometers per day); total time length: statistics of the total driving time of all matching travel records (such as 3 hours per day); proportion analysis: mileage proportion: electric vehicle travel mileage / vehicle owner's total travel mileage (including other transportation modes); time length proportion: electric vehicle travel time length / vehicle owner's total travel time length. Preference modeling: high-frequency scenario: identify the typical scenarios of the vehicle owner using the electric vehicle (such as weekday commuting, weekend long-distance travel); energy consumption efficiency: combine mileage and charging data to calculate the average energy consumption per unit of mileage; charging habit: statistics of the charging time period distribution. Further, according to the information of travel behavior statistics and preference generation and preference modeling, the travel behavior preference of the target vehicle owner driving the electric vehicle is generated.

[0128] The application reduces the misjudgment rate through double signaling data cross verification, such as temporary car interference; supports updating the preference model by period (day / week / month) to adapt to behavior changes; can provide electric vehicle usage habit analysis for vehicle enterprises to optimize product design; and can plan the layout of charging facilities and subsidy policies based on the proportion of electric vehicle travel.

[0129] It can also be understood that the mobile phone signaling data can reflect the whole process of the driver's 7x24 hour travel trajectory, and the Internet of Vehicles card signaling data can reflect the trajectory of the electric vehicle travel. The two are analyzed by time and space superposition, that is, by calculating the time and space proximity between the driver and the vehicle signaling position points, the trajectory section of the driver choosing to drive the electric vehicle for travel can be obtained, and the proportion of the distance and time of driving the electric vehicle for travel to all travel distance and time, etc. can be calculated, so as to realize the characterization of the electric vehicle user's travel behavior characteristics from the driver side. The specific steps include: based on the driver's mobile phone location signaling data, and according to the parking identification and OD extraction method, the driver's travel OD is obtained. The driver's travel OD and the travel OD of the vehicle under his name are analyzed by time and space superposition. If the spatial positions corresponding to the O points of the two travel ODs are the same, the spatial positions corresponding to the D points are also the same, and the time difference t(O) of the O point departure time and the time difference t(D) of the D point arrival time are both less than the set threshold t1, it is considered that the two travel ODs have high time and space proximity, and the driver drives the electric vehicle for travel. The optional time threshold t1 in the application is set to 5 minutes. Based on the above travel OD, the distance and time of the driver driving the electric vehicle for travel and the distance and time of all travel of the driver in a given period, such as a week or a month, are calculated, and the proportions of the two are calculated, which can reflect the preference characteristics of the driver choosing to drive the electric vehicle for travel.

[0130] Optionally, according to the vehicle identification result, the travel behavior characteristics of the target electric vehicle are analyzed, further comprising:

[0131] According to the vehicle identification result, the micro analysis result and the macro analysis result of the target area passed by the target electric vehicle are determined; the micro analysis result is used to represent the result of statistical grid vehicle flow and parking behavior of the target electric vehicle after the target area is divided into grids; and the macro analysis result is used to represent the result of the overall travel rule corresponding to the total daily travel amount, the average travel distance or the average travel time of the target electric vehicle in the target area;

[0132] The micro analysis result and the macro analysis result are used as the comprehensive analysis result of the target electric vehicle.

[0133] In the embodiments of the present application, the travel behavior analysis of the target electric vehicle is realized by combining micro-grid and macro statistics for comprehensive evaluation. At the micro level, the target area is divided into 500m x 500m grid units, the number of vehicles in each grid, the number of vehicles in each grid, and the distribution of the number of vehicles in each grid are counted, and the time and space heat map is generated to identify the charging demand hot spot; at the macro level, the global OD data is aggregated, the total number of daily trips, the average distance and other indicators are calculated, and the overall charging demand trend is predicted by the time series model. The micro analysis results and the macro analysis results are used as the comprehensive analysis results of the target electric vehicle, and the two are cooperated to form a "grid-level accurate positioning and regional situation awareness" analysis closed loop, which provides decision basis for electric vehicle behavior analysis, charging facility layout and power grid dispatching.

