A method, device, equipment and related system for determining vehicle refueling day

Abstract

The present invention discloses a method, device, equipment and related system for determining a vehicle refueling day. The method may include: based on the current date, obtaining the current vehicle fuel level, the current vehicle location, and the vehicle identification; inputting the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day. The present invention predicts the user's refueling date by analyzing the user's refueling behavior habits, and can remind the user to refuel at a more appropriate time, place and when the oil price is favorable, thereby providing the user with a more personalized service and improving the user experience.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, equipment and related system for determining a vehicle refueling day. Background Art

[0002] The deep integration of new technologies and traditional industries has brought about tremendous changes in the automotive industry. Networked smart travel has gradually become a focus of people's driving, and has also allowed cars to shift from a means of transportation to a mobile living space. Smart travel is a new business form and new model with "rational allocation of online resources and efficient and high-quality offline operation" formed through the close integration of the construction of industrial Internet and new infrastructure. The basic categories of smart travel are as follows: unmanned driving, vehicle-road collaboration, smart cockpit, smart car service, etc.

[0003] In addition, as people's demand for vehicles increases, in addition to the demand for smart cockpits in the vehicles themselves, car owners' demand for smart car services is also increasing, such as smart parking, smart refueling, and other smart car services combined with infrastructure services.

[0004] It is worth mentioning that the smart car-use function combined with infrastructure services is relatively complex. For example, when a smart refueling car owner wants to use the smart refueling service, he needs to download an app on his mobile phone or check the official account whether the destination gas station supports the smart refueling service, which greatly increases the complexity of the user's use of the function. Summary of the invention

[0005] The inventors found that current refueling reminders are generally based on vehicle fuel tank alarm reminders, which are then synchronized to the vehicle's onboard terminal or warning device to remind the driver. With the development of smart refueling, this method has shown obvious delay disadvantages and cannot accurately provide personalized services to drivers.

[0006] In view of the above problems, the present invention is proposed to provide a method, device, equipment and related system for determining a vehicle refueling day that overcomes the above problems or at least partially solves the above problems.

[0007] In a first aspect, an embodiment of the present invention provides a method for determining a vehicle refueling day, which may include:

[0008] Based on the current date, obtain the current vehicle fuel level, current vehicle location, and vehicle identification;

[0009] The current date, the current vehicle fuel level, the current vehicle location and the vehicle identification are input into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day.

[0010] Optionally, if the current date is a vehicle refueling day, the method may further include:

[0011] Obtain gas stations on the route ahead, and determine whether there is a gas station preferred by the user among the gas stations based on the user preferences in the user profile;

[0012] If so, navigation is performed with the gas station preferred by the user as the end point and the current location as the starting point.

[0013] Optionally, the method may further include: comparing the navigation distance between the current location and the gas station preferred by the user with the maximum driving distance that the vehicle can travel with the current amount of fuel;

[0014] If the navigation distance is less than the maximum driving distance, navigation is performed with the gas station preferred by the user as the end point and the current location as the starting point.

[0015] Optionally, the method may further include: obtaining gas stations on the route ahead and information on preferential fuel prices of the gas stations;

[0016] Sorting the navigation distances and fuel price discounts of the fuel stations to determine the fuel station with the best fuel price discount within the driving distance that the vehicle can travel with the current fuel volume;

[0017] Navigation is performed with the current location as the starting point and the gas station with the best oil price discount as the end point.

[0018] In a second aspect, an embodiment of the present invention provides a method for determining a vehicle refueling time period, which may include:

[0019] Based on the current date, obtain the current vehicle fuel level, current vehicle location, and vehicle identification, and determine the future time period;

[0020] The current date, the future time period, the current vehicle fuel level, the current vehicle location and the vehicle identification are input into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period.

[0021] In a third aspect, an embodiment of the present invention provides a machine learning model training method, which may include:

[0022] Acquire a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the historical date;

[0023] The recursive neural network model is trained using the sample data in the training sample set, the vehicle oil level, the vehicle location, the historical date and the vehicle identification are input into the recursive neural network model, and whether the vehicle has been refueled on the historical date is output; so as to determine the model training parameters in the recursive neural network model.

[0024] Optionally, determining whether the vehicle has performed a refueling operation on a historical date may include:

[0025] Based on the vehicle position when the vehicle stops, the amount of fuel when the vehicle stops, and the vehicle stop time, as well as the vehicle position when the vehicle restarts, the amount of fuel when the vehicle restarts, and the vehicle restart time, it is determined whether the vehicle has been refueled.

[0026] Optionally, it may also include:

[0027] Performing at least one of the following processing on the sample data: data preprocessing, data normalization processing, and data balancing processing;

[0028] Preprocessing the sample data includes: removing duplicate sample data in the sample data; and / or filling missing items in the sample data;

[0029] Normalizing the sample data includes: assigning consistent weights to features of different dimensions in the sample data;

[0030] The balancing process for the sample data includes: screening the sample data based on the category of the sample data; and / or evenly dividing the training samples and the test samples in the sample data set.

[0031] In a fourth aspect, an embodiment of the present invention provides a machine learning model training method, which may include:

[0032] Acquire a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the date;

[0033] Determining a historical time period based on the historical dates in the training sample set;

[0034] The recursive neural network model is trained using the sample data in the training sample set, the vehicle fuel volume, the historical time period, the vehicle location and the vehicle identification are input into the recursive neural network model, and whether the vehicle has been refueled in the historical time period is output; so as to determine the model training parameters in the recursive neural network model.

[0035] In a fifth aspect, an embodiment of the present invention provides a device for determining a vehicle refueling day, which may include:

[0036] A first acquisition module is used to acquire the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date;

[0037] The first determination module is used to input the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day.

[0038] In a sixth aspect, an embodiment of the present invention provides a device for determining a vehicle refueling time period, which may include:

[0039] A first acquisition module is used to acquire the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date;

[0040] The first determination module is used to determine a future time period based on a current date; and input the current date, the future time period, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period.

