Equipment control method and system and intelligent ring

Abstract

The invention discloses a device control method and system and an intelligent ring, and relates to the technical field of wearable devices, the device control method and system are applied to the intelligent ring, the intelligent ring comprises a motion sensor, the intelligent ring is connected with a terminal device, and the method comprises the following steps: obtaining motion data collected by the motion sensor, and performing feature extraction on the motion data to obtain feature data; inputting the feature data into a pre-trained gesture classification model to obtain a gesture classification result; and obtaining a control instruction corresponding to the gesture classification result, and controlling the terminal equipment based on the control instruction. The practicability of the intelligent ring is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wearable devices, in particular to a device control method, system and smart ring. BACKGROUND

[0002] With the continuous development of the smart wearable device market, smart rings as a new type of wearable product have gradually attracted market attention. However, the current smart rings mostly focus on a single function, such as NFC (Near Field Communication) payment or simple activity counting, and it is difficult to meet the diversified needs of users in complex scenarios, resulting in low practicality of smart rings.

[0003] Therefore, how to improve the practicality of the smart ring is a technical problem to be solved at present. SUMMARY

[0004] The main purpose of the present application is to provide a device control method, system and smart ring, aiming to solve the technical problem of how to improve the practicality of the smart ring.

[0005] To achieve the above purpose, the present application provides a device control method, which is applied to a smart ring, the smart ring comprising a motion sensor, the smart ring being connected with a terminal device, and the device control method comprising the following steps: acquiring motion data collected by the motion sensor, and performing feature extraction on the motion data to obtain feature data; inputting the feature data into a pre-trained gesture classification model to obtain a gesture classification result; acquiring a control instruction corresponding to the gesture classification result, and controlling the terminal device based on the control instruction.

[0006] In an embodiment, the step of acquiring the control instruction corresponding to the gesture classification result comprises: acquiring a device identifier of the terminal device; searching for the control instruction corresponding to the gesture classification result in a preset instruction library according to the device identifier.

[0007] In an embodiment, the motion data comprises acceleration data and angular velocity data, and the step of performing feature extraction on the motion data to obtain feature data comprises: acquiring acceleration mean value, acceleration variance and acceleration peak value based on the acceleration data; performing integral calculation on the angular velocity data to obtain an angle change amount; acquiring a gesture trajectory based on the acceleration data, and calculating signal energy of the gesture trajectory in a frequency domain to obtain frequency spectrum energy; combining the acceleration mean value, the acceleration variance, the acceleration peak value, the angle change amount and the spectral energy to obtain feature data.

[0008] In an embodiment, the step of inputting the feature data into a pre-trained gesture classification model to obtain a gesture classification result comprises: inputting the feature data into a pre-trained gesture classification model to output a gesture classification probability vector; obtaining a preset probability threshold, and searching for a maximum classification probability in the gesture classification probability vector; if the maximum classification probability is greater than or equal to the probability threshold, determining a gesture category indicated by the maximum classification probability as the gesture classification result; if the maximum classification probability is less than the probability threshold, determining that the current gesture recognition fails.

[0009] In an embodiment, the feature data comprises an acceleration mean value, an acceleration variance, an angle change amount and a spectral energy of a gesture trajectory, and the step of obtaining a preset probability threshold comprises: selecting at least one parameter from the acceleration mean value, the acceleration variance and the angle change amount as time domain reference data; performing weighted summation calculation on the time domain reference data and the spectral energy to obtain gesture energy; setting a probability threshold based on the gesture energy, wherein the probability threshold and the gesture energy have a positive correlation trend.

[0010] In an embodiment, the step of setting a probability threshold based on the gesture energy comprises: if the motion data indicates that the smart ring is in a stationary state, calculating noise energy based on the acceleration variance, wherein the noise energy is positively correlated with the acceleration variance; setting a probability threshold based on the gesture energy and the noise energy, wherein the probability threshold and the noise energy have a negative correlation trend.

[0011] In an embodiment, the smart ring further comprises a physiological sensor, and the device control method further comprises: obtaining physiological data collected by the physiological sensor, and calculating a physiological parameter based on the physiological data; if the physiological parameter matches a preset health abnormality parameter threshold, controlling the terminal device to perform health abnormality prompting.

