Maintenance reminding method and system based on motorcycle internet of vehicles time series data

Abstract

The application provides a motorcycle network timing data-based maintenance reminding method and system, which comprises the following steps: obtaining a motorcycle identification code and multi-dimensional time sequence data according to timing data packets periodically sent by a motorcycle on-board terminal, constructing the multi-dimensional time sequence data and a plurality of previously obtained multi-dimensional time sequence data into a timing data sample, inputting the timing data sample into a maintenance diagnosis model corresponding to a motorcycle model determined according to the motorcycle identification code to perform maintenance diagnosis, obtaining a maintenance diagnosis result, and determining whether to send a maintenance reminder to the motorcycle on-board terminal according to the maintenance diagnosis result. Therefore, the application can use a corresponding maintenance diagnosis model to perform maintenance diagnosis on relevant timing data sent by the motorcycle on-board terminal, so that the maintenance reminder is more accurate and more in line with the actual working conditions of the motorcycle.

Description

Technical Field

[0001] The present invention relates to the technical field of motorcycle data analysis, and in particular to a maintenance reminder method and system based on motorcycle vehicle-to-vehicle network time series data. Background Art

[0002] To stand out in the fiercely competitive market, motorcycle manufacturers and related companies are increasing their R&D and investment in connected vehicle technology. By launching motorcycles with connected vehicle functionality, they are enhancing their product's added value and competitiveness, attracting more consumers. Furthermore, connected vehicles provide companies with more opportunities for interaction and communication with users, helping them better understand user needs and improve their products and services.

[0003] The Internet of Vehicles (IoV) system collects and analyzes various motorcycle data to provide users with services such as navigation, vehicle condition monitoring, safety warnings, and personalized customization. In particular, motorcycle engines operate at relatively high workloads, resulting in shorter maintenance cycles. Furthermore, because motorcycle heat dissipation and lubrication systems are less sophisticated than those of automobiles, and given the varying driving environments and habits, even the same type of motorcycle can experience significant variations in maintenance schedules. Therefore, the current maintenance reminder system, which uses mileage intervals and time periods, fails to account for the impact of the motorcycle's actual operating conditions on maintenance timing. If maintenance is performed according to the IoV system's maintenance reminders, it is easy to miss the optimal maintenance opportunity, impacting the motorcycle's lifespan and used-vehicle value, and even increasing the probability of motorcycle failure. Therefore, to help motorcycle users accurately identify maintenance opportunities, extend their motorcycle's lifespan, and ensure driving safety, it is necessary to conduct big data analysis based on the motorcycle vehicle network to provide precise maintenance reminders. Summary of the Invention

[0004] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a maintenance reminder method based on motorcycle vehicle-to-vehicle time series data, by configuring a pre-trained maintenance diagnosis model for each motorcycle model, and using the corresponding maintenance diagnosis model to perform targeted maintenance diagnosis on the relevant time series data sent by the motorcycle's on-board terminal, so that the maintenance reminder is more accurate and more in line with the maintenance needs of the motorcycle's actual working conditions.

[0005] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0006] The maintenance reminder method based on motorcycle vehicle network time series data includes the following steps:

[0007] S1: Parse the time series data packets periodically sent by the motorcycle onboard terminal to obtain the motorcycle identification code and multi-dimensional time series data;

[0008] S2: constructing the currently obtained multidimensional time series data and the previously obtained multiple multidimensional time series data into a time series data sample;

[0009] S3: Determine the motorcycle model based on the motorcycle identification code, input the sequence data sample into a maintenance diagnosis model corresponding to the determined motorcycle model, perform maintenance diagnosis, and output a maintenance diagnosis result; wherein each maintenance diagnosis model is trained using the last sequence data sample constructed before each maintenance in the historical maintenance records of a motorcycle model;

[0010] S4: Determine whether to send a maintenance reminder to the motorcycle onboard terminal based on the maintenance diagnosis result.

[0011] According to a specific embodiment, the present invention provides a maintenance reminder method based on motorcycle vehicle network time series data. In step S1, after obtaining the multidimensional time series data, the multidimensional time series data is placed in a FIFO cache queue established when registering the motorcycle identification code; wherein the queue length of the FIFO cache queue is adapted to the time window size of the time series data sample.

