Analysis method, system, device and storage medium based on shared battery data

Abstract

The present application relates to an analysis method, system, device and storage medium based on shared battery data, including: obtaining battery information, user information, geographic information and user demand information of a shared battery terminal; preprocessing, data integration and characterization of the battery information, user information and geographic information to obtain feature data; analyzing the feature data and user demand information through a pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information; performing de-characterization on the predicted feature information to obtain the user's travel planning results and output them to the user. The present application makes full use of the big model's excellent analysis and data mining capabilities for complex multi-dimensional information to provide electric vehicle operators and users with more effective user portraits, route planning, macro scheduling and many customized user services involving user behavior.

Description

Technical Field

[0001] The present application relates to the technical field of tram travel analysis, and in particular to an analysis method, system, device and storage medium based on shared battery data. Background Art

[0002] At present, electric vehicles are becoming a popular mode of travel for the public due to their energy-saving, economical and environmentally friendly advantages. However, unlike fuel vehicles, electric vehicles have the problem of long charging time, which makes users' "mileage anxiety" always a difficult problem to solve. In recent years, with the continuous innovation of battery technology and the continuous improvement of charging network infrastructure, the cost of purchasing and using batteries has become lower and lower, so the business model based on shared batteries has become more and more popular. Especially in the field of electric two-wheeled vehicles, the battery replacement mode based on shared batteries has become a rapidly popular form of endurance, providing an extremely effective method for scenarios that require timeliness, such as instant delivery and car rental travel. It can be foreseen that with the development of battery technology and standardization, the endurance technology based on shared batteries will inevitably become the mainstream endurance method for electric vehicles in the future.

[0003] Since shared batteries are a new form of business that is still in the ascendant, there are many aspects to be optimized and improved in terms of user support and services. For example, in the face of a large number of electric two-wheeled vehicle users who use shared batteries, how to conduct data mining and describe their user portraits to achieve more accurate user services and advertising recommendations; in the face of instant delivery and logistics that require timeliness, how to reasonably analyze and plan user routes under the condition of shared battery replacement; for a large number of shared battery users in road traffic, how to layout and optimize the charging and replacement infrastructure related to shared batteries based on the user's riding habits. In response to the above problems, existing algorithms such as portrait analysis and path planning usually use models or rules with specific functions to solve them, but these methods usually only consider information in a single dimension or local dimensional space, so they are insufficient in terms of versatility and functionality. In recent years, the rapid development of AI big models has enabled all walks of life to make great progress in data analysis and intelligence. This idea can solve the above problems well, but AI big models are still a blank in areas such as user travel services involved in shared batteries. Summary of the invention

[0004] The embodiments of the present application provide an analysis method, system, device and storage medium based on shared battery data, so as to at least solve the problem in the related art of lacking trajectory prediction using shared battery data.

[0005] In a first aspect, an embodiment of the present application provides an analysis method based on shared battery data, including:

[0006] Collecting data from the user terminal and the shared battery terminal to obtain user terminal information and battery terminal information, wherein the user terminal information includes function requirement information;

[0007] Preprocessing and characterizing the user-side information and the battery-side information in sequence to obtain initial characteristic data;

[0008] Analyze the initial feature data and functional requirement information through a pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the functional requirement information;

[0009] The predicted feature information is subjected to de-feature processing to obtain data analysis results and feed them back to the user.

[0010] In one embodiment, the preprocessing of the user-side information and the battery-side information includes:

[0011] Recording the user-end information and the battery-end information as raw data, filtering the raw data, removing abnormal points and noise points, and obtaining first data;

[0012] Completing the missing data in the first data by interpolation or extension to obtain second data;

[0013] The second data is compressed to obtain data to be characterized.

[0014] In one embodiment, characterization processing is performed on the data to be characterized, including:

[0015] Processing the data to be characterized by normalization or label mapping table to obtain data to be processed;

[0016] The data to be processed are sequentially subjected to feature extraction and feature fusion processing through unsupervised learning to obtain the initial feature data.

[0017] In one embodiment, the construction of the shared battery trajectory AI big model includes:

[0018] Select the neural network architecture according to user needs and determine the training strategy to obtain the initial AI large model;

[0019] Acquire historical data of the user terminal and the shared battery terminal, and divide the historical data into a training set, a validation set, and a test set;

[0020] The initial AI large model is trained according to the training strategy and the training set, and the parameters of the initial AI large model are adjusted through the verification set to obtain an intermediate AI large model;

[0021] The test set is applied to evaluate the performance of the intermediate AI large model, and when the evaluation result meets the preset user requirements, the shared battery trajectory AI large model is obtained.

[0022] In one embodiment, the neural network architecture includes multiple feature extraction modules and / or multiple feature restoration modules; wherein,

[0023] The feature extraction module includes a first multi-head attention module, a first feedforward network module and a first normalization module; the feature restoration module includes a masked multi-head attention module, a second multi-head attention module, a second feedforward network module and a second normalization module;

[0024] The multiple feature extraction modules or the multiple feature restoration modules are combined in the form of stacking multiple attention modules.

[0025] In one embodiment, the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information, including:

[0026] Processing the initial feature data through a first multi-head self-attention module to obtain a first feature vector sequence;

[0027] Processing the sum of the first feature vector sequence and the initial feature data through a first feedforward network module with a residual network and a first normalization module to obtain first high-level feature information;

[0028] Processing the initial feature data through a masked multi-head self-attention module to obtain a second feature vector sequence;

[0029] The second feature vector sequence and the first high-level feature information are processed sequentially through a second multi-head attention module, a second feedforward network module, and a second normalization module to obtain high-level feature restoration information;

[0030] The height feature restoration information is predicted according to the functional requirement information to obtain the predicted feature information.