[0012] In an embodiment, the physiological data comprises a PPG signal, an ECG signal and a PCG signal, the physiological parameter comprises a heart rate of the user, and the step of calculating the physiological parameter based on the physiological data comprises: calculating a signal quality of the PPG signal, the ECG signal and the PCG signal, respectively; setting a fusion weight of the PPG signal, the ECG signal and the PCG signal based on the signal quality, respectively, wherein the fusion weight is positively correlated with the signal quality; calculating a first heart rate based on the PPG signal, a second heart rate based on the ECG signal, and a third heart rate based on the PCG signal; performing weighted sum calculation on the first heart rate, the second heart rate and the third heart rate according to the fusion weight corresponding to the PPG signal, the ECG signal and the PCG signal, to obtain the heart rate of the user.

[0013] In addition, to achieve the above object, the present application further provides a device control system, which comprises a smart ring and a terminal device in communication connection, wherein the smart ring comprises a motion sensor. The smart ring is configured to: acquire motion data collected by the motion sensor, and perform feature extraction on the motion data to obtain feature data; input the feature data into a pre-trained gesture classification model to obtain a gesture classification result; acquire a control instruction corresponding to the gesture classification result, and send the control instruction to the terminal device; The terminal device is configured to: execute the control instruction.

[0014] In addition, to achieve the above object, the present application further provides a smart ring, which comprises a motion sensor and a processor, wherein the motion sensor is configured to collect motion data, and the processor is configured to execute the device control method as described above.

[0015] The one or more technical solutions provided by the present application have at least the following technical effects: The embodiments of the present application realize gesture-based device control by integrating a motion sensor in the smart ring and combining a terminal device, and improve the application value and practicality of the smart ring in complex scenarios. Specifically, the motion data collected by the motion sensor in the smart ring is input into a gesture classification model after feature extraction, and then the gesture classification result and the corresponding control instruction are obtained, thereby realizing flexible control of the terminal device through gestures. This makes the smart ring no longer limited to single functions such as simple NFC payment or activity counting, but can realize various function controls according to the gesture operations of the user in different scenarios, so that a ring can issue control commands to the connected terminal devices such as mobile phones and computers in various scenarios such as multimedia playing, page browsing and smart home control, and the smart ring is transformed from a function-fixed special device into a "control center" that can understand user intentions and realize diversified control of terminal devices, thereby greatly enriching the functional extension of the smart ring, making it better adapt to the diversified needs of users in various complex scenarios, and effectively improving the practicality of the smart ring, providing users with a more convenient and intelligent use experience. BRIEF DESCRIPTION OF DRAWINGS

[0016] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and serve to explain the principles of the present application together with the specification.

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, those skilled in the art can obtain other drawings according to these drawings without any creative effort.

[0018] Figure 1 A flowchart of a first embodiment of the device control method of the present application; Figure 2 A device control flowchart related to an embodiment of the device control method of the present application; Figure 3 A flowchart of a second embodiment of the device control method of the present application; Figure 4 A flowchart of a third embodiment of the device control method of the present application Figure 5 A system architecture diagram of the device control system of the present application.

[0019] The object implementation, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0020] In order to make the above objectives, characteristics and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0021] Existing smart rings mainly focus on a single function, such as NFC payment or simple activity counting, and it is difficult to achieve high-precision gesture recognition and cross-device interaction. Especially in virtual reality (VR), augmented reality (AR) and mobile office scenarios, users urgently need a light, accurate and low-latency non-contact control method.

[0022] Based on this, the main solution of the present application is to provide a smart ring provided with a motion sensor, acquire motion data collected by the motion sensor, perform feature extraction on the motion data to obtain feature data, input the feature data into a pre-trained gesture classification model to obtain a gesture classification result, acquire a control instruction corresponding to the gesture classification result, and control a terminal device based on the control instruction.

[0023] The motion data collected by the motion sensor in the smart ring is input into the gesture classification model after feature extraction, and then the gesture classification result and the corresponding control instruction are obtained, so that flexible control of the terminal device by gestures is realized. This makes the smart ring no longer limited to simple NFC payment or activity counting and other single functions, but can realize various function controls according to the gesture operations of the user in different scenarios, so that a ring can issue control commands to the connected terminal devices such as mobile phones and computers in multimedia playback, page browsing, smart home control and other scenarios, so that the smart ring changes from a function-fixed special device to a "control center" that can understand the user's intention and realize diversified control of the terminal device, thereby greatly enriching the functional extension of the smart ring, making it better adapt to the diversified needs of users in various complex scenarios, thereby effectively improving the practicality of the smart ring and providing users with a more convenient and intelligent use experience.