[0012] According to a specific embodiment, the maintenance reminder method based on motorcycle Internet of Vehicles time series data provided by the present invention, in step S2, the method of constructing the currently obtained multidimensional time series data and the previously obtained multiple multidimensional time series data into a time series data sample is:

[0013] When new multidimensional time series data is added to the FIFO cache queue, all the multidimensional time series data cached in the FIFO cache queue are merged according to the same dimension to obtain a time series data set corresponding to different dimensions;

[0014] After processing the time series data sets corresponding to different dimensions into a data set with data features of a mean of 0 and a standard deviation of 1, principal component analysis is performed on each of the time series data sets to obtain the principal component time series data corresponding to different dimensions, and then the principal component time series data corresponding to all dimensions are used as the time series data samples.

[0015] According to a specific embodiment, the maintenance reminder method based on motorcycle vehicle networking time series data provided by the present invention, in step S1, the time series data packet periodically sent by the motorcycle vehicle-mounted terminal is parsed to obtain the multi-dimensional time series data in the following manner:

[0016] Extracting time series data of power system state related data, brake system state related data, driving behavior state related data, and driving environment related data according to the data identifier of each time series data in the time series data packet;

[0017] The time series data of the power system state association data, the brake system state association data, the driving behavior state association data and the driving environment association data are respectively assigned different dimension labels to obtain the multidimensional time series data.

[0018] According to a specific embodiment, the present invention provides a maintenance reminder method based on motorcycle vehicle-to-vehicle time series data, wherein the maintenance diagnosis model is constructed based on a long short-term memory network; and the maintenance diagnosis model includes: an input layer having four neurons, a hidden layer having several LSTM units, and an output layer having several neurons; wherein the diagnosis results of the maintenance diagnosis model are the maintenance probabilities of each specific maintenance item of the power system and the maintenance probabilities of each specific maintenance item of the braking system.

[0019] According to a specific embodiment, the maintenance reminder method based on motorcycle vehicle networking time series data provided by the present invention, in step S4, determines whether to issue a maintenance reminder to the motorcycle vehicle-mounted terminal according to the maintenance diagnosis result:

[0020] If the maintenance probability of any specific maintenance item of the power system and the braking system reaches a set threshold, a maintenance reminder for the specific maintenance item will be issued.

[0021] Based on the same inventive concept, the present invention also provides a maintenance reminder system based on motorcycle vehicle network time series data, which includes:

[0022] The motorcycle onboard terminal is used to periodically send time-sequence data packets to the cloud server;

[0023] The cloud server interacts with the motorcycle onboard terminal to implement the maintenance reminder method based on motorcycle vehicle network time series data provided by the present invention.

[0024] The present invention also provides a computer-readable storage medium having one or more programs stored thereon, which, when executed by one or more processors, implement the maintenance reminder method based on motorcycle vehicle network time series data provided by the present invention.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. The present invention provides a maintenance reminder method based on motorcycle vehicle-to-vehicle network time series data. After obtaining the motorcycle vehicle identification code and multidimensional time series data based on the time series data packets periodically transmitted by the motorcycle's onboard terminal, this multidimensional time series data is combined with multiple previously obtained multidimensional time series data to form a time series data sample. This time series data sample is then input into a maintenance diagnosis model corresponding to the motorcycle model determined based on the motorcycle's vehicle identification code to perform maintenance diagnosis, obtain a maintenance diagnosis result, and determine whether to issue a maintenance reminder to the motorcycle's onboard terminal based on the maintenance diagnosis result. Therefore, the present invention can use the corresponding maintenance diagnosis model to perform maintenance diagnosis on the relevant time series data transmitted by the motorcycle's onboard terminal in a targeted manner, making the maintenance reminder more accurate and more in line with the maintenance needs of the motorcycle's actual operating conditions.

[0027] 2. The maintenance reminder method based on motorcycle Internet of Vehicles time series data provided by the present invention utilizes the multidimensional time series data obtained by caching the FIFO cache queue, and when new multidimensional time series data is added to the FIFO cache queue, all the multidimensional time series data in the FIFO cache queue are merged to obtain time series data sets of different dimensions. Finally, principal component analysis is performed on the time series data sets of each dimension to obtain corresponding principal component time series data, and then obtain time series data samples; therefore, the present invention can reduce the complexity of the maintenance diagnosis model and improve the accuracy of the maintenance diagnosis results and the convergence speed of the maintenance diagnosis model during training by integrating the multidimensional time series data accumulated for a period of time and extracting valuable data.