[0041] In a seventh aspect, an embodiment of the present invention provides a machine learning model training device, which may include:

[0042] A second acquisition module is used to acquire a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the historical date;

[0043] The training module is used to train the recursive neural network model with the sample data in the training sample set, input the vehicle oil volume, the vehicle location, the historical date and the vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled on the historical date; so as to determine the model training parameters in the recursive neural network model.

[0044] In an eighth aspect, an embodiment of the present invention provides a machine learning model training device, which may include:

[0045] A second acquisition module is used to acquire a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the historical date;

[0046] A second determination module is used to determine a historical time period based on the historical dates in the training sample set;

[0047] A training module is used to train the recursive neural network model using the sample data in the training sample set, input the vehicle fuel volume, the historical time period, the vehicle location and the vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled in the historical time period; so as to determine the model training parameters in the recursive neural network model.

[0048] In the ninth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, it implements the method for determining the vehicle refueling day as described in the first aspect, or implements the method for determining the vehicle refueling time period as described in the second aspect, or implements the machine learning model training method as described in the third aspect, or implements the machine learning model training method as described in the fourth aspect.

[0049] In the tenth aspect, an embodiment of the present invention provides a vehicle-mounted terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for determining a vehicle refueling day as described in the first aspect is implemented, or the method for determining a vehicle refueling time period as described in the second aspect is implemented.

[0050] In the eleventh aspect, an embodiment of the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the machine learning model training method described in the third aspect is implemented, or the machine learning model training method described in the fourth aspect is implemented.

[0051] In a twelfth aspect, an embodiment of the present invention provides a machine learning model training system, which may include a terminal and a server;

[0052] The terminal is used to collect a training sample set, wherein the sample data in the training sample set includes: the amount of vehicle oil, the vehicle location, the vehicle identification, and whether a refueling operation was performed on a historical date;

[0053] The server includes the machine learning model training device as described in claim 12, which is used to receive the training sample set sent by the terminal, and use the sample data in the training sample set to train the recursive neural network model, input the vehicle oil amount, the vehicle location, the historical date and the vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled on the historical date; so as to determine the model training parameters in the recursive neural network model.

[0054] In a thirteenth aspect, an embodiment of the present invention provides a machine learning model training system, which may include a terminal and a server;

[0055] The terminal is used to collect a training sample set, wherein the sample data in the training sample set includes: the amount of vehicle oil, the vehicle location, the vehicle identification, and whether a refueling operation was performed on a historical date;

[0056] The server includes a training device for the machine learning model as described in claim 13, which is used to receive the training sample set sent by the terminal, and determine the historical time period based on the historical date in the training sample set; use the sample data in the training sample set to train the recursive neural network model, input the vehicle oil amount, the historical time period, the vehicle location and the vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled during the historical time period; to determine the model training parameters in the recursive neural network model.

[0057] In a fourteenth aspect, an embodiment of the present invention provides a vehicle refueling day determination system, which may include: a vehicle-mounted terminal and a server;

[0058] The vehicle terminal is used to obtain the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date; and send the current date, the current vehicle fuel level, the current vehicle location, and the vehicle identification to the server;

[0059] The server is used to input the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day; and send the determination result to the vehicle terminal.

[0060] In a fifteenth aspect, an embodiment of the present invention provides a vehicle refueling time period determination system, which may include: a vehicle-mounted terminal and a server;

[0061] The vehicle terminal is used to obtain the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date; and send the current date, the current vehicle fuel level, the current vehicle location, and the vehicle identification to the server;

[0062] The server is used to determine a future time period based on a current date, and input the current date, the future time period, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period; and send the determination result to the vehicle terminal.

[0063] The beneficial effects of the above technical solution provided by the embodiment of the present invention include at least:

[0064] The embodiment of the present invention provides a method, device, equipment and related system for determining a vehicle refueling day. The method may include: based on the current date, obtaining the current vehicle fuel level, the current vehicle location, and the vehicle identification; inputting the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day. The present invention predicts the user's refueling date by analyzing the user's refueling behavior habits, and can remind the user to refuel at a more appropriate time, place and when the oil price is favorable, thereby providing the user with more personalized services and improving the user experience.

[0065] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0066] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0068] Figure 1 A schematic diagram of the process of the machine learning model training method provided in Example 1 of the present invention;

[0069] Figure 2 This is a schematic diagram of the structure of the machine learning model training device provided in Example 1 of the present invention;

[0070] Figure 3 This is a schematic diagram of the structure of the machine learning model training system provided in Example 1 of the present invention;

[0071] Figure 4 A schematic diagram of a flow chart of another machine learning model training method provided in Embodiment 2 of the present invention;

[0072] Figure 5 This is a flow chart of a method for determining a vehicle refueling day provided in Embodiment 3 of the present invention;

[0073] Figure 6 This is a flow chart of a specific method for determining a vehicle refueling day provided in Embodiment 3 of the present invention;

[0074] Figure 7 This is a schematic diagram of the structure of the vehicle refueling day determination device provided in Embodiment 3 of the present invention;

[0075] Figure 8 This is a schematic diagram of the structure of the vehicle refueling day determination system provided in Embodiment 3 of the present invention;

[0076] Fig. 9 This is a flow chart of the method for determining a vehicle refueling time period provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0077] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0078] Example 1

[0079] Embodiment 1 of the present invention provides a method for training a machine learning model. The method in the embodiment of the present invention can be implemented by a device for determining a habitual refueling day. Figure 1 As shown, the method may include the following steps:

[0080] Step S11, obtaining a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on that date.

[0081] Step S12: train the recursive neural network model with the sample data in the training sample set, input the vehicle fuel level, vehicle location, historical date and vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled on the historical date; to determine the model training parameters in the recursive neural network model.