[0031] In one embodiment, the pre-built shared battery trajectory AI big model is used to analyze the feature data and user demand information to obtain predicted feature information corresponding to the user demand information, and further includes:

[0032] Acquire a plurality of labeled user customized attributes, and fuse the user customized attributes with the initial feature data to obtain new initial feature data;

[0033] Processing the new initial feature data through a first multi-head self-attention module to obtain a third feature vector sequence;

[0034] Processing the sum of the third feature vector sequence and the initial feature data through a first feedforward network module with a residual network and a first normalization module to obtain second high-level feature information;

[0035] The probability of the user-customized attribute to which the second highly characteristic information belongs is calculated through a linear layer and a softmax function to obtain a classification prediction probability, and the classification prediction probability is recorded as the predicted characteristic information.

[0036] In one embodiment, when the user-side information includes map information and travel start and end point location information, the neural network architecture includes multiple feature extraction modules, and the data analysis result is a predicted trajectory path; the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information, and also includes:

[0037] Processing the initial feature data through a first multi-head self-attention module to obtain a fourth feature vector sequence, wherein the initial feature data includes features of the map information and the travel start and end point location information;

[0038] The sum of the fourth feature vector sequence and the initial feature data is processed sequentially by a first feedforward network module with a residual network and a first normalization module to obtain third high-level feature information, that is, the third high-level feature information is the predicted feature information;

[0039] or,

[0040] When the collected data also includes multimodal data, the neural network architecture includes multiple feature restoration modules; the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information, and also includes:

[0041] Performing characterization and feature extraction processing on the multimodal data in sequence to obtain multimodal feature data;

[0042] Processing the initial feature data through a masked multi-head self-attention module to obtain a fifth feature vector sequence;

[0043] The fifth feature vector sequence and the multimodal feature data are processed in sequence by a second multi-head attention module, a second feedforward network module and a second normalization module to obtain a feature output sequence; the feature output sequence is recorded as the predicted feature information.

[0044] In a second aspect, an embodiment of the present application provides an analysis system based on shared battery data, which is used to execute the analysis method based on shared battery data described in any of the above embodiments, including:

[0045] A data collection unit, used to collect data from the user terminal and the shared battery terminal to obtain user terminal information and battery terminal information, wherein the user terminal information includes function requirement information;

[0046] A data processing unit, used to perform preprocessing, data integration and characterization processing operations on the user-side information and the battery-side information respectively to obtain initial characteristic data;

[0047] A large model calculation unit, used to analyze the initial feature data and functional requirement information through a pre-built shared battery trajectory AI large model to obtain predicted feature information corresponding to the functional requirement information;

[0048] The result output unit is used to perform de-feature processing on the predicted feature information, obtain data analysis results and feed them back to the user.

[0049] In one embodiment, the large model calculation unit is arranged in a shared battery terminal or a cloud server, and the large model calculation unit includes:

[0050] The feature extraction module is used to process the initial feature data through a first multi-head self-attention module to obtain a first feature vector sequence; and process the sum of the first feature vector sequence and the feature data through a first feedforward network module with a residual network and a first normalization module to obtain first high-level feature information;

[0051] The feature restoration module is used to process the initial feature data through a masked multi-head self-attention module to obtain a second feature vector sequence; process the second feature vector sequence and the first high-level feature information through a second multi-head attention module, a second feedforward network module and a second normalization module to obtain high-level feature restoration information;

[0052] The feature prediction module is used to predict the height feature restoration information according to the functional requirement information to obtain the predicted feature information.

[0053] In a third aspect, an embodiment of the present application 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 computer program, the analysis method based on shared battery data as described in the first aspect above is implemented.

[0054] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the analysis method based on shared battery data as described in the first aspect above.

[0055] The analysis method, system, device and storage medium based on shared battery data provided by the embodiments of the present application have at least the following technical effects:

[0056] This application uses shared batteries as the core carrier, shared battery trajectories and related attribute information as driving data, and adopts a self-developed shared battery trajectory large model as the core algorithm, and uses a terminal-cloud collaborative model as a service provision channel. In addition, this application makes full use of the large model's excellent analysis and data mining capabilities for complex multi-dimensional information to provide electric vehicle operators and users with more effective user portraits, path planning, macro scheduling, and many customized user services involving user behavior.

[0057] Details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0059] Figure 1 is a flow chart of an analysis method based on shared battery data in one embodiment of the present application;

[0060] Figure 2 is a structural block diagram of an analysis system based on shared battery data in one embodiment of the present application;

[0061] Figure 3 is a flow chart of preprocessing and characterization in one embodiment of the present application;

[0062] Figure 4 This is a basic architecture diagram based on a shared battery trajectory model in an embodiment of the present application;

[0063] Figure 5 This is an example diagram of a fine-tuning structure of a shared battery trajectory large model for a trajectory classification task in one embodiment of the present application;

[0064] Figure 6 This is an example diagram of a fine-tuning structure of a shared battery trajectory large model for trajectory generation and planning tasks in one embodiment of the present application;

[0065] Figure 7This is an example diagram of a fine-tuning structure of a shared battery trajectory large model for multi-modal trajectory analysis in an embodiment of the present application;

[0066] Figure 8 It is a user service data processing and analysis flow diagram based on the end-cloud collaboration of the shared battery trajectory big model in one embodiment of the present application;

[0067] Fig. 9 It is a basic architecture diagram of an electronic device in one embodiment of the present application. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0069] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0070] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0071] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a", "the" and the like involved in this application do not indicate a quantity limitation, and may indicate the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like involved in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The "multiple" involved in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that there may be three relationships, for example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific ordering of the objects.