[0028] 3. The present invention provides a maintenance reminder method based on motorcycle vehicle networking time series data. The multi-dimensional time series data includes time series data of brake system status correlation data, driving behavior status correlation data, and driving environment correlation data. In this way, by introducing the time series data of driving behavior status correlation data and driving environment correlation data, the actual working condition of the motorcycle is incorporated into the maintenance diagnosis, so that the maintenance diagnosis results are more in line with the actual working condition of the motorcycle, further ensuring the accuracy of the maintenance diagnosis and meeting the maintenance needs of the actual working condition of the motorcycle. Description of the drawings:

[0029] Figure 1 Schematic diagram of the process of the present invention;

[0030] Figure 2 This is a schematic diagram of the processing flow of multidimensional time series data of the method of the present invention;

[0031] Figure 3 Schematic diagram of the structure of the system of the present invention. DETAILED DESCRIPTION

[0032] The present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments. However, this should not be construed as limiting the scope of the present invention to the following embodiments, as all technologies implemented based on the present invention fall within the scope of the present invention.

[0033] like Figure 1 As shown, the present invention provides a maintenance reminder method based on motorcycle vehicle network time series data, which includes the following steps:

[0034] S1: Parse the time series data packets periodically sent by the motorcycle onboard terminal to obtain the motorcycle identification code and multi-dimensional time series data;

[0035] S2: constructing the currently obtained multidimensional time series data and the previously obtained multiple multidimensional time series data into a time series data sample;

[0036] S3: Determine the motorcycle model based on the motorcycle identification code, input the sequence data sample into a maintenance diagnosis model corresponding to the determined motorcycle model, perform maintenance diagnosis, and output a maintenance diagnosis result; wherein each maintenance diagnosis model is trained using the last sequence data sample constructed before each maintenance in the historical maintenance records of a motorcycle model;

[0037] S4: Determine whether to send a maintenance reminder to the motorcycle onboard terminal based on the maintenance diagnosis result.

[0038] Specifically, the present invention provides a maintenance reminder method based on motorcycle IoV time-series data. This method is primarily a value-added service provided by motorcycle manufacturers, facilitating precise maintenance for motorcycle users and providing targeted maintenance services. Each time a motorcycle manufacturer develops and produces a new model, it establishes a maintenance diagnosis model corresponding to that model and uses test data and maintenance and repair data from the R&D phase for initial training to produce a preliminary maintenance diagnosis model. As subsequent motorcycles of that model are sold, the historical maintenance records and associated time-series data samples increase, leading to retraining of the maintenance diagnosis model for iterative optimization. Furthermore, after the motorcycle manufacturer ships the motorcycles for quality inspection, it stores a record of the motorcycle's identification code. Once the motorcycle is sold to a user and the user registers their identity using the motorcycle identification code, the manufacturer activates a cloud storage space based on the motorcycle identification code. This cloud storage space uses a FIFO cache queue mechanism to store multidimensional time-series data generated based on time-series data packets periodically transmitted by the motorcycle's onboard terminal.

[0039] Furthermore, in step S1, after obtaining the multidimensional time series data, the multidimensional time series data is placed into a FIFO buffer queue established when registering the motorcycle identification code; wherein the queue length of the FIFO buffer queue is adapted to the time window size of the time series data samples. In implementation, both the queue length of the FIFO buffer queue and the time window size of the time series data samples can be adjusted based on iterative optimization of the maintenance diagnosis model.

[0040] like Figure 2 As shown, in step S2, the method of constructing the currently obtained multidimensional time series data and the previously obtained multiple multidimensional time series data into a time series data sample is:

[0041] When new multidimensional time series data is added to the FIFO cache queue, all the multidimensional time series data cached in the FIFO cache queue are merged according to the same dimension to obtain a time series data set corresponding to different dimensions;

[0042] After processing the time series data sets corresponding to different dimensions into a data set with data features of a mean of 0 and a standard deviation of 1, principal component analysis is performed on each of the time series data sets to obtain the principal component time series data corresponding to different dimensions, and then the principal component time series data corresponding to all dimensions are used as the time series data samples.