[0082] It should be noted that the above-mentioned recursive neural network model in the embodiment of the present invention is an artificial neural network (ANN) with a tree-like hierarchical structure and network nodes recursively process input information according to their connection order. It is one of the deep learning algorithms.

[0083] The sample data collected in the embodiment of the present invention is collected based on historical dates, and the vehicle fuel volume, vehicle location, vehicle identification, and whether refueling was performed at a certain time point in the historical date are collected.

[0084] It should be noted here that the vehicle fuel level can be obtained through the fuel gauge on the vehicle, and can also be obtained through the liquid level gauge set in the fuel tank. The embodiment of the present invention does not make specific limitations on this. The above-mentioned vehicle fuel level in the embodiment of the present invention is the remaining fuel level of the vehicle, which is represented by a number between 0 and 1, wherein 0 indicates that the fuel tank is empty, and 1 indicates that the fuel tank is full. The above-mentioned vehicle position is obtained through the positioning device carried by the vehicle, for example, it can be longitude and latitude coordinates, or it can be place name coordinates, etc.; the above-mentioned vehicle identification is mainly a user identification and an identification of the vehicle's own fuel tank capacity, etc. Since each user has different refueling habits, it is necessary to mark the identification of different users and collect data; and the identification of the vehicle's own fuel tank capacity combined with the remaining fuel level can predict the maximum driving distance that the vehicle can still travel, so as to facilitate the later judgment of whether the distance of the gas station recommended to the user meets the maximum driving distance limit, etc.

[0085] In an optional embodiment, the sample data in the training sample set in the embodiment of the present invention may be collected when the vehicle is started or turned off. Whether a refueling operation has been performed needs to be determined by judgment. The inventor of this embodiment uses the vehicle oil volume obtained (detected) next time to determine whether a refueling operation has been performed in the sample data collected last time. The specific implementation method is as follows:

[0086] Based on the vehicle position when the vehicle stops, the amount of fuel when the vehicle stops, the vehicle stop time, and the vehicle position when the vehicle restarts, the amount of fuel when the vehicle restarts, and the vehicle restart time, it is determined whether the vehicle has been refueled.

[0087] In a specific example, referring to Table 1, historical sample data is collected.

[0088] Table 1

[0089]

[0090]

[0091] In the above historical sample data provided in the embodiment of the present invention, as shown in Table 1, the data collected are from different users. Of course, the embodiment of the present invention is not limited to the historical data of two users, and a large amount of user data is collected, and then after screening, the screened sample data is used for training. The above vehicle position in the embodiment of the present invention can be a location or a coordinate position. As long as it is easy for the computer device to read, the embodiment of the present invention does not specifically limit the identification method of the above position; the above date "2003030730" means 07:30 on March 3, 2020.

[0092] In the embodiment of the present invention, whether the above-mentioned vehicle has been refueled is determined by the oil volume, position, etc. obtained next time. For example, sample data 1 is the oil volume, vehicle position, and vehicle start time after the vehicle is turned on, sample data 2 is the oil volume, vehicle position, and vehicle shutdown time before the vehicle is parked and turned off, and sample data 3 is the oil volume, vehicle position, and vehicle start time when it is turned on again. By comparing the changes in the oil volume in sample data 2 and sample data 1, which show a decreasing trend and different vehicle positions, it can be judged that no refueling operation was performed in sample data 1 and sample data 2; by judging sample data 2 and sample data 3, the oil volume has not changed, and the vehicle position has not changed, it is judged that sample data 3 has not been refueled either. Similarly, by judging sample data 5, sample data 6 and sample data 7, it can be clearly judged that no refueling operation was performed in sample data 5, while the oil amount in sample data 7 increased relative to sample data 6, and the vehicle position is the same, so it can be judged that the vehicle refueled at position 3, and it can be judged that sample data 6 was collected before the vehicle stopped and turned off at the gas station, and sample data 7 was collected when the vehicle was turned on after refueling was completed.

[0093] It should be noted that the sample data in the training sample set obtained in the embodiment of the present invention not only includes sample data for which refueling operation has been performed, but also includes sample data for which refueling operation has not been performed. In the sample data for which refueling operation has been performed, the amount of oil in the vehicle (the signal amount of the fuel tank) and the vehicle position (trajectory data) in two consecutive sample data can be used to count the user's refueling behavior. The sample data can reflect the remaining oil amount of the vehicle before refueling, the remaining oil amount of the vehicle after refueling is completed, the start time of refueling, the completion time of refueling, etc. The trajectory data of the vehicle position can reflect the starting position of the vehicle before driving to the gas station, the position of the gas station, and the end position of the vehicle after leaving the gas station, etc.

[0094] In a specific example, the fuel level of a vehicle is based on the remaining fuel level before the vehicle starts to refuel and the remaining fuel level after the vehicle is refueled. This can be done by recording the remaining fuel level each time the vehicle stops and the remaining fuel level at the time the vehicle starts. The refueling time also records the time when the vehicle stops and the time when the vehicle starts again, as well as the location of each stop and the starting location when arriving at this place and the location of the destination after leaving. For example: each time the user starts the vehicle, the current location of the vehicle, the starting time and the remaining fuel level are recorded. After the user arrives at the destination, the current location of the vehicle, the stopping time and the remaining fuel level are recorded. These data are collected according to a fixed frequency and mechanism.

[0095] In another alternative embodiment, referring to Figure 1As shown, the method may also include: performing at least one of the following processing on the sample data: data preprocessing, data normalization processing and data balancing processing.

[0096] The preprocessing of the sample data includes: removing duplicate sample data from the sample data; and / or filling in missing items in the sample data. For example, if the sample data has duplicate data or missing data, the duplicate data or missing data may be removed; or only the duplicate data in the sample data may be removed, and the missing data in the sample data may be filled, for example, a filling algorithm may be used to fill in the missing data.

[0097] Normalizing the sample data includes assigning consistent weights to features of different dimensions in the sample data. For example, each feature dimension of the sample data is set to a consistent weight to improve the convergence speed of the iterative solution of the model.

