Riding model construction method and device, electronic equipment and storage medium

Abstract

The invention provides a riding model construction method and device, electronic equipment and a storage medium. The method comprises the following steps: selecting a target open source model as a basic framework; obtaining riding data, performing data preprocessing, and constructing a riding data set based on the riding data after data preprocessing; pre-training the target open source model based on the riding data set to obtain a trained riding model; and obtaining a specific task data set corresponding to the user demand, and adjusting the trained riding model based on the specific task data set to obtain an adjusted riding model. The invention provides a construction scheme of a customized riding model, so that the riding model better meets user requirements, the applicability and accuracy of the riding model are improved, the riding model is adjusted according to different requirements, the performance of the riding model is improved, and personalized riding experience is provided for a user through the adjusted riding model.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a riding model construction method, device, electronic device and storage medium. Background Art

[0002] With the development of intelligent transportation systems, cycling, as an environmentally friendly and healthy way of travel, is becoming more and more popular. However, the existing cycling models have several problems and limitations in practical applications, which seriously restrict the optimization of cycling experience and the further development of intelligent transportation systems.

[0003] Currently, most existing models rely on general data sets for training, which makes it impossible for the models to accurately capture the characteristics and needs of specific cycling scenarios. For example, the cycling environment (such as road conditions, traffic rules, and cycling culture) in different cities or regions is significantly different, and general models often cannot adapt to these differences and thus cannot provide personalized cycling suggestions; in addition, existing cycling models often lack sufficient user feedback mechanisms, resulting in a gradual decline in model performance over time.

[0004] The preceding description is intended to provide general background information and does not necessarily constitute prior art. Summary of the invention

[0005] In response to the above technical problems, the present application provides a riding model construction method, device, electronic device and storage medium to solve the problems of poor applicability and low accuracy of existing riding assistance models, so that the riding assistance model can better meet user needs, effectively improve the performance of the riding assistance model, and provide users with a personalized riding experience.

[0006] In order to solve the above technical problems, the present application provides a method for constructing a cycling model, comprising:

[0007] Select the target open source model as the basic framework;

[0008] Obtain riding data and perform data preprocessing, and build a riding data set based on the preprocessed riding data;

[0009] Pre-training the target open source model based on the cycling data set to obtain a trained cycling model;

[0010] A specific task data set corresponding to the user's needs is obtained, and the trained cycling model is adjusted based on the specific task data set to obtain an adjusted cycling model.

[0011] Furthermore, in some embodiments of the present application, the acquiring of riding data and performing data preprocessing, and constructing a riding data set based on the riding data after the data preprocessing, includes:

[0012] Acquire riding data through various sensors;

[0013] Performing data preprocessing on the riding data, wherein the data preprocessing includes at least one of data cleaning, removing outliers, and removing incomplete records;

[0014] Label the riding data after data preprocessing to obtain riding data marked as different types of riding scenes;

[0015] A cycling data set is constructed based on the cycling data of the different types of cycling scenes and their corresponding scene classification labels.

[0016] Further, in some embodiments of the present application, pre-training the target open source model based on the cycling data set to obtain a trained cycling model includes:

[0017] Based on the user requirements, setting the loss function and training parameters of the target open source model;

[0018] Extracting riding features corresponding to the target open source model based on the target open source model, and determining corresponding riding analysis results based on the riding features;

[0019] Calculating the difference loss of the riding analysis result based on the loss function, and adjusting the model parameters of the target open source model;

[0020] The target open source model is continuously trained based on the adjusted model parameters until the training conditions are met, thereby obtaining a trained cycling model.

[0021] Furthermore, in some embodiments of the present application, the step of obtaining a specific task data set corresponding to the user's needs and adjusting the trained riding model based on the specific task data set to obtain the adjusted riding model includes:

[0022] Responding to user needs and determining specific tasks corresponding to the user needs;

[0023] Acquire a specific task data set corresponding to the specific task;

[0024] The network structure and weight of the riding model are adjusted based on the specific task data to obtain an adjusted riding model, so as to perform the specific task according to the adjusted riding model.

[0025] Furthermore, in some embodiments of the present application, the method further includes:

[0026] Recommending a route to the user based on the cycling data by using the cycling model to generate a recommended cycling route;

[0027] Obtaining user feedback on the recommended cycling route;

[0028] The riding model is adjusted based on the feedback result to obtain an adjusted riding model.

[0029] Furthermore, in some embodiments of the present application, the method further includes:

[0030] Performing a riding safety analysis on the user according to the riding data by using the riding model to obtain a riding safety analysis result;

[0031] Based on the riding safety analysis result, corresponding safety recommendations are generated.

[0032] Furthermore, in some embodiments of the present application, the method further includes:

[0033] Performing a cycling health analysis on the user according to the cycling data by using the cycling model to obtain a health analysis result;

[0034] Based on the health analysis results, corresponding health recommendations are generated.