[0024] Based on this, the present application proposes a device control method of the first embodiment, which is applied to a smart ring, the smart ring includes a motion sensor, and the smart ring is connected with a terminal device. As shown in Figure 1 The device control method includes the following steps: Step S10, acquiring motion data collected by the motion sensor, performing feature extraction on the motion data to obtain feature data; The smart ring refers to an electronic device that can be worn on a finger by a user, which is similar in shape to a traditional ring, but its internal processing chip and communication module are usually integrated to perform data calculation and wireless communication with external devices, so as to realize specific interaction and control functions.

[0025] The terminal device refers to a computing device directly operated by a user to access network services or resources, such as a mobile phone, a VR glasses, a computer, etc., which is not specifically limited in the embodiment.

[0026] The motion sensor refers to a sensor for detecting and measuring the motion of an object, for example, an IMU (Inertial Measurement Unit), which can include one or more combinations of a three-axis accelerometer and a three-axis gyroscope, for collecting acceleration and / or angular velocity data of the smart ring in three-dimensional space, thereby forming the motion data. For example, the motion sensor is an IMU unit including a three-axis accelerometer and a three-axis gyroscope, and the corresponding motion data is the acceleration data and angular velocity data collected by the IMU unit.

[0027] The IMU unit of the smart ring collects acceleration and angular velocity data based on a configurable sampling period. In actual application, in order to balance power consumption and performance, a dynamic sampling strategy can be used: after the ring is worn, it automatically enters a low-power monitoring mode to continuously collect data at a lower sampling rate; when a specific wake-up gesture (such as a quick double-click of the fingers) is detected or a wake-up instruction sent by the terminal device is received, it switches to a full-function mode to collect data at a higher sampling rate, ensuring the accuracy required for subsequent gesture recognition.

[0028] After obtaining the acceleration data and angular velocity data, feature extraction is performed. The extracted feature data includes but is not limited to the mean, variance, and peak value of the three-axis acceleration; the angle change obtained by integrating the three-axis angular velocity; and the time domain and frequency domain features such as gesture trajectory spectrum energy obtained by fast Fourier transform. These feature data can effectively represent the characteristics of the user's gesture and provide reliable basis for subsequent gesture classification.

[0029] Step S20, inputting the feature data into a pre-trained gesture classification model to obtain a gesture classification result; The gesture classification model is a lightweight machine learning model deployed to the smart ring after training on a server or high-performance computing device, such as a neural network model optimized based on the TinyML (Tiny Machine Learning) framework. The training data set contains sample data of various user gestures, which are labeled and classified to train the model to establish a mapping relationship between gesture features and gesture categories. The well-trained model can classify the input feature data and recognize gestures such as waving, pinching, rotating, and tapping, and classify them into corresponding predefined categories, such as the "waving" category, the "pinching" category, etc.

[0030] In step S30, the control instruction corresponding to the gesture classification result is obtained to control the terminal device based on the control instruction.

[0031] The control instruction corresponding to the gesture classification result can be obtained from a preset instruction library. The control instruction is predefined according to the user's needs and application scenarios, and is used to instruct the terminal device to perform specific operations. For example, if the gesture classification result is "waving", the corresponding control instruction may be "play the next song"; if the gesture classification result is "pinching", the corresponding control instruction may be "zoom in the picture"; if the gesture classification result is "rotation", the corresponding control instruction may be "adjust the volume".

[0032] The control instruction can be sent to the terminal device connected to the smart ring through the communication module of the smart ring. After receiving the control instruction, the terminal device parses and executes the corresponding operation, thereby realizing the user's convenient control of the terminal device through gestures. This control method provides a natural, intuitive and efficient interaction experience for the user, greatly improving the practicality and user experience of the smart ring.

[0033] Further, to improve communication security, the smart ring can perform encryption processing on the control instruction before sending it, such as using the AES-128 encryption algorithm to encrypt the instruction data, and sending the encrypted control instruction to the terminal device through the communication module to ensure the security and privacy protection of data transmission. After receiving the encrypted instruction, the terminal device decrypts and parses to execute the corresponding operation, thereby realizing the user's safe and convenient control of the terminal device through natural gestures. This provides an intuitive, efficient and secure interaction experience for the user, and enhances the practical value and user experience of the smart ring through a rich gesture instruction library and reliable security mechanism.