[0098] Balancing the sample data includes: screening the sample data based on the category of the sample data; and / or evenly dividing the training samples and test samples in the sample data set. When the sample data is used as training data for the pre-built user refueling habit prediction model, the data volume between categories is balanced to avoid unsatisfactory training results due to the skewed sample data volume during subsequent model training.

[0099] It should also be noted that in order to make the sample data more balanced, if the training sample set includes both training data and test data, the sample data is divided according to a certain ratio. For example, the sample data can be divided according to a ratio of 8:2, 80% of the sample data is used as training data to train the prediction model; 20% of the sample data is used as test data to test the trained prediction model and perform model evaluation.

[0100] It should be noted that the embodiment of the present invention does not make any specific limitation on the order of executing the above-mentioned data preprocessing, data normalization processing and data balancing processing. Data preprocessing can be performed first, data normalization processing can be performed first, or data balancing processing can be performed first. Of course, they can also be performed simultaneously. The embodiment of the present invention does not make any specific limitation on this.

[0101] In another optional embodiment, before training the recursive neural network model with the sample data in the training sample set, the method may further include: performing noise processing on the sample data using a binning algorithm and / or a clustering algorithm.

[0102] Since the sample data in the training sample set in the embodiment of the present invention is sent through different smart terminals, it is necessary to select whether the data is usable training data after decompression. The information represented by the data screening in the embodiment of the present invention is the amount of oil, time, and vehicle location collected and persistently stored in the background service software, and the conditional screening or the information with problems, such as data that failed to be decompressed. The next step is to extract, group, arrange and fuse the data content according to the rules and the structured data format after screening and processing, convert the format, and persist it in the data warehouse. Finally, the model is trained through the fused persistent data, and the user refueling habit model is completed synchronously.

[0103] It is understandable that data screening is mainly to solve the problems or meaningless data in the collected data, so that the model can be trained through accurate feature information. The main categories of screening are as follows:

[0104] First, unusable data, which includes situations where the data cannot be decompressed according to the standard gzip format defined by the protocol. Data information that cannot be parsed according to the agreed data definition format after decompression. During the parsing process, the necessary transmission information agreed upon is missing, or the transmission information is not used according to the agreed rules. For example, the data definition range is 0 to 1, but the information provided is 2, which exceeds the defined range. The data described above is defined as unusable data.

[0105] Second, meaningless data. This includes the situation where the parking position when the vehicle was last turned off is very different from the position where it was started again, which makes it impossible to determine whether the user has refueled, so the data is meaningless. The starting and stopping positions of the vehicle are the same, and the fuel level has not changed. This type of data shows that the user has not made any meaningful travel behavior, so the data is meaningless and needs to be deduplicated or de-noised. There is data that does not meet the standards in the data. For example, the remaining fuel level remains unchanged after driving 50KM, which also means that the data is meaningless. The data described above are all defined as meaningless data.

[0106] The main purpose of feature fusion is to integrate the training samples needed for learning the user refueling habit model, reduce the data dimension, merge the feature data that can be classified and fused, extract effective feature data, and reduce the multi-dimensional features to reduce the dimension and reduce the resource consumption of model training. The main method is to divide the collected user data into data without oil volume change and data with oil volume change according to the change of oil volume. The refueling location of the data with oil volume change forms location set data, and the clustering algorithm is used to cluster the location set data to generate clustered data of the location set data.

[0107] Among them, the clustering algorithm can adopt a density-based clustering method with noise, and the clustering of the location set data can be specifically, through the clustering algorithm, at least a plurality of location points in the location set data whose distances between the locations are within a preset threshold range are clustered into a pile to generate cluster data. Among them, the preset threshold of the distance can be preset by relevant technical personnel according to an empirical value or an experimental value, for example, it can be 500 meters. Exemplarily, at least 4 location points in the location set data whose distances between the locations are within a range of 500 meters are clustered into a pile through the DBSCAN clustering algorithm, and can be clustered into multiple piles to generate cluster data. According to the cluster data, non-cluster data is defined as noise data for elimination, and the refueling date and time data in the cluster data is converted into periodic data from Monday to Sunday as a sample of training data, for example: User 1 refuels on Monday, and such samples are then subjected to data standardization and normalization.

[0108] The data that is filtered and discarded in the embodiment of the present invention can also be used to reversely verify whether there are problems with the vehicle hardware equipment, and the generated filtered data can also be synchronized to other data services. For example, based on the filtered and discarded data, analysis is performed. If the data is unusable, the smart device with such data should have some unidentified vulnerabilities, such as network transmission vulnerabilities, scenario mechanism vulnerabilities, or execution environment vulnerabilities. For example, there are vulnerabilities in the execution environment, resulting in byte loss after compression, or transmission modification, resulting in failure to parse the data. Or there is meaningless data, and it may be identified that there is a problem with the instrument or the signal protocol provided, resulting in meaningless data. For example: after traveling 50KM, but the vehicle's fuel volume data has not changed, there may be a problem with the vehicle's fuel volume signal.

[0109] In another optional embodiment, the method may also include: verifying the recursive neural network model using a test sample in a training sample set; wherein, the recursive neural network model is verified by comparing the accuracy, precision, recall rate and / or area under the receiver operating characteristic curve of the recursive neural network model with a preset threshold to determine whether the recursive neural network model has been trained.

[0110] In the embodiment of the present invention, the stopping condition of training may be that the number of training iterations exceeds a preset number threshold; or, the error of the output result of the current model meets the preset error requirement. For example, the area under the receiver operating characteristic curve (AUC) has a value between 0.5 and 1.0, and the larger the AUC value, the better the performance of the model.

[0111] In an optional embodiment, in the above step S12, the recursive neural network model is trained using sample data in the training sample set to determine model training parameters in the recursive neural network model.

