A model training and energy supplement intention recognition method, device, equipment and medium

Abstract

The embodiment of the application discloses a model training and energy supplement intention recognition method, device, equipment and medium. The model training method comprises the following steps: extracting sample energy supplement reference data from sample vehicle data; determining energy supplement behavior data according to the sample energy supplement reference data; constructing portrait data of the sample vehicle according to the energy supplement behavior data; and training an energy supplement intention recognition model according to the energy supplement behavior data and the portrait data. The technical scheme of the embodiment of the application adopts two-stage data screening operations, that is, preliminarily screening energy supplement related data and secondarily screening energy supplement behavior data corresponding to an energy supplement moment, thereby improving the accuracy and efficiency of vehicle energy supplement behavior data extraction. The portrait data of the sample vehicle and the energy supplement behavior data are used as model training data together, so that the model training data is more comprehensive, and the accuracy and generalization of the energy supplement intention recognition model training are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of intelligent transportation technology, and in particular to a method, apparatus, device and medium for model training and power replenishment intention recognition. Background Technology

[0002] With the rapid development of intelligent transportation technology, the demand for more segmented intelligent service scenarios and predictive proactive services is increasing. For example, by predicting a user's intention to recharge, services such as recharge reminders or recharge service recommendations can be provided.

[0003] Currently, existing technologies typically set rules based on the user's vehicle's remaining fuel or battery charge, and when a set threshold is triggered, it is considered that there is an intention to refuel. However, this method has low predictive accuracy and urgently needs improvement. Summary of the Invention

[0004] This invention provides a method, apparatus, device, and medium for model training and energy replenishment intention recognition, which addresses the deficiencies in the prior art and improves the accuracy of prediction.

[0005] In a first aspect, embodiments of the present invention provide a model training method, comprising:

[0006] Extract sample energy replenishment reference data from sample vehicle data;

[0007] Based on the sample energy replenishment reference data, determine the energy replenishment behavior data;

[0008] Based on the refueling behavior data, construct a profile of the sample vehicles;

[0009] A power replenishment intention recognition model is trained based on power replenishment behavior data and profile data.

[0010] Secondly, embodiments of the present invention also provide a method for recognizing a power replenishment intention, comprising:

[0011] Obtain current energy replenishment reference data from current vehicle data;

[0012] Input the current energy replenishment reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment;

[0013] The energy replenishment intention recognition model is trained using any of the methods described in this embodiment of the invention.

[0014] Thirdly, embodiments of the present invention also provide a model training apparatus, comprising:

[0015] The sample data extraction module is used to extract sample energy replenishment reference data from sample vehicle data;

[0016] The behavior data determination module is used to determine the energy replenishment behavior data based on the sample energy replenishment reference data;

[0017] The profile data construction module is used to construct profile data for sample vehicles based on the refueling behavior data;

[0018] The model training module is used to train a power replenishment intention recognition model based on power replenishment behavior data and profile data.

[0019] Fourthly, embodiments of the present invention also provide a power replenishment intention recognition device, comprising:

[0020] The current data acquisition module is used to obtain the current energy replenishment reference data from the current vehicle data;

[0021] The energy replenishment intention recognition module is used to input the current energy replenishment reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

[0022] The energy replenishment intention recognition model is trained through the energy replenishment intention recognition model training module.

[0023] Fifthly, embodiments of the present invention also provide an electronic device, comprising:

[0024] One or more processors;

[0025] Memory, used to store one or more programs;

[0026] When one or more programs are executed by one or more processors, the one or more processors implement the model training or power replenishment intention recognition method of any embodiment of the present invention.

[0027] In a sixth aspect, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the model training or power replenishment intention recognition method of any embodiment of the present invention.

[0028] The technical solution of this invention extracts sample charging reference data from sample vehicle data; determines charging behavior data based on the sample charging reference data; constructs a profile of the sample vehicle based on the charging behavior data; and trains a charging intent recognition model based on the charging behavior data and the profile data. This technical solution employs a two-stage data filtering operation: initial filtering of charging-related data and secondary filtering of charging behavior data corresponding to the charging time, improving the accuracy and efficiency of vehicle charging behavior data extraction. Using both the sample vehicle profile data and the charging behavior data as model training data makes the model training data more comprehensive, thereby improving the accuracy and generalization of the charging intent recognition model training. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 This is a flowchart of a model training method provided in Embodiment 1 of the present invention.

[0031] Figure 2 This is a schematic diagram of the energy replenishment intention data selection principle provided in Embodiment 1 of the present invention.

[0032] Figure 3 This is a flowchart of a model training method provided in Embodiment 2 of the present invention.

[0033] Figure 4 This is a flowchart of a model training method provided in Embodiment 3 of the present invention.

[0034] Figure 5 This is a statistical distribution diagram of refueling time intervals provided in Embodiment 3 of the present invention.

[0035] Figure 6 This is a statistical distribution chart of the amount of fuel dispensed in a single refueling operation, provided in Embodiment 3 of the present invention.

[0036] Figure 7 This is a flowchart of a method for identifying energy replenishment intentions provided in Embodiment 4 of the present invention.

[0037] Figure 8 This is a schematic diagram of the training and application scenario of a power replenishment intention recognition model provided in Embodiment 5 of the present invention.

[0038] Figure 9 This is a system framework diagram for energy replenishment intention recognition provided in Embodiment 5 of the present invention.

[0039] Figure 10 This is a schematic diagram of the structure of a model training device provided in Embodiment Six of the present invention.

[0040] Figure 11 This is a schematic diagram of a power replenishment intention recognition device provided in Embodiment 7 of the present invention.

[0041] Figure 12 This is a schematic diagram of the structure of an electronic device provided in Embodiment 8 of the present invention. Detailed Implementation

[0042] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0043] Example 1

[0044] Figure 1 This is a flowchart of a model training method provided in Embodiment 1 of the present invention. This embodiment is applicable to the training of vehicle refueling intention models. The method can be executed by a model training device, which can be implemented in hardware and / or software and can be integrated into an electronic device with model training capabilities.

[0045] like Figure 1 As shown, the method includes:

[0046] S110. Extract sample energy replenishment reference data from sample vehicle data.

[0047] The sample vehicles can be vehicles that provide data for model training. To ensure the accuracy of model training, the number of sample vehicles in this embodiment is usually multiple.

[0048] According to any embodiment of the present invention, the model training method includes, in any case, sample vehicle data including: vehicle terminal log data and vehicle sensor data.

[0049] The vehicle terminal log data can be embedded log data generated by user behavior within the vehicle infotainment system. For example, vehicle terminal log data may include recorded data such as user navigation searches or screen clicks. Vehicle terminal log data can also be referred to as color data. "Vehicle infotainment system" can be a shorthand for in-vehicle infotainment products installed in vehicles, which functionally enable information communication between people and vehicles, between vehicles and the outside world, or between vehicles.

[0050] Vehicle sensor data can be data generated by various vehicle sensors. For example, vehicle sensor data may include data such as vehicle speed or remaining fuel level. Vehicle sensor data can also be referred to as gray data.

[0051] By using vehicle terminal log data and vehicle sensor data, preliminary data preparation can be done to determine vehicle charging-related data, determine vehicle charging behavior, and identify vehicle charging intentions.