[0034] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as the above embodiment one can be referred to the above introduction, and the subsequent will not be described. On this basis, the motion data includes acceleration data and angular velocity data, the step of extracting features from the motion data to obtain feature data, comprising: Step A10, based on the acceleration data to obtain acceleration mean, acceleration variance and acceleration peak value; For the X, Y, Z three axis acceleration data collected by three-axis accelerometer, the statistical characteristics of each axis are calculated respectively: the acceleration mean is used to represent the average acceleration level in the process of gesture motion, which is obtained by calculating the average value of acceleration data in a certain time window (for example, 200 milliseconds). The acceleration variance reflects the fluctuation degree of acceleration data, which is obtained by calculating the average value of the square deviation of acceleration data from the mean value, which can reflect the stability and complexity of gesture motion. The acceleration peak value represents the maximum value of acceleration in the process of gesture motion, which is obtained by detecting the maximum value of acceleration data, which can be used to identify the rapid action or impact characteristics in gesture.

[0035] Step A20, integrating the angular velocity data to obtain the angle change; By numerical integration of the angular velocity data collected by the three-axis gyroscope in the time dimension, the rotation angle change around the X, Y, Z three coordinate axes is calculated respectively. This angle change can accurately reflect the rotation posture change of the hand in the process of gesture execution, and provide key motion features for recognizing rotation gestures.

[0036] Step A30, based on the acceleration data to obtain gesture trajectory, and calculate the signal energy of the gesture trajectory in frequency domain to obtain frequency spectrum energy; According to the acceleration data to reconstruct the gesture trajectory: by integrating the acceleration data in time series, the velocity information of gesture motion can be obtained, and then the velocity information is integrated, that is, the spatial trajectory of gesture is obtained, that is, the gesture trajectory is obtained.

[0037] The gesture trajectory reflects the motion path of the gesture in three-dimensional space. Then, the gesture trajectory signal is converted from time domain to frequency domain, and the frequency spectrum energy of the gesture trajectory is calculated by Fourier transform and other methods. The frequency spectrum energy can reflect the energy distribution of different frequency components in the gesture trajectory, and the high frequency component is usually related to the rapid change or complex motion of the gesture, while the low frequency component is related to the overall motion trend of the gesture. This frequency domain feature can effectively capture the periodic motion pattern and frequency characteristics of the gesture, and enhance the recognition ability of gestures with specific rhythm or repetition characteristics.

[0038] Step A40, combine the acceleration mean, the acceleration variance, the acceleration peak, the angle change and the spectral energy to obtain feature data.

[0039] The extracted features are combined to form complete feature data. Specifically, the acceleration mean, acceleration variance, acceleration peak, angle change and spectral energy are combined into a feature vector. This feature vector describes the acceleration, rotation and frequency characteristics of the gesture, and can represent the features of the gesture from multiple dimensions.

[0040] This embodiment extracts and fuses the feature parameters of linear motion and rotational motion from two dimensions of time domain and frequency domain, constructs a feature vector with complementary information and comprehensive representation, and realizes high-precision and high-robustness gesture feature representation on the resource-limited smart ring. Specifically, the mean, variance and peak extracted from the acceleration data accurately describe the linear motion characteristics of the gesture from three different aspects of overall trend, motion stability and instantaneous impact, ensuring the basic distinguishing ability for actions with different motion amplitudes, rhythms and intensities. The angle change obtained by integrating the angular velocity introduces three-dimensional space rotation information, which can accurately capture complex posture changes such as wrist turning, and greatly expands the category range of identifiable gestures. More importantly, the spectral energy feature obtained by frequency domain analysis of the gesture trajectory can effectively capture the periodic pattern and rhythm characteristics of the gesture, which forms a strong complement to the aforementioned time domain statistical features, so that subtle gesture differences with similar time domain patterns but different frequency distributions can also be effectively identified. Finally, these features representing linear acceleration, rotation angle and frequency characteristics are combined into a unified multi-dimensional feature vector to produce a synergistic effect. This feature set not only ensures computational efficiency to adapt to embedded devices, but also provides high discrimination and strong anti-interference input data for subsequent classification models, laying a technical foundation for precise and reliable device control of the smart ring.