[0072] The user service method and system based on the large model of shared battery trajectories fully fills the imperfections in the current services in this field. The existing technologies and methods used in this field usually do not take into account the particularity of the shared battery user group. For example, the traditional navigation that relies solely on spatial location does not take into account the problem of battery life, which will cause serious "mileage anxiety" problems for instant delivery and long-distance users. For another example, there is a lack of a portrait description method for the user group using shared batteries, which further limits the service quality and targeted service for such users. In particular, there are currently almost no relevant methods and strategies for the management and guidance of electric two-wheeled vehicle users. In addition, there is currently no suitable method for reference on how to analyze the group behavior of shared battery users to obtain the best charging and swapping layout network. As such, services involving shared battery users lack relevant technologies to provide services with specific functions. Based on the above situation, the present application provides an analysis method and system based on shared battery data.

[0073] On the first aspect, an embodiment of the present application provides a user service analysis method based on shared battery data. The system is driven by shared battery data and uses an AI big model as the core algorithm. It aims to solve the current incompleteness of user services in the shared battery field, including: portrait analysis and recommendation problems for electric vehicle users; path planning problems including charging and swapping and other endurance infrastructure; macro-scheduling problems for charging and swapping infrastructure for electric vehicles; and user-related services associated with specific shared battery technologies.

[0074] The analysis method based on shared battery data of the present application is executed by an analysis system based on shared battery data, which includes a shared battery user terminal, a shared battery terminal, and a cloud server. The schematic diagram of the entire system module and the information flow relationship are as follows: Figure 2 As shown. Combined Figure 1 The analysis system of this application is described in detail. Figure 1 is a flow chart of the analysis method based on shared battery data, such as Figure 1 As shown, the method includes the following steps.

[0075] Step S1, collecting data from a user terminal and a shared battery terminal to obtain user terminal information and battery terminal information, wherein the user terminal information includes function requirement information.

[0076] Among them, the shared battery user terminal (ie, user terminal) is the main platform for providing user services and interacting with users, and can be based on the user's mobile phone or other personalized user-portable devices. This part mainly provides user-side information, mainly including user information of the user (such as age, gender, occupation, vehicle, etc.), geographic information related to the user's environment (such as map information), and functional requirement information. Among them, the functional requirement information represents the services required by the user, which corresponds to the different application functions of the subsequent use of the AI ​​large model. The shared battery terminal is used to provide battery-side information. In an embodiment of the present application, the battery end (ie, the shared battery terminal) can first send the trajectory information and battery information to the user end, and then the user end sends these data to the cloud (ie, the cloud server) for processing; the battery end can also directly send the trajectory data and battery information to the cloud for processing; when the computing power of the battery end is strong, the battery end can also process the trajectory data and battery information by itself.

[0077] Step S2, preprocessing, data integration and characterization processing are performed on the user-side information and the battery-side information in sequence to obtain initial characteristic data. In this embodiment, the data preprocessing, data integration and characterization operations can be performed in the shared battery terminal or in the cloud server. As for which hardware is used for processing, it can be selected according to the actual situation of the user.

[0078] Specifically, in one embodiment, during the preprocessing process, the user-side information and the battery-side information are first recorded as original data, the original data are screened, abnormal points and noise points are removed, and the first data are obtained; the missing data in the first data are supplemented by interpolation or extension method to obtain the second data; the second data is compressed to obtain the data to be characterized.

[0079] After preprocessing, the data is integrated. Data integration is the process of merging data from different sources and formats into a unified data store to provide a consistent, integrated view. Figure 2 After the data is integrated in the shared battery terminal, it can be characterized in the shared battery terminal or the cloud server, while the data that has been integrated in the cloud server can only be characterized in the cloud server. After the data is integrated, the shared battery terminal will temporarily store the battery trajectory data, and then transmit the data back to the cloud server for storage in a timely manner to form a battery trajectory data set.

[0080] In one embodiment, the integrated battery information, user information and geographic information are characterized, including: processing the data to be characterized by normalization or label mapping table to obtain the data to be processed; extracting features and fusing features on the data to be processed in sequence by unsupervised learning to obtain the initial feature data.

[0081] In one embodiment, the processing of battery trajectory data includes preprocessing and characterization of battery trajectory data. Figure 3 . This trajectory data is a set of multi-dimensional vector sequences with time sequence, which mainly contains battery trajectory sequence, user and battery attribute information and geographic information. The trajectory sequence information includes timestamp, longitude, latitude, speed, stay time and other information. User and battery attribute data include: age, gender, occupation, vehicle type, battery type, battery power and other information. Map attribute data includes: longitude and latitude information of road network, POI information, longitude and latitude information of electric two-wheeled vehicle related business points (charging, battery replacement, maintenance). In terms of data preprocessing, it is necessary to remove abnormal points and noise points of all data. These points are usually data with obvious abnormalities, such as the large-scale drift of latitude and longitude information of thousands of kilometers; at the same time, the missing data is supplemented by various interpolation and extension methods. For example, for the incomplete trajectory data of one day, the trajectory data is supplemented for one day by the default static method; the preprocessing of data also needs to use some data point optimization methods to compress the data. For example, long-term stay points do not need to be represented by hundreds of points, but can be represented by merging them into one point and the stay time.