[0112] The training parameters may include: prediction labels, intercept terms, and number of iterations, etc. Among them, the prediction label is the user's target behavior, which can be the user refueling, or not refueling, as well as at least one of the other extended behaviors. The intercept term can improve the accuracy of the trained user refueling habit prediction model, and can be specifically set by relevant technical personnel as required. The number of iterations can be determined based on the actual situation of the model, for example, it can be set to 100 times. In an embodiment of the present invention, since the prediction model is updated periodically, the updated training model data will be integrated, marked, and published. This allows subsequent functions to be updated and used in the latest user refueling habit daily recursive neural network model.

[0113] Based on the same inventive concept, an embodiment of the present invention further provides a vehicle refueling day determination device, referring to Figure 2 As shown, the device may include: a first acquisition module 11 and a first determination module 12, and its working principle is as follows:

[0114] The first acquisition module 11 is used to obtain the current vehicle fuel level, the current vehicle position, and the vehicle identification based on the current date; the first determination module 12 is used to input the current date, the current vehicle fuel level, the current vehicle position, and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day.

[0115] In another optional embodiment, the first determination module 12 is further used to determine whether the vehicle has been refueled based on the vehicle position when the vehicle stops, the amount of fuel in the vehicle when the vehicle stops, the time the vehicle stops, and the vehicle position when the vehicle is restarted, the amount of fuel in the vehicle when the vehicle is restarted, and the time the vehicle is restarted.

[0116] In another alternative embodiment, referring to Figure 2 As shown, the device may also include: a data preprocessing module 13, a data normalization processing module 14 and a data balancing processing module 15; wherein the data preprocessing module 13 is used to remove duplicate sample data in the sample data; and / or, fill in missing items in the sample data; the data normalization processing module 14 is used to assign consistent weights to features of different dimensions in the sample data; the data balancing processing module 15 is used to screen the sample data based on the category of the sample data; and / or, evenly divide the training samples and test samples in the sample data set.

[0117] In another alternative embodiment, referring to Figure 2 As shown, it may also include: a noise processing module 16, which is used to perform noise processing on the sample data by adopting a binning algorithm and / or a clustering algorithm.

[0118] In another alternative embodiment, referring to Figure 2 As shown, the device may also include: a verification module 17, the verification module 17 is used to verify the recursive neural network model using the test samples in the training sample set; wherein, the recursive neural network model is verified, and the accuracy, precision, recall rate and / or area under the receiver operating characteristic curve of the recursive neural network model are compared with a preset threshold to determine whether the recursive neural network model is trained.

[0119] Regarding the machine learning model training device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method and will not be elaborated here.

[0120] Based on the same inventive concept, an embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned machine learning model training method is implemented.

[0121] Based on the same inventive concept, a computer device is also provided in an embodiment of the present invention, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned machine learning model training method when executing the program.

[0122] Based on the same inventive concept, an embodiment of the present invention further provides a machine learning model training system, referring to Figure 3 As shown, the system may include: a terminal 31 and a server 32;

[0123] The terminal 31 is used to collect a training sample set, and the sample data in the training sample set includes: the amount of oil in the vehicle, the vehicle location, the vehicle identification, and whether the refueling operation was performed on the historical date;

[0124] The server 32 includes the above-mentioned machine learning model training device, which is used to receive the training sample set sent by the terminal, and use the sample data in the training sample set to train the recursive neural network model, input the vehicle fuel level, vehicle location, historical date and vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled on the historical date; so as to determine the model training parameters in the recursive neural network model.

[0125] The terminal in the embodiment of the present invention may be an intelligent device installed on a vehicle for collection, such as a vehicle-mounted terminal, and a large amount of sample data is sent to a backend server through the intelligent device for processing.

[0126] The terminal (vehicle intelligent device) compresses the collected data and related required information as a whole, and transmits the compressed data to the background server through the communication device. The upload frequency is divided into two cases: one is based on the amount of data, and the other is a fixed mechanism. After the background server synchronizes to the information, it will classify the data and persist the information to the relevant intelligent devices according to different channels. It should be noted that the content of the data collection can be added or deleted by the technician according to actual needs, and the embodiment of the present invention does not impose any restrictions on this.

[0127] The data compression scheme uses the standard gzip compression method. If the current data size has reached 16KB, the collected data will be compressed using the gzip standard method and synchronized to the backend server. If the vehicle has not been started for a long time, for more than 240 hours, the collected data will be compressed using the gzip standard method and synchronized to the backend server when it is started again. Data synchronization uses the standard https network protocol, which is received via a secure network protocol to the backend server. The backend server will parse the data according to the defined protocol, parse the data type, and synchronize the time. The data will be persistently stored in the smart device according to the data type and synchronization time.

[0128] Example 2

[0129] Based on the same inventive concept, another machine learning model training method is also provided in the embodiment of the present invention, referring to Figure 4 As shown, the method may include the following steps:

[0130] Step S41, obtaining a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the date.

[0131] Step S42: Determine the historical time period based on the historical dates in the training sample set.

[0132] Step S43, train the recursive neural network model with the sample data in the training sample set, input the vehicle fuel level, historical time period, vehicle location and vehicle identification into the recursive neural network model, and output whether the vehicle has been refueled in the historical time period; to determine the model training parameters in the recursive neural network model.

[0133] It should be noted that the main difference between the present embodiment 2 and the above embodiment 1 is that step S42 is added to divide the historical date into different time periods; for example, the time periods are divided into 00:00-06:00, 06:00-09:00, 09:00-12:00, 12:00-15:00, 15:00-18:00, 18:00-21:00, 21:00-24:00, etc. The time period is used as a training data to determine the specific time period for refueling. Compared with the training for refueling days in embodiment 1, this can more accurately recommend a specific refueling time period to the user according to the user's refueling habits.