[0041] In a possible implementation, the step of obtaining the control instruction corresponding to the gesture classification result comprises: Step B10, obtaining a device identifier of the terminal device; Obtain the device identifier of the terminal device connected with the smart ring. The device identifier is a unique identifier of the terminal device, used to distinguish different terminal devices and their functions. For example, the device identifier can be the MAC address of the device, the device model, the device name customized by the user, or other unique identification codes. By obtaining the device identifier, the specific type and function of the terminal device currently connected with the smart ring can be determined, thereby providing a basis for subsequent control instruction matching. For example, if the device identifier corresponds to a mobile phone, the control instruction set related to the mobile phone will be called; if the device identifier corresponds to a VR glasses, the control instruction set related to the VR device will be called.

[0042] Step B20, according to the device identifier, in the preset instruction library, find the control instruction corresponding to the gesture classification result.

[0043] According to the obtained device identifier, find the control instruction corresponding to the device identifier and the gesture classification result from the preset instruction library. The instruction library is a pre-defined database that stores the control instructions corresponding to different terminal devices under different gesture classification results. For example, if the gesture classification result is "wave hand", and the terminal device corresponding to the device identifier is a mobile phone, the control instruction of "wave hand" gesture on the mobile phone will be found from the instruction library, which may be "play the next song"; if the terminal device corresponding to the device identifier is a VR glasses, the control instruction may be "switch VR scene". By matching the device identifier and the gesture classification result, it is ensured that the generated control instruction matches the function of the terminal device and the operation intention of the user, thereby realizing precise control.

[0044] The embodiment introduces the device identifier recognition and dynamic instruction mapping mechanism, realizes the adaptive and precise control of a unified gesture interaction system on multiple heterogeneous terminal devices, and thus supports seamless cross-platform user experience. The control instruction corresponding to the gesture classification result is found based on the device identifier of the terminal device. This mechanism enables the same gesture action of the user (such as "wave hand") to automatically map and execute completely different but contextually logical functions (such as song switching on a mobile phone, scene switching in VR, or page turning on a computer) on different types of terminal devices (such as mobile phones, VR glasses, or computers), thereby realizing the universal application of a gesture logic across multiple device platforms. This not only eliminates the burden of repeated learning of multiple gesture commands for users to manipulate different devices, greatly reducing the learning cost and operation complexity, but also builds a natural interaction paradigm of "gesture learning once, consistent everywhere" from the bottom, improving the practicality and consistency of user experience of the smart ring as a universal control center in a complex multi-device environment.

[0045] Based on the first embodiment and / or the second embodiment of the present application, in the third embodiment of the present application, the same or similar contents as the above-mentioned embodiments one and two can be referred to the above description, and the subsequent will not be described again. On this basis, the step of inputting the feature data into the pre-trained gesture classification model to obtain the gesture classification result includes: Step C10, inputting the feature data into the pre-trained gesture classification model to output a gesture classification probability vector; The feature data extracted from the motion data is input into a pre-trained gesture classification model. The model outputs a gesture classification probability vector, which contains the probability value of each predefined gesture category. For example, if the gesture categories include "wave", "pinch", and "rotate", the probability vector may be a three-dimensional vector, and each dimension corresponds to the probability value of a gesture category. These probability values reflect the confidence of the model for each gesture category.

[0046] Step C20, obtaining a preset probability threshold, and finding the maximum classification probability in the gesture classification probability vector; A preset probability threshold is obtained, which is used to determine whether the probability output by the model is high enough to ensure the reliability of gesture recognition. Moreover, the maximum classification probability value in the gesture classification probability vector is found, that is, the highest probability value in the probability vector. The category corresponding to this maximum classification probability value is the gesture category that the model considers most likely. For example, if the probability vector is [0.7, 0.2, 0.1], the maximum classification probability is 0.7, and the corresponding category is "wave".

[0047] Step C30, if the maximum classification probability is greater than or equal to the probability threshold, the gesture category indicated by the maximum classification probability is determined as the gesture classification result; The maximum classification probability value is compared with the preset probability threshold. If the maximum classification probability value is greater than or equal to the probability threshold, it means that the model has a high confidence in recognizing the gesture, and at this time the gesture category indicated by the maximum classification probability is determined as the final gesture classification result. For example, if the preset probability threshold is 0.6 and the maximum classification probability is 0.7, "wave" can be determined as the gesture classification result. This result will be used for subsequent control instruction generation and control operation of the terminal device.

