Smoking cessation intervention methods, devices, smart wearable devices and storage media

Abstract

This invention relates to the field of healthcare technology, providing a method, device, smart wearable device, and storage medium for smoking cessation intervention. The method, applied to the smart wearable device, includes: collecting multimodal data such as physiological, behavioral, and environmental data from the user, and inputting this data into an intelligent monitoring model to obtain an intervention strategy output by the model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data consists of historical data from the multimodal data and labeled with intervention strategies indicating intervention levels. Smoking cessation intervention is then performed on the user based on the intervention strategy. By using smart wearable devices, the limitations of traditional smoking cessation interventions are overcome, achieving end-to-end intelligent management from data collection and risk prediction to precise intervention. This constructs a dynamic intervention mechanism integrating environment, physiology, and behavior, providing users with efficient, accurate, and personalized smoking cessation support and enhancing the effectiveness of smoking cessation interventions.

Description

Technical Field

[0001] This invention relates to the field of healthcare technology, and in particular to a method, device, smart wearable device, and storage medium for smoker cessation intervention. Background Technology

[0002] Smoking is a major factor harming human health. Smokers are dependent on nicotine in tobacco, making quitting a highly challenging process that requires sustained motivation, effective supervision, and personalized guidance. Traditional smoking cessation methods, such as nicotine replacement therapy and behavioral interventions, rely on the user's self-discipline and lack real-time monitoring and dynamic intervention, resulting in extremely low success rates.

[0003] With the popularization and development of smart devices, using them to assist in smoking cessation has become a new possibility. Existing intervention devices for assisting smoking cessation have limited functions, only recording the number of cigarettes smoked, and rely on the user's active recording. They cannot accurately identify the user's craving for tobacco and the risk of relapse, and the intervention methods are too mechanical (only vibration reminders), making it difficult to provide accurate and effective smoking cessation support, and the intervention effect is not significant. Summary of the Invention

[0004] This invention provides a smoking cessation intervention method, device, smart wearable device, and storage medium to address the shortcomings of existing smart devices for assisting smoking cessation, such as limited functionality, overly mechanical intervention methods, difficulty in providing precise and effective smoking cessation support, and limited intervention effects.

[0005] This invention provides a smoking cessation intervention method for smokers, applied to smart wearable devices, comprising the following steps:

[0006] Collect multimodal data from users; the multimodal data includes physiological data, behavioral data, and environmental data;

[0007] The multimodal data is input into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels.

[0008] The user was given a smoking cessation intervention based on the intervention strategy.

[0009] According to the smoking cessation intervention method for smokers provided by the present invention, the process by which the intelligent monitoring model processes the multimodal data to obtain an intervention strategy includes:

[0010] Based on the physiological data in the multimodal data, the changing trends of the user's physiological characteristics related to smoking cravings are determined;

[0011] Based on the behavioral data in the multimodal data, the user's pose changes are identified; the pose changes include the user's arm movement trajectory and arm raising frequency;

[0012] The environmental data in the multimodal data is classified into scenarios to determine the scenario category of the user's environment; the scenario category is used to characterize the degree of influence of the environment on the user's desire to smoke;

[0013] By comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the category of the scene, the probability and degree of the user's craving for smoking are determined within a future preset time period.

[0014] An intervention level is determined based on the probability and the degree of smoking craving, and an intervention strategy is generated based on the intervention level; the intervention strategy includes intervention content under the intervention level.

[0015] According to the smoking cessation intervention method for smokers provided by the present invention, after collecting the user's multimodal data, it further includes:

[0016] The multimodal data is sent to the terminal application, and the terminal application sends the multimodal data to the server.

[0017] Based on the multimodal data, the server periodically iteratively trains the intelligent engine corresponding to the intelligent monitoring model to obtain updated model parameters.

[0018] The system receives updated model parameters sent by the server and optimizes the intelligent monitoring model using the updated model parameters.

[0019] According to the smoking cessation intervention method for smokers provided by the present invention, the smart wearable device is equipped with a multimodal sensor array, which is used to collect the multimodal data; wherein:

[0020] The multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor;

[0021] The biosensor is used to collect the physiological data, which includes at least one of heart rate, blood pressure, blood glucose, blood oxygen, psychological state data, sleep data, and skin conductivity.

[0022] The motion sensor is used to collect the behavioral data, which includes at least one of the user's motion data, hand movements, hand raising frequency, and arm movement trajectory.

[0023] The carbon monoxide sensor and the temperature sensor are used to collect the environmental data, which includes at least the ambient temperature and the ambient carbon monoxide concentration; the ambient nicotine concentration is calculated based on the ambient carbon monoxide concentration.

[0024] According to the smoking cessation intervention method for smokers provided by the present invention, the smart wearable device is further provided with at least one environmental monitoring port for extracting air samples from the user's environment and collecting exhaled gas samples from the user; the physiological data also includes the concentration of nicotine in the body; the collection of the user's multimodal data includes:

[0025] Based on a pre-generated timed sampling task, an air sample is extracted through the environmental monitoring hole, and the first concentration of carbon monoxide in the air sample is detected.

[0026] Output sampling prompt information; the sampling prompt information is used to prompt the user to perform a blowing operation on the environmental monitoring hole for a preset duration;

[0027] When the user's exhalation is detected to be complete, a sample of the user's exhaled gas is obtained, and a second concentration of carbon monoxide in the exhaled gas sample is detected.

[0028] The user's nicotine concentration is estimated based on the difference between the first concentration and the second concentration.

[0029] The smart wearable device is also equipped with a microphone hole, which is used to collect ambient sound for smoke detection.

[0030] The smart wearable device also includes a display screen, which is used to display part or all of the multimodal data.

[0031] According to the smoking cessation intervention method for smokers provided by the present invention, after conducting smoking cessation intervention on the user based on the intervention strategy, the method further includes:

[0032] Monitor the effectiveness of the intervention strategy on the user;

[0033] Based on the intervention effect and the recorded logs, a smoking cessation report is generated and sent to the terminal application for the user to view. The recorded logs are generated based on the user's smoking record operations and / or detected smoking behavior, using a timed reminder task to remind the user to record smoking and learn smoking cessation techniques. The smoking cessation report includes smoking behavior statistics, physiological change trends, nicotine addiction assessment results, nicotine concentration change curves in the body, and smoking cessation achievements.

[0034] If the user meets the reward conditions based on the smoking cessation results, a virtual object is issued to the user's account; the virtual object is used to redeem products or services.

[0035] The smoking cessation results are pushed to the terminal application of the user's interactive users; the interactive users are users who have social relationships with the user.

[0036] The present invention also provides a smoking cessation intervention device for smokers, applied to a smart wearable device, the smoking cessation intervention device comprising the following modules:

[0037] The data acquisition module is used to collect the user's multimodal data, which includes physiological data, behavioral data, and environmental data.

[0038] A smoking monitoring module is used to input the multimodal data into an intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels.

[0039] The smoking cessation intervention module is used to conduct smoking cessation intervention on the user based on the intervention strategy.

[0040] The present invention also provides a smart wearable device for smoking cessation intervention. The main body of the smart wearable device includes a back shell and a display screen. The back shell includes a concave surface and a convex surface. The display screen is embedded in the concave surface of the back shell. A multimodal sensor array is arranged on the convex surface of the back shell.

[0041] The smart wearable device further includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the smoking cessation intervention methods described above.