[0035] Accordingly, the present application also provides a riding model building device, comprising:

[0036] The selection module is used to select the target open source model as the basic framework;

[0037] The data module is used to obtain and preprocess the riding data, and to construct a riding data set based on the preprocessed riding data;

[0038] A training module, used for pre-training the target open source model based on the cycling data set to obtain a trained cycling model;

[0039] The adjustment module is used to obtain a specific task data set corresponding to user needs, and adjust the trained riding model based on the specific task data set to obtain an adjusted riding model.

[0040] The present application also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the steps of the riding model building method as described above when executing the program.

[0041] The present application also provides a storage medium storing a computer program that can be loaded by a processor and execute the riding model building method as described above.

[0042] Implementing the embodiments of the present application has the following beneficial effects:

[0043] As described above, the present application provides a method, device, electronic device and storage medium for constructing a cycling model, and the method for constructing a cycling model includes: selecting a target open source model as a basic framework; obtaining cycling data and performing data preprocessing, and constructing a cycling data set based on the cycling data after data preprocessing; pre-training the target open source model based on the cycling data set to obtain a trained cycling model; obtaining a specific task data set corresponding to user needs, and adjusting the trained cycling model based on the specific task data set to obtain an adjusted cycling model. The present application proposes a construction scheme for a customized cycling model, so that the cycling model is more in line with user needs, improves the applicability and accuracy of the cycling model, and adjusts the cycling model according to different needs to improve the performance of the cycling model, thereby providing users with a personalized cycling experience through the adjusted cycling model. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0045] Figure 1 It is a flowchart of a riding model construction method provided in an embodiment of the present application;

[0046] Figure 2 is another flow chart of the riding model construction method provided in an embodiment of the present application;

[0047] Figure 3 is a schematic diagram of the structure of a riding model building device provided in an embodiment of the present application;

[0048] Figure 4 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0049] The realization of the purpose, functional features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. The above-mentioned drawings have shown clear embodiments of this application, which will be described in more detail later. These drawings and textual descriptions are not intended to limit the scope of the concept of this application in any way, but to illustrate the concept of this application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0050] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0051] It should be noted that, in this article, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, components, features, and elements with the same name in different embodiments of the present application may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context of the specific embodiment.

[0052] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present application, and have no specific meanings. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0054] With the increase of urban traffic pressure and people's pursuit of a healthy lifestyle, cycling has gradually been widely praised as an environmentally friendly and beneficial way of travel. However, despite the many benefits brought by the popularity of cycling, cyclists still face many challenges in the actual cycling process. For example, the complexity of urban traffic, changeable weather conditions, personal health and safety issues have a significant impact on the cycling experience.

[0055] On the technical level, existing cycling assistance systems often rely on standardized algorithms and data models, which are unable to cope with personalized needs and dynamic environmental changes. Specifically, traditional cycling navigation systems may not be able to provide customized route suggestions based on the rider's personal preferences and real-time traffic information. In addition, existing systems also have limitations in data collection and processing, often ignoring the rider's behavioral patterns and environmental factors, resulting in the inability to provide accurate cycling suggestions.

[0056] In addition, existing riding assistance systems are not perfect in terms of user interaction and feedback mechanisms. Changes in user needs and preferences cannot be fed back to the system in a timely manner, resulting in the system being unable to continuously learn and optimize to adapt to the changing needs of users. These problems limit the potential of riding assistance systems in improving the riding experience.

[0057] In response to the above problems, the present application proposes a cycling model construction method, device, electronic device and storage medium, providing a more intelligent, flexible and personalized cycling model construction solution. By introducing advanced data processing technology and machine learning algorithms, it can dynamically adapt to changes in the cycling environment and provide a personalized cycling experience, providing cyclists with more intelligent and accurate cycling assistance services.

[0058] See also Figure 1 , Figure 1 : is a flow chart of a method for constructing a cycling model provided in an embodiment of the present application. A method for constructing a cycling model provided in this embodiment specifically includes the following steps:

[0059] S1. Select the target open source model as the basic framework;

[0060] Specifically, for step S1, an open source model that meets the characteristics of the user's riding data is selected as the basic framework. The open source model has good generalization and scalability so as to be able to adapt to the characteristics of the riding data and subsequent customization requirements.

[0061] S2. Obtain the riding data and perform data preprocessing, and construct a riding data set based on the riding data after data preprocessing;

[0062] Specifically, for step S2, data during riding is collected through a variety of sensors (such as GPS, accelerometer, heart rate monitor, etc.); then data preprocessing is performed, including data cleaning (removing noise and irrelevant information), removing outliers (identifying and processing outliers in the data), and removing incomplete records (ensuring data integrity).

[0063] In a specific embodiment, the user's physiological parameters, including heart rate, blood oxygen saturation, body temperature, respiratory rate, etc., are monitored in real time through smart wearable devices such as smart watches, heart rate belts or health trackers. These physiological data are essential for assessing the user's health and riding intensity. The user's riding speed, riding acceleration and riding posture during riding are obtained in real time through riding tools; sensors on riding tools, such as speedometers, accelerometers and gyroscopes, can provide detailed information about the user's riding performance. These sensors can measure riding speed, acceleration changes, and posture adjustments during riding, helping to analyze the user's riding skills and potential safety risks. The user's environmental data during riding is obtained in real time through smart mobile terminals; smart mobile terminals such as smart phones or tablets can use built-in sensors or connect external sensors to collect environmental data, including temperature, humidity, air pressure, altitude, etc.