[0052] The sample refueling reference data can be refueling-related data from the sample vehicle data. Optionally, the sample refueling reference data may include, but is not limited to, at least one of the following: engine status data, motor status data, vehicle speed data, vehicle energy data, vehicle mileage data, vehicle information data (e.g., vehicle identification number), time data, and location data (e.g., latitude and longitude information). Specifically, engine status data may be data indicating whether the engine is started or stopped; motor status data may be data indicating whether the motor is powered on or off; vehicle speed data may include current vehicle speed and speed changes; vehicle energy data may include remaining energy and percentage of remaining energy; and vehicle mileage data may include the vehicle's total mileage or the remaining mileage that the vehicle can travel.

[0053] It should be noted that the sample energy replenishment reference data in this embodiment includes both the data during energy replenishment and the data without energy replenishment.

[0054] Extracting sample refueling reference data from sample vehicle data can be a process of initially filtering out refueling-related data from all sample vehicle data. For example, relevant script code can be called to read fields from a database to extract the sample refueling reference data.

[0055] By extracting reference data for energy replenishment from the sample vehicle data, the sample vehicle data was initially screened, removing data unrelated to energy replenishment and initially identifying data related to energy replenishment. This laid the groundwork for the subsequent rapid and accurate determination of energy replenishment behavior data.

[0056] S120. Determine the energy replenishment behavior data based on the sample energy replenishment reference data.

[0057] The charging behavior data can be data corresponding to the charging behavior at specific times, selected from sample charging reference data. Optionally, the charging behavior data can be a set of sample charging reference data combined with the charging behavior at specific times; the charging behavior data can be mapped one-to-one by determining the charging behavior at specific times. For example, the charging behavior data may include: "Vehicle ID", "Engine Status", "Remaining Fuel Level", "Remaining Fuel Percentage", "Mileage that the Vehicle Can Travel with Remaining Fuel Level", "Total Mileage", "Current Speed", "Season of the Day", "Month of the Day", "Day of the Week", "Time of Day", "Hour of the Day", "Minute of the Day", and "Whether it is a Special Holiday or Public Holiday".

[0058] Based on the above technical solution, preferably, the energy replenishment behavior data is determined according to the sample energy replenishment reference data, including: directly locating the energy replenishment time according to the sample energy replenishment reference data, and then determining the energy replenishment behavior data.

[0059] The refueling time can be any time when the vehicle engages in refueling activities, for example, when it engages in refueling or charging activities.

[0060] Optionally, the refueling time can be extracted from changes in vehicle energy data, both during and before / after the data changes, to determine the refueling behavior data corresponding to each refueling time. For example, when a sample vehicle refuels, the vehicle's remaining energy, remaining energy percentage, and remaining driving range in the sample refueling reference data will increase. Changes in the sample refueling reference data can be used to determine if a refueling time has occurred, and thus, to determine the corresponding refueling behavior data.

[0061] Alternatively, the vehicle location information can be used to extract the charging time corresponding to the data of the vehicle at the charging location, and then determine the charging behavior data corresponding to the charging time.

[0062] Optionally, the refueling time corresponding to the vehicle being at a refueling location and the vehicle energy data changing can be extracted from vehicle energy data changes and vehicle location information, thereby determining the refueling behavior data corresponding to the refueling time.

[0063] Based on the above technical solution, preferably, the energy replenishment behavior data is determined according to the sample energy replenishment reference data, including: extracting energy replenishment intention data related to energy replenishment intention from the sample energy replenishment reference data; and extracting energy replenishment behavior data related to energy replenishment behavior from the energy replenishment intention data.

[0064] The energy replenishment intention data can be sample energy replenishment reference data corresponding to the moment containing the energy replenishment intention. Optionally, the energy replenishment intention data can correspond to at least one moment containing the energy replenishment intention. For example, Figure 2 This is a schematic diagram illustrating the principle of energy replenishment intention data selection provided in Embodiment 1 of the present invention. Figure 2 As shown, the energy replenishment behavior is set to "refuel," the horizontal axis represents time, and the circular dots represent the time points when the user expresses the intention to refuel. Assume that between "intention t" and "refueling t," the user maintains the intention to "refuel," and the energy replenishment reference data within this interval is considered positive samples; between "refueling t-1" and "intention t," the user maintains the intention not to refuel, and the energy replenishment reference data within this interval is considered negative samples. Energy replenishment intention data can be the energy replenishment reference data corresponding to the time between "intention t" and "refueling t," i.e., positive sample data.

[0065] Extracting charging intention data related to the charging intention from the sample charging reference data can be achieved by determining the point in time when the charging intention arises or the operation performed when the charging intention arises, thus defining the charging intention data between the generation of the charging intention and the charging action. Optionally, determining the point in time when the charging intention arises can be the midpoint between two charging actions or other time values ​​within the two charging actions. The operation performed when the charging intention arises can be related to going to a gas station or a battery swapping station. Once the point in time when the charging intention arises is determined, the interval of the charging intention action data can be identified, and thus the charging intention action data can be extracted.

[0066] Extracting data related to refueling behavior from refueling intention data can be a process of determining whether a vehicle is engaged in refueling behavior by analyzing changes in the data of sample vehicles. For example, when the remaining fuel level of a gasoline vehicle increases and the vehicle is turned off, it can be considered that the gasoline vehicle has engaged in refueling behavior.

[0067] Based on the above scheme, the preferred approach further filters the sample energy replenishment reference data by extracting energy replenishment intention data related to the energy replenishment intention. Simultaneously, it transforms the user's energy replenishment intention into obtainable energy replenishment intention data, making the user's energy replenishment intention visible. By extracting energy replenishment behavior data related to the energy replenishment behavior from the energy replenishment intention data, the obtained energy replenishment behavior data becomes more accurate through further filtering of the energy replenishment intention data.

[0068] S130. Based on the refueling behavior data, construct a profile of the sample vehicles.

[0069] The profile data can be generated from the behavioral data of the sample vehicles, producing label data describing the sample vehicles. Optionally, the profile data may include at least one of the following: parking location, refueling location, driving route, and energy data. For example, the profile data may include data such as "distance from the last gas station" and "distance from frequently visited gas stations".

[0070] Based on the above technical solution, preferably, profile data is constructed based on the refueling behavior data, including: determining the refueling location based on the refueling behavior data; and constructing profile data of the sample vehicle based on the refueling location.

[0071] The locations for energy replenishment can include gas stations, charging stations, charging piles, or battery swapping stations.

[0072] Determining the refueling location based on refueling behavior data can be achieved by filtering the latitude and longitude coordinates of the refueling location from the data. For example, a refueling and charging behavior discovery algorithm can be used to identify the latitude and longitude coordinates of each refueling and charging session of a sample vehicle from the refueling behavior data.

[0073] Based on the refueling locations, a profile of the sample vehicle can be constructed. This can be done by determining the frequently visited refueling locations and / or the last refueling location visited by the sample vehicle from refueling behavior data using the latitude and longitude coordinates of the refueling locations. For example, based on the latitude and longitude coordinates of refueling and charging, a "reverse geospatial analysis" service can be used to obtain gas stations or charging stations near the current coordinate point, and the results can be stored in a database to obtain the location and name of each refueling and charging visit of the sample vehicle. Among them, the refueling location that appears most frequently is the "frequently visited gas station"; the most recent refueling location is the "last visited gas station".

[0074] By determining the refueling locations and constructing a profile of the sample vehicles based on these locations, we can identify the frequently visited or last visited refueling locations of the sample vehicles. Furthermore, we can determine the distance of the sample vehicles from the refueling locations based on these locations, making the auxiliary judgment of the sample vehicles' refueling intentions more targeted.