[0042] According to the intelligent wearable device for smoking cessation intervention provided by the present invention, the multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor.

[0043] The back cover also includes a first side and a second side arranged opposite to each other, as well as a third side and a fourth side arranged opposite to each other; the first side and the second side are provided with environmental monitoring holes for extracting gas samples required for detection; a microphone hole is provided on the first side or the second side, a magnetic charging contact is provided on the first side or the second side, and a fixing component for wearing is provided on the third side and the fourth side.

[0044] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the smoking cessation intervention method as described above.

[0045] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the smoking cessation intervention method as described above.

[0046] The present invention provides a smoking cessation intervention method, device, smart wearable device, and storage medium. Through the collection of multimodal data by the smart wearable device and the identification of multimodal data by the intelligent monitoring model, it achieves real-time and comprehensive monitoring of the user's physiological state, behavioral changes, and environment. Based on intervention strategies, it provides dynamic and personalized intervention for the user's smoking cessation. By overcoming the limitations of traditional smoking cessation interventions through smart wearable devices, it achieves end-to-end intelligent management from data collection and risk prediction to precise intervention, constructing a dynamic intervention mechanism integrating environment, physiology, and behavior. This provides users with efficient, accurate, and personalized smoking cessation support in real time, enhancing the effectiveness of smoking cessation interventions. Attached Figure Description

[0047] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0048] Figure 1 This is a flowchart illustrating the smoking cessation intervention method for smokers provided by the present invention.

[0049] Figure 2 This is a schematic diagram of the data flow provided by the present invention.

[0050] Figure 3 This is a schematic diagram of the smoking cessation intervention device provided by the present invention.

[0051] Figure 4 This is a front view of the smart wearable device provided by the present invention.

[0052] Figure 5 This is a rear view of the smart wearable device provided by the present invention.

[0053] Figure 6 This is another structural schematic diagram of the smart wearable device provided by the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0055] This invention provides a smoking cessation intervention method for smokers, applied to smart wearable devices, particularly smart wearable devices within a smoking cessation assistance system. These smart wearable devices include smartwatches, smart bracelets, or smart glasses, and the smoking cessation assistance system also includes terminal applications and servers. Through these smart wearable devices, the user's physiological and psychological state can be monitored in real-time and comprehensively to determine the user's degree of dependence on smoking, thereby enabling dynamic and personalized interventions. This provides precise and effective smoking cessation support, enhancing the effectiveness of the smoking cessation assistance intervention.

[0056] Specifically, Figure 1 This is a flowchart illustrating the smoking cessation intervention method for smokers provided by the present invention, as shown below. Figure 1 As shown, the method includes the following steps:

[0057] Step 100: Collect the user's multimodal data; the multimodal data includes physiological data, behavioral data, and environmental data;

[0058] Step 200: Input the multimodal data into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model; the intelligent monitoring model is built based on a neural network model and trained with sample data, the sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels, and the sample labels are marked with intervention levels;

[0059] Step 300: Conduct smoking cessation intervention on the user based on the intervention strategy.

[0060] First, multimodal data of the user is collected. This multimodal data includes at least the user's physiological data, behavioral data, and environmental data. Physiological data includes blood pressure, blood glucose, blood oxygen, heart rate, body temperature, psychological state data, and sleep data. Sleep data may further include sleep depth and sleep duration. Psychological state data includes stress index. Behavioral data characterizes the user's movement, especially the movement of the parts of the user wearing the smart wearable device. Environmental data includes ambient temperature, ambient humidity, and air quality. Air quality may further include the composition and concentration of gaseous compounds, including but not limited to carbon monoxide, characterizing the user's external environment.

[0061] Optionally, the multimodal data includes physiological data reflecting the user's internal physiological state and emotions, behavioral data reflecting changes in the user's external physical behavior, and environmental data reflecting the external environment of the user's space. Physiological data can reflect the user's physiological state before and after smoking and in daily life, analyzing the correlation between the user's craving for cigarettes (i.e., the onset of nicotine addiction) and specific biological characteristics, providing a basis for predicting the occurrence of cravings. Behavioral data can capture the user's hand movements, changes in body posture, etc., and by matching them with specific smoking action patterns, such as the frequency of raising the hand and the trajectory of arm movements, it is possible to monitor and identify the user's impending or ongoing smoking behavior in real time. Environmental data can be used to monitor the carbon monoxide concentration in the user's environment, helping to understand the impact of the user's environment on the occurrence of cravings. For example, in environments with high carbon monoxide concentrations, users are more likely to experience cravings for cigarettes, and it can also remind users to avoid environments with severe secondhand smoke pollution.

[0062] In one embodiment, the smart wearable device is equipped with multimodal sensors for collecting multimodal data from the user. These multimodal sensors include at least biosensors, motion sensors, and environmental sensors. The biosensors collect the user's physiological data, including a blood pressure sensor for collecting blood pressure data, and further including a heart rate sensor and a skin conductance sensor. These sensors collect physiological data such as heart rate and skin conductivity. The biosensors also include other sensors for collecting blood glucose, blood oxygen, sleep data, and psychological state data, which will not be listed here. The environmental sensors include carbon monoxide sensors and temperature sensors. The temperature sensor collects the air temperature of the user's environment, and the carbon monoxide sensor detects the carbon monoxide concentration in the air. Based on the carbon monoxide concentration detected by the carbon monoxide sensor, the nicotine concentration in the air can be estimated, thus obtaining the environmental nicotine concentration. Optionally, the environmental data in the multimodal data also includes environmental audio data. The environmental sensors also include an audio sensor for collecting environmental audio data. By matching the audio characteristics of the collected environmental audio data with the environmental audio characteristics when the user smokes, it is used to detect in real time whether the user is smoking.

[0063] The smart wearable device incorporates an intelligent monitoring model. This model is built upon artificial intelligence (AI) models such as deep neural networks or machine learning models and trained using sample data. The sample data consists of historical multimodal data and user intervention strategies, each labeled with an intervention level. The intelligent monitoring model can monitor a user's smoking probability in real time. Specifically, the collected multimodal data is input into the intelligent monitoring model to obtain the user's intervention strategy output by the model.

[0064] In one embodiment, the intelligent monitoring model is built upon a neural network model, which includes an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer comprises multiple (D, where D is a positive integer greater than 1) neurons. The first hidden layer comprises H1 neurons and a first activation function. The second hidden layer comprises H2 neurons and a second activation function. The output layer comprises one neuron and a third activation function. For example, H1 is 128, H2 is 64, the first and second activation functions are the same (ReLU function), and the third activation function is a sigmoid function. When multimodal data is input into the intelligent monitoring model, it is specifically transformed into feature vectors, which serve as input to the input layer of the intelligent monitoring model. The feature vectors corresponding to the multimodal data may include one or more, which can be determined based on the dimension of each data item in the multimodal data; no specific limitation is made in this regard.

[0065] Furthermore, based on the intervention strategy output by the intelligent monitoring model, smoking cessation intervention is carried out for users. Optionally, the intervention strategy is generated based on the predicted probability of the user's smoking craving occurring within a preset time period in the future. The intervention strategy includes intervention methods and intervention levels. Intervention methods include pushing smoking cessation reminders, reminding users of their current smoking cessation effect and progress, and pushing alternative activities for attention diversion, etc. The intervention level is determined according to the predicted probability of the user's smoking craving occurring. Based on the intervention level in the intervention strategy, different levels of smoking craving can be addressed to achieve graded intervention for users' smoking cessation.