[0064] S3. Pre-train the target open source model based on the cycling dataset to obtain a trained cycling model;

[0065] Specifically, for step S3, the preprocessed riding data is labeled and marked as different riding scenes (such as urban riding, mountain riding, etc.). Based on the data of different types of riding scenes and their corresponding scene classification labels, a riding data set for model training is constructed.

[0066] S4. Obtain a specific task data set corresponding to the user's needs, and adjust the trained riding model based on the specific task data set to obtain an adjusted riding model;

[0067] Specifically, for step S4, the selected open source model is pre-trained based on the cycling data set so that the model can learn cycling-related features. According to user needs, a specific task data set is obtained, and the pre-trained model is adjusted, including the optimization of the network structure and weights, to adapt to the specific task.

[0068] It can be seen that the cycling model construction method provided in this embodiment, through pre-training and adjustment, enables the model to better understand and predict cycling scenarios, thereby improving the accuracy and applicability of the model in practical applications; the pre-training process enables the model to learn a wider range of cycling features, thereby enhancing the generalization ability of the model in different cycling scenarios; through data preprocessing and labeling, the quality of the data set is improved, making model training more efficient, and reducing the time and resource consumption of invalid data processing; through the adjustment of specific task data sets, the model can better meet the personalized needs of users, such as route recommendation, safety analysis, etc.; the model can be adjusted according to user feedback, so that users can participate in the optimization process of the model, thereby improving user satisfaction and the practicality of the model.

[0069] In a specific embodiment, Figure 2 As shown, this embodiment also provides a specific implementation method of the riding model construction method, and the specific process is as follows:

[0070] S101. Select a suitable open source large model as the basic model;

[0071] S102. Data collection: Collect a large amount of cycling data, including speed, heart rate, altitude, terrain, etc., as well as corresponding professional suggestions;

[0072] S103. Training model: training a basic model on the collected data so that it can understand and analyze the riding data;

[0073] S104. Tune the expected results of specific tasks;

[0074] S105. Deployed on the client side based on the TensorFlow Lite framework;

[0075] S106. Cyclic tuning based on users’ daily usage data.

[0076] This embodiment selects a suitable open source model, prepares cleaned and labeled professional cycling data and cycling knowledge base data; performs pre-training on the cycling data so that the model can learn cycling-related features; and performs fine-tuning for specific tasks based on pre-training. For example, if you want to develop a cycling route recommendation system, you can use the pre-trained model to further train on the labeled cycling data set so that the model can accurately predict the cycling routes that users may like; collect user feedback for tuning and updating, adjust model parameters, data processing methods or model architecture, and gradually improve model performance.

[0077] Further, in some embodiments, step S2 of "obtaining riding data and performing data preprocessing, and constructing a riding data set based on the riding data after the data preprocessing" may specifically include:

[0078] S21. Obtain riding data through various sensors;

[0079] S22. Performing data preprocessing on the riding data, the data preprocessing including at least one of data cleaning, removing outliers, and removing incomplete records;

[0080] S23. Labeling the pre-processed riding data to obtain riding data marked as different types of riding scenes;

[0081] S24. Construct a cycling dataset based on the cycling data of different types of cycling scenarios and their corresponding scene classification labels.

[0082] Specifically, for step S2, a variety of sensors (such as GPS, speedometer, heart rate monitor, cadence sensor, etc.) are used to collect various data of the rider during the riding process. Sensor data includes but is not limited to location information, speed, heart rate, cadence, tilt angle, acceleration, etc. Then the riding data is preprocessed, including: data cleaning, removing noise and inaccurate records in sensor data to ensure data accuracy; removing outliers, identifying and processing abnormal points in the data, such as sudden speed changes or unreasonable heart rate readings, to reduce interference during model training; removing incomplete records, ensuring that all records are complete, and eliminating or completing data that lacks key information. The preprocessed riding data is classified and different riding scenes are marked, such as urban riding, mountain riding, indoor training, etc. The marking can also include riding difficulty, weather conditions, road conditions, etc., to increase the richness and practicality of the data set. The preprocessed and labeled data is organized into a structured riding data set for model training and testing. The data set should contain rich scenes and conditions to ensure that the model can learn diverse riding characteristics.

[0083] This embodiment ensures that the data used for model training is accurate and reliable through the data preprocessing step, thereby improving the effect of model training; by labeling different riding scenes and conditions, the model can learn a wider range of riding characteristics, thereby enhancing the generalization ability of the model in different riding environments; through effective data preprocessing and labeling, invalid calculations in the model training process are reduced, and training efficiency is improved; the constructed riding data set can support the learning of multiple tasks, such as route recommendation, safety analysis, health advice, etc., which increases the practicality of the model; through detailed data labeling, the model can better understand the user's riding habits and preferences, and provide users with more personalized services; detailed data labeling contributes to the interpretability of the model, making the model's prediction and analysis results easier to be understood by users and professionals.