[0043] During implementation, the time series data sets corresponding to different dimensions are respectively processed into a data set with data characteristics of a mean of 0 and a standard deviation of 1, and the Z-score normalization method is used to convert the data points contained in the time series data set into a distribution with a mean of 0 and a standard deviation of 1.

[0044] Furthermore, in step S1, the time series data packets periodically sent by the motorcycle onboard terminal are parsed to obtain the multi-dimensional time series data in the following manner:

[0045] Extracting time series data of power system state related data, brake system state related data, driving behavior state related data, and driving environment related data according to the data identifier of each time series data in the time series data packet;

[0046] The time series data of the power system state association data, the brake system state association data, the driving behavior state association data and the driving environment association data are respectively assigned different dimension labels to obtain the multidimensional time series data.

[0047] During implementation, the power system status-related data of a motorcycle include: engine data (speed: reflects the number of engine rotations per minute, and is a key indicator for measuring the engine's working intensity and operating status; temperature: includes coolant temperature or oil temperature; oil pressure: appropriate oil pressure can ensure good circulation of oil in various parts of the engine, and insufficient oil pressure may cause increased wear of components; intake volume: intake volume directly affects the engine's combustion efficiency and power output, and is related to factors such as throttle opening and the smoothness of the air filter; exhaust temperature: reflects the temperature of exhaust gas after engine combustion, and excessively high exhaust temperature may indicate incomplete combustion or excessive engine load), fuel system data (fuel pressure: ensures that fuel can be stably supplied to the engine for combustion; fuel injection volume of the fuel injector: matches the engine's operating conditions (such as speed and load)), and transmission system data (clutch status: includes the degree of clutch engagement and wear; transmission oil temperature: transmission oil temperature affects the smoothness of gear shifting and transmission efficiency).

[0048] The motorcycle's brake system status-related data includes: brake pressure data (front wheel brake pressure and rear wheel brake pressure), brake pad wear data (front wheel brake pad thickness and rear wheel brake pad thickness), brake disc status data, and brake fluid data (brake fluid level: a low brake fluid level will cause air to enter the brake system, affecting the braking effect; brake fluid water content: brake fluid is hygroscopic and will absorb water over time and in changing usage environments).

[0049] Motorcycle driving behavior status-related data includes: operation data (throttle opening: reflecting the driver's control over power output; brake operation force and frequency: recording the force and number of times the driver presses the brake pedal or pulls the brake handle; clutch operation: including the frequency of clutch disengagement and engagement, and the duration of each operation; gear shift operation data: recording information such as the timing of gear shifts and the order of gear changes), driving style data (acceleration mode: determined by the speed change rate and throttle opening changes; deceleration mode: determined by the speed change rate and brake operation force changes; vehicle speed stability: observing the fluctuation of vehicle speed over a period of time).

[0050] The motorcycle driving environment-related data includes: road data (road type: different road types have different road conditions, including highways, urban roads, rural roads, mountain roads, etc.; road surface conditions: divided into dry, wet, flooded, icy, muddy, etc.; slope information: records the slope size and direction of the road), meteorological data (weather conditions: such as sunny, cloudy, rainy, snowy, foggy, etc. Different weather conditions have an impact on driving safety and vehicle performance; temperature: low temperatures may affect battery performance, tire pressure and rubber elasticity; high temperatures may cause problems such as engine heat dissipation difficulties and fuel evaporation; latitude and longitude coordinates and altitude: different altitudes affect engine power, torque output, oil pressure and tire performance, etc.)

[0051] Considering that each dimension of the time series data set also contains a variety of different types of data, it cannot be directly input into the maintenance diagnosis model. Therefore, it is necessary to further process the time series data set of each dimension. Since the principal component analysis method (PCA) has the property of dimensionality reduction, it can not only retain most of the information of the original data, but also reduce the amount and complexity of subsequent data analysis and modeling, thereby improving computational efficiency. Therefore, after completing the standardization processing of the time series data set of each dimension using the Z-score standardization method, the principal component analysis of each dimension of the time series data set is further performed. The analysis process is as follows:

[0052] 1. The calculation of the covariance matrix is ​​the key step: Let the multi-dimensional time series data matrix be X n×p (t), where n is the length of the time series, p is the dimension of the data, and t represents time; the covariance matrix is:

[0053]

[0054] Among them, x ki (t) is the data of the i-th dimension at the k-th time point, is the data of the i-th dimension. By calculating the covariance matrix, we can obtain the correlation information between the dimensions, laying the foundation for the subsequent extraction of principal components.