[0066] Optionally, the intelligent monitoring model detects the probability of a user's smoking intention or behavior based on multimodal data, and generates an intervention strategy based on this probability. This intervention strategy includes intervention methods and levels. Intervention methods include sending reminders to the user via vibration, sound, flashing lights, etc. The reminders include smoking cessation goals, warnings about the health hazards of current smoking behavior, and suggested coping strategies, helping users to recognize their cravings and take appropriate measures to quit. Considering the important role of psychological factors in the smoking cessation process, the reminders may also include stress management methods, such as exercise, relaxation (music, movies), deep breathing exercises, meditation guidance, and drinking ice water, to help users alleviate negative emotions and stress caused by withdrawal.

[0067] For example, if the probability of a user smoking within a preset time period is less than a preset low-risk threshold, content of interest can be pushed to the user, or a push notification can be sent to the application to instruct the application to push content of interest to the user, thereby attracting the user's attention through content of interest and enabling the user to maintain the current low-probability smoking state. If the probability of a user smoking within a preset time period is greater than or equal to a preset high-risk threshold, smoking alternative activities can be pushed to the user, or a push notification can be sent to the application to instruct the application to push smoking alternative activities to the user, thereby diverting the user's attention through smoking alternative activities, or reminding the user to engage in attention-diversion activities instead of smoking when experiencing the urge to smoke. When the user is detected in an environment containing nicotine, the user can be reminded to leave the current area immediately.

[0068] In one embodiment, the intelligent monitoring model can be deployed in a smart wearable device using edge computing. The smart wearable device inputs the collected multimodal data into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model.

[0069] In another embodiment, the intelligent monitoring model can also be deployed on the application side of a mobile terminal. The intelligent wearable device sends the collected multimodal data to the application side, and the application side inputs the received multimodal data into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. Alternatively, the intelligent wearable device performs feature extraction on the collected multimodal data to obtain multimodal features, sends the multimodal features to the application side, and the application side inputs the received multimodal features into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The application side sends the intervention strategy to the intelligent wearable device, and the intelligent wearable device receives the intervention strategy sent by the application side to intervene in the user's smoking cessation.

[0070] In some embodiments, the intelligent monitoring model can also be deployed on a server. The smart wearable device sends the collected multimodal data to the application, the application sends the received multimodal data to the server, and the server inputs the received multimodal data into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. Alternatively, the smart wearable device performs feature extraction on the collected multimodal data to obtain multimodal features, sends the multimodal features to the application, the application sends the received multimodal features to the server, and the server inputs the received multimodal features into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The server sends the intervention strategy output by the intelligent monitoring model to the smart wearable device, or sends it to the smart wearable device through the application. The smart wearable device then performs smoking cessation intervention on the user based on the received intervention strategy.

[0071] The intelligent monitoring model comprehensively analyzes users' multimodal data, predicting the probability of cravings for smoking within a preset timeframe based on changes in the user's internal physiological state and external behavior, combined with the user's environment. It then generates intervention strategies based on the user's current smoking cessation stage. The smoking cessation stages mainly include the preparatory stage (the stage before quitting), the action stage (the stage of implementing smoking cessation behaviors), and the maintenance stage (the stage of successfully quitting and preventing relapse).

[0072] In this embodiment, by collecting multimodal data through smart wearable devices and identifying the multimodal data through intelligent monitoring models, real-time monitoring of the user's physiological state, behavioral changes, and environment is achieved. Based on the intervention strategies generated from this real-time monitoring, dynamic and personalized interventions are provided for the user's smoking cessation. By overcoming the limitations of traditional smoking cessation interventions through smart wearable devices, this approach achieves end-to-end intelligent management from data collection and risk prediction to precise intervention. It constructs a dynamic intervention mechanism integrating environment, physiology, and behavior, providing users with efficient, accurate, and personalized smoking cessation support in real time, thereby enhancing the effectiveness of smoking cessation interventions.

[0073] The intelligent monitoring model comprehensively analyzes multimodal data to predict the probability of a user experiencing cravings for cigarettes in the future. Based on the user's current stage of smoking cessation, it generates intervention strategies according to the probability of these cravings. Specifically, the process by which the intelligent monitoring model processes multimodal data to obtain intervention strategies includes:

[0074] Based on the physiological data in the multimodal data, the changing trends of the user's physiological characteristics related to smoking cravings are determined;

[0075] Based on the behavioral data in the multimodal data, the user's pose changes are identified; the pose changes include the user's arm movement trajectory and arm raising frequency;

[0076] The environmental data in the multimodal data is classified into scenarios to determine the scenario category of the user's environment; the scenario category is used to characterize the degree of influence of the environment on the user's urge to smoke;

[0077] By comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the category of the scene, the probability and degree of the user's craving for smoking are determined within a future preset time period.

[0078] An intervention level is determined based on the probability and the degree of smoking craving, and an intervention strategy is generated based on the intervention level; the intervention strategy includes intervention content under the intervention level.

[0079] The intelligent monitoring model processes multimodal data, including extracting multimodal features from the multimodal data, performing comprehensive analysis based on the extracted multimodal features, determining the probability and degree of smoking cravings of users within a preset time period in the future, and then generating intervention strategies.

[0080] Specifically, for physiological data in multimodal data, the changing trend of physiological characteristics related to smoking cravings of users is determined based on the physiological data. The changing trend of these physiological characteristics represents the changes in the physiological state related to smoking cravings of users within a future preset time period. The probability of users having smoking cravings is predicted based on the changing trend of these physiological characteristics.

[0081] For behavioral data in multimodal data, the user's posture changes are analyzed by analyzing the user's behavioral data, including the user's arm movement trajectory and the frequency of raising the hand. By matching these changes with the user's specific smoking action patterns, it can be determined whether the user is about to or is currently engaging in smoking behavior.

[0082] For environmental data in multimodal data, the environmental data is classified into scenarios. Based on factors such as the concentration of nicotine in the environment, the scenario category of the user's environment is determined. This scenario category is used to characterize the degree to which the user's environment affects the user's urge to smoke.

[0083] By comprehensively analyzing the changing trends of biological characteristics related to smoking cravings, changes in user posture, and the scene categories of the user's environment, the probability and degree of smoking cravings that users will experience in the future within a preset time period are determined, and intervention strategies are generated based on the probability and degree of smoking cravings.

[0084] Furthermore, regarding the intervention strategy, the intervention level is first determined based on the probability of smoking cravings and the degree of craving. An intervention strategy is then generated based on this intervention level, which includes the intervention method and the intervention level (also known as the degree of intervention). The intervention strategy further includes the intervention content at each intervention level. In other words, the intervention strategy represents the use of a specific intervention method and the output of corresponding intervention content at each intervention level. Different probabilities and / or degrees of smoking cravings result in different generated intervention strategies, and different intervention strategies will have at least one difference in the intervention method, intervention level, and intervention content.

[0085] Optional interventions may include a variety of approaches such as positive guidance, exercise reminders, and hazard warnings.

[0086] In one embodiment, the intelligent monitoring model processes data from different modalities in multimodal data in parallel, or sequentially. When the intelligent monitoring model processes data from different modalities in multimodal data in parallel, it includes multiple network branches. These multiple network branches are obtained by partitioning the input layer, first hidden layer, and second hidden layer of the neural network model. Each network branch corresponds to one modality of data; that is, the intelligent monitoring model includes at least a first network branch corresponding to physiological data, a second network branch corresponding to behavioral data, and a third network branch corresponding to environmental data. This partitioning of data from different modalities is achieved through multiple independent network branches. The intelligent monitoring model may also include a feature fusion layer, which fuses the multimodal features output from different network branches and predicts the probability and intensity of a user's craving for smoking within a predetermined time period based on the fused multimodal features.