[0134] Based on the same inventive concept, another machine learning model training method is also provided in the embodiment of the present invention, which can be referred to Figure 2 As shown, it may include: a first acquisition module 11 and a first determination module 12; its working principle is as follows:

[0135] The first acquisition module 11 acquires a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether refueling operation was performed on the date; the first determination module 12 determines the historical time period based on the historical date in the training sample set; and the first determination module 12 uses the sample data in the training sample set to train the recursive neural network model, inputs the vehicle fuel volume, the historical time period, the vehicle location, and the vehicle identification into the recursive neural network model, and outputs whether the vehicle has refueled in the historical time period; to determine the model training parameters in the recursive neural network model.

[0136] Based on the same inventive concept, an embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned machine learning model training method is implemented.

[0137] Based on the same inventive concept, a computer device is also provided in an embodiment of the present invention, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned machine learning model training method when executing the program.

[0138] Based on the same inventive concept, another machine learning model training system is also provided in the embodiment of the present invention, which can also be referred to Figure 3 As shown, the system may include: a terminal 1 and a server 2;

[0139] Terminal 1 is used to collect training sample sets, and the sample data in the training sample sets include: the amount of vehicle fuel, vehicle location, vehicle identification and whether refueling operations were performed on historical dates on historical dates; server 2 includes a training device for the above-mentioned machine learning model, which is used to receive the training sample sets sent by terminal 1, and determine the historical time period based on the historical dates in the training sample sets; use the sample data in the training sample sets to train the recursive neural network model, input the amount of vehicle fuel, historical time period, vehicle location and vehicle identification into the recursive neural network model, and output whether the vehicle has refueled in the historical time period; to determine the model training parameters in the recursive neural network model.

[0140] Example 3

[0141] Based on the same inventive concept, a method for determining a vehicle refueling day is provided in an embodiment of the present invention. Figure 5 As shown, the method may include:

[0142] Step S51: Based on the current date, obtain the current vehicle fuel level, current vehicle location, and vehicle identification.

[0143] Step S52: input the current date, the current fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a refueling day for the vehicle.

[0144] It should be noted that the recursive neural network model pre-trained in the embodiment of the present invention can be obtained by training in the manner of embodiment 1, or can be obtained by training in other manners, and the embodiment of the present invention does not specifically limit this. In the embodiment of the present invention, the above determination of whether the current date is a refueling day for the vehicle is used as a message prompt to the user to facilitate the user's travel strategy.

[0145] It should also be noted that the triggering scenario of the above method in the embodiment of the present invention can be any travel scenario of the user, including but not limited to the following scenarios: commuting scenario, travel scenario, travel scenario for visiting relatives and friends, travel scenario for gathering, travel scenario for banquet, etc. The scenario includes scenarios of user driving behaviors such as starting the vehicle, driving, and parking. Among them, the embodiment of the present invention will especially trigger the timing in the scenario where the user starts the vehicle under the predefined premise of commuting to work or commuting from get off work.

[0146] In a specific example, the user's commuting information can be provided by commuting information, which can be the existing commuting module information. According to the unified commuting standard, it can provide commuting prediction labels between home and company. According to whether the current vehicle status is to start the car machine to obtain the status of KL15on, the basic scenario of commuting information push is established.

[0147] The embodiment of the present invention analyzes the user's refueling behavior habits to predict the user's refueling date, and can remind the user to refuel at a more appropriate time and place and when the oil price is preferential, thereby providing the user with more personalized services and improving the user experience.

[0148] In another specific example, the vehicle identification obtained after the vehicle is started is used to communicate over the network through the standard https protocol. The prediction model published by the user's habitual refueling day obtains the user's unique identification, and the time of the week is calculated by time. The prediction model of the user's habitual refueling day is used to predict the result, which is either a refueling day or a non-refueling day, 1 or 0 result data, and the result data is synchronized to the vehicle smart device to make the next decision. Example: User A refuels every Wednesday after get off work or Friday at work. Today is Wednesday. According to the prediction model, the result will be 1, that is, the result data of the refueling day, and it will be displayed to the user.

[0149] In another alternative embodiment, referring to Figure 6 As shown, the method may also include:

[0150] Step S53: Obtain gas stations on the route ahead, and determine whether there is a gas station that the user likes among the gas stations based on the user preferences in the user portrait.

[0151] The gas stations on the above-mentioned forward route in the embodiment of the present invention may be acquired through navigation data in the navigation route, or may be acquired through a pathfinding function, and the embodiment of the present invention does not specifically limit this.

[0152] Step S54: If so, navigate with the user's favorite gas station as the end point and the current location as the starting point.

[0153] In an embodiment of the present invention, by identifying the gas station brands preferred by the user, targeted personalized services are provided to the user. For example, by identifying whether there are gas stations of the user's favorite brand on the route the user often takes home, and under the condition that the above-mentioned gas station day reminder is met, the user can be intelligently recommended gas stations of the user's favorite brand on the way home from get off work.

[0154] In another optional embodiment, also refer to Figure 6 As shown, the method may also include:

[0155] Step S55: Compare the navigation distance from the current location to the gas station preferred by the user with the maximum driving distance that the vehicle can travel with the current amount of fuel.

[0156] Step S54: If the navigation distance is less than the maximum driving distance, navigation is performed with the user's favorite gas station as the end point and the current location as the starting point.

[0157] In an embodiment of the present invention, the distance between the current location and the gas station is determined, and whether the current remaining fuel level of the vehicle supports the required mileage to reach the gas station is determined. If it supports reaching the gas station, a navigation route to the gas station can be provided to the user.

[0158] In another optional embodiment, also refer to Figure 6 As shown, the method may also include:

[0159] Step S56: Obtain the gas stations on the route ahead and their fuel price discount information.

[0160] Step S57: Sort the navigation distances and fuel price discounts of the fuel stations to determine the fuel station with the best fuel price discount within the range of the driving distance that the vehicle can travel with the current fuel volume.

[0161] Step S58: Navigate with the current location as the starting point and the gas station with the best oil price discount as the end point.