[0134] Step 15 of the present application also includes that for the behavior of driving electric vehicles, its characteristics can be described from two levels of macro and micro. Specifically, at the micro level, the target area can be divided into traffic zones or grids, and the 24-hour electric vehicle inflow, outflow and parking characteristics in each traffic zone or grid are counted, which is helpful to evaluate the charging demand of local area and assist the site selection and layout planning of charging facilities. Taking grid scale analysis as an example, the specific steps include: dividing the target area into 500m x 500m grids, then mapping the electric vehicle trip OD (target trip data) to the grid, i.e. respectively mapping the O point (starting point) and D point (ending point) to the grid according to their latitude and longitude. 24 hours are divided into 15-minute intervals, then for each time slice, based on the trip OD mapped to the grid in the previous step, the number of electric vehicles in different grids is counted and summarized to form a grid-scale electric vehicle inflow and outflow characteristic statistical table, as shown in Table 4.

[0135] Table 4: Grid-scale electric vehicle inflow and outflow characteristic statistical table.

[0136]

[0137] For each grid, further according to the parking time of electric vehicles at O point and D point, the number of electric vehicles with different parking time in the grid is counted and summarized to form a grid-scale electric vehicle parking characteristic statistical table, as shown in Table 5.

[0138] Table 5: Grid-scale electric vehicle parking characteristic statistical table.

[0139]

[0140] Based on the above two statistical tables (Table 4 and Table 5), on the one hand, the grid-scale electric vehicle departure, arrival and parking heat map can be drawn to intuitively show its spatial distribution characteristics, and on the other hand, its time sequence change characteristics in a day can be analyzed.

[0141] For the macro level, based on the electric vehicle trip OD, the total daily trip quantity (OD quantity) of electric vehicles in the target area, the average trip distance, the average trip duration, and other characteristics can be summarized and statistically analyzed, as shown in Table 6, which helps to perceive the traffic situation and predict the total charging demand.

[0142] Table 6: Grid-scale electric vehicle parking feature statistical table.

[0143]

[0144] The application realizes accurate depiction of charging demand from local to global through the dual perspectives of micro-grid analysis and macro trend statistics, provides multi-dimensional data support for urban planning, energy management and policy making, and has technical feasibility and practical landing value.

[0145] Referring to Figure 2 The embodiment of the application provides a specific method of an electric vehicle trip feature analysis method, including: step 101, acquiring location signaling and online log data corresponding to a communication operator's vehicle networking card and mobile card; step 102, constructing an electric vehicle identification model based on the vehicle networking card data; step 103, constructing a potential electric vehicle owner identification model based on the mobile card data; step 104, constructing a person-vehicle matching model to realize matching verification of the potential electric vehicle and owner identification results; and step 105, analyzing electric vehicle trip behavior characteristics based on the electric vehicle and owner identification results.

[0146] In summary, the scheme of the application provides full-quantity electric vehicle identification and data aggregation: based on the online log / location signaling data of the communication operator's vehicle networking card, different electric vehicles of different car companies and brands are identified through an unsupervised clustering model, which breaks through the limitations of traditional sampling surveys or single car company data and realizes full-quantity electric vehicle trip OD data aggregation. Accurate identification and correlation verification of the owner: the potential owner is identified by combining the behavior rule model of the mobile user accessing the charging / car company APP; the owner's identity is verified by correlating the vehicle networking card and the mobile card using real-name authentication information, and the all-weather trip trajectory of the owner missing in the traditional vehicle networking data is supplemented. Multi-dimensional trip feature analysis: the driving preference is identified through the spatio-temporal superposition analysis (O / D point proximity) of the owner's dynamic trajectory and the vehicle trajectory; the electric vehicle spatio-temporal features are depicted by combining the micro-grid-level traffic statistics and the macro-level trip total quantity, distance and duration statistics table. Data fusion advantage: the vehicle networking card (vehicle) and mobile card (owner) data are integrated to comprehensively analyze the trip behavior from two dimensions, which improves the result reliability and coverage range.