[0082] Then, the physical quantities and label quantities in the data are digitized using statistical normalization and label mapping methods respectively. Then, the characterization network module obtained through unsupervised learning of the trajectory large model is used to integrate the features of the shared battery digitization related data. Then, a set of multidimensional vector data carrying all the relevant data of the shared battery is used for subsequent continuous training and fine-tuning of the model.

[0083] Specifically, data characterization will go through two steps. The first is the digitization of the original data. For physical quantity information, a normalization method based on the theoretical maximum value is used. This mode allows all physical quantity related information to be adjusted in proportion without losing the meaning of the data, such as longitude and latitude, speed, power, etc.; for attribute label information, a label number mapping method is used for digitization, such as battery type, poi information, vehicle type, etc.; for map data, feature extraction is first performed, geographic entities are classified and encoded according to attributes, the spatial relationship between geographic entities is extracted, and represented by topological information. The specific feature encoding method can be represented by a coordinate sequence based on point, line, and surface entities; or a raster data encoding method can be used to encode geographic features according to the characteristics of each raster. The second step of data characterization is to further extract and fuse features of the data in combination with unsupervised learning. In this step, the initial digitization results of the original data will be further embedded in the unsupervised training model, which belongs to the pre-module of large model training. The input trajectory sequence information, user and battery attribute information, and related geographic attribute information are integrated into a set of battery trajectory data sequences through a series of multi-layer perceptrons, providing a basis for subsequent trajectory large model training and fine-tuning.

[0084] Step S3: Analyze the initial feature data and functional requirement information through the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the functional requirement information. Figure 2 In the embodiment of the present application, the trajectory AI large model can run on a shared battery terminal or on a cloud server, but since model training requires a large amount of data, the trajectory AI large model training process is carried out in the cloud server.

[0085] In an embodiment of the present application, the construction of the shared battery trajectory AI big model includes: selecting a neural network architecture according to user needs, and determining a training strategy to obtain an initial AI big model; obtaining historical data of user terminals and shared battery terminals, and dividing the historical data into a training set, a validation set, and a test set; training the initial AI big model according to the training strategy and the training set, and adjusting the parameters of the initial AI big model through the validation set to obtain an intermediate AI big model; applying the test set to evaluate the performance of the intermediate AI big model, and when the evaluation result meets the preset user requirements, the shared battery trajectory AI big model is obtained.

[0086] In this embodiment, the neural network architecture includes multiple feature extraction modules and / or multiple feature restoration modules. The feature extraction module includes a first multi-head attention module, a first feedforward network module and a first normalization module; the feature restoration module includes a masked multi-head attention module, a second multi-head attention module, a second feedforward network module and a second normalization module, wherein the multiple feature extraction modules or multiple feature restoration modules are combined in the form of module stacking. Specifically, the AI ​​large model algorithm based on shared battery trajectory is the key to the implementation of the solution. Reference Figure 4 ,The basic architecture of the model adopts a multi-layer data feature extraction and ,data feature restoration module, in which the data feature extraction module is used to ,extract features from the data in the input feature sequence, and the feature restoration ,module is used to restore features of the extracted data.

[0087] In a preferred embodiment, when the user's functional requirement is to perform basic analysis on the data, the neural network architecture of this embodiment is a transformer architecture, including an input module (linear layer fusion), an output module (linear layer and prediction features), N feature extraction modules, and N feature restoration modules. The input module obtains the data to be processed, and the output module outputs the processing results of the feature restoration module. In this embodiment, the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI large model to obtain the predicted feature information corresponding to the user demand information, including: the initial feature data is processed by the first multi-head self-attention module to obtain a first feature vector sequence; the first feedforward network module with a residual network and the first normalization module are used to process the sum of the first feature vector sequence and the initial feature data to obtain the first high-level feature information; the initial feature data is processed by the masked multi-head self-attention module to obtain a second feature vector sequence; the second feature vector sequence and the first high-level feature information are processed by the second multi-head attention module, the second feedforward network module and the second normalization module in turn to obtain high-level feature restoration information; the high-level feature restoration information is predicted according to the functional requirement information to obtain the predicted feature information. Preferably, the feature extraction module is composed of a multi-head attention module, a layer normalization module, a feedforward network and a layer normalization module; the feature restoration module is composed of a masked multi-head attention module, a multi-head attention module, a layer normalization module, a feedforward network and a layer normalization module.

[0088] Specific reference Figure 2 , the data to be processed include trajectory data, user attributes, map attributes and other information, and then these data are processed by data characterization, classification features, and map characterization respectively. After characterization, they are input into the input module (i.e., linear layer) of the neural network architecture for fusion processing. The fused series of features are then position-encoded. The purpose of this step is to give the sequence of each dimension of the vector an obvious order of precedence, which is conducive to the model's judgment of the order of information at different times. The feature data with position encoding are sent to the feature extraction module and feature restoration module of the first layer respectively. After the feature data is processed by N feature extraction modules, the processing results are sent to the multi-head self-attention module of the feature restoration module. The feature data is sent to its own multi-head attention module after the mask multi-head attention module and layer normalization. After being processed by N feature restoration modules, the feature restoration results are obtained. The restoration results are processed by linear layer and prediction to obtain the predicted feature information related to the user's functional requirements. Finally, the predicted feature information is de-featured and the data analysis results are output.