[0075] By constructing profile data of sample vehicles, the data on the refueling behavior of sample vehicles is made more specific, and the data on the refueling intention of sample vehicles is more comprehensive.

[0076] S140. Based on the power replenishment behavior data and profile data, train a power replenishment intention recognition model.

[0077] The charging intention recognition model is a model that performs the task of recognizing charging intentions based on charging behavior data and profile data. Specifically, by inputting the charging behavior data and profile data into the charging intention recognition model, the model can provide a result indicating whether the sample vehicle has the intention to charge.

[0078] It should be noted that the charging behavior data and profile data in this embodiment correspond to the data of multiple sample vehicles. For each sample vehicle, a mapping relationship can be established between the vehicle ID and the charging behavior data and profile data of that vehicle. Subsequently, the corresponding charging behavior data and profile data of each vehicle ID can be used as a set of training data and input into the charging intention recognition model to train the model.

[0079] The process of training the charging intention recognition model can be as follows: input charging behavior data and profile data into the charging intention recognition model. The charging intention recognition model can then predict whether the data set contains a charging intention through an intention prediction algorithm. The prediction results of the charging intention recognition model are compared with the actual charging intention results of the sample vehicles. The loss function is calculated, and then the parameter configuration of the model is adjusted based on the loss function. By iterating the charging intention recognition model multiple times according to the above scheme, a well-trained charging intention recognition model can be obtained.

[0080] The intent prediction algorithm can be Linear Support Vector Classification (LinearSVC), Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), or Gradient Boosting Decision Tree (GBDT), etc. Preferably, the GBDT algorithm can be selected.

[0081] Optionally, the battery replenishment intention recognition model can be further processed before training the battery replenishment behavior data and profile data. For example, further processing based on feature engineering, including feature extraction, aggregation, and formatting, can be used to improve the saliency of the model's feature recognition.

[0082] For example, assuming there are 39,000 data points on power replenishment behavior, firstly, the 39,000 data points are randomly shuffled, and the random number seed can be set to 42. Then, the training set, validation set, and test set are selected according to the ratio of "training set: validation set: test set" equal to "7:1:2". Then, the power replenishment intention recognition model is trained using the training set and validated using the validation set. The power replenishment intention recognition model is then continuously adjusted as needed to obtain the optimized power replenishment intention recognition model. Finally, the final power replenishment intention recognition model is evaluated using the test set. If the evaluation passes, the training process of the power replenishment intention recognition model is complete.

[0083] The technical solution of this invention extracts sample charging reference data from sample vehicle data; determines charging behavior data based on the sample charging reference data; constructs a profile of the sample vehicle based on the charging behavior data; and trains a charging intent recognition model based on the charging behavior data and the profile data. This technical solution employs a two-stage data filtering operation: initial filtering of charging-related data and secondary filtering of charging behavior data corresponding to the charging time, improving the accuracy and efficiency of vehicle charging behavior data extraction. Using both the sample vehicle profile data and the charging behavior data as model training data makes the model training data more comprehensive, thereby improving the accuracy and generalization of the charging intent recognition model training.

[0084] Example 2

[0085] Figure 3 This is a flowchart of a model training method provided in Embodiment 2 of the present invention. Based on the above embodiments, this embodiment further optimizes the method by extracting energy supplementation intention data related to the energy supplementation intention from the sample energy supplementation reference data. For example... Figure 3 As shown, the method includes:

[0086] S310. Extract sample energy replenishment reference data from sample vehicle data.

[0087] S320. Determine the recharge time based on the vehicle sensor data in the sample recharge reference data.

[0088] Determining the refueling time based on vehicle sensor data in the sample refueling reference data can be achieved by identifying refueling times when refueling activities occur using the vehicle sensor data in the sample refueling reference data. For example, the refueling time of vehicle refueling / charging can be identified using data such as "remaining fuel / battery level" and "current speed" from the vehicle sensor data.

[0089] By determining the recharge time based on the vehicle sensor data in the sample recharge reference data, and mapping the vehicle sensor data to the recharge time, the corresponding vehicle sensor data can be found by confirming the recharge time.

[0090] S330. Determine the intended energy replenishment period based on the energy replenishment time and reference driving duration.

[0091] The "power replenishment intention period" can be any time period during which a power replenishment intention is expressed. Specifically, it can be the time period from the generation of the power replenishment intention to the execution of the power replenishment action.

[0092] Based on the above technical solution, after determining the energy replenishment time, the model training method also includes:

[0093] If there are at least two refueling opportunities, the average driving time is determined based on the at least two refueling opportunities; the reference driving time is determined based on the average driving time and the preset interval percentage.

[0094] The average driving time can be the average of the time intervals between adjacent refueling times. Optionally, different sample vehicles may have different refueling frequencies, and thus different average driving times. For example, if refueling times are set to t1, t2, t3, and t4, and the time intervals between adjacent refueling times are (t2-t1), (t3-t2), and (t4-t3), then the average driving time is the average of these time intervals.

[0095] The preset interval percentage can be a percentage of the driving time interval from the generation of the refueling intention until the vehicle begins to perform the refueling action. Optionally, the preset interval percentage can be a value based on empirical assumptions.

[0096] The reference driving time can be the driving time interval from the generation of the refueling intention until the vehicle begins to perform the refueling behavior. Optionally, the reference driving time can be a value assumed through experience, or it can be a fixed time interval. Optionally, the reference driving time can also be a value determined by a preset interval percentage and an average driving time. For example, the average driving time is set to t, and the preset interval percentage is... Reference driving time is

[0097] Preferably, determining the reference driving time by using a preset interval percentage and average driving time allows for the statistical analysis of the average driving time of different sample vehicles, making the average driving time data more accurate. Simultaneously, it also makes the reference driving time data more accurate. Statistically obtaining the reference driving time based on data from different sample vehicles makes the resulting reference driving time data more targeted.

[0098] Determining the charging intention period can be achieved by setting the charging time as the point in time when the charging action is performed, and setting the point in time corresponding to the reference driving time back from the charging time as the point in time when the charging intention is generated. The time period between these two points in time is the charging intention period. For example, let the charging time be T, the reference driving time be ΔT, the point in time when the charging intention is generated be (T-ΔT), and the charging intention period be the time period from (T-ΔT) to T.

[0099] By determining the time period for the intended refueling based on the refueling time and the reference driving duration, the time period for the intended refueling is specified, which facilitates data extraction.

[0100] S340. Extract the sample energy replenishment reference data located in the energy replenishment intention period as energy replenishment intention data.

[0101] The sample refueling reference data includes vehicle information and time data. Based on the vehicle information and time data, the corresponding sample refueling reference data for a given vehicle can be found within the intended refueling time period. This sample refueling reference data can then serve as the refueling intention data. For example, the sample refueling data for a single sample vehicle at a single moment can be compiled into a dataset containing the sample refueling reference data for that vehicle at that moment. The corresponding sample refueling reference data can be found by confirming the vehicle ID and the corresponding time.

[0102] By extracting sample energy replenishment reference data during the energy replenishment intention period as energy replenishment intention data, the energy replenishment intention data is extracted, which provides a guarantee for the subsequent rapid and accurate extraction of energy replenishment behavior data.

[0103] S350. Extract energy replenishment behavior data related to energy replenishment behavior from the energy replenishment intention data.

[0104] S360. Based on the refueling behavior data, construct a profile of the sample vehicles.

[0105] S370. Based on the power replenishment behavior data and profile data, train a power replenishment intention recognition model.