[0084] Furthermore, in some embodiments, step S3 “pre-training the target open source model based on the cycling data set to obtain a trained cycling model” may specifically include:

[0085] S31. Based on user needs, set the loss function and training parameters of the target open source model;

[0086] S32. Extracting riding features corresponding to the target open source model based on the target open source model, and determining corresponding riding analysis results based on the riding features;

[0087] S33. Calculate the difference loss of the riding analysis result based on the loss function, and adjust the model parameters of the target open source model;

[0088] S34. Continue to train the target open source model based on the adjusted model parameters until the training conditions are met to obtain a trained cycling model.

[0089] Specifically, for the model training process, first, select an open source model suitable for cycling data analysis, which should have good learning ability and adaptability. According to user needs and the characteristics of cycling data, set appropriate loss functions, such as cross entropy loss, mean square error, etc., to optimize the learning objectives of the model. Set training parameters, including learning rate, batch size, number of iterations, etc., to ensure that the model can effectively learn the characteristics of cycling data. Then, use the target open source model to identify and extract key features in cycling data, such as cycling speed, heart rate changes, pedaling frequency, etc., analyze the extracted features, and determine their relationship with cycling scenes and cycling effects. Based on the set loss function, calculate the difference loss between the model prediction results and the actual cycling data. According to the loss calculation results, adjust the model parameters and optimize the model performance. Under the premise of meeting the training conditions, continue to train the model until the model performance reaches a satisfactory level. Finally, a trained cycling model is obtained, which can effectively process and analyze cycling data.

[0090] Specifically, for feature extraction, the corresponding original riding features are extracted from the riding data set, and the corresponding derived riding features are calculated based on the original riding features. The extraction of original riding features includes location information, physiological data, environmental data and equipment data, for example, obtaining the longitude and latitude of the rider for tracking the riding path; calculating the rider's speed through GPS data; using a heart rate monitor to obtain the rider's heart rate data; obtaining the number of pedals per minute of the rider through a cadence sensor; obtaining the ambient temperature during riding through a temperature sensor; obtaining the ambient humidity through a humidity sensor; obtaining the acceleration change during riding through an accelerometer; and obtaining the tilt angle of the bicycle through a gyroscope. The calculation of derived cycling features includes energy consumption, cycling rhythm, cycling efficiency, heart rate recovery rate, cycling intensity index, path slope analysis and weather adaptability. For example, based on cycling speed, heart rate and cadence, combined with known energy consumption models, the energy consumption during cycling is calculated; combined with cadence and heart rate data, the cyclist's rhythm changes, such as steady riding, accelerated riding, etc., are analyzed; by comparing cycling speed with energy consumption, the cycling efficiency is calculated; the speed at which the heart rate drops after riding is calculated to evaluate the cyclist's recovery ability; combined with heart rate, speed and cadence, the cycling intensity index is calculated to evaluate the difficulty of riding; the slope of the cycling path is calculated using the altitude changes in GPS data, which is particularly important for mountain riding; combined with temperature, humidity and wind speed data, the cyclist's adaptability to current weather conditions is evaluated. By analyzing derived cycling features, personalized training plans can be developed for cyclists.

[0091] Through pre-training and parameter adjustment, the model in this embodiment can learn and predict riding data more accurately, thus improving the performance of the model; the trained model can analyze riding data more effectively and provide more accurate riding analysis results. Through continuous training and parameter adjustment, the model can adapt to different riding scenarios and user needs, thus enhancing the adaptability of the model; accurate riding data analysis and prediction can provide users with a better riding experience and improve user satisfaction; the trained riding model can be applied to a variety of scenarios, such as riding route recommendations, riding safety analysis, riding health advice, etc., to support diverse application needs.

[0092] Further, in some embodiments, step S4 "obtaining a specific task data set corresponding to the user's needs, and adjusting the trained riding model based on the specific task data set to obtain an adjusted riding model" may specifically include:

[0093] S41. Respond to user needs and determine specific tasks corresponding to user needs;

[0094] S42. Obtaining a specific task data set corresponding to a specific task;

[0095] S43. Adjust the network structure and weight of the cycling model based on the specific task data to obtain an adjusted cycling model, so as to perform a specific task according to the adjusted cycling model.

[0096] Specifically, for step S4, clarify the specific needs of the user, such as cycling route recommendation, cycling safety analysis or cycling health analysis, etc.; determine the specific tasks that the model needs to complete based on user needs, such as developing a cycling route recommendation system; collect data related to specific tasks, for example, if the task is route recommendation, it is necessary to collect the user's cycling history, preferred routes, terrain data, etc. According to the needs of specific tasks, adjust the network structure of the model, including adding or deleting certain layers, to better handle specific task data. Based on the task-specific data set, fine-tune the model's weights to optimize the model's performance on specific tasks. Finally, a cycling model optimized for a specific task is obtained, which can perform specific tasks more accurately, such as recommending cycling routes that users may like.