[0087] Furthermore, the first network branch outputs changes in the user's biological characteristics related to smoking cravings; the second network branch outputs changes in the pose of the user's target body parts related to smoking behavior, including arm movement trajectories and hand-raising frequencies; and the third network branch outputs the scene category of the user's environment to identify the degree to which the user's smoking cravings are influenced by the environment. Optionally, predictions are made based on the output features of the first, second, and third network branches to obtain the probability and degree of smoking cravings for the user in a future preset time period corresponding to each network branch. This includes the first probability and degree of smoking cravings corresponding to the first network branch, the second probability and degree of smoking cravings corresponding to the second network branch, and the third probability and degree of smoking cravings corresponding to the third network branch. Finally, the first, second, and third probabilities are weighted and summed or weighted averaged to obtain the final probability of the user having a smoking craving in a future preset time period, and the first, second, and third degrees of smoking cravings are weighted and summed or weighted averaged to obtain the degree of smoking cravings for the user in a future preset time period. Alternatively, the output features of each network branch can be fused to obtain fused multimodal features. Based on the fused multimodal features, predictions can be made to obtain the probability and degree of a user's craving for smoking within a preset time period in the future.

[0088] The intelligent monitoring model uses nicotine concentrations in the user's environment and body, combined with physiological data such as heart rate, blood pressure, and stress index, as well as multi-scenario classification of the user's environment, to predict the probability and degree of the user's craving for smoking. It then generates intervention strategies in real time to provide dynamic and personalized intervention for the user's smoking cessation.

[0089] Furthermore, smart wearable devices have broken through the limitations of traditional smoking cessation interventions, realizing full-chain intelligence from data collection, risk prediction and precise intervention, and constructing a dynamic intervention mechanism integrating "environment-physiology-behavior" to provide smokers with efficient, accurate and personalized smoking cessation support.

[0090] In one embodiment, the intelligent monitoring model is deployed in a smart wearable device, and the model parameters are optimized in the cloud via a server. This ensures the model's performance while enabling real-time intervention and support for users' smoking cessation. The use of distributed edge computing ensures user privacy and data security, and compared to deploying the intelligent monitoring model on a server, reduces intervention latency, thus improving the real-time nature of smoking cessation intervention. Therefore, after step 100, the following may also be included:

[0091] Step 110: The multimodal data is sent to the terminal application. The terminal application iteratively optimizes the intelligent monitoring model based on the multimodal data to obtain the first model parameters, and sends the first model parameters to the server. The server periodically iteratively optimizes the intelligent engine corresponding to the intelligent monitoring model based on the first model parameters sent by the terminal applications of different users to obtain the second model parameters.

[0092] Step 120: Receive the second model parameters sent by the server and use the second model parameters to optimize and update the intelligent monitoring model.

[0093] The collected multimodal data is sent to the terminal application. The terminal application stores the multimodal data and iteratively optimizes the intelligent monitoring model based on the multimodal data to obtain the first model parameters, which are then sent to the server. The server periodically iteratively optimizes the intelligent engine corresponding to the intelligent monitoring model based on the first model parameters received from the terminal applications of different users to obtain the second model parameters, which are then sent to the intelligent wearable devices.

[0094] The smart wearable device receives the second model parameters sent by the server and uses these parameters to optimize and update the model parameters of its local smart monitoring model. Optionally, the server can send the second model parameters directly to the smart wearable device, or the server can send the second model parameters to the smart wearable device through a terminal application.

[0095] In one embodiment, reference is made to Figure 2 The diagram illustrates the data flow during the optimization and updating of the intelligent monitoring model. The smart wearable device sends collected multimodal data to the terminal application. The terminal application iteratively optimizes its local model based on this multimodal data, obtaining first model parameters which are then sent to the server. The server's cloud-based intelligent engine contains global model parameters. These global model parameters are obtained through federated training of the smart wearable device's local intelligent monitoring model, with different users' terminal applications as participants. Based on the first model parameters received from different users' terminal applications, the server periodically iteratively trains the intelligent engine's global model parameters to obtain second model parameters. These second model parameters are then sent to the smart wearable device, or the terminal application can send them to the smart wearable device. This optimizes and updates the model parameters of the local intelligent monitoring model on each smart wearable device, simultaneously optimizing the intelligent monitoring model and protecting the privacy of users' multimodal data.

[0096] The smart wearable device is equipped with a multimodal sensor array for collecting multimodal data. The multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor. The temperature sensor is used to collect at least one of the user's body temperature and the ambient temperature.

[0097] Furthermore, the carbon monoxide sensor and temperature sensor are used to collect environmental data. The collected environmental data includes at least the ambient carbon monoxide concentration and the ambient temperature collected by the temperature sensor, used for real-time monitoring of air quality, such as nicotine concentration and ambient temperature, in the user's surrounding environment. The ambient nicotine concentration is calculated based on the carbon monoxide concentration detected by the carbon monoxide sensor. Optionally, the multimodal sensor array may also include a humidity sensor to collect ambient humidity, or a temperature and humidity sensor may be used instead of the temperature sensor to simultaneously collect both ambient temperature and humidity. Optionally, the multimodal sensor array may also include an air quality sensor, which can be a gas sensor or an electrochemical sensor, used to collect the types and concentrations of gaseous compounds in the air. Based on this, the collected environmental data may also include the types and concentrations of gaseous compounds, particularly the concentrations of harmful gases, including but not limited to carbon oxides and nitrogen oxides.

[0098] Biosensors are used to collect users' physiological data, which includes at least one of heart rate, blood pressure, blood glucose, blood oxygen, psychological state data, sleep data, and skin conductivity. Psychological state data includes stress index, which characterizes the user's emotions, and sleep data includes sleep depth and sleep duration.

[0099] Motion sensors are used to collect user behavior data. Motion sensors include accelerometers and gyroscopes, which are used to capture behavioral data such as changes in the user's body posture, especially changes in the posture of the target parts of the body worn with smart wearable devices, including but not limited to the user's motion data, hand movements, frequency of raising hands, and arm movement trajectory.

[0100] In one embodiment, the smart wearable device further includes at least one environmental monitoring port for extracting air samples from the user's environment and collecting samples of the user's exhaled breath. The user's physiological data also includes the concentration of nicotine in their body. Based on this, step 100 includes:

[0101] Step 101: Based on the pre-generated timed sampling task, an air sample is extracted through the environmental monitoring hole, and the first concentration of carbon monoxide in the air sample is detected.

[0102] Step 102: Output sampling prompt information; the sampling prompt information is used to prompt the user to perform a blowing operation on the environmental monitoring hole for a preset duration;

[0103] Step 103: When the user's exhalation is detected to be complete, an exhaled gas sample of the user is obtained, and the second concentration of carbon monoxide in the exhaled gas sample is detected.

[0104] Step 104: Calculate the user's nicotine concentration based on the difference between the first concentration and the second concentration.