[0089] In one embodiment, the feature data and user demand information are analyzed by a pre-built shared battery trajectory AI large model to obtain predicted feature information corresponding to the user demand information, and also includes: obtaining multiple labeled user customized attributes, and fusing the user customized attributes with the initial feature data to obtain new initial feature data; processing the new initial feature data through a first multi-head self-attention module to obtain a third feature vector sequence; processing the sum of the third feature vector sequence and the initial feature data through a first feedforward network module with a residual network and a first normalization module to obtain second high-level feature information; performing probability calculation on the user customized attributes to which the second high-level feature information belongs through a linear layer and a softmax function to obtain a classification prediction probability; the classification prediction probability is recorded as the predicted feature information. Reference Figure 5 , Figure 5 exist Figure 4 On the basis of, it is also necessary to obtain the custom attributes set by the user and input the custom attributes into the linear layer fusion. The purpose of this embodiment is to analyze the input data and determine which of the custom attributes these data belong to, so as to obtain the data analysis results. Figure 4 The neural network structure is fine-tuned on this basis, that is, this embodiment only runs the feature extraction module to perform data analysis, and in the output part, a linear layer and a softmax function are added to process the results output by the feature extraction module to obtain the classification prediction results of the customized attributes.

[0090] In one embodiment, when the user-side information includes map information and travel start and end point location information, the neural network architecture includes multiple feature extraction modules, and the data analysis result is a predicted trajectory path; the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information, and also includes:

[0091] The initial feature data is processed by the first multi-head self-attention module to obtain a fourth feature vector sequence, wherein the initial feature data includes features of the map information and the location information of the starting and ending points of the trip; the sum of the fourth feature vector sequence and the initial feature data is processed by the first feedforward network module with a residual network and the first normalization module in turn to obtain third high-level feature information, that is, the third high-level feature information is the predicted feature information. Figure 6 In this embodiment, Figure 4The neural network structure is fine-tuned based on the above, that is, only the feature extraction module is run. In this embodiment, there is no need to input trajectory data and user attribute data, only map attribute data (such as map data and user's starting coordinates) is input. By analyzing the shared battery trajectory AI large model, the optimal driving trajectory can be generated and recommended to the user.

[0092] In one embodiment, when the collected data also includes multimodal data, the neural network architecture includes multiple feature restoration modules; the feature data and user demand information are analyzed by a pre-built shared battery trajectory AI large model to obtain predicted feature information corresponding to the user demand information, and also includes: characterizing and extracting the multimodal data in sequence to obtain multimodal feature data; processing the initial feature data through a masked multi-head self-attention module to obtain a fifth feature vector sequence; processing the fifth feature vector sequence and the multimodal feature data through a second multi-head attention module, a second feedforward network module, and a second normalization module in sequence to obtain a feature output sequence; the feature output sequence is recorded as the predicted feature information. Reference Figure 7 , Figure 7 exist Figure 4 Fine-tune on the basis of, that is, only run the feature restoration module. In addition, it is necessary to collect data from other modalities (such as images, sounds, etc.), and then perform data characterization processing. After characterization, they are sent to the data feature extraction network layer and the linear layer for fusion. The data feature extraction network layer performs feature extraction to obtain multimodal feature data; the linear layer fuses the characterized multimodal data and other data to obtain the initial feature data. The special feature of this embodiment is that the multimodal feature data also needs to be input into the multi-head attention module in the feature restoration module for processing. Finally, the analysis results related to the multimodal data are output.

[0093] For the AI ​​large model provided in this application, the trajectory data is first characterized before entering the entire model, in which the battery trajectory sequence, user and battery attributes, and map information are characterized and fused through a series of linear layers. The fused features are presented as a set of vector sequences of specific dimensions. This series of features is then position-encoded. The purpose of this step is to give the sequence of each dimension of the vector an obvious order of precedence, which is beneficial for the model to determine the order of information at different times. After characterization and position encoding, the vector sequence will be input into the first feature extraction module.

[0094] In the feature extraction module, the input vector sequence is first processed using a set of multi-head self-attention modules. After that, the vector sequence is summed with the input features, and then passes through a feedforward network with a residual network and a normalization process. At this point, the data has completed the process of the first feature extraction module. When the data has completed multiple feature extraction modules, it has highly characterized information covering the trajectory data.

[0095] This information is then fed into the feature restoration module in turn. The feature restoration module and the feature extraction module basically have the same data processing mode. The difference is that the feature results obtained through a series of feature extraction modules are used as one of the input items in each multi-head attention input. Mask information is added to the data input part to facilitate training and prediction and generation of specific point and dimension data. The AI ​​large model based on shared battery trajectory is trained with massive battery trajectory data. The training process uses random partial masking for the input data. The difference between the restored data and the input data at the mask is calculated as the loss function, and the model parameters are optimized according to the gradient descent method.

[0096] Step S4: De-characterize the predicted feature information to obtain data analysis results and feed them back to the user. Figure 2 In this embodiment, the cloud server or shared battery terminal obtains the output result after performing inverse feature processing on the output result of the large model, and then sends the output result to the shared battery user terminal.

[0097] Based on the shared battery trajectory AI big model, this application can make full use of the big model to achieve rapid application and implementation of specific functions through targeted fine-tuning. In terms of trajectory classification related services, user portraits, user behavior tendencies and other characteristics can be labeled, and only the characterization module and feature extraction module using the trajectory data in the big model can be fine-tuned. The feature restoration module after the feature extraction module is replaced with a linear layer, and the classification prediction probability is given by the softmax function. This fine-tuning model architecture can be seen in Figure 5 ,The fine-tuning process can also add customized attribute data information to the ,input data.

[0098] In terms of trajectory planning services, we can use the large model's ability to restore data features and train the model's ability to generate specific sequence segments under the premise of masking the trajectory input. This fine-tuning model architecture can be seen in Figure 6 In specific applications, the map information and the start and end point information can be input on the input side, and the model can plan the entire trajectory and realize trajectory-related services.