[0055] 2. Solving the eigenvalues ​​and eigenvectors of the covariance matrix is ​​the core step of PCA: solving the characteristic equation ∑-λI | =0, we can get p eigenvalues ​​λ1,λ2,λ3···λ p , arranged in descending order, these eigenvalues ​​represent the variance explained by each principal component. i , by solving the linear equations (∑-λ i I)v i =0, we can get the corresponding eigenvector vi , the eigenvector represents the direction of the principal component, that is, the projection direction of the original data on the new coordinate axis. In multi-dimensional time series data, these eigenvectors convert the original time series data into a set of new principal component time series, thereby obtaining principal component time series data corresponding to different dimensions.

[0056] Specifically, the present invention provides a maintenance reminder method based on motorcycle vehicle-to-vehicle time series data, wherein the maintenance diagnosis model is constructed based on a long short-term memory network; moreover, the maintenance diagnosis model includes: an input layer with four neurons, a hidden layer with several LSTM units, and an output layer with several neurons; wherein the diagnosis results of the maintenance diagnosis model are the maintenance probabilities of each specific maintenance item of the power system and the maintenance probabilities of each specific maintenance item of the braking system.

[0057] During implementation, the maintenance diagnosis model was trained by first determining the architecture of the long short-term memory network, which consists of an input layer with 4 neurons, a hidden layer with approximately 500 LSTM units, and an output layer with 5 neurons (considering only core maintenance items such as changing the engine oil and oil filter, cleaning or replacing the air filter, checking and adjusting the chain, checking the brake system, and checking the electrical system).

[0058] Next, we collect training samples. As the number of users continues to increase, we can collect more historical maintenance records from users. Of course, we can also collect the last time series data sample constructed by the user before a specific maintenance project. In this way, as time goes by, the collected training samples gradually approach the needs of maintenance projects under the user's actual working conditions.

[0059] After collecting training samples, training begins. The LSTM units in the hidden layer process the four input features. Using their internal gating mechanisms (input gate, forget gate, and output gate), they control the flow and memory of information. These units can learn the long-term dependencies between these features in specific maintenance projects. To convert the outputs into probabilistic form, a Softmax activation function is typically used in the output layer. The Softmax function converts the output values ​​of the output layer neurons into a probability distribution, such that the sum of the two outputs is 1.

[0060] During the training process, the LSTM model is trained by defining a suitable loss function (such as the cross-entropy loss function, which is used to measure the difference between the predicted probability and the true probability) and selecting an appropriate optimization algorithm (such as Adam, RMSProp, etc.). During the training process, the model will continuously adjust its weights to minimize the loss function. At the same time, validation data can be used to prevent overfitting, and the performance of the model can be optimized by adjusting the hyperparameters of the model (such as the number of hidden layers, the number of LSTM units, the learning rate, etc.). It should be pointed out that the present invention does not make any technical contribution to the long short-term memory network (LSTM), but rather uses the long short-term memory network (LSTM) to implement maintenance diagnosis for motorcycles and output the probability distribution of specific maintenance items.

[0061] Specifically, the maintenance reminder method based on motorcycle vehicle networking time series data provided by the present invention, in step S4, determines whether to issue a maintenance reminder to the motorcycle vehicle-mounted terminal according to the maintenance diagnosis result:

[0062] If the maintenance probability of any specific maintenance item of the power system and the braking system reaches a set threshold, a maintenance reminder for the specific maintenance item will be issued.

[0063] like Figure 3 As shown, the present invention provides a maintenance reminder system based on motorcycle vehicle network time series data, which includes:

[0064] The motorcycle onboard terminal is used to periodically send time-sequence data packets to the cloud server;

[0065] The cloud server interacts with the motorcycle onboard terminal to implement the maintenance reminder method based on motorcycle vehicle network time series data provided by the present invention.