[0048] Step C40, if the maximum classification probability is less than the probability threshold, it is determined that this gesture recognition fails.

[0049] If the maximum classification probability value is less than the preset probability threshold, it indicates that the model has insufficient confidence in gesture recognition, and a reliable gesture category cannot be determined. At this time, this gesture recognition is determined to be a failure. In this case, control instructions will not be generated to avoid incorrect control operations. For example, if the maximum classification probability is 0.5 and the probability threshold is 0.6, it is determined that the gesture recognition fails, and the user can be prompted to re-perform the gesture operation or check the wearing state of the smart ring and environmental interference factors.

[0050] In one possible implementation, the feature data includes acceleration mean, acceleration variance, angle change amount, and spectral energy of the gesture trajectory. The step of obtaining the preset probability threshold includes: Step D10, selecting at least one parameter from the acceleration mean, the acceleration variance, and the angle change amount as time domain reference data; At least one parameter is selected from the acceleration mean, the acceleration variance, and the angle change amount in the feature data as time domain reference data. These parameters can reflect the motion characteristics of the gesture in the time domain, for example, the acceleration mean can represent the overall motion intensity of the gesture, the acceleration variance can reflect the motion stability of the gesture, and the angle change amount can embody the rotation characteristics of the gesture. The purpose of selecting time domain reference data is to comprehensively consider the performance of the gesture in the time domain, so as to more comprehensively evaluate the energy characteristics of the gesture. For example, in a preferred embodiment, the acceleration mean and the angle change amount can be selected as the time domain reference data to comprehensively reflect the strength and rotation amplitude of the gesture.

[0051] Step D20, performing weighted sum calculation on the time domain reference data and the spectral energy to obtain gesture energy; The selected time domain reference data and the spectral energy of the gesture trajectory are weighted and summed to obtain a comprehensive gesture energy value. The spectral energy reflects the characteristics of the gesture in the frequency domain, and can embody the frequency components and complexity of the gesture. By weighted sum, the characteristics in the time domain and the frequency domain can be fused to obtain a more comprehensive gesture energy evaluation. The weighting coefficients can be adjusted according to the importance of the actual application scene and the gesture characteristics. For example, assuming that the weight of the acceleration mean is 0.4, the weight of the angle change amount is 0.3, and the weight of the spectral energy is 0.3, then the gesture energy calculation formula is: gesture energy = 0.4 x acceleration mean + 0.3 x angle change amount + 0.3 x spectral energy.

[0052] Step D30, setting a probability threshold based on the gesture energy, wherein the probability threshold and the gesture energy have a positive correlation trend.

[0053] The probability threshold is set according to the calculated gesture energy value. The probability threshold is used to determine whether the probability output by the gesture classification model is high enough to ensure the reliability of gesture recognition. There is a positive correlation trend between the probability threshold and the gesture energy, that is, the higher the gesture energy, the higher the probability threshold. This is because a high gesture energy usually means that the gesture action is more obvious and strong, and the model's recognition confidence for the gesture should also be higher. For example, if the gesture energy value is high, the probability threshold can be set to 0.8; if the gesture energy value is low, the probability threshold can be set to 0.6. This way of dynamically adjusting the probability threshold can flexibly adjust the recognition standard according to the strength and complexity of the gesture, thereby improving the accuracy and reliability of gesture recognition.

[0054] In one possible implementation, the step of setting the probability threshold based on the gesture energy comprises: Step E10, if the motion data indicates that the smart ring is in a stationary state, calculating the noise energy based on the acceleration variance, wherein the noise energy is positively correlated with the acceleration variance; Determine whether the smart ring is in a stationary state. This can be achieved by analyzing the acceleration data in the motion data, for example, if the acceleration data is close to zero or fluctuates within a small range within a long time window, it can be considered that the smart ring is in a stationary state. When the smart ring is in a stationary state, the acceleration variance in the motion data mainly reflects the noise level of the environment or the sensor itself. Therefore, the noise energy is calculated based on the acceleration variance, and the noise energy is positively correlated with the acceleration variance. For example, the noise energy can be calculated using the following formula: noise energy = k x acceleration variance, where k is a positive proportionality constant used to adjust the calculation scale of the noise energy. In this way, the noise level in the stationary state can be quantified to provide a reference for the subsequent setting of the probability threshold.