[0099] In addition, for situations where additional multimodal data is needed for analysis, Figure 7The fine-tuning strategy shown in the figure features the image, audio and other dimensional information accordingly, and integrates them into the input of the feature extraction module. According to the characteristics of multimodal data, the data is processed by the targeted data extraction network layer and then given to the feature extraction module. The entire model finally outputs information based on multimodal data.

[0100] In this application, the shared battery terminal is a hardware module at the battery end, which includes sensors related to the battery position and some main control systems with the ability to calculate, store, and send and receive data. The battery trajectory information and battery attribute information are formed in this part. For shared battery terminals with certain data processing capabilities, the previous data is directly preprocessed at this terminal, and the preprocessing results are merged through a data integration method to convert the original data into an array based on trajectory time series. If the shared battery terminal has a more powerful edge computing capability, the merged time series data will be calculated directly at the terminal for further characterization, and then calculated through a large model deployed at the terminal. The large model selects different function fine-tuning modules for calculation according to different input functional requirements. After the result of the calculation is de-characterized, the final output result is obtained. The output results include user portraits, user trajectory planning results, abnormal prediction results, etc., which are finally presented on the shared battery user terminal.

[0101] The cloud server is a remote hardware support for user services. This part has powerful computing power and storage capabilities, providing the final storage record for user data and calculations. Given that most shared battery terminals do not have the ability to process and store large amounts of data, in most cases, the user information, geographic-related information, and functional requirements returned from the user terminal will be sent to the cloud server together with the battery terminal trajectory information and battery information for data preprocessing. The data is integrated after preprocessing and then characterized. The large model selects different functional fine-tuning modules for calculation according to different input functional requirements. After the calculation results are de-characterized, the final analysis results are given to the shared battery user terminal. The results of the analysis can also be given to B-end users for large-scale analysis. The storage of user trajectory raw data is also located in the cloud server. All data collected by shared battery terminals, whether temporarily stored by the terminal or through the cloud, will eventually be collected in the shared battery trajectory data set. The trajectory data set is sorted and combined with the feedback from front-line users, and the original large model based on the shared battery trajectory and its fine-tuning module are corrected to continuously iterate and optimize the algorithm function.

[0102] In a preferred embodiment, the shared battery terminal and the cloud server form a set of terminal-cloud collaborative operation strategies. The specific process is shown in Figure 8 When the shared battery terminal has a small amount of computing power, this part of the module only completes the initial screening and combination of data. Large-scale feature processing and large model calculations are handed over to the cloud server ( Figure 2 and Figure 8 When the shared battery terminal has powerful computing capabilities, such as an embedded system with a GPU or a high-performance CPU chip (DSP, FPGA), characterization and large model computing can be performed directly on the shared battery terminal ( Figure 2 and Figure 8 The data will be temporarily stored in the shared battery terminal, and the battery trajectory data set will be transmitted back to the cloud server and summarized in a timely manner. For battery terminals with powerful computing capabilities, a situation-based switching strategy can also be used to operate. For example, when the terminal network signal is not smooth or the cloud computing resources are tight, the shared battery terminal computing power is used for calculations first; when the battery terminal computing resources are insufficient, it can also be switched to the cloud for calculations in a timely manner. In short, the timeliness of user services is ensured. When the shared battery terminal has no storage and computing analysis capabilities, all data processing and analysis processes are performed through cloud services ( Figure 2 and Figure 8 (see process ② in the figure).

[0103] In summary, the analysis method based on shared battery data provided by the embodiment of the present application and the end-cloud collaboration mechanism proposed by the patent of the present invention from the service system level can well ensure that the service can be obtained reliably even in the case of poor network communication. From the service method level, the trajectory, attributes and geographic map information of shared battery users are integrated, and the sequence data of such users are learned and output through the AI ​​big model, which can provide services to shared battery users in a very targeted manner. For example, in terms of path planning, the big model method combines charging and swapping infrastructure. In addition to considering the shortest path problem of users, it also combines the charging and swapping locations to solve the user's "mileage anxiety" problem. In terms of the analysis of shared battery user portraits, due to the integration of multi-dimensional battery, user, and geographic related information, it can more accurately analyze the user's travel habits, user occupation and other user portrait information. With the support of massive trajectory data, the trajectory big model can layout the optimal charging and swapping network in a specific area, thereby providing more reliable recommendations for shared battery operation teams and B-end users. In addition, all recommendation service functions and analysis and prediction functions that can establish connections with shared battery users can use the versatility of this large model to fine-tune and expand functions, providing a more powerful method tool for the future development of the shared battery industry. In terms of the systematic approach of this method to provide services to users, a computing power balance model based on end-cloud collaboration is proposed, which can make full use of the edge computing capabilities of shared batteries and maximize the reliability and real-time performance of the trajectory large model service method.

[0104] In a second aspect, an embodiment of the present application provides an analysis system based on shared battery data, which is used to execute the analysis method based on shared battery data described in any of the above embodiments, and the system includes:

[0105] A data collection unit, used to collect data from the user terminal and the shared battery terminal to obtain user terminal information and battery terminal information, wherein the user terminal information includes function requirement information;

[0106] A data processing unit, used to perform preprocessing, data integration and characterization processing on the user-side information and the battery-side information respectively to obtain initial characteristic data;

[0107] A large model calculation unit, used to analyze the initial feature data and functional requirement information through a pre-built shared battery trajectory AI large model to obtain predicted feature information corresponding to the functional requirement information;

[0108] The result output unit is used to perform de-feature processing on the predicted feature information, obtain data analysis results and feed them back to the user.