[0147] The application is based on operator vehicle networking card data, covers real and continuous trajectories of full-brand electric vehicles, and avoids the deviation of traditional sampling or simulation. The vehicle driving trajectory and the user's full-time travel data are fused to realize the multi-dimensional characterization of travel characteristics from time, space, and user behavior. The application supports traffic situation awareness and charging demand prediction scenarios, and breaks through the limitations of traditional vehicle networking serving only a single vehicle enterprise. Through operator big data fusion and multi-model collaborative analysis, the application realizes accurate identification of electric vehicles and users, multi-dimensional analysis of travel characteristics, and has comprehensive data, deep analysis, and commercial landing potential, providing innovative solutions for the transportation, energy, and automobile industries.

[0148] The above introduces various methods of embodiments of the application. The following will further provide a device for implementing the above method.

[0149] Please refer to Figure 3 The embodiments of the application also provide an electric vehicle travel characteristic analysis device, which comprises:

[0150] The acquisition module 31 is configured to acquire first information of a vehicle terminal for communication service and second information of a user terminal for communication service.

[0151] The first determination module 32 is configured to determine electric vehicle information corresponding to the first information according to the first information.

[0152] The second determination module 33 is configured to determine electric vehicle owner information corresponding to the second information according to the second information.

[0153] The first processing module 34 is configured to match and connect and verify the electric vehicle information and the electric vehicle owner information to obtain a vehicle identification result.

[0154] The second processing module 35 is configured to analyze travel behavior characteristics of a target electric vehicle according to the vehicle identification result.

[0155] Optionally, the first determination module 32 comprises:

[0156] The first acquisition unit is configured to acquire vehicle travel behavior characteristics and online behavior characteristics of a vehicle terminal.

[0157] The first construction unit is configured to construct an electric vehicle identification model for identifying electric vehicle information according to the vehicle travel behavior characteristics, the online behavior characteristics, and a preset initial model using an unsupervised algorithm.

[0158] The first determination unit is configured to input the first information into the electric vehicle identification model to determine electric vehicle information corresponding to the first information.

[0159] Optionally, the first obtaining unit is specifically configured to:

[0160] obtain first sample data corresponding to the Internet of Vehicles card data;

[0161] obtain location signaling data and online log data according to a preset identifier in the first sample data;

[0162] obtain the vehicle travel behavior feature according to the location signaling data; the vehicle travel behavior feature includes at least one of a maximum distance of single travel, an average distance of single travel, a maximum speed of single travel, an average speed of single travel, a number of visits to a gas station within a preset time, and a number of visits to a charging station within a preset time;

[0163] obtain the online behavior feature according to the online log data; the online behavior feature includes an application type feature and a traffic type feature.

[0164] Optionally, the first constructing unit is specifically configured to:

[0165] splicing the preprocessed vehicle travel behavior feature and the online behavior feature to obtain a fusion feature vector;

[0166] obtain second sample data of vehicle travel of an unlabeled vehicle type;

[0167] input the second sample data and the fusion feature vector into an initial model using a preset unsupervised algorithm, train the initial model, and obtain an electric vehicle identification model for identifying electric vehicle information after training.

[0168] Optionally, the second determining module 33 includes:

[0169] a second obtaining unit configured to obtain third sample data corresponding to mobile phone card data;

[0170] a second determining unit configured to determine an application program association rule and a time duration rule of a user terminal for communication services;

[0171] a second constructing unit configured to construct an electric vehicle owner identification model for identifying electric vehicle owner information based on the third sample data, the application program association rule, and the time duration rule;

[0172] a third determining unit configured to input the second information into the electric vehicle owner identification model to determine electric vehicle owner information corresponding to the second information.

[0173] Optionally, the second processing module 35 includes:

[0174] A fourth determining unit is configured to determine, according to the vehicle identification result, a target vehicle owner associated with the vehicle identification result and the target electric vehicle;

[0175] A third obtaining unit is configured to obtain mobile phone location signaling data of the target vehicle owner and vehicle networking card signaling data corresponding to the target electric vehicle;

[0176] A first processing unit is configured to extract, based on the mobile phone location signaling data and the vehicle networking card signaling data, a first traffic travel amount sequence in the mobile phone location signaling data and a second traffic travel amount sequence in the vehicle networking card signaling data, respectively.

[0177] A second processing unit is configured to determine, in a case where a starting point spatial position and an ending point spatial position of the first traffic travel amount sequence and the second traffic travel amount sequence are the same, and a starting point departure time difference and an ending point arrival time difference are less than a preset time length, target travel data of the target vehicle owner driving the target electric vehicle.