[0106] The technical solution of this invention involves extracting sample charging reference data from sample vehicle data; determining the charging time based on vehicle sensor data within the sample charging reference data; determining the charging intention period based on the charging time and reference driving duration; extracting sample charging reference data within the charging intention period as charging intention data; extracting charging behavior data related to the charging behavior from the charging intention data; constructing a profile of the sample vehicle based on the charging behavior data; and training a charging intention recognition model based on the charging behavior data and the profile data. By using the charging time and reference driving duration, the charging intention period is confirmed temporally, and the charging intention data is confirmed and extracted through the charging intention period, thus improving the accuracy of charging intention recognition.

[0107] Optionally, based on the above embodiments, another possible method for extracting energy replenishment intention data related to energy replenishment intention from sample energy replenishment reference data may include the following steps:

[0108] Step A: Filter keywords related to the intention to recharge from the vehicle terminal log data in the sample recharge reference data.

[0109] The keywords related to the intention to replenish energy can be keywords related to the energy replenishment behavior. For example, they could be information about the energy replenishment location, such as the specific address or name of the battery swapping station. Alternatively, they could be general terms for energy replenishment locations, such as battery swapping station, charging station, or gas station.

[0110] Filtering keywords related to the intention to recharge from vehicle terminal log data in the sample recharge reference data can be the process of selecting keywords related to the intention to recharge from the search records in the vehicle terminal log data. For example, one can view the historical search records in the vehicle terminal log data and look for keywords corresponding to "gas station," "battery swapping station," or the address of a gas station or battery swapping station; these can be used to filter keywords related to the intention to recharge.

[0111] Step B: From the vehicle sensor data in the sample recharge reference data, filter the associated sensor data of recharge intent keywords.

[0112] Among them, the associated sensor data can be related data of vehicle sensor data containing the keyword of refueling intention. For example, if the keyword of refueling intention is set as "gas station", the vehicle sensor data contains the location data of the gas station. The other vehicle sensor data corresponding to this location data, such as engine status data, vehicle speed data, vehicle energy data, vehicle mileage data, vehicle information data and time data, are all associated sensor data.

[0113] Step C: From the vehicle sensor data in the sample recharge reference data, filter the associated sensor data of the recharge intention keyword. This can be done by identifying the field of the recharge intention keyword in the vehicle sensor data, and then confirming the associated sensor data by identifying the corresponding vehicle sensor data.

[0114] Based on the keywords of the power replenishment intention and the associated sensor data, the power replenishment intention data is determined.

[0115] Specifically, the replenishment intention data is determined based on the replenishment intention keywords and associated sensor data. This can be achieved by identifying the vehicle information data or time data corresponding to the keywords in the vehicle terminal log data through the replenishment intention keywords, and then finding the associated sensor data corresponding to the vehicle information data or time data from the associated sensor data. This determines the replenishment intention data at the time the replenishment intention is generated. Combined with the replenishment behavior, the replenishment intention data can be determined.

[0116] The advantage of this configuration in this embodiment is that by associating the keywords of the power replenishment intention in the vehicle terminal log data with the vehicle sensor data, the time when the power replenishment intention is generated can be determined more accurately, thus improving the accuracy of the power replenishment intention determination.

[0117] Example 3

[0118] Figure 4 This is a flowchart of a model training method provided in Embodiment 3 of the present invention. Based on the above embodiments, this embodiment further optimizes the method by extracting energy supplementation behavior data related to energy supplementation behavior from the energy supplementation intention data. For example... Figure 4 As shown, the method includes:

[0119] S410. Extract sample refueling reference data from sample vehicle data.

[0120] S420. Extract energy replenishment intention data related to energy replenishment intention from the sample energy replenishment reference data.

[0121] S430: Iterate through the energy replenishment intention data at each time step, and determine whether the stack push condition is met at that time step. If yes, execute S440; otherwise, return to execute S430.

[0122] The conditions for pushing data onto the stack can be conditions that satisfy the energy replenishment intention data. Optionally, when the vehicle is a gasoline-powered vehicle, the conditions for pushing data onto the stack can be engine status data, vehicle speed data, and event type; when the vehicle is an electric vehicle, the conditions for pushing data onto the stack can be motor status data, vehicle speed data, and event type. The event type can be "vehicle off" or "vehicle started"; the event type can also be "vehicle powered down" or "vehicle powered on". For example, assuming the vehicle is a gasoline-powered vehicle, the conditions for pushing data onto the stack could be that the vehicle speed is 0, the engine is off, and the event type is "vehicle off," which might be a scenario where the vehicle is stopped to prepare for energy replenishment; the conditions for pushing data onto the stack could also be that the vehicle speed is 0, the engine is started, and the event type is "vehicle started," which might be a scenario where the vehicle leaves the gas station after energy replenishment.

[0123] This embodiment can sequentially determine whether the stacking condition is met for the energy replenishment intention data at each time moment. If it is met, the subsequent operation S440 is executed. If it is not met, the next energy replenishment intention data is executed, and this step continues.

[0124] S440. If satisfied, push the energy replenishment intention data at that moment onto the stack, and determine whether the number of data entries in the stack meets the preset number. If yes, execute S450; otherwise, return to execute S430.

[0125] The number of data entries in the stack can be the number of power replenishment intent data entries already pushed onto the stack. The number of data entries in the stack cannot exceed a preset number. For example, if the preset number is set to 2, and after power replenishment intent data is pushed onto the stack, the process returns to operation S430 to continue retrieving power replenishment intent data that meets the push conditions. If after power replenishment intent data is pushed onto the stack, the process stops pushing data onto the stack and the subsequent operation S450 is executed.

[0126] S450. If satisfied, determine whether there is an energy growth event based on the energy replenishment intention data in the stack. If yes, execute S460. If no, clear the energy replenishment intention data in the stack and return to execute S430.

[0127] Among them, energy growth events are events in which vehicle energy data increases. Optionally, this can be an increase in remaining fuel or an increase in remaining battery power.

[0128] To determine whether an energy growth event exists, the energy replenishment intention data in the stack is further filtered to confirm whether energy replenishment behavior occurs during the process of vehicle status data changes, thus further ensuring the accuracy of energy replenishment behavior acquisition.

[0129] Based on the above technical solution, preferably, determining whether an energy growth event exists based on the energy replenishment intention data in the stack may include: determining the energy replenishment amount and / or energy replenishment time interval based on the energy replenishment intention data in the stack; and determining whether an energy growth event exists based on the energy replenishment amount and / or energy replenishment time interval.

[0130] The increase in energy replenishment can be the change in vehicle energy obtained by comparing the energy replenishment intention data in the stack. For example, it could be the increase in remaining fuel or the increase in remaining battery power.

[0131] The energy replenishment interval can be the time interval data obtained by comparing the energy replenishment intention data in the stack. For example, if there are two data items in the stack, the energy replenishment interval can be the time interval between the two data items.

[0132] For example, Table 1 below is a historical data statistics table of in-stack refueling intention data. As shown in Table 1, the time from "2020-12-01 18:16:49" to "2020-12-01 18:19:19" corresponds to refueling event 1, which is a normal refueling event with a stable change in the current fuel level. The time "2020-12-01 18:19:59" corresponds to refueling event 2, which is an abnormal refueling event with a brief fluctuation in the current fuel level. Due to limitations in sensor accuracy, the refueling data sequence has brief fluctuations. To ensure data accuracy, the refueling data needs to be cleaned.