[0162] In an embodiment of the present invention, if there are multiple gas stations that meet the conditions on the route ahead, they are sorted according to the mileage from the current location and the degree of discount, and the information of the gas station with the shortest mileage and the greatest discount is displayed for the user to choose.

[0163] Based on the same inventive concept, an embodiment of the present invention further provides a vehicle refueling day determination device, referring to Figure 7 As shown, the device may include: a first acquisition module 71 and a first determination module 72, and its working principle is as follows:

[0164] The first acquisition module 71 is used to acquire the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date. The first determination module 72 is used to determine a future time period based on the current date; and input the current date, the future time period, the current vehicle fuel level, the current vehicle location, and the vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period.

[0165] In an alternative embodiment, referring to Figure 7 As shown, it also includes: a judgment module 73 and a navigation module 74. The first acquisition module 71 is also used to obtain gas stations on the route ahead. The judgment module 73 determines whether there is a gas station that the user likes among the gas stations based on the user preference in the user portrait. If so, the navigation module 74 takes the gas station that the user likes as the end point and the current position as the starting point for navigation.

[0166] In another optional embodiment, also refer to Figure 7 As shown, the judgment module 73 is also used to compare the navigation distance between the current position and the gas station preferred by the user and the maximum driving distance that the vehicle can travel with the current amount of fuel; if the navigation distance is less than the maximum driving distance, the navigation module 74 takes the gas station preferred by the user as the end point and the current position as the starting point for navigation.

[0167] In another optional embodiment, it also includes a sorting module 75, and the above-mentioned first acquisition module 71 is also used to obtain gas stations on the route ahead and the oil price discount information of the gas stations; the sorting module 75 is used to sort the navigation distances and oil price discounts of the gas stations, so that the first determination module 72 can determine the gas station with the best oil price discount within the driving distance that the current fuel volume of the vehicle can travel; the navigation module 74 takes the current position as the starting point and navigates to the gas station with the best oil price discount as the end point.

[0168] Regarding the vehicle refueling day determination device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0169] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned method for determining a vehicle refueling day is implemented.

[0170] Based on the same inventive concept, an embodiment of the present invention also provides a vehicle-mounted terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned vehicle refueling day determination method when executing the program.

[0171] Based on the same inventive concept, a vehicle refueling day determination system is also provided in an embodiment of the present invention. Figure 8 As shown, the system may include: a vehicle terminal 81 and a server 82;

[0172] The vehicle terminal 81 is used to obtain the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date; and send the current date, the current vehicle fuel level, the current vehicle location, and the vehicle identification to the server 82;

[0173] The server 82 is used to input the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day; and send the determination result to the vehicle terminal.

[0174] The specific implementation methods of the above-mentioned devices, media, vehicle-mounted terminals and related systems in the embodiments of the present invention have been described in detail in the embodiments of the method, and will not be elaborated here.

[0175] Example 4

[0176] Embodiment 4 of the present invention provides a method for determining a vehicle refueling time period, referring to Fig. 9 As shown, the method may include the following steps:

[0177] Step S91: Based on the current date, obtain the current vehicle fuel level, current vehicle location, and vehicle identification.

[0178] Step S92: Determine a future time period based on the current date.

[0179] In the embodiment of the present invention, a time period of 2 to 4 hours in the future can be determined based on the current time to determine whether refueling is required in the time period of 2 to 4 hours in the future; or a time period between the current time and the next 3 hours can be determined based on the current time to determine whether refueling is required in this time period. In the embodiment of the present invention, the method for determining the future time period is to estimate in advance, that is, the embodiment of the present invention is not limited to predicting whether refueling is required at the current time, but needs to predict whether refueling is required in a certain time period in the future. However, the embodiment of the present invention does not specifically limit how to estimate the future time period based on the current time, which can be set by the user, or the prediction method can be triggered when the oil level reaches 1 / 3 of the grid.

[0180] Step S93: input the current date, future time period, current vehicle fuel level, current vehicle location and vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period.

[0181] It should be noted that the difference between Example 4 of the present invention and Example 3 is that what is determined is not the refueling day, but the refueling time period, and accordingly, the pre-trained recurrent neural network model used is no longer the same as the model in Example 3. The recurrent neural network model in the embodiment of the present invention may be pre-trained in the manner of Example 2, or may be pre-trained in other ways, and this embodiment does not specifically limit this.

[0182] The above method provided in the embodiment of the present invention can more accurately determine the refueling time period, so that a specific refueling time period can be more accurately recommended to the user based on the user's refueling habits, further improving the user experience.

[0183] Based on the same inventive concept, a vehicle refueling day determination device is also provided in the embodiment of the present invention, which can be referred to Figure 7 As shown, the device may include: a first acquisition module 71 and a first determination module 72, and its working principle is as follows:

[0184] The first acquisition module 71 is used to obtain the current vehicle fuel level, the current vehicle position, and the vehicle identification based on the current date; the first determination module 71 is used to determine a future time period based on the current date; the first determination module 71 inputs the current date, the future time period, the current vehicle fuel level, the current vehicle position, and the vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period.

[0185] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned method for determining a vehicle refueling time period is implemented.

[0186] Based on the same inventive concept, an embodiment of the present invention also provides a vehicle-mounted terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for determining a vehicle refueling time period when executing the program.

[0187] Based on the same inventive concept, a vehicle refueling time period determination system is also provided in an embodiment of the present invention, which can be referred to as Figure 8 As shown, the system may include: a vehicle terminal 81 and a server 82;

[0188] The vehicle terminal 81 is used to obtain the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date; and send the current date, the current vehicle fuel level, the current vehicle location, and the vehicle identification to the server;

[0189] The server 82 is used to determine a future time period based on the current date, and input the current date, the future time period, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period; and send the determination result to the vehicle terminal.

[0190] The specific implementation methods of the above-mentioned devices, media, vehicle-mounted terminals and related systems in the embodiments of the present invention have been described in detail in the embodiments of the method, and will not be elaborated here.