[0178] A third processing unit is configured to count a total mileage, a total time length of vehicle owner driving electric vehicle travel in all target travel data in a preset period, and a proportion of the total mileage and the total time length in all travel mileages and time lengths, and generate travel behavior preferences of the target vehicle owner driving electric vehicles.

[0179] Optionally, the second processing module 35 further includes:

[0180] A fifth determining unit is configured to determine, according to the vehicle identification result, a micro analysis result and a macro analysis result of a target area passed through by the target electric vehicle; the micro analysis result is used to represent a result of counting vehicle flow and parking behavior in a grid after the target area is divided into the grid; and the macro analysis result is used to represent a result of overall travel regularity in the target area.

[0181] A fourth processing unit is configured to take the micro analysis result and the macro analysis result as a comprehensive analysis result of the target electric vehicle.

[0182] It should be noted that the device in this embodiment corresponds to the device applied to the electric vehicle travel feature analysis method described above, and the implementation manners in the above embodiments are applicable to the embodiments of the device and can achieve the same technical effects. The device provided in the embodiments of the present application can realize all method steps achieved by the method embodiments and achieve the same technical effects. Therefore, the same parts and beneficial effects in the method embodiments will not be described in detail.

[0183] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the steps of the method for analyzing the characteristics of electric vehicle travel. The processor readable storage medium can be any available medium or data storage device that the processor can access, including but not limited to a magnetic storage (such as a floppy disk, a hard disk, a magnetic tape, a magneto-optical disk (MO) and the like), an optical storage (such as a CD, a DVD, a BD, a HVD and the like), and a semiconductor storage (such as a ROM, an EPROM, an EEPROM, a non-volatile memory (NAND FLASH), a solid state disk (SSD)) and the like.

[0184] The embodiment of the present application further provides a computer program product, which comprises computer instructions, the computer instructions are executed by a processor to implement each process in the above method embodiments, and can achieve the same technical effects, to avoid repetition, which will not be described here.

[0185] It should be noted that in the technical solutions of the present disclosure, the collection, collection, updating, analysis, processing, use, transmission, storage and the like of user personal information are in line with the relevant legal regulations, are used for legal purposes, and do not violate public order and good customs. Necessary measures are taken on personal information to prevent illegal access to user personal information data and maintain user personal information security and network security.

[0186] It should be noted that in this paper, the term "comprise", "comprise" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the sentence "includes a" does not exclude the presence of another identical element in the process, method, article or device including the element.

[0187] From the above description of the embodiments, those skilled in the art can clearly understand that the above embodiment method can be realized by software plus necessary general hardware platform, of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), which includes a plurality of instructions to make a terminal (which can be a mobile phone, computer, server, air conditioner, or network equipment, etc.) execute the method described in each embodiment of the present application.

[0188] The embodiments of the present application are described above with reference to the accompanying drawings, but the present application is not limited to the specific embodiments described above, and the specific embodiments described above are merely illustrative, but not restrictive, and a person of ordinary skill in the art can make many forms under the inspiration of the present application without departing from the purpose of the present application and the scope protected by the claims.

Claims

1. A method for analyzing electric vehicle travel characteristics, characterized in that: include: Acquire first information for a vehicle-mounted terminal to perform a communication service and second information for a user terminal to perform a communication service; determining, based on the first information, electric vehicle information corresponding to the first information; Determining, based on the second information, information about the electric vehicle owner corresponding to the second information; Matching and connecting the electric vehicle information and the electric vehicle owner information to obtain a vehicle identification result; Based on the vehicle identification result, the travel behavior characteristics of the target electric vehicle are analyzed.

2. The method according to claim 1, characterized in that Determining electric vehicle information corresponding to the first information according to the first information includes: Obtain vehicle travel behavior characteristics and vehicle terminal Internet access behavior characteristics; Constructing an electric vehicle identification model for identifying electric vehicle information based on the vehicle travel behavior characteristics, the Internet behavior characteristics, and a preset initial model using an unsupervised algorithm; The first information is input into the electric vehicle identification model to determine the electric vehicle information corresponding to the first information.