[0133] According to the technical solution of the present invention, the energy replenishment intention data in the stack is used to determine the energy replenishment growth amount and / or the energy replenishment time interval; based on the energy replenishment growth amount and / or the energy replenishment time interval, it is determined whether the energy replenishment growth amount reaches the growth amount threshold and / or the energy replenishment time interval reaches the energy replenishment time interval threshold, thereby determining whether there is an energy growth event, thus realizing the cleaning of the energy replenishment intention data.

[0134] Among them, the growth threshold is the minimum set value for the energy replenishment growth.

[0135] The energy replenishment time interval threshold is the minimum set value for the energy replenishment time interval.

[0136]

[0137] Optionally, the energy replenishment time interval threshold can be adjusted by setting a fluctuation tolerance time window to smooth the data. The specific selection method for the energy replenishment time interval threshold is as follows:

[0138] Calculate the time interval between two refueling operations, selecting the time interval as T1-T. n .

[0139] Draw a cumulative distribution percentage chart: determine the time interval in which the cumulative percentage reaches 90% or more;

[0140] This time interval is selected as the energy replenishment time interval threshold.

[0141] For example, Figure 5 This is a statistical distribution diagram of refueling time intervals provided in Embodiment 3 of the present invention.

[0142] like Figure 5 As shown, the refueling interval is set as the refueling interval, the horizontal axis represents the time interval between two refuelings, and the vertical axis represents the cumulative percentage of normal data. The selection method for the refueling interval threshold is as follows:

[0143] Calculate the time interval between two refueling operations, and select a time interval of (10-600s).

[0144] Plotting the cumulative distribution percentage: When the time interval between two refueling sessions is 300 seconds, the cumulative percentage reaches 92%.

[0145] Therefore, a threshold of 300 seconds was selected for the time interval between two refueling operations.

[0146] By setting a threshold for the energy replenishment time interval, energy replenishment intention data with short-term abnormal fluctuations is removed, ensuring the accuracy of the energy replenishment intention reference data.

[0147] Optionally, the selection of the growth threshold and the data processing procedure are as follows:

[0148] For calculating vehicle energy data for energy replenishment behavior, only retain vehicle energy data greater than 1.

[0149] Plot a frequency graph of vehicle energy data: Determine the minimum value of vehicle energy data that accounts for an abnormally high proportion;

[0150] The vehicle's energy data represents the threshold for the increase.

[0151] For example, Figure 6 This is a statistical distribution chart of the amount of fuel dispensed in a single refueling operation, provided in Embodiment 3 of the present invention.

[0152] like Figure 6As shown, the energy replenishment increase is defined as the refueling amount of a refueling event, the horizontal axis represents the refueling amount of each refueling event, and the vertical axis represents the frequency of refueling amounts. The process of cleaning outliers in a single refueling event involves selecting a threshold for the single refueling amount. The data processing procedure is as follows:

[0153] Calculate the amount of fuel added in each refueling event, and only retain fuel amounts greater than 1L.

[0154] A frequency chart of refueling volume was drawn: the proportion of refueling volumes of 2L was unusually high.

[0155] Therefore, the effective refueling volume threshold is selected as >2L.

[0156] By setting a threshold for the amount of growth, interference from abnormal data from a single energy replenishment is avoided, further ensuring the accuracy of the data.

[0157] S460. If it exists, use the power replenishment intention data in the stack as power replenishment behavior data, and clear the power replenishment intention data in the stack.

[0158] Optionally, in this embodiment, the energy replenishment intention data in the stack can be extracted and used as energy replenishment behavior data.

[0159] Clearing the energy replenishment intention data in the stack can be the process of clearing the energy replenishment intention data in the stack, so that the number of energy replenishment intention data in the stack becomes 0.

[0160] By using the energy replenishment intention data in the stack as energy replenishment behavior data and clearing the energy replenishment intention data in the stack, the energy replenishment behavior data is determined, and the extraction of energy replenishment behavior data is completed. At the same time, clearing the energy replenishment intention data in the stack provides conditions for the subsequent judgment of energy replenishment intention data, and realizes the cyclic judgment of data.

[0161] S470. Based on the refueling behavior data, construct a profile of the sample vehicles.

[0162] S480. Based on the power replenishment behavior data and profile data, train a power replenishment intention recognition model.

[0163] The technical solution of this invention extracts sample energy replenishment reference data from sample vehicle data; extracts energy replenishment intention data related to energy replenishment intention from the sample energy replenishment reference data; sequentially traverses the energy replenishment intention data at each time point, determining whether the stacking condition is met at that time point; if met, the energy replenishment intention data at that time point is pushed onto the stack, and the number of data entries in the stack is determined to be within a preset limit; if met, the existence of an energy increase event is determined based on the energy replenishment intention data in the stack; if present, the energy replenishment intention data in the stack is used as energy replenishment behavior data, and the stack is cleared; a profile of the sample vehicle is constructed based on the energy replenishment behavior data; and a energy replenishment intention recognition model is trained based on the energy replenishment behavior data and the profile data. This step-by-step filtering of the energy replenishment intention data enables accurate judgment of the energy replenishment behavior data. Simultaneously, clearing the energy replenishment reference data in the stack enables cyclical judgment of the data.

[0164] Example 4

[0165] Figure 7 This is a flowchart of a charging intention recognition method provided in Embodiment 4 of the present invention. This embodiment can be applied to the case of vehicle charging intention recognition. The method can be executed by a charging intention recognition device, which can be implemented in hardware and / or software. The charging intention recognition device can be integrated into an electronic device configured with a charging intention recognition model, and the electronic device can be an in-vehicle terminal.

[0166] like Figure 7 As shown, the method includes:

[0167] S710: Obtain current energy replenishment reference data from current vehicle data.

[0168] Among them, the current vehicle data can be all the real-time data generated by the current vehicle at the current moment.

[0169] According to any embodiment of the present invention, the method for identifying a recharge intention includes current vehicle data including: vehicle terminal log data and vehicle sensor data.

[0170] By mapping the vehicle terminal log data and vehicle sensor data one by one, preparations are made for the power replenishment intention recognition process.

[0171] The current energy replenishment reference data can be all related data related to energy replenishment in the current vehicle data.

[0172] Obtaining the current energy replenishment reference data from the current vehicle data is similar to the process of extracting sample energy replenishment reference data from sample vehicle data as described in the above embodiments, and will not be repeated here.

[0173] By obtaining current energy replenishment reference data from current vehicle data, preliminary data filtering was achieved, removing data that was irrelevant to energy replenishment.

[0174] S720. Input the current energy replenishment reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

[0175] The energy replenishment intention recognition model can be trained using the model training method of any embodiment of the present invention. The energy replenishment intention recognition model is used to analyze the input energy replenishment reference data and provide a result indicating whether an energy replenishment intention exists at the current time.

[0176] The result of the charging intention recognition can be a determination of whether the current vehicle intends to charge. For example, it may or may not intend to charge.

[0177] Optionally, the data can be further processed before being input into the energy replenishment intent recognition model. For example, further processing based on feature engineering, including feature extraction, aggregation, and formatting, can be used to improve the saliency of the model's feature recognition.

[0178] The refueling intention recognition model solves the problem of not being able to predict a vehicle's refueling intention in advance, and realizes the recognition and prediction of the current vehicle's refueling intention based on the current refueling reference data.

[0179] Based on the above technical solution, preferably, after obtaining the current energy replenishment reference data from the current vehicle data, the energy replenishment intention recognition method further includes:

[0180] Obtain supplementary reference data for the current energy replenishment reference data from historical energy replenishment reference data.