[0097] This embodiment improves the accuracy and reliability of the model when performing specific tasks by adjusting the model for data sets of specific tasks; the model can better meet the personalized needs of users, such as recommending cycling routes that meet user preferences, thereby improving user satisfaction; through the optimization of specific tasks, the model can more efficiently utilize computing resources, reduce unnecessary calculations, and improve overall performance; the adjusted model can be applied to a variety of different cycling-related tasks, such as safety analysis, health advice, etc., which increases the practicality and flexibility of the model; by continuously collecting user feedback and cycling data, the model can be continuously optimized and adjusted to adapt to changing user needs and environmental conditions; the adjusted model can provide more powerful data support for cycling-related decisions, such as recommending the best cycling route to help users make better cycling decisions; the model can quickly adapt to new specific tasks, showing good adaptability and flexibility, which is particularly important in the rapidly developing field of cycling technology.

[0098] Furthermore, in some embodiments, the riding model building method further includes:

[0099] S51. Recommending routes to users based on cycling data through a cycling model to generate a recommended cycling route;

[0100] S52. Obtaining user feedback on recommended cycling routes;

[0101] S53. Adjust the riding model based on the feedback result to obtain an adjusted riding model.

[0102] Specifically, the cycling model construction method provided in this embodiment may further include a route recommendation function.

[0103] Using the trained cycling model, analyze cycling data, including the user's historical cycling routes, preferences, terrain and weather conditions, to generate personalized cycling recommendation routes. Design a user interface to allow users to provide feedback on the recommended cycling routes, including route satisfaction, actual experience, suggestions, etc. Adjust the cycling model based on user feedback. This may include adjusting the recommendation algorithm, optimizing route selection logic, or updating the user preference model. Retrain the model using user feedback and new cycling data to improve the model's prediction accuracy and user satisfaction. Establish a continuous iterative optimization process to ensure that the model can continuously adapt to changes in user needs and changes in the external environment.

[0104] Through personalized route recommendations, users can obtain cycling routes that better suit their preferences and needs, thereby enhancing the fun and satisfaction of cycling. The user feedback mechanism encourages users to participate in the optimization process of the model, thereby increasing user participation and loyalty to the system. Regular adjustment and retraining of the model based on user feedback helps to improve the prediction accuracy and overall performance of the model. Continuous iterative optimization of the model enables it to adapt to changes in user needs and changes in the external environment, such as seasonal weather changes or updates to urban infrastructure. By combining user feedback and cycling data, the model can provide higher-quality decision support and help users make better cycling plans. Optimized cycling route recommendations help users engage in safer and healthier cycling activities, which may include avoiding high-risk areas or recommending healthier cycling routes.

[0105] Furthermore, in some embodiments, the riding model building method further includes:

[0106] S61. Perform a cycling safety analysis on the user based on the cycling data through the cycling model to obtain a cycling safety analysis result;

[0107] S62. Generate corresponding safety recommendations based on the riding safety analysis results.

[0108] Specifically, the riding model construction method provided in this embodiment may also include a riding safety analysis function, using the riding model to analyze riding data, including speed, route, environmental conditions, etc., to identify potential safety risks. The model analyzes riding behaviors, such as sudden braking and sharp turns, and evaluates the safety risks of these behaviors. Based on the results of the riding safety analysis, targeted safety suggestions are generated. Safety suggestions include the use of riding equipment, safety tips for specific routes, suggestions for improving riding behaviors, etc. Real-time monitoring of riding status is achieved, and when potential risks are detected, early warnings are issued to riders in a timely manner. The early warning system can be integrated into riding applications or smart devices (such as smart riding glasses) to provide instant feedback. In-depth analysis of historical riding data is performed to identify common safety issues and accident patterns. Using machine learning algorithms, learn from data and predict possible safety risks. Provide user education and training resources to help riders enhance safety awareness and riding skills. Customize personalized training plans and safety guidance based on model analysis results.

[0109] This embodiment significantly enhances the safety awareness and safety of riders through riding safety analysis and real-time warnings; effectively reduces the occurrence of riding accidents by identifying and warning of potential safety risks in advance; provides safety suggestions and real-time monitoring functions to enhance users' trust in riding applications and services; helps riders avoid unsafe behaviors and environments and improves the overall riding experience through safety analysis and suggestions; provides education and training resources to help cultivate healthy riding habits and culture among riders; uses data analysis and pattern recognition to achieve evidence-based riding safety management and decision-making; and continuously optimizes and innovates riding safety solutions through continuous data analysis and user feedback.

[0110] Furthermore, in some embodiments of the present application, the method further includes:

[0111] S71. Perform a cycling health analysis on the user based on the cycling data using the cycling model to obtain a health analysis result;

[0112] S72. Generate corresponding health advice based on the health analysis results.