[0105] The smart wearable device outputs sampling prompts at pre-produced timed sampling tasks. These prompts instruct the user to blow air into the environmental monitoring port for a preset duration. The prompts may include, but are not limited to, vibration combined with text prompts, such as "Please blow air into the monitoring port." Before outputting the sampling prompts, an air sample is drawn through the environmental monitoring port, and a carbon monoxide sensor detects the first concentration of carbon monoxide in the air. Upon detecting that the user has finished blowing air, an exhaled air sample is obtained, and the second concentration of carbon monoxide in the exhaled air sample is detected.

[0106] Furthermore, based on the difference between the first and second carbon monoxide concentrations detected by the carbon monoxide sensor, the nicotine concentration in the user's body is determined, thus obtaining the user's nicotine concentration. Optionally, at least two environmental monitoring ports are included, each corresponding to a carbon monoxide sensor. These ports simultaneously sample air and the user's exhaled air, detecting carbon monoxide concentrations to ensure the accuracy of the detection results.

[0107] In one embodiment, based on a pre-generated timed sampling task, the user is reminded to measure the concentration of nicotine in their body at a fixed time every day to maintain data stability. For example, a sampling reminder message is output one hour after waking up every morning to remind the user to sample, which can avoid the morning CO peak.

[0108] The concentration of carbon monoxide in an air sample detected by a carbon monoxide sensor can be used to calculate the concentration of nicotine in the air, i.e., the environmental nicotine concentration. Similarly, the concentration of carbon monoxide in an exhaled breath sample detected by a carbon monoxide sensor can be used to calculate the concentration of nicotine in the exhaled breath. However, since the exhaled breath sample contains air, the concentration of nicotine in the user's body is calculated based on the difference between the concentration of carbon monoxide in the exhaled breath and that in the air, ensuring the accuracy of the data.

[0109] Alternatively, the concentration of nicotine in the environment can be estimated using the following method:

[0110] Environmental nicotine concentration (ppm) = [CO]_{ppm}×K×MW_{nicotine} / MW_{CO}×CF;

[0111] Wherein, [CO]_{ppm} represents the concentration of carbon monoxide (CO) in the air detected by the carbon monoxide sensor (ppm represents concentration), K represents the tobacco release ratio coefficient (default 0.08, range 0.06-0.12), MW_{nicotine} represents the molecular weight of nicotine (typically 162.23 g / mol), MW_{CO} represents the molecular weight of CO (typically 28 g / mol), and CF is the environmental calibration factor. For example, CF=0.3 for ventilated spaces and CF=1.2 for enclosed spaces.

[0112] The concentration of nicotine in a user's body can be estimated by combining the metabolic relationship of nicotine with the concentration relationship between carbon monoxide and nicotine. Specifically, during the metabolism of nicotine, the proportion of CO bound to hemoglobin is strongly correlated with nicotine intake, which can be expressed as:

[0113] {Exhaled CO (ppm)}≈0.92×{Blood carboxyhemoglobin (COHb %)};

[0114] The relationship between COHb% and nicotine dosage can be expressed as:

[0115] COHb% = 0.16 × nicotine intake (mg);

[0116] Therefore, the concentration of nicotine in a user's body can be calculated as follows:

[0117] Nicotine concentration in the body (ppm) = (exhaled CO - ambient CO) ÷ (0.92 × 0.16) × α;

[0118] Here, exhaled CO represents the average CO concentration at the end of the user's breath, ambient CO represents the CO concentration in the air measured simultaneously, and α is an individual metabolic factor with a default value of 1.0, and a value of 1.2 to 1.5 for smokers. The ambient CO concentration is dynamically subtracted from the exhaled CO concentration to ensure accurate detection of the user's CO concentration.

[0119] In one embodiment, during the user's breath sampling process for a preset duration, the smart wearable device provides a sampling countdown and real-time airflow intensity feedback. The preset duration is, for example, 5 or 10 seconds. After the user completes the breath sampling, a sampling completion prompt informs the user of the sampling result, indicating whether the sampling was successful or unsuccessful. For example, in the case of successful sampling, the sampling completion prompt can alert the user to successful sampling by flashing a green light and a short vibration, and can also output and display the detected nicotine concentration in the user's body. In the case of unsuccessful sampling, the prompt can prompt the user to re-blow by flashing a red light and outputting a re-sampling prompt message, such as "Insufficient airflow, please blow again."

[0120] The extracted air samples are also used as detection samples for the temperature sensor in the multimodal sensor array, allowing the temperature sensor to detect the ambient temperature. Optionally, the temperature sensor can also be exposed to the air for real-time ambient temperature detection; this is not specifically limited.

[0121] Furthermore, the smart wearable device also features a microphone hole and a display screen. The microphone hole is used to collect ambient sounds for smoke detection (such as the sound of a lighter or cough) to determine whether the user is smoking or whether there are smokers around the user. The display screen is used to display some or all of the collected multimodal data.

[0122] Optionally, the multimodal sensor array may also include an image sensor, which can be a camera. The acquired multimodal data may also include environmental images, which can be acquired with the user's authorization. These environmental images can be used to further assist in determining the scene category of the user's environment. For example, detecting the sound of a lighter igniting through ambient sound collected via a microphone hole, or identifying the image features of a burning cigarette from an environmental image, can help determine if the user is smoking, thus improving the accuracy of smoking behavior recognition. Optionally, the image sensor is positioned at a target location on the smart wearable device. This target location ensures that the user's hand movements can be captured when the user is smoking, even when the user is wearing the smart wearable device normally.

[0123] For intervention strategies generated by intelligent monitoring models, it is necessary to monitor the intervention effects on users. These effects are determined by monitoring whether users subsequently engage in smoking behavior, or whether they promptly cease smoking. Furthermore, to enhance users' motivation to quit smoking, social interaction features and smoking record functions should be provided.

[0124] Following step 300, the following may also be included:

[0125] Step 401: Monitor the intervention effect of the intervention strategy on the user;

[0126] Step 402: Based on the intervention effect and the recorded log, a smoking cessation report is generated and sent to the terminal application for the user to view; the recorded log is generated based on the user's smoking record operation and / or detected smoking behavior, according to a timed reminder task; the timed reminder task is used to remind the user to record smoking and learn smoking cessation techniques at regular intervals; the smoking cessation report includes smoking behavior statistics, physiological change trends, nicotine addiction assessment results, nicotine concentration change curve in the body, and smoking cessation results;

[0127] Step 403: If the user meets the reward conditions based on the smoking cessation results, a virtual object is issued to the user's account; the virtual object is used to redeem products or services.

[0128] Step 404: Push the smoking cessation results to the terminal application of the user's interactive user; the interactive user is a user who has a social relationship with the user.

[0129] The monitoring and intervention strategy is used to assess the effectiveness of the intervention on the user. Based on the intervention results and recorded logs, a smoking cessation report is generated and sent to the user's terminal application for review. The recorded logs are generated based on scheduled reminders, taking into account the user's smoking activity and / or detected smoking behavior. These scheduled reminders are used to periodically remind the user to record their smoking activity and learn smoking cessation techniques. The generated smoking cessation report includes smoking behavior statistics, physiological trends, nicotine addiction assessment results, nicotine concentration changes in the body, and smoking cessation achievements.

[0130] In the terminal application, user information management can be achieved, including displaying smoking cessation results, setting personalized smoking cessation plans, and social interaction. On the one hand, the smart wearable device provides users with a log recording interface for operation based on scheduled reminders. Users can also actively trigger log recording through buttons, manually recording the number of cigarettes smoked, feedback on smoking cessation intervention strategies, and feedback on intervention effects on the log recording interface provided by the smart wearable device. On the other hand, the smart wearable device monitors the user's smoking intentions or smoking behaviors in real time through collected multimodal data, generates intervention strategies and provides reminders, ensuring that users can receive smoking cessation intervention and care information in a timely manner.