[0109] In one embodiment, the large model calculation unit is arranged in a shared battery terminal or a cloud server, and the large model calculation unit includes:

[0110] The feature extraction module is used to process the initial feature data through a first multi-head self-attention module to obtain a first feature vector sequence; and process the sum of the first feature vector sequence and the feature data through a first feedforward network module with a residual network and a first normalization module to obtain first high-level feature information;

[0111] The feature restoration module is used to process the initial feature data through a masked multi-head self-attention module to obtain a second feature vector sequence; process the second feature vector sequence and the first high-level feature information through a second multi-head attention module, a second feedforward network module and a second normalization module to obtain high-level feature restoration information;

[0112] The feature prediction module is used to predict the height feature restoration information according to the functional requirement information to obtain the predicted feature information.

[0113] In addition, this application focuses on protecting a series of services based on shared battery information, using shared batteries as terminals, using shared battery positioning information and related surrounding information as user service information sources, and providing users with related services through a series of data processing, characterization methods, and algorithm calculation processes. These modes all belong to the service mode of this application. Shared batteries, as a storage carrier of electrical energy, have the function of continuously providing energy to user devices. Devices that are renamed in disguise or have the same function all belong to the scope of the objects described in this application, and the service modes based on trajectories and related data as drivers that match them all belong to the protection scope of this application.

[0114] In addition to the digitization and deep learning-based feature fusion methods mentioned in the patent, the processing of time-related data such as the longitude and latitude, speed, and power of shared batteries and map attributes, using similar features through a series of linear layers, multi-layer perceptrons, matrix operations, convolution calculations, numerical calculations and other feature processing methods all belong to the same processing mode as the present invention. In the process of analyzing the shared battery trajectory, this application adopts a large model method based on a multi-head attention mechanism and feature extraction and feature restoration, but the number of layers of the large model is reduced, and the small model or regular algorithm degenerated into it also belongs to the alternative solution of this application.

[0115] The core algorithm of this application is based on the shared battery trajectory big model, which is mainly used to provide services to users or merchants of shared batteries. All fine-tuning strategies for specific functions based on it fall within the scope of the big model algorithm service described in this application, and are not limited to the several model fine-tuning architectures mentioned in this application.

[0116] In order to ensure the stable operation of the shared battery service system, the computing power coordination between the shared battery terminal and the cloud server proposed in this application is also a major feature. The shared battery terminal always uses the battery as the main energy supply core, and is equipped with a series of sensors and hardware and control units that can meet its functions. Vehicle-integrated or independently equipped power storage modules with similar characteristics belong to the shared battery terminal mentioned in this application.

[0117] In a third aspect, an embodiment of the present application provides an electronic device, Fig. 9 FIG. 1 is a block diagram of an electronic device according to an exemplary embodiment. Fig. 9 As shown, the electronic device may include a processor 11 and a memory 12 storing computer program instructions.

[0118] Specifically, the processor 11 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0119] Among them, the memory 12 may include a large-capacity memory for data or instructions. By way of example and not limitation, the memory 12 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 12 may include a removable or non-removable (or fixed) medium. Where appropriate, the memory 12 may be inside or outside the data processing device. In a specific embodiment, the memory 12 is a non-volatile memory. In a specific embodiment, the memory 12 includes a read-only memory (ROM) and a random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), wherein the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data output dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0120] The memory 12 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program instructions executed by the processor 11 .

[0121] The processor 11 implements any one of the analysis methods based on shared battery data in the above embodiments by reading and executing computer program instructions stored in the memory 12 .

[0122] In one embodiment, the electronic device may further include a communication interface 13 and a bus 10. Fig. 9 As shown, the processor 11, the memory 12, and the communication interface 13 are connected via a bus 10 and communicate with each other.

[0123] The communication interface 13 is used to realize the communication between the modules, devices, units and / or equipment in the embodiment of the present application. The communication port 13 can also realize data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage and image / data processing workstations.

[0124] The bus 10 includes hardware, software or both, and couples the components of the electronic device to each other. The bus 10 includes but is not limited to at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 10 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, bus 10 may include one or more buses. Although embodiments of the present application describe and illustrate a particular bus, the present application contemplates any suitable bus or interconnect.

[0125] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a program stored thereon, and when the program is executed by a processor, the analysis method based on shared battery data provided in the first aspect is implemented.

[0126] The readable storage medium may include but is not limited to: a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device or any suitable combination of the above.

[0127] In a possible implementation, the present invention can also be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of the analysis method based on shared battery data provided in the first aspect.

[0128] The program code for executing the present invention may be written in any combination of one or more programming languages, and may be executed entirely on a user device, partially on a user device, as an independent software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0129] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0130] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. An analysis method based on shared battery data, characterized in that: include: Collecting data from the user terminal and the shared battery terminal to obtain user terminal information and battery terminal information, wherein the user terminal information includes function requirement information; Preprocessing and characterizing the user-side information and the battery-side information in sequence to obtain initial characteristic data; Analyze the initial feature data and functional requirement information through a pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the functional requirement information; Performing de-feature processing on the predicted feature information to obtain data analysis results and feeding back to the user; The construction of the shared battery trajectory AI model includes: Select the neural network architecture according to user needs and determine the training strategy to obtain the initial AI large model; Acquire historical data of the user terminal and the shared battery terminal, and divide the historical data into a training set, a validation set, and a test set; The initial AI large model is trained according to the training strategy and the training set, and the parameters of the initial AI large model are adjusted through the verification set to obtain an intermediate AI large model; The test set is used to evaluate the performance of the intermediate AI large model, and when the evaluation result meets the preset user requirements, the shared battery trajectory AI large model is obtained; The neural network architecture includes multiple feature extraction modules and / or multiple feature restoration modules; wherein, The feature extraction module includes a first multi-head attention module, a first feedforward network module and a first normalization module; the feature restoration module includes a masked multi-head attention module, a second multi-head attention module, a second feedforward network module and a second normalization module; The multiple feature extraction modules and the multiple feature restoration modules are combined in the form of module stacking.