[0113] Specifically, the cycling model construction method provided in this embodiment may also include a cycling health analysis function, using the cycling model to analyze cycling data, including the rider's physiological data (such as heart rate, pedaling frequency), environmental data (such as temperature, humidity), and behavioral data (such as cycling speed, duration). Through model analysis, the rider's exercise intensity, energy consumption, and possible health benefits are evaluated. Based on the results of the cycling health analysis, targeted health suggestions are generated. Health suggestions include appropriate exercise intensity, recommended cycling frequency, and personalized health goal setting. Provide personalized health guidance, and customize personalized cycling plans according to the rider's age, health status, and exercise goals. Through model analysis, identify the rider's exercise habits and potential health risks, and provide corresponding adjustment suggestions. Realize real-time monitoring of the rider's health status, track exercise effects and changes in health indicators. Use the model to predict the rider's health trends and adjust health suggestions and cycling plans in a timely manner. Provide education and training resources to help riders understand how to improve their health through cycling. Customize personalized health education and training content through model analysis results.

[0114] This embodiment helps cyclists achieve health goals more effectively and improves the health benefits of cycling through cycling health analysis and suggestions; timely discovers potential health risks by monitoring and analyzing cycling data, takes preventive measures, and reduces the occurrence of health problems; provides personalized health guidance and suggestions to enhance users' interest and participation in cycling activities; encourages cyclists to form healthy cycling habits and promotes overall lifestyle improvements through education and training; improves user satisfaction with cycling applications and services by providing effective health suggestions and guidance; uses data analysis and model predictions to achieve evidence-based health decisions and cycling plan adjustments; and continuously optimizes and innovates cycling health solutions through continuous data analysis and user feedback.

[0115] Specifically, this embodiment can also generate corresponding riding records and riding reports based on the riding data analysis results and sports indicators. The generation of riding records and reports includes integrating the data collected during riding and the analysis results into structured reports. These reports may include riding summaries, health advice, sports trend analysis, etc. These records and reports can be stored in the user's personal account, or accessed and shared through mobile applications or cloud services. When the riding model is subsequently adjusted or trained, the riding records and riding reports can be directly imported as the user's historical riding data.

[0116] In summary, the cycling model construction method provided in this embodiment includes: selecting a target open source model as a basic framework; obtaining cycling data and performing data preprocessing, and constructing a cycling data set based on the cycling data after data preprocessing; pre-training the target open source model based on the cycling data set to obtain a trained cycling model; obtaining a specific task data set corresponding to user needs, and adjusting the trained cycling model based on the specific task data set to obtain an adjusted cycling model. This embodiment provides a cycling model construction scheme, through pre-training and adjustment, the model can better understand and predict cycling scenarios, improve the accuracy and applicability of the model in practical applications; the pre-training process enables the model to learn a wider range of cycling features, and enhances the generalization ability of the model in different cycling scenarios; through data preprocessing and annotation, the quality of the data set is improved, making the model training more efficient, and reducing the time and resource consumption of invalid data processing; through the adjustment of the specific task data set, the model can better meet the personalized needs of users, such as route recommendation, safety analysis, etc.; the model can be adjusted according to user feedback, so that users participate in the optimization process of the model, and improve user satisfaction and the practicality of the model.

[0117] In order to better implement the riding model construction method of the embodiment of the present application, the embodiment of the present application also provides a riding model construction device. The meanings of the terms are the same as those in the above riding model construction method, and the specific implementation details can refer to the description in the method embodiment.

[0118] See also Figure 3 , Figure 3 The schematic diagram of the structure of the riding model building device provided in the embodiment of the present application, wherein the riding model building device may specifically include a selection module 201, a data module 202, a training module 203 and an adjustment module 204, which may be specifically as follows:

[0119] A selection module 201 is used to select a target open source model as a basic framework;

[0120] The data module 202 is used to obtain the riding data and perform data preprocessing, and construct a riding data set based on the riding data after the data preprocessing;

[0121] A training module 203 is used to pre-train the target open source model based on the cycling data set to obtain a trained cycling model;

[0122] The adjustment module 204 is used to obtain a specific task data set corresponding to the user's needs, and adjust the trained riding model based on the specific task data set to obtain an adjusted riding model.

[0123] In a specific embodiment, the riding model building device provided in this embodiment may further include:

[0124] Data collection module, used to collect riding data in real time through sensors based on mobile phone apps;

[0125] Data preprocessing module, used to clean and extract features from raw riding data;

[0126] Model inference module, used to call the optimized model to generate cycling suggestions;

[0127] Communication module, used to push cycling suggestions to AR glasses display via BLE,

[0128] Furthermore, in some embodiments, the data module 202 may specifically include:

[0129] A data acquisition unit, used for acquiring riding data through a variety of sensors;

[0130] A preprocessing unit, used for performing data preprocessing on the riding data, wherein the data preprocessing includes at least one of data cleaning, removing outliers, and removing incomplete records;

[0131] A labeling unit, used for labeling the riding data after data preprocessing to obtain riding data labeled as different types of riding scenes;

[0132] The data set construction unit is used to construct a cycling data set based on cycling data of different types of cycling scenes and their corresponding scene classification labels.