[0131] Specifically, based on a user's smoking cessation results, and upon determining that the user meets the reward distribution conditions, a virtual object is sent to the user's account. This virtual object can be used to redeem products or services. The smoking cessation results are then pushed to the user's interactive user terminal application. Interactive users are those with whom the user has a social relationship. The reward distribution conditions refer to positive smoking cessation results, such as a significant decrease in the user's smoking frequency during the smoking cessation phase, or a reduction in the number of times the user experiences cravings for cigarettes within a monitoring period. Through the establishment of a social interaction mechanism, users can invite other users to create smoking cessation groups. Group members can view the user's smoking cessation progress and results, providing encouragement and support. Alternatively, group members can view each other's smoking cessation results for mutual encouragement. Simultaneously, through the setting of an incentive mechanism, based on the user's smoking cessation results, such as completed smoking cessation tasks and reduced smoking, corresponding virtual objects are issued as rewards. These virtual objects can be points, badges, or coupons, which users can redeem for smoking cessation or health-related products or services, motivating users to persist in quitting smoking.

[0132] In some embodiments, smoking cessation reports can also be uploaded to a server via a terminal application for secure storage and in-depth analysis. The server summarizes and analyzes the smoking cessation data of different users to uncover common problems and patterns in the smoking cessation process, thereby further optimizing the intelligent engine and improving the generation of smoking cessation intervention strategies. At the same time, cloud storage can ensure the security and traceability of user data, and users can obtain complete smoking cessation data at any time when changing devices or needing to review their smoking cessation journey.

[0133] It should be noted that during communication and data transmission, smart wearable devices, terminal applications, and servers encrypt and control the transmitted data, and de-identify sensitive user information. Users have complete control over their own data and can choose to delete or modify any data, thereby ensuring the privacy and security of user data.

[0134] In this embodiment, a smart wearable device is used, allowing users to conveniently wear it anytime, anywhere for real-time monitoring of smoking behavior and physiological state, as well as intervention for smoking cessation. It is portable and real-time, reducing manual recording by users, lowering usage costs and burdens, and improving data timeliness and accuracy. Furthermore, the device's functionality is enriched through real-time reminders, emotion regulation, social support, and incentive mechanisms. The intervention methods are flexible and diverse, providing users with comprehensive care and support, making the smoking cessation process easier and more manageable, enhancing user adherence and persistence to the cessation plan, and improving the user experience.

[0135] Furthermore, multimodal data fusion reduces the misjudgment rate of smoking intentions or behaviors. Through more accurate and comprehensive monitoring of smoking behavior and the development of personalized smoking cessation programs, customized interventions are provided based on each user's tobacco dependence characteristics and individual circumstances. This effectively helps users overcome physiological and psychological dependence and improves the success rate of quitting smoking. Combined with continuous optimization of intelligent monitoring algorithms based on data-driven approaches, the collected and analyzed user smoking cessation data provides a basis for continuous optimization and improvement of real-time monitoring, thereby enabling more accurate and effective smoking cessation intervention services to meet the diverse needs of different users.

[0136] Smart wearable devices have broken through the limitations of traditional smoking cessation interventions, achieving end-to-end intelligent management from data collection and risk prediction to precise intervention. They construct a three-pronged intervention mechanism encompassing environment, physiology, and behavior, generating intervention strategies in real time based on the predicted probability and intensity of cravings. This allows for dynamic, tiered, and personalized intervention for users seeking to quit smoking. Furthermore, through diverse intervention content such as positive guidance, exercise reminders, and hazard warnings, they provide smokers with efficient, precise, and personalized smoking cessation support, demonstrating significant application value in the field of intelligent healthcare.

[0137] The following describes the smoking cessation intervention device for smokers provided by the present invention. The smoking cessation intervention device described below can be referred to in correspondence with the smoking cessation intervention method described above.

[0138] Reference Figure 3 The smoking cessation intervention device provided in this embodiment of the invention is applied to a smart wearable device. The smoking cessation intervention device includes:

[0139] The data acquisition module 10 is used to collect the user's multimodal data; the multimodal data includes physiological data, behavioral data, and environmental data.

[0140] The smoking monitoring module 20 is used to input the multimodal data into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels.

[0141] The smoking cessation intervention module 30 is used to conduct smoking cessation intervention on the user based on the intervention strategy.

[0142] In one embodiment, the intelligent monitoring model includes a multimodal recognition module, used for:

[0143] Based on the physiological data in the multimodal data, the changing trends of the user's physiological characteristics related to smoking cravings are determined;

[0144] Based on the behavioral data in the multimodal data, the user's pose changes are identified; the pose changes include the user's arm movement trajectory and arm raising frequency;

[0145] The environmental data in the multimodal data is classified into scenarios to determine the scenario category of the user's environment; the scenario category is used to characterize the degree of influence of the environment on the user's urge to smoke;

[0146] By comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the category of the scene, the probability and degree of the user's craving for smoking are determined within a future preset time period.

[0147] An intervention level is determined based on the probability and the degree of smoking craving, and an intervention strategy is generated based on the intervention level; the intervention strategy includes intervention content under the intervention level.

[0148] In one embodiment, the smoker cessation intervention device further includes a model optimization module for:

[0149] The multimodal data is sent to the terminal application, which iteratively optimizes the intelligent monitoring model based on the multimodal data to obtain the first model parameters, and then sends the first model parameters to the server.

[0150] The server periodically iterates and optimizes the intelligent engine corresponding to the intelligent monitoring model based on the first model parameters sent by different users' terminal applications to obtain the second model parameters.

[0151] The system receives the second model parameters sent by the server and uses the second model parameters to optimize and update the intelligent monitoring model.

[0152] In one embodiment, the data acquisition module 10 includes a multimodal sensor array, which is used to acquire the multimodal data; wherein:

[0153] The multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor;

[0154] The biosensor is used to collect the physiological data, which includes at least one of heart rate, blood pressure, blood glucose, blood oxygen, psychological state data, sleep data, and skin conductivity.

[0155] The motion sensor is used to collect the behavioral data, which includes at least one of the user's motion data, hand movements, hand raising frequency, and arm movement trajectory.

[0156] The carbon monoxide sensor and the temperature sensor are used to collect the environmental data, which includes at least the ambient temperature and the ambient carbon monoxide concentration; the ambient nicotine concentration is calculated based on the ambient carbon monoxide concentration.

[0157] In one embodiment, the data acquisition module 10 further includes at least one environmental monitoring port for extracting air samples from the user's environment and collecting exhaled gas samples from the user; the physiological data also includes the concentration of nicotine in the body; the data acquisition module 10 is also used for;

[0158] Based on a pre-generated timed sampling task, an air sample is extracted through the environmental monitoring hole, and the first concentration of carbon monoxide in the air sample is detected.

[0159] Output sampling prompt information; the sampling prompt information is used to prompt the user to perform a blowing operation on the environmental monitoring hole for a preset duration;

[0160] When the user's exhalation is detected to be complete, a sample of the user's exhaled gas is obtained, and a second concentration of carbon monoxide in the exhaled gas sample is detected.