2. The method according to claim 1, characterized in that The preprocessing of the user-side information and the battery-side information includes: Recording the user-end information and the battery-end information as raw data, filtering the raw data, removing abnormal points and noise points, and obtaining first data; Completing the missing data in the first data by interpolation or extension to obtain second data; The second data is compressed to obtain data to be characterized.

3. The method according to claim 2, characterized in that Performing characterization processing on the data to be characterized includes: Processing the data to be characterized by normalization or label mapping table to obtain data to be processed; The data to be processed are sequentially subjected to feature extraction and feature fusion processing through unsupervised learning to obtain the initial feature data.

4. The method according to claim 1, characterized in that: The pre-built shared battery trajectory AI big model is used to analyze the feature data and user demand information to obtain predicted feature information corresponding to the user demand information, including: Processing the initial feature data through a first multi-head self-attention module to obtain a first feature vector sequence; Processing the sum of the first feature vector sequence and the initial feature data through a first feedforward network module with a residual network and a first normalization module to obtain first high-level feature information; Processing the initial feature data through a masked multi-head self-attention module to obtain a second feature vector sequence; The second feature vector sequence and the first high-level feature information are processed sequentially through a second multi-head attention module, a second feedforward network module, and a second normalization module to obtain high-level feature restoration information; The height feature restoration information is predicted according to the functional requirement information to obtain the predicted feature information.

5. The method according to claim 1, characterized in that The method of analyzing the feature data and user demand information by using a pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information further includes: Acquire a plurality of labeled user customized attributes, and fuse the user customized attributes with the initial feature data to obtain new initial feature data; Processing the new initial feature data through a first multi-head self-attention module to obtain a third feature vector sequence; Processing the sum of the third feature vector sequence and the initial feature data through a first feedforward network module with a residual network and a first normalization module to obtain second high-level feature information; The probability of the user-customized attribute to which the second highly characteristic information belongs is calculated through a linear layer and a softmax function to obtain a classification prediction probability, and the classification prediction probability is recorded as the predicted characteristic information.

6. The method according to claim 1, characterized in that When the user-side information includes map information and travel start and end point location information, the neural network architecture includes multiple feature extraction modules, and the data analysis result is a predicted trajectory path; the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information, and also includes: Processing the initial feature data through a first multi-head self-attention module to obtain a fourth feature vector sequence, wherein the initial feature data includes features of the map information and the travel start and end point location information; The sum of the fourth feature vector sequence and the initial feature data is processed sequentially by a first feedforward network module with a residual network and a first normalization module to obtain third high-level feature information, that is, the third high-level feature information is the predicted feature information; or, When the collected data also includes multimodal data, the neural network architecture includes multiple feature restoration modules; the feature data and user demand information are analyzed by the pre-built shared battery trajectory AI big model to obtain predicted feature information corresponding to the user demand information, and also includes: Performing characterization and feature extraction processing on the multimodal data in sequence to obtain multimodal feature data; Processing the initial feature data through a masked multi-head self-attention module to obtain a fifth feature vector sequence; The fifth feature vector sequence and the multimodal feature data are processed in sequence by a second multi-head attention module, a second feedforward network module and a second normalization module to obtain a feature output sequence; the feature output sequence is recorded as the predicted feature information.

7. An analysis system based on shared battery data, characterized in that: The system for executing the analysis method based on shared battery data according to any one of claims 1 to 6 comprises: A data collection unit, used to collect data from the user terminal and the shared battery terminal to obtain user terminal information and battery terminal information, wherein the user terminal information includes function requirement information; A data processing unit, used to perform preprocessing, data integration and characterization processing on the user-side information and the battery-side information respectively to obtain initial characteristic data; A large model calculation unit, used to analyze the initial feature data and functional requirement information through a pre-built shared battery trajectory AI large model to obtain predicted feature information corresponding to the functional requirement information; The result output unit is used to perform de-feature processing on the predicted feature information, obtain data analysis results and feed back to the user.

8. The analysis system according to claim 7, characterized in that The large model calculation unit is arranged in a shared battery terminal or a cloud server, and the large model calculation unit includes: A feature extraction module, configured to process the initial feature data through a first multi-head self-attention module to obtain a first feature vector sequence; and process the sum of the first feature vector sequence and the feature data through a first feedforward network module with a residual network and a first normalization module to obtain first high-level feature information; A feature restoration module is used to process the initial feature data through a masked multi-head self-attention module to obtain a second feature vector sequence; and process the second feature vector sequence and the first high-level feature information through a second multi-head attention module, a second feedforward network module and a second normalization module to obtain high-level feature restoration information; The feature prediction module is used to predict the height feature restoration information according to the functional requirement information to obtain the predicted feature information.

9. An electronic device, characterized in that: The invention comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the analysis method based on shared battery data as claimed in any one of claims 1 to 6 when executing the computer program.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the analysis method based on shared battery data according to any one of claims 1 to 6 is implemented.

Images