3. The method according to claim 2, characterized in that Obtain vehicle travel behavior characteristics and vehicle terminal Internet behavior characteristics, including: Obtaining first sample data corresponding to the Internet of Vehicles card data; Acquire location signaling data and Internet access log data of the vehicle terminal according to a preset identifier in the first sample data; Acquire the vehicle travel behavior characteristics according to the location signaling data; the vehicle travel behavior characteristics include at least one of the following: maximum distance of a single trip, average distance of a single trip, maximum speed of a single trip, average speed of a single trip, number of visits to gas stations within a preset time, and number of visits to charging stations within a preset time; The online behavior characteristics are obtained according to the online log data; the online behavior characteristics include access application characteristics and traffic characteristics.

4. The method according to claim 2, characterized in that Based on the vehicle travel behavior characteristics, the Internet behavior characteristics, and a preset initial model using an unsupervised algorithm, an electric vehicle identification model for identifying electric vehicle information is constructed, including: The pre-processed vehicle travel behavior features and Internet behavior features are spliced ​​together to obtain a fused feature vector; Obtain second sample data of vehicle trips without a labeled vehicle type; The second sample data and the fused feature vector are input into an initial model preset with an unsupervised algorithm, the initial model is trained, and an electric vehicle recognition model for recognizing electric vehicle information after training is obtained.

5. The method according to claim 1, wherein Determining, based on the second information, the electric vehicle owner information corresponding to the second information, including: Obtaining third sample data corresponding to the mobile phone card data; Determine application association rules and time continuity rules for user terminals to perform communication services; constructing an electric vehicle owner identification model for identifying electric vehicle owner information based on the third sample data, the application association rule, and the time persistence rule; The second information is input into the electric vehicle owner identification model to determine the electric vehicle owner information corresponding to the second information.

6. The method according to claim 1, characterized in that Based on the vehicle identification results, the travel behavior characteristics of the target electric vehicle are analyzed, including: Determining, according to the vehicle identification result, a target vehicle owner and the target electric vehicle associated with the vehicle identification result; Obtaining the target vehicle owner's mobile phone location signaling data and the vehicle network card signaling data corresponding to the target electric vehicle; Based on the mobile phone location signaling data and the vehicle network card signaling data, extracting a first traffic volume sequence in the mobile phone location signaling data and a second traffic volume sequence in the vehicle network card signaling data respectively; When the starting point spatial position and the ending point spatial position of the first traffic volume sequence and the second traffic volume sequence are the same, and the departure time difference of the starting point and the arrival time difference of the ending point are less than a preset time length, determining the target travel data of the target vehicle owner driving the target electric vehicle; The total mileage and total duration of the electric vehicle trips of the target vehicle owner in all target travel data within a preset period, as well as the proportion of the total mileage and the total duration to all travel mileage and duration, are counted to generate the travel behavior preference of the target vehicle owner in driving the electric vehicle.

7. The method according to claim 1, characterized in that Analyzing the travel behavior characteristics of the target electric vehicle based on the vehicle identification result further includes: Determine, based on the vehicle identification result, a micro-analysis result and a macro-analysis result of a target area passed by the target electric vehicle; the micro-analysis result is used to represent the result of dividing the target area into grids and counting the vehicle flow within the grids and the parking behavior of the target electric vehicle; the macro-analysis result is used to represent the result of the overall travel pattern corresponding to the total daily trip volume, average trip distance, or average trip duration of the target electric vehicle in the target area; The microscopic analysis result and the macroscopic analysis result are used as the comprehensive analysis result of the target electric vehicle.

8. An electric vehicle travel characteristics analysis device, characterized in that: include: An acquisition module, configured to acquire first information for the vehicle-mounted terminal to perform a communication service and second information for the user terminal to perform a communication service; A first determining module, configured to determine electric vehicle information corresponding to the first information based on the first information; A second determining module, configured to determine, based on the second information, information about the electric vehicle owner corresponding to the second information; A first processing module is used to match and connect the electric vehicle information and the electric vehicle owner information to obtain a vehicle identification result; The second processing module is used to analyze the travel behavior characteristics of the target electric vehicle according to the vehicle identification result.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 7.

Images