[0181] The historical charging reference data can be all relevant charging data from the current vehicle's historical data. Optionally, the historical charging reference data is the vehicle's historical charging experience value, which records relevant data from the vehicle's historical charging records.

[0182] Supplementary reference data can be historical energy replenishment reference data that is correlated with historical energy replenishment reference data and current energy replenishment reference data, or it can be the historical energy replenishment reference data of the current vehicle directly used as energy replenishment reference data.

[0183] Accordingly, the current power replenishment reference data is input into the power replenishment intention recognition model to obtain the power replenishment intention recognition result at the current moment, including:

[0184] Input the current energy replenishment reference data and supplementary reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

[0185] Specifically, inputting the current energy replenishment reference data and supplementary reference data into the energy replenishment intention recognition model can be achieved by inputting the current energy replenishment reference data and supplementary reference data as a whole into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

[0186] By acquiring supplementary reference data, the current energy replenishment reference data is supplemented from historical data, providing more sufficient data support for subsequent predictions. The current energy replenishment reference data and supplementary reference data are input into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment. The input data is more comprehensive, ensuring the accuracy of the prediction.

[0187] Based on the above technical solution, preferably, the power replenishment intention recognition method inputs the current power replenishment reference data into the power replenishment intention recognition model to obtain the power replenishment intention recognition result at the current moment, including:

[0188] Obtain the current vehicle's profile data.

[0189] The current vehicle profile data can be generated in advance using the current vehicle's historical behavior data to produce label data describing the current vehicle.

[0190] The current vehicle profile data and current charging reference data are input into the charging intention recognition model to obtain the charging intention recognition result at the current moment.

[0191] The current vehicle profile data and current refueling reference data can be input into the refueling intention recognition model as a data set. The model can then calculate and determine whether there is a refueling intention at the current moment.

[0192] By acquiring profile data, the input data for the energy replenishment intention recognition model becomes more comprehensive, further ensuring the accuracy of energy replenishment intention recognition.

[0193] The technical solution of this invention obtains current charging reference data from current vehicle data; inputs the current charging reference data into a charging intention recognition model to obtain the charging intention recognition result at the current moment. By applying the charging intention recognition model, the function of directly predicting and recognizing charging intentions using current charging reference data is realized, improving the efficiency and accuracy of charging intention recognition.

[0194] Example 5

[0195] Figure 8 This is a schematic diagram illustrating the training and application scenario of a power replenishment intention recognition model provided in Embodiment 5 of the present invention. This embodiment can implement the model training method and power replenishment intention recognition method provided in the above embodiments of the present invention. Figure 8As shown, the model is first trained and the algorithm is repeatedly evaluated / optimized to obtain the intent recognition model. Then, the model is applied to realize the recognition process of power replenishment intent.

[0196] Optionally, the model training process includes: source data acquisition, selection of energy replenishment intention samples, profile extraction, construction of training dataset, modeling, and algorithm evaluation / optimization.

[0197] In this context, source data acquisition is equivalent to the process of extracting sample energy supplementation reference data.

[0198] Selecting energy replenishment intention samples is equivalent to extracting energy replenishment intention data related to energy replenishment intentions.

[0199] Image extraction is equivalent to constructing image data.

[0200] Building the training dataset is equivalent to extracting energy replenishment behavior data and constructing profile data.

[0201] Modeling can be a process of constructing a power replenishment intention recognition model, which is equivalent to training the power replenishment intention recognition model.

[0202] Optionally, the power replenishment intent recognition process includes: real-time data source, profile extraction, feature set construction, model and intent. The power replenishment intent recognition process is the process of recognizing power replenishment intent using the power replenishment intent recognition model.

[0203] The real-time data source is equivalent to obtaining the current energy replenishment reference data from the current vehicle data.

[0204] The profile data is equivalent to obtaining the profile data of the current vehicle.

[0205] Feature set construction is equivalent to obtaining current energy replenishment reference data and current vehicle profile data.

[0206] The model is essentially inputting the current vehicle profile data and current charging reference data into the charging intention recognition model.

[0207] The intent is equivalent to obtaining the current energy replenishment intent recognition result.

[0208] Figure 9 This is a system framework diagram for power replenishment intention recognition provided in Embodiment 5 of the present invention. This system framework diagram can be integrated into an in-vehicle terminal. This embodiment can implement the power replenishment intention recognition method provided in the above embodiments of the present invention. Figure 9 As shown, the system framework consists of a data access and storage foundation layer, a real-time / offline computing engine layer and a technical implementation layer, and a backend service application layer, from bottom to top.

[0209] The data source can be real-time collected vehicle terminal log data and vehicle sensor data. The data source is transmitted via a data bus; for example, the data bus can be a distributed publish / subscribe messaging system. Data storage can be implemented using a database, optionally using multiple databases. For example, the database types can include: a database management system based on a distributed file system, an open-source non-relational distributed database management system, or a relational database management system, etc.

[0210] Real-time computing can be implemented using a distributed processing engine (Flink) designed for both streaming and batch data. Vehicle terminal log data and vehicle sensor data can be processed in real-time using the real-time computing engine, and the processed data can be stored in a database. Offline computing can be performed using an offline computing engine to process vehicle terminal log data and vehicle sensor data, generating vehicle profile data, which can then be stored in a database.

[0211] Feature engineering is used to further process data before it is input into the model, including feature extraction, aggregation, and formatting, to help the model improve the saliency of feature recognition. Recharge recognition can be used to obtain current recharge reference data from current vehicle data. Intent prediction can predict recharge intent based on vehicle terminal log data and sensor data using a recharge intent recognition model and output the recharge intent prediction result.

[0212] This system can also provide corresponding services based on the recognition of refueling intentions. For example, it can recommend suitable promotional activities based on refueling intentions. It can also recommend refueling points of interest (POIs), providing personalized recommendations based on geographic information, public service stations, and information about buildings or service stations that can provide services, such as bus stops. It can also provide refueling reminders; when a user's refueling intention is determined, a refueling reminder can be sent through the vehicle's infotainment system. These refueling reminders can include audio, pop-up windows, or image information reminders.

[0213] By constructing a system framework for power replenishment intent recognition, the application of the power replenishment intent recognition model was realized, and the power replenishment intent recognition service was extended to achieve more personalized performance.

[0214] Example 6

[0215] Figure 10 This is a schematic diagram of a model training device provided in Embodiment Six of the present invention. This device can implement the model training method provided in the above embodiments of the present invention. This embodiment is applicable to the training of vehicle refueling intention models. The device can be implemented by software and / or hardware and can be integrated into an electronic device with model training capabilities. Figure 10As shown, the device includes:

[0216] The sample data extraction module 1010 is used to extract sample energy replenishment reference data from sample vehicle data;

[0217] The behavior data determination module 1020 is used to determine the energy replenishment behavior data based on the sample energy replenishment reference data;

[0218] The profile data construction module 1030 is used to construct profile data of sample vehicles based on the energy replenishment behavior data;

[0219] The model training module 1040 is used to train a power replenishment intention recognition model based on power replenishment behavior data and profile data.

[0220] The technical solution of this invention extracts sample charging reference data from sample vehicle data; determines charging behavior data based on the sample charging reference data; constructs a profile of the sample vehicle based on the charging behavior data; and trains a charging intent recognition model based on the charging behavior data and the profile data. This technical solution employs a two-stage data filtering operation: initial filtering of charging-related data and secondary filtering of charging behavior data corresponding to the charging time, improving the accuracy and efficiency of vehicle charging behavior data extraction. Using both the sample vehicle profile data and the charging behavior data as model training data makes the model training data more comprehensive, thereby improving the accuracy and generalization of the charging intent recognition model training.