[0191] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0192] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0193] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0194] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0195] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for determining a vehicle refueling day, characterized in that: include: Based on the current date, obtain the current vehicle fuel level, current vehicle location, and vehicle identification; Inputting the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day; The recursive neural network model is pre-trained by the following method: obtaining a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether refueling operation was performed on the historical date; using the sample data in the training sample set to train the recursive neural network model, inputting the vehicle fuel volume, vehicle location, historical date, and vehicle identification into the recursive neural network model, and outputting whether the vehicle was refueled on the historical date to determine the model training parameters in the recursive neural network model.

2. The method according to claim 1, characterized in that If the current date is a refueling day for the vehicle, the method further includes: Obtain gas stations on the route ahead, and determine whether there is a gas station preferred by the user among the gas stations based on the user preferences in the user profile; If so, navigation is performed with the gas station preferred by the user as the end point and the current location as the starting point.

3. The method according to claim 2, characterized in that Also includes: Comparing the navigation distance from the current location to the gas station preferred by the user with the maximum driving distance that the vehicle can travel with the current amount of fuel; If the navigation distance is less than the maximum driving distance, navigation is performed with the gas station preferred by the user as the end point and the current location as the starting point.

4. The method according to any one of claims 1 to 3, characterized in that Also includes: Obtain information about gas stations on the route ahead and their gas price discounts; Sorting the navigation distances and fuel price discounts of the fuel stations to determine the fuel station with the best fuel price discount within the driving distance that the vehicle can travel with the current fuel volume; Navigation is performed with the current location as the starting point and the gas station with the best oil price discount as the end point.

5. A method for determining a vehicle refueling time period, characterized in that: include: Based on the current date, obtain the current vehicle fuel level, current vehicle location, and vehicle identification, and determine the future time period; Inputting the current date, the future time period, the current vehicle fuel level, the current vehicle location, and the vehicle identification into a pre-trained recursive neural network model to determine whether the future time period is a vehicle refueling time period; Wherein, the recursive neural network model is pre-trained by the following method: obtaining a training sample set, the sample data in the training sample set including: vehicle fuel volume, vehicle location, historical date, vehicle identification and whether refueling operation was performed on the date; determining the historical time period based on the historical date in the training sample set; training the recursive neural network model with the sample data in the training sample set, inputting the vehicle fuel volume, the historical time period, the vehicle location and the vehicle identification into the recursive neural network model, and outputting whether refueling operation was performed on the vehicle in the historical time period; to determine the model training parameters in the recursive neural network model.

6. A machine learning model training method, characterized in that: include: Acquire a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the historical date; The recursive neural network model is trained using the sample data in the training sample set, the vehicle oil level, the vehicle location, the historical date and the vehicle identification are input into the recursive neural network model, and whether the vehicle has been refueled on the historical date is output; so as to determine the model training parameters in the recursive neural network model.

7. The method according to claim 6, characterized in that Determining whether the vehicle has been refueled on a historical date, including: Based on the vehicle position when the vehicle stops, the amount of fuel when the vehicle stops, and the vehicle stop time, as well as the vehicle position when the vehicle restarts, the amount of fuel when the vehicle restarts, and the vehicle restart time, it is determined whether the vehicle has been refueled.

8. The method according to claim 6, characterized in that Also includes: Performing at least one of the following processing on the sample data: data preprocessing, data normalization processing, and data balancing processing; Preprocessing the sample data includes: removing duplicate sample data in the sample data; and / or filling missing items in the sample data; Normalizing the sample data includes: assigning consistent weights to features of different dimensions in the sample data; The balancing process for the sample data includes: screening the sample data based on the category of the sample data; and / or evenly dividing the training samples and the test samples in the sample data set.

9. A machine learning model training method, characterized in that: include: Acquire a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether a refueling operation was performed on the date; Determining a historical time period based on the historical dates in the training sample set; The recursive neural network model is trained using the sample data in the training sample set, the vehicle fuel volume, the historical time period, the vehicle location and the vehicle identification are input into the recursive neural network model, and whether the vehicle has been refueled in the historical time period is output; so as to determine the model training parameters in the recursive neural network model.

10. A device for determining a vehicle refueling date, characterized in that: include: A first acquisition module is used to acquire the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date; A first determination module is used to input the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day; The recursive neural network model is pre-trained by the following method: obtaining a training sample set, wherein the sample data in the training sample set includes: vehicle fuel volume, vehicle location, historical date, vehicle identification, and whether refueling operation was performed on the historical date; using the sample data in the training sample set to train the recursive neural network model, inputting the vehicle fuel volume, vehicle location, historical date, and vehicle identification into the recursive neural network model, and outputting whether the vehicle was refueled on the historical date to determine the model training parameters in the recursive neural network model.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, it implements the method for determining a vehicle refueling day as described in any one of claims 1 to 4, or implements the method for determining a vehicle refueling time period as described in claim 5, or implements the machine learning model training method as described in any one of claims 6 to 8, or implements the machine learning model training method as described in claim 9.

12. A vehicle-mounted terminal, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for determining a vehicle refueling day according to any one of claims 1 to 4 is implemented, or the method for determining a vehicle refueling time period according to claim 5 is implemented.

13. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the machine learning model training method as described in any one of claims 6 to 8, or implements the machine learning model training method as described in claim 9.

14. A vehicle refueling day determination system, characterized in that: include: Vehicle-mounted terminals and servers; The vehicle terminal is used to obtain the current vehicle fuel level, the current vehicle location, and the vehicle identification based on the current date; and send the current date, the current vehicle fuel level, the current vehicle location, and the vehicle identification to the server; The server is used to input the current date, the current vehicle fuel level, the current vehicle location and the vehicle identification into a pre-trained recursive neural network model to determine whether the current date is a vehicle refueling day; And send the determination result to the vehicle-mounted terminal.

Images