[0133] Furthermore, in some embodiments, the training module 203 may specifically include:

[0134] Model setting unit, used to set the loss function and training parameters of the target open source model based on user needs;

[0135] A feature extraction unit, used to extract riding features corresponding to the target open source model based on the target open source model, and determine corresponding riding analysis results based on the riding features;

[0136] A parameter adjustment unit, used to calculate the difference loss of the riding analysis results based on the loss function and adjust the model parameters of the target open source model;

[0137] The training unit is used to continue training the target open source model based on the adjusted model parameters until the training conditions are met to obtain a trained cycling model.

[0138] Furthermore, in some embodiments, the adjustment module 204 may specifically include:

[0139] A task determination unit, used to respond to user needs and determine specific tasks corresponding to the user needs;

[0140] A task data unit is used to obtain a specific task data set corresponding to a specific task;

[0141] The model adjustment unit is used to adjust the network structure and weight of the cycling model based on specific task data to obtain an adjusted cycling model, so as to perform a specific task according to the adjusted cycling model.

[0142] Furthermore, in some embodiments, the cycling model building device may further include a route recommendation module, and the route recommendation module specifically includes:

[0143] A route generation unit, used to recommend a route to a user based on the cycling data using a cycling model, and generate a recommended cycling route;

[0144] A first feedback unit, used to obtain feedback results from users regarding the recommended cycling route;

[0145] The first adjustment unit is used to adjust the riding model based on the feedback result to obtain an adjusted riding model.

[0146] Furthermore, in some embodiments of the present application, the riding model building device may further include a safety analysis module, and the safety analysis module specifically includes:

[0147] A safety analysis unit, used to perform a riding safety analysis on the user according to the riding data through a riding model to obtain a riding safety analysis result;

[0148] The safety suggestion unit is used to generate corresponding safety suggestions based on the riding safety analysis results.

[0149] Furthermore, in some embodiments of the present application, the cycling model building device may further include a health analysis module, and the health analysis module specifically includes:

[0150] A health analysis unit, used to perform a cycling health analysis on the user according to the cycling data through a cycling model to obtain a health analysis result;

[0151] The health advice unit is used to generate corresponding health advice based on the health analysis results.

[0152] The riding model construction device provided in this embodiment is used to select the target open source model as the basic framework through the selection module 201; obtain the riding data and perform data preprocessing through the data module 202, and construct the riding data set based on the riding data after data preprocessing; pre-train the target open source model based on the riding data set through the training module 203 to obtain the trained riding model; obtain the specific task data set corresponding to the user's needs through the adjustment module 204, and adjust the trained riding model based on the specific task data set to obtain the adjusted riding model. A riding model construction device provided in this embodiment, through pre-training and adjustment, enables the model to better understand and predict the riding scene, improves the accuracy and applicability of the model in practical applications; the pre-training process enables the model to learn a wider range of riding features, and enhances the generalization ability of the model in different riding scenes; through data preprocessing and annotation, the quality of the data set is improved, the model training is more efficient, and the time and resource consumption of invalid data processing are reduced; through the adjustment of the specific task data set, the model can better meet the personalized needs of users, such as route recommendation, safety analysis, etc.; the model can be adjusted according to user feedback, so that the user participates in the optimization process of the model, and improves the user satisfaction and the practicality of the model.

[0153] For the specific definition of the riding model construction device, please refer to the definition of the riding model construction method above, which will not be repeated here. Each module in the above-mentioned riding model construction device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0154] In addition, the present application also provides an electronic device, such as Figure 4 , which shows a schematic diagram of the structure of an electronic device involved in an embodiment of the present application. Specifically, the electronic device may include a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, a power supply 303, an input unit 304 and other components. Those skilled in the art can understand that Figure 4 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0155] The processor 301 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. By running or executing software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, it executes various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole. Optionally, the processor 301 may include one or more processing cores; preferably, the processor 301 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 301.

[0156] The memory 302 can be used to store software programs and modules. The processor 301 executes various functional applications and riding model building methods by running the software programs and modules stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 302 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 302 may also include a memory controller to provide the processor 301 with access to the memory 302.

[0157] The electronic device also includes a power supply 303 for supplying power to various components. Preferably, the power supply 303 can be logically connected to the processor 301 through a power management system, so as to manage charging, discharging, power consumption and other functions through the power management system. The power supply 303 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators and other arbitrary components.

[0158] The electronic device may further include an input unit 304, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0159] Although not shown, the electronic device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 301 in the electronic device will load the executable files corresponding to the processes of one or more application programs into the memory 302 according to the following instructions, and the processor 301 will run the application programs stored in the memory 302, thereby realizing various functions, as follows:

[0160] Select the target open source model as the basic framework; obtain the riding data and perform data preprocessing, and build a riding data set based on the riding data after data preprocessing; pre-train the target open source model based on the riding data set to obtain a trained riding model; obtain a specific task data set corresponding to user needs, and adjust the trained riding model based on the specific task data set to obtain an adjusted riding model.