[0221] Optionally, the behavior data determination module 1020 includes:

[0222] The intent data extraction unit is used to extract energy replenishment intent data related to energy replenishment intent from the sample energy replenishment reference data;

[0223] The behavior data extraction unit is used to extract energy replenishment behavior data related to energy replenishment behavior from the energy replenishment intention data.

[0224] Optionally, the intent data extraction unit may include:

[0225] The keyword filtering subunit is used to filter keywords related to the intention of recharging from the vehicle terminal log data in the sample recharging reference data;

[0226] The associated data filtering subunit is used to filter associated sensor data with keywords related to the recharge intention from vehicle sensor data in the sample recharge reference data;

[0227] The intent data determination subunit is used to determine the power replenishment intent data based on the power replenishment intent keywords and associated sensor data.

[0228] Optionally, the intent data extraction unit may also include:

[0229] The recharge timing determination subunit is used to determine the recharge timing based on the vehicle sensor data in the sample recharge reference data;

[0230] The intention period determination subunit is used to determine the intention period for recharging based on the recharging time and the reference driving duration;

[0231] The intention data determination sub-unit is used to extract sample energy replenishment reference data located during the energy replenishment intention period, as energy replenishment intention data.

[0232] Furthermore, after determining the energy replenishment time, the energy replenishment time determination sub-unit also includes:

[0233] The average duration is determined by the sub-unit. If there are at least two refueling times, the average driving time is determined based on the at least two refueling times.

[0234] The reference duration determination subunit is used to determine the reference driving duration based on the average driving time and the preset interval percentage.

[0235] Optionally, the energy replenishment behavior data extraction unit includes:

[0236] The condition judgment subunit is used to iterate through the energy replenishment intention data at each time step and determine whether the push condition is met at that time step.

[0237] The data count judgment subunit is used to determine whether the number of data items in the stack meets the preset number after the energy replenishment intention data at that moment is pushed onto the stack if the preset number is met.

[0238] The growth determination subunit is used to determine whether an energy growth event exists based on the energy replenishment intention data in the stack if the conditions are met.

[0239] The data determination sub-unit is used to, if it exists, take the energy replenishment intention data in the stack as the energy replenishment behavior data and clear the energy replenishment intention data in the stack.

[0240] Furthermore, the growth determination subunit can be used for:

[0241] Based on the energy replenishment intention data in the stack, determine the energy replenishment growth amount and / or energy replenishment time interval;

[0242] Determine whether an energy growth event exists based on the amount of energy replenishment and / or the time interval between energy replenishments.

[0243] Optional, the profile data construction module includes:

[0244] The energy replenishment location determination unit is used to determine the energy replenishment location based on energy replenishment behavior data;

[0245] The profile data construction unit is used to construct profile data of sample vehicles based on the refueling location.

[0246] In any model training device of this invention, the sample vehicle data includes: vehicle terminal log data and vehicle sensor data.

[0247] The model training apparatus provided in this embodiment of the invention can execute the model training method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0248] Example 7

[0249] Figure 11 This is a schematic diagram of a charging intention recognition device provided in Embodiment 7 of the present invention. This device can implement the charging intention recognition method provided in the above embodiments of the present invention. This embodiment is applicable to vehicle charging intention recognition. The device can be implemented by software and / or hardware and can be integrated into an electronic device configured with a charging intention recognition model, which can be a vehicle-mounted terminal. Figure 11 As shown, the device includes:

[0250] The current data acquisition module 1110 is used to obtain the current energy replenishment reference data from the current vehicle data;

[0251] The energy replenishment intention recognition module 1120 is used to input the current energy replenishment reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment;

[0252] The energy replenishment intention recognition model is trained through the energy replenishment intention recognition model training module.

[0253] The technical solution of this invention obtains current charging reference data from current vehicle data; inputs the current charging reference data into a charging intention recognition model to obtain the charging intention recognition result at the current moment. By applying the charging intention recognition model, the function of directly predicting and recognizing charging intentions using current charging reference data is realized, improving the efficiency and accuracy of charging intention recognition.

[0254] Accordingly, after the current energy replenishment data acquisition module 1110 obtains the current energy replenishment reference data from the current vehicle data, it also includes:

[0255] The supplementary data acquisition module is used to obtain supplementary reference data for the current energy replenishment reference data from historical energy replenishment reference data.

[0256] Accordingly, the power replenishment intention recognition module 1120 includes:

[0257] Input the current energy replenishment reference data and supplementary reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

[0258] Accordingly, the current data acquisition module 1110 includes:

[0259] The profile data acquisition unit is used to acquire profile data of the current vehicle.

[0260] The recognition result acquisition unit is used to input the current vehicle profile data and the current energy replenishment reference data into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

[0261] In any of the devices for identifying recharging intentions in this embodiment of the invention, the current vehicle data includes: vehicle terminal log data and vehicle sensor data.

[0262] The power replenishment intention recognition device provided in the embodiments of the present invention can execute the power replenishment intention recognition method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method execution.

[0263] Example 8

[0264] Figure 12 This is a schematic diagram of the structure of an electronic device provided in Embodiment 8 of the present invention. Figure 12 A block diagram is shown that is suitable for implementing embodiments of the present invention. Figure 12 The device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0265] like Figure 12 As shown, the electronic device 1200 is presented in the form of a general-purpose computing device. The components of the electronic device 1200 may include, but are not limited to: one or more processors or processing units 1210, system memory 1220, and bus 1230 connecting different system components (including system memory 1220 and processing unit 1210).

[0266] Bus 1230 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0267] Electronic device 1200 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 1200, including volatile and non-volatile media, removable and non-removable media.

[0268] System memory 1220 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 1221 and / or cache memory (cache 1222). Electronic device 1200 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 1223 may be used to read and write non-removable, non-volatile magnetic media (… Figure 12 Not shown; usually referred to as a "hard drive"). Although Figure 12 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 1230 via one or more data media interfaces. System memory 1220 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of various embodiments of the present invention.

[0269] A program / utility 1225 having a set (at least one) of program modules 1224 may be stored, for example, in system memory 1220. Such program modules 1224 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 1224 typically perform the functions and / or methods described in the embodiments of this invention.

[0270] Electronic device 1200 can also communicate with one or more external devices 1300 (e.g., keyboard, pointing device, display 1310, etc.), and with one or more devices that enable a user to interact with electronic device 1200, and / or with any device that enables electronic device 1200 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed via input / output (I / O) interface 1240. Furthermore, electronic device 1200 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1250. As shown, network adapter 1250 communicates with other modules of electronic device 1200 via bus 1230. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 1200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0271] The processing unit 1210 executes various functional applications and data processing by running programs stored in the system memory 1220, such as implementing the model training method and energy replenishment intention recognition method provided in the embodiments of the present invention.

[0272] Example 8

[0273] Embodiment 8 of the present invention also provides a computer-readable storage medium storing a computer program (or computer-executable instructions) thereon, which, when executed by a processor, is used to perform the model training method and the energy replenishment intention recognition method provided in the embodiments of the present invention.