[0161] The user's nicotine concentration is estimated based on the difference between the first concentration and the second concentration.

[0162] The data acquisition module 10 also includes a microphone hole, which is used to collect ambient sound for smoke detection.

[0163] The smoker cessation intervention device also includes a display module, which includes a display screen for displaying part or all of the multimodal data.

[0164] In one embodiment, the smoker cessation intervention device further includes an interaction and management module for:

[0165] Monitor the effectiveness of the intervention strategy on the user;

[0166] Based on the intervention effect and the recorded logs, a smoking cessation report is generated and sent to the terminal application for the user to view. The recorded logs are generated based on the user's smoking record operations and / or detected smoking behavior, using a timed reminder task to remind the user to record smoking and learn smoking cessation techniques. The smoking cessation report includes smoking behavior statistics, physiological change trends, nicotine addiction assessment results, nicotine concentration change curves in the body, and smoking cessation achievements.

[0167] If the user meets the reward conditions based on the smoking cessation results, a virtual object is issued to the user's account; the virtual object is used to redeem products or services.

[0168] The smoking cessation results are pushed to the terminal application of the user's interactive users; the interactive users are users who have social relationships with the user.

[0169] In this embodiment, the data acquisition module collects multimodal data through a multimodal sensor array. The smoking monitoring module identifies and analyzes the multimodal data collected by the data acquisition module based on an intelligent monitoring model to determine the probability and degree of the user's craving for cigarettes. Based on the probability and degree of the user's craving for cigarettes, an intervention strategy is generated to achieve real-time and comprehensive monitoring of the user's physiological state, behavior, and environment, as well as dynamic and personalized intervention based on the intervention strategy. This realizes intelligent management of the entire chain from data acquisition and risk prediction to precise intervention, and constructs a dynamic intervention mechanism integrating environment, physiology, and behavior to provide users with efficient, accurate, and personalized smoking cessation support and enhance the intervention effect.

[0170] This invention also provides a smart wearable device for smoking cessation intervention. The main body of the smart wearable device includes a back shell and a display screen. The back shell includes a concave surface and a convex surface. The display screen is embedded in the concave surface of the back shell, and a multimodal sensor array is arranged on the convex surface of the back shell. The multimodal sensor array is used to collect multimodal data.

[0171] The smart wearable device also includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the computer program, it implements the smoking cessation intervention methods for smokers as described in the above embodiments, which will not be repeated here.

[0172] In one embodiment, the multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor. Optionally, the multimodal sensor array may also include an air quality sensor and an image sensor.

[0173] The back cover includes a first side and a second side, as well as a third side and a fourth side, arranged opposite to each other. Environmental monitoring holes for extracting gas samples required for detection are provided on the first and second sides, and these holes are covered with a hydrophobic film. A microphone hole and a magnetic charging contact are located on either the first or second side. Wearing fastening components are provided on the third and fourth sides. Optionally, if the smart wearable device is a smartwatch, the fastening components are watch straps arranged opposite to each other on the third and fourth sides.

[0174] In this design, the convex surface of the back shell is the side that fits snugly against the skin of the wearing area, while the first, second, third, and fourth sides are the sides between the convex and concave surfaces of the back shell. In one embodiment, refer to... Figure 4 The image shows a front view of the smart wearable device. The display screen is embedded in the concave surface of the back cover, and the power button and emergency button are also located on the fourth side of the back cover.

[0175] Further reference Figure 5 The image shows a rear view of a smart wearable device. A multimodal sensor array is arranged on the convex surface of the device's back cover. Figure 5 The example provided illustrates a blood pressure sensor, a carbon monoxide sensor, and a temperature sensor in a multimodal sensor array. For example... Figure 5 As shown, environmental monitoring holes for sampling are provided on both the first and second sides of the back cover. These holes are covered with a hydrophobic membrane to prevent saliva and water vapor from entering during gas sampling. The hydrophobic membrane can be made of polytetrafluoroethylene (PTFE), and it can be applied to the environmental monitoring holes, either covering the openings or located on the inner surface of the holes; there is no specific limitation in this regard. Optionally, the back cover of the smart wearable device is made of ceramic material, and the back cover also includes a metal frame.

[0176] Figure 6 This example illustrates another physical structure diagram of a smart wearable device, such as... Figure 6 As shown, the smart wearable device may include: a processor 610, a communications interface 620, a memory 630, and a communication bus 640. The processor 610, communications interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions in the memory 630 to execute steps of a smoker cessation intervention method, such as:

[0177] Collect multimodal data from users; the multimodal data includes physiological data, behavioral data, and environmental data;

[0178] The multimodal data is input into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels.

[0179] The user was given a smoking cessation intervention based on the intervention strategy.

[0180] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0181] On the other hand, the present invention also provides a computer program product comprising a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to perform the steps of the smoking cessation intervention methods provided by the above methods, for example including:

[0182] Collect multimodal data from users; the multimodal data includes physiological data, behavioral data, and environmental data;

[0183] The multimodal data is input into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels.

[0184] The user was given a smoking cessation intervention based on the intervention strategy.

[0185] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the smoking cessation intervention methods provided by the above methods, including, for example:

[0186] Collect multimodal data from users; the multimodal data includes physiological data, behavioral data, and environmental data;

[0187] The multimodal data is input into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels.

[0188] The user was given a smoking cessation intervention based on the intervention strategy.

[0189] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0190] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0191] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smoking cessation intervention method for smokers, characterized in that, The smoking cessation intervention method, applied to smart wearable devices, includes: Collect multimodal data from users; the multimodal data includes physiological data, behavioral data, and environmental data; the environmental data includes environmental audio data; The multimodal data is input into the intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels. The user is given a smoking cessation intervention based on the aforementioned intervention strategy; the intervention strategy includes the intervention method, the intervention level, and the intervention content under the intervention level. The smoking cessation intervention for the user based on the intervention strategy includes: Using the aforementioned intervention method, the intervention content at the aforementioned intervention level is output; the intervention content includes positive guidance, exercise reminders, and hazard warnings; The process by which the intelligent monitoring model processes the multimodal data to obtain an intervention strategy includes: Based on the physiological data in the multimodal data, the changing trends of the user's physiological characteristics related to smoking cravings are determined; Based on the behavioral data in the multimodal data, the user's pose changes are identified; the pose changes include the user's arm movement trajectory and arm raising frequency; The environmental data in the multimodal data is classified into scenarios to determine the scenario category of the user's environment; the scenario category is used to characterize the degree of influence of the environment on the user's urge to smoke; By comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the category of the scene, the probability and degree of the user's craving for smoking are determined within a future preset time period. The intervention level is determined based on the probability and the degree of smoking craving, and an intervention strategy is generated based on the intervention level. The step of classifying the environmental data in the multimodal data to determine the scene category of the user's environment includes: Based on the environmental audio data, smoke sound detection is performed to determine whether the user is in a smoking environment; The intelligent monitoring model includes a first network branch, a second network branch, and a third network branch: the first network branch outputs the changing trends of the user's physiological characteristics related to smoking cravings; the second network branch outputs the changes in the user's posture; and the third network branch outputs the scene category of the user's environment. The method of comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the scene category to determine the probability and degree of the user's craving for smoking within a preset time period includes: Based on the output features of the first network branch, the second network branch, and the third network branch, predictions are made to obtain the first probability and the first degree of smoking craving corresponding to the first network branch, the second probability and the second degree of smoking craving corresponding to the second network branch, and the third probability and the third degree of smoking craving corresponding to the third network branch. The first probability, the second probability, and the third probability are weighted and summed or weighted and averaged to obtain the probability that the user will have a craving for smoking in the future within a preset time period. The user's smoking craving level is obtained by weighted summation or weighted average of the first smoking craving level, the second smoking craving level, and the third smoking craving level over a future preset time period.