[0161] In a specific embodiment, the electronic device may be smart cycling glasses, which include: a communication interface for receiving data transmitted from a cycling model; a display unit for displaying recommended cycling routes, safety recommendations, and health recommendations; a storage unit for storing the user's exercise records and health reports; and a processor for parsing the received data and controlling the display content of the display unit.

[0162] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0163] The embodiments of the present application enable the model to better understand and predict riding scenarios through pre-training and adjustment, thereby improving the accuracy and applicability of the model in practical applications; the pre-training process enables the model to learn a wider range of riding characteristics, thereby enhancing the generalization ability of the model in different riding scenarios; through data pre-processing and labeling, the quality of the data set is improved, the model training is made more efficient, and the time and resource consumption of invalid data processing is reduced; through the adjustment of specific task data sets, the model can better meet the personalized needs of users, such as route recommendation, safety analysis, etc.; the model can be adjusted according to user feedback, so that users can participate in the optimization process of the model, thereby improving user satisfaction and the practicality of the model.

[0164] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0165] To this end, an embodiment of the present application provides a storage medium in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the steps in any one of the riding model building methods provided in the embodiment of the present application. For example, the instructions can execute the following steps:

[0166] Select the target open source model as the basic framework; obtain the riding data and perform data preprocessing, and build a riding data set based on the riding data after data preprocessing; pre-train the target open source model based on the riding data set to obtain a trained riding model; obtain a specific task data set corresponding to user needs, and adjust the trained riding model based on the specific task data set to obtain an adjusted riding model.

[0167] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0168] The storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc. Since the instructions stored in the storage medium can execute the steps in any of the riding model building methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any of the riding model building methods provided in the embodiments of the present application can be achieved. For details, please refer to the previous embodiments and will not be repeated here.

[0169] The above is a detailed introduction to a riding model construction method, device, electronic device and storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for constructing a cycling model, characterized in that: The steps include: Select the target open source model as the basic framework; Obtain riding data and perform data preprocessing, and build a riding data set based on the preprocessed riding data; Pre-training the target open source model based on the cycling data set to obtain a trained cycling model; A specific task data set corresponding to the user's needs is obtained, and the trained cycling model is adjusted based on the specific task data set to obtain an adjusted cycling model.

2. The riding model construction method according to claim 1, characterized in that: The step of obtaining the riding data and performing data preprocessing, and constructing a riding data set based on the riding data after the data preprocessing, includes: Acquire riding data through various sensors; Performing data preprocessing on the riding data, wherein the data preprocessing includes at least one of data cleaning, removing outliers, and removing incomplete records; Label the riding data after data preprocessing to obtain riding data marked as different types of riding scenes; A cycling data set is constructed based on the cycling data of the different types of cycling scenes and their corresponding scene classification labels.

3. The riding model construction method according to claim 2, characterized in that: The pre-training the target open source model based on the cycling data set to obtain a trained cycling model includes: Based on the user requirements, setting the loss function and training parameters of the target open source model; Extracting riding features corresponding to the target open source model based on the target open source model, and determining corresponding riding analysis results based on the riding features; Calculating the difference loss of the riding analysis result based on the loss function, and adjusting the model parameters of the target open source model; The target open source model is continuously trained based on the adjusted model parameters until the training conditions are met, thereby obtaining a trained cycling model.

4. The riding model construction method according to claim 1, characterized in that: The acquiring of a specific task data set corresponding to the user's needs, and adjusting the trained riding model based on the specific task data set to obtain an adjusted riding model, includes: Responding to user needs and determining specific tasks corresponding to the user needs; Acquire a specific task data set corresponding to the specific task; The network structure and weight of the riding model are adjusted based on the specific task data to obtain an adjusted riding model, so as to perform the specific task according to the adjusted riding model.

5. The riding model construction method according to claim 1, characterized in that: The method further comprises: Recommending a route to the user based on the cycling data by using the cycling model to generate a recommended cycling route; Obtaining user feedback on the recommended cycling route; The riding model is adjusted based on the feedback result to obtain an adjusted riding model.

6. The riding model construction method according to claim 1, characterized in that: The method further comprises: Performing a riding safety analysis on the user according to the riding data by using the riding model to obtain a riding safety analysis result; Based on the riding safety analysis result, corresponding safety recommendations are generated.

7. The riding model construction method according to claim 1, characterized in that: The method further comprises: Performing a cycling health analysis on the user according to the cycling data by using the cycling model to obtain a health analysis result; Based on the health analysis results, corresponding health recommendations are generated.

8. A riding model building device, characterized in that: include: The selection module is used to select the target open source model as the basic framework; The data module is used to obtain and preprocess the riding data, and to construct a riding data set based on the preprocessed riding data; A training module, used for pre-training the target open source model based on the cycling data set to obtain a trained cycling model; The adjustment module is used to obtain a specific task data set corresponding to user needs, and adjust the trained riding model based on the specific task data set to obtain an adjusted riding model.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program according to the steps of the cycling model building method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the riding model building method as described in any one of claims 1 to 7.

Images