[0274] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0275] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0276] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0277] Computer program code for performing the operations of embodiments of the present invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0278] The acquisition, storage, and application of sample vehicle data, sample recharging reference data, recharging behavior data, sample vehicle profile data, vehicle terminal log data, vehicle sensor data, associated sensor data, recharging intent data, current vehicle data, current recharging reference data, historical recharging reference data, supplementary reference data, and current vehicle profile data involved in the technical solution disclosed herein all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0279] "It should be noted that the power replenishment intention recognition model in this embodiment is not targeted at a specific user and cannot reflect the personal information of a specific user."

[0280] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the embodiments of the present invention have been described in detail above, the embodiments of the present invention are not limited to the above embodiments. Many other equivalent embodiments may be included without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.

Claims

1. A model training method, characterized in that, include: Sample energy replenishment reference data is extracted from sample vehicle data; wherein, the sample energy replenishment reference data includes: engine status data, motor status data, vehicle speed data, vehicle energy data, vehicle mileage data, vehicle information data, time data, and location data; Extract energy replenishment intention data related to the energy replenishment intention from the sample energy replenishment reference data; From the replenishment intention data, replenishment behavior data related to the replenishment behavior is extracted, including: sequentially traversing the replenishment intention data at each time moment and determining whether the push condition is met at that time moment; if met, the replenishment intention data at that time moment is pushed onto the stack, and then it is determined whether the number of data entries in the stack meets the preset number; if met, based on the replenishment intention data in the stack, it is determined whether there is an energy growth event; if so, the replenishment intention data in the stack is used as replenishment behavior data, and the replenishment intention data in the stack is cleared; wherein, the replenishment behavior data is a set of sample replenishment reference data combined with the time moment corresponding to the replenishment behavior; the energy growth event is an event in which the vehicle energy data increases; the push condition is the condition that the replenishment intention data meets the replenishment behavior condition; Based on the energy replenishment behavior data, the energy replenishment location is determined; Based on the refueling location, a profile of the sample vehicle is constructed; wherein, the profile data is generated from the sample vehicle's behavioral data to produce label data describing the sample vehicle, and the profile data includes parking location, refueling location, driving route and energy data; The power replenishment behavior data and the profile data are input into the power replenishment intention recognition model for training.

2. The method according to claim 1, characterized in that, From the sample energy replenishment reference data, energy replenishment intention data related to the energy replenishment intention is extracted, including: From the vehicle terminal log data in the sample power replenishment reference data, filter out keywords related to power replenishment intent; From the vehicle sensor data in the sample recharge reference data, filter the associated sensor data of the recharge intention keywords; Based on the power replenishment intention keywords and the associated sensor data, the power replenishment intention data is determined.

3. The method according to claim 1, characterized in that, From the sample energy replenishment reference data, energy replenishment intention data related to the energy replenishment intention is extracted, including: The recharge time is determined based on the vehicle sensor data in the sample recharge reference data; Based on the refueling time and reference driving duration, determine the intended refueling period; Extract sample energy replenishment reference data located during the energy replenishment intention period, and use it as energy replenishment intention data.

4. The method according to claim 3, characterized in that, After determining the energy replenishment time, the following is also included: If there are at least two refueling times, the average driving time is determined based on the at least two refueling times. A reference driving time is determined based on the average driving time and the preset interval percentage; wherein the preset interval percentage is the percentage of the driving time interval from the generation of the refueling intention until the vehicle begins to perform the refueling behavior.

5. The method according to claim 1, characterized in that, Based on the energy replenishment intention data in the stack, determine whether an energy growth event exists, including: Based on the energy replenishment intention data in the stack, determine the energy replenishment growth amount and / or energy replenishment time interval; Based on the energy replenishment amount and / or the energy replenishment time interval, determine whether an energy growth event exists.

6. The method according to any one of claims 1-5, characterized in that, The sample vehicle data includes: vehicle terminal log data and vehicle sensor data.

7. A method for recognizing energy replenishment intentions, characterized in that, include: Obtain current energy replenishment reference data from current vehicle data; The current power replenishment reference data is input into the power replenishment intention recognition model to obtain the power replenishment intention recognition result at the current moment; The energy replenishment intention recognition model is trained using the method described in any one of claims 1-6.

8. The method according to claim 7, characterized in that, After obtaining the current refueling reference data from the current vehicle data, it also includes: Obtain supplementary reference data for the current energy replenishment reference data from historical energy replenishment reference data; Accordingly, the current power replenishment reference data is input into the power replenishment intention recognition model to obtain the power replenishment intention recognition result at the current moment, including: The current energy replenishment reference data and the supplementary reference data are input into the energy replenishment intention recognition model to obtain the energy replenishment intention recognition result at the current moment.

9. The method according to claim 7, characterized in that, The current power replenishment reference data is input into the power replenishment intention recognition model to obtain the power replenishment intention recognition result at the current moment, including: Obtain the current vehicle's profile data; The current vehicle profile data and the current charging reference data are input into the charging intention recognition model to obtain the charging intention recognition result at the current moment.

10. The method according to any one of claims 7-9, wherein, The current vehicle data includes: vehicle terminal log data and vehicle sensor data.

11. A model training device, characterized in that, include: The sample data extraction module is used to extract sample energy replenishment reference data from sample vehicle data; wherein, the sample energy replenishment reference data includes: engine status data, motor status data, vehicle speed data, vehicle energy data, vehicle mileage data, vehicle information data, time data, and location data; The behavior data determination module includes: an intent data extraction unit and a behavior data extraction unit; The intent data extraction unit is used to extract energy replenishment intent data related to energy replenishment intent from the sample energy replenishment reference data; The behavior data extraction unit is used to extract energy replenishment behavior data related to energy replenishment behavior from the energy replenishment intention data; wherein, the energy replenishment behavior data is a set of sample energy replenishment reference data combined with the corresponding time of the energy replenishment behavior; The behavioral data extraction unit includes: The condition judgment subunit is used to sequentially traverse the power replenishment intention data at each time point and determine whether the push condition is met at that time point; wherein, the push condition is the condition that the power replenishment intention data meets the power replenishment behavior. The data count judgment subunit is used to determine whether the number of data items in the stack meets the preset number after the energy replenishment intention data at that moment is pushed onto the stack if the preset number is met. The growth determination subunit is used to determine whether an energy growth event exists based on the energy replenishment intention data in the stack if the condition is met; wherein, the energy growth event is an event in which the vehicle energy data increases. The data determination sub-unit is used to, if it exists, take the energy replenishment intention data in the stack as the energy replenishment behavior data and clear the energy replenishment intention data in the stack. The profile data construction module includes: a power replenishment location determination unit and a profile data construction unit; The energy replenishment location determination unit is used to determine the energy replenishment location based on the energy replenishment behavior data; The profile data construction unit is used to construct profile data of the sample vehicle based on the refueling location; wherein, the profile data is generated by producing label data describing the sample vehicle through the sample vehicle's behavior data, and the profile data includes parking location, refueling location, driving route and energy data; The model training module is used to input the power replenishment behavior data and the portrait data into the power replenishment intention recognition model to train the power replenishment intention recognition model.

12. A device for recognizing a power replenishment intention, characterized in that, include: The current data acquisition module is used to obtain the current energy replenishment reference data from the current vehicle data; The power replenishment intention recognition module is used to input the current power replenishment reference data into the power replenishment intention recognition model to obtain the power replenishment intention recognition result at the current moment; The energy replenishment intention recognition model is trained using the method described in any one of claims 1-6.

13. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the model training method as described in any one of claims 1-6, or the energy replenishment intention recognition method as described in any one of claims 7-10.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the model training method as described in any one of claims 1-6, or the energy replenishment intention recognition method as described in any one of claims 7-10.

Images