2. The smoking cessation intervention method for smokers according to claim 1, characterized in that, Following the collection of the user's multimodal data, the following is also included: The multimodal data is sent to the terminal application, which iteratively optimizes the intelligent monitoring model based on the multimodal data to obtain the first model parameters, and then sends the first model parameters to the server. The server periodically iterates and optimizes the intelligent engine corresponding to the intelligent monitoring model based on the first model parameters sent by different users' terminal applications to obtain the second model parameters. The system receives the second model parameters sent by the server and uses the second model parameters to optimize and update the intelligent monitoring model.

3. The smoking cessation intervention method for smokers according to claim 1, characterized in that, The smart wearable device is equipped with a multimodal sensor array, which is used to collect the multimodal data; wherein: The multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor; The biosensor is used to collect the physiological data, which includes at least one of heart rate, blood pressure, blood glucose, blood oxygen, psychological state data, sleep data, and skin conductivity. The motion sensor is used to collect the behavioral data, which includes at least one of the user's motion data, hand movements, hand raising frequency, and arm movement trajectory. The carbon monoxide sensor and the temperature sensor are used to collect the environmental data, which includes at least the ambient temperature and the ambient carbon monoxide concentration; the ambient nicotine concentration is calculated based on the ambient carbon monoxide concentration.

4. The smoking cessation intervention method for smokers according to claim 3, characterized in that, The smart wearable device also has at least one environmental monitoring port for extracting air samples from the user's environment and collecting samples of the user's exhaled breath; the physiological data also includes the concentration of nicotine in the body; the collection of the user's multimodal data includes: Based on a pre-generated timed sampling task, an air sample is extracted through the environmental monitoring hole, and the first concentration of carbon monoxide in the air sample is detected. Output sampling prompt information; the sampling prompt information is used to prompt the user to perform a blowing operation on the environmental monitoring hole for a preset duration; When the user's exhalation is detected to be complete, a sample of the user's exhaled gas is obtained, and a second concentration of carbon monoxide in the exhaled gas sample is detected. The user's nicotine concentration is estimated based on the difference between the first concentration and the second concentration. The smart wearable device is also equipped with a microphone hole, which is used to collect ambient sound for smoke detection. The smart wearable device also includes a display screen, which is used to display part or all of the multimodal data.

5. The smoking cessation intervention method for smokers according to claim 1, characterized in that, After implementing the smoking cessation intervention on the user based on the intervention strategy, the process further includes: Monitor the effectiveness of the intervention strategy on the user; Based on the intervention effect and the recorded logs, a smoking cessation report is generated and sent to the terminal application for the user to view. The recorded logs are generated based on the user's smoking record operations and / or detected smoking behavior, using a timed reminder task to remind the user to record smoking and learn smoking cessation techniques. The smoking cessation report includes smoking behavior statistics, physiological change trends, nicotine addiction assessment results, nicotine concentration change curves in the body, and smoking cessation achievements. If the user meets the reward conditions based on the smoking cessation results, a virtual object is issued to the user's account; the virtual object is used to redeem products or services. The smoking cessation results are pushed to the terminal application of the user's interactive users; the interactive users are users who have social relationships with the user.

6. A smoking cessation intervention device for smokers, characterized in that, The smoker cessation intervention device, applied to smart wearable devices, includes: The data acquisition module is used to collect the user's multimodal data; the multimodal data includes physiological data, behavioral data, and environmental data; the environmental data includes environmental audio data. A smoking monitoring module is used to input the multimodal data into an intelligent monitoring model to obtain the intervention strategy output by the intelligent monitoring model. The intelligent monitoring model is built based on a neural network model and trained with sample data. The sample data is constructed using historical data of the multimodal data as samples and intervention strategies as sample labels. The sample labels are marked with intervention levels. A smoking cessation intervention module is used to conduct smoking cessation intervention on the user based on the intervention strategy. The intervention strategy includes intervention methods, intervention levels, and intervention content at each intervention level; the smoking cessation intervention module is specifically used for: Using the aforementioned intervention method, the intervention content at the aforementioned intervention level is output; the intervention content includes positive guidance, exercise reminders, and hazard warnings; The intelligent monitoring model includes a multimodal recognition module, used for: Based on the physiological data in the multimodal data, the changing trends of the user's physiological characteristics related to smoking cravings are determined; Based on the behavioral data in the multimodal data, the user's pose changes are identified; the pose changes include the user's arm movement trajectory and arm raising frequency; The environmental data in the multimodal data is classified into scenarios to determine the scenario category of the user's environment; the scenario category is used to characterize the degree of influence of the environment on the user's urge to smoke; By comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the category of the scene, the probability and degree of the user's craving for smoking are determined within a future preset time period. The intervention level is determined based on the probability and the degree of smoking craving, and an intervention strategy is generated based on the intervention level. The step of classifying the environmental data in the multimodal data to determine the scene category of the user's environment includes: Based on the environmental audio data, smoke sound detection is performed to determine whether the user is in a smoking environment; The intelligent monitoring model includes a first network branch, a second network branch, and a third network branch: the first network branch outputs the changing trends of the user's physiological characteristics related to smoking cravings; the second network branch outputs the changes in the user's posture; and the third network branch outputs the scene category of the user's environment. The method of comprehensively analyzing the changing trends of the physiological characteristics, the changes in posture, and the scene category to determine the probability and degree of the user's craving for smoking within a preset time period includes: Based on the output features of the first network branch, the second network branch, and the third network branch, predictions are made to obtain the first probability and the first degree of smoking craving corresponding to the first network branch, the second probability and the second degree of smoking craving corresponding to the second network branch, and the third probability and the third degree of smoking craving corresponding to the third network branch. The first probability, the second probability, and the third probability are weighted and summed or weighted and averaged to obtain the probability that the user will have a craving for smoking in the future within a preset time period. The user's smoking craving level is obtained by weighted summation or weighted average of the first smoking craving level, the second smoking craving level, and the third smoking craving level over a future preset time period.

7. A smart wearable device for smoking cessation intervention, characterized in that, The main body of the smart wearable device includes a back shell and a display screen. The back shell includes a concave surface and a convex surface, and the display screen is embedded in the concave surface of the back shell. A multimodal sensor array is arranged on the convex surface of the back shell. The smart wearable device further includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the smoking cessation intervention method for smokers as described in any one of claims 1 to 5.

8. The smart wearable device for smoking cessation intervention according to claim 7, characterized in that, The multimodal sensor array includes at least one carbon monoxide sensor, at least one temperature sensor, at least one biosensor, and at least one motion sensor; The back cover also includes a first side and a second side arranged opposite to each other, as well as a third side and a fourth side arranged opposite to each other; the first side and the second side are provided with environmental monitoring holes for extracting gas samples required for detection, and the environmental monitoring holes are covered with a hydrophobic film; a microphone hole is provided on the first side or the second side, a magnetic charging contact is provided on the first side or the second side, and a fixing component for wearing is provided on the third side and the fourth side.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the smoking cessation intervention method as described in any one of claims 1 to 5.

Images