[0066] During implementation, the cloud server built several maintenance and diagnosis models corresponding to each motorcycle model. Furthermore, a cloud storage space was activated for each motorcycle identification code. This cloud storage space used a FIFO cache queue mechanism to store multidimensional time series data generated from the time-series data packets periodically transmitted by the motorcycle's onboard terminal. Furthermore, the cloud server performed standardization, data set merging, and principal component analysis on the multidimensional time series data.

[0067] In addition, the present invention also provides a computer-readable storage medium on which one or more programs are stored. When the one or more programs are executed by one or more processors, the maintenance reminder method based on motorcycle vehicle network time series data provided by the present invention is implemented.

[0068] It should be understood that the system disclosed herein can be implemented in other ways. For example, the module division described above is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, the communication connections between modules can be through interfaces, indirect coupling or communication connections between devices or units, and can be electrical or otherwise.

[0069] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each module may exist physically separately, or two or more modules may be integrated into a single processing unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0070] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0071] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A maintenance reminder method based on motorcycle vehicle networking time series data, characterized in that: The following steps are involved: S1: Parsing the time series data packets periodically sent by the motorcycle on-board terminal to obtain a motorcycle identification code and multi-dimensional time series data; wherein, based on the data identifier of each time series data item in the time series data packet, extracting the time series data of power system state correlation data, brake system state correlation data, driving behavior state correlation data, and driving environment correlation data, and assigning different dimension labels to each of them to obtain the multi-dimensional time series data; S2: constructing the currently obtained multidimensional time series data and the previously obtained multiple multidimensional time series data into a time series data sample; S3: Determine the motorcycle model based on the motorcycle identification code, input the time series data sample into a maintenance diagnosis model corresponding to the determined motorcycle model, perform maintenance diagnosis, and output a maintenance diagnosis result; wherein each maintenance diagnosis model is trained using the last time series data sample constructed before each maintenance in the historical maintenance records of a motorcycle model; the maintenance diagnosis model is constructed based on a long short-term memory network; S4: Determining whether to send a maintenance reminder to the motorcycle onboard terminal based on the maintenance diagnosis result; Wherein, in step S1, after obtaining the multidimensional time series data, the multidimensional time series data is placed in a FIFO cache queue established when registering the motorcycle identification code; wherein, the queue length of the FIFO cache queue is adapted to the time window size of the time series data sample; in step S2, the currently obtained multidimensional time series data and the multiple multidimensional time series data obtained previously are constructed into a time series data sample in the following manner: when new multidimensional time series data is added to the FIFO cache queue, all the multidimensional time series data cached in the FIFO cache queue are merged according to the same dimension to obtain time series data sets corresponding to different dimensions; after processing the time series data sets corresponding to different dimensions into a data set with a data feature of a mean of 0 and a standard deviation of 1, principal component analysis is performed on each time series data set to obtain principal component time series data corresponding to different dimensions, and then the principal component time series data corresponding to all dimensions are used as the time series data samples.

2. The maintenance reminder method based on motorcycle vehicle networking time series data according to claim 1, characterized in that: The maintenance diagnosis model includes: an input layer with four neurons, a hidden layer with multiple LSTM units, and an output layer with multiple neurons; wherein the diagnosis results of the maintenance diagnosis model are the maintenance probabilities of each specific maintenance item of the power system and the maintenance probabilities of each specific maintenance item of the braking system.

3. The maintenance reminder method based on motorcycle vehicle networking time series data according to claim 2, characterized in that: In step S4, based on the maintenance diagnosis result, it is determined whether to send a maintenance reminder to the motorcycle onboard terminal: If the maintenance probability of any specific maintenance item of the power system and the braking system reaches a set threshold, a maintenance reminder for the specific maintenance item will be issued.

4. A maintenance reminder system based on motorcycle vehicle network time series data, characterized in that: include: The motorcycle onboard terminal is used to periodically send time-sequence data packets to the cloud server; The cloud server interacts with the motorcycle on-board terminal to implement the maintenance reminder method based on motorcycle vehicle network time series data as described in any one of claims 1 to 3.

5. A computer-readable storage medium, characterized in that One or more programs are stored thereon, characterized in that when the one or more programs are executed by one or more processors, the maintenance reminder method based on motorcycle vehicle network time series data as described in any one of claims 1 to 3 is implemented.

Images