[0055] Step E20, setting the probability threshold based on the gesture energy and the noise energy, wherein the probability threshold and the noise energy have a negative correlation trend.

[0056] The probability threshold is set according to the calculated noise energy and the previously calculated gesture energy. The setting of the probability threshold takes into account the combined influence of gesture energy and noise energy. Specifically, the probability threshold has a positive correlation trend with the gesture energy, that is, the higher the gesture energy, the higher the probability threshold, because a stronger gesture signal requires a higher confidence to confirm; at the same time, the probability threshold has a negative correlation trend with the noise energy, that is, the higher the noise energy, the lower the probability threshold, in order to avoid misjudging low-confidence gestures in a high-noise environment. For example, the probability threshold can be set using the following formula: probability threshold = basic threshold + a x gesture energy β×noise energy, where the base threshold is a preset baseline value, and α and β are positive adjustment coefficients used to adjust the influence of gesture energy and noise energy on the probability threshold. In this way, the probability threshold can be dynamically adjusted to adapt to different gesture intensities and noise environments, thereby improving the accuracy and robustness of gesture recognition.

[0057] It should be noted that if the motion data indicates that the smart ring is in motion, the probability threshold can be set based solely on the gesture energy.

[0058] In recent years, smart wearable devices have rapidly developed in the fields of health monitoring and human-computer interaction. Smartwatches, bracelets, and other products can already monitor basic health indicators such as heart rate, blood oxygen, and sleep, but they still have limitations in terms of wearing comfort, long-term continuous monitoring, and fine gesture interaction. In contrast, smart rings are small, discreet, and imperceptible when worn, making them more suitable for all-day use.

[0059] However, existing smart rings mostly focus on single functions, such as NFC payments or simple activity counting, making it difficult to achieve high-precision gesture recognition and cross-device interaction. Especially in virtual reality (VR), augmented reality (AR), and mobile office scenarios, users urgently need a lightweight, accurate, and low-latency contactless control method. Currently, there is no effective way to integrate high-precision physiological monitoring and high-degree-of-freedom gesture interaction on a single micro-ring platform.

[0060] Based on this, and the first, second, and / or third embodiments of this application, in the fourth embodiment of this application, the content that is the same as or similar to the above-described embodiments one, two, and three can be referred to the above description and will not be repeated hereafter. In addition, the smart ring further includes a physiological sensor, and the device control method further includes: Step F10: Acquire the physiological data collected by the physiological sensor, and calculate the physiological parameters based on the physiological data; The smart ring's physiological sensors collect the user's physiological data, including but not limited to PPG (Photoplethysmography), ECG (Electrocardiogram), and PCG (Phonocardiogram) signals. These signals are preprocessed and analyzed to calculate specific physiological parameters.

[0061] For example, in a specific embodiment, the physiological sensor specifically includes a PPG sensor, an ECG sensor, and a PCG sensor. The PPG sensor adopts a combination architecture of multi-wavelength LED light sources (red light, infrared light, and green light) and a high-sensitivity photodetector, measures heart rate, blood oxygen saturation, and perfusion index, and the like by means of photoplethysmography; the ECG sensor precisely arranges multiple contact electrodes in the inner and outer rings of the ring, forms a bipolar or tri-polar lead system, and is used to collect high-quality electrocardio signals; and the PCG sensor adopts a specially designed micro MEMS microphone or vibration sensor, effectively collects heart sound signals by closely adhering to the phalanx. Each sensor adopts a synchronous sampling mechanism, the time alignment accuracy reaches the millisecond level, and the time consistency of multi-modal data is ensured. In the signal processing stage, the original physiological data collected is processed as follows: For the PPG signal, first, a 0.5 Hz-5.0 Hz band-pass filter is used to eliminate baseline drift and high-frequency noise interference, and then an adaptive threshold algorithm is used to accurately detect the pulse wave peak point, and the instantaneous heart rate value is calculated accordingly; For the ECG signal, the processing procedure includes differential amplification enhancement, 50 / 60 Hz hardware band-stop filter to eliminate power frequency interference, 0.5 Hz-40 Hz digital band-pass filter to suppress electromyographic noise, and finally Pan-Tompkins algorithm is applied to detect QRS complex in real time, and extract heart rate and heart rate variability parameters; For the PCG signal, first, a high-pass filter with a cutoff frequency of 20 Hz is used to remove bone conduction low-frequency noise, and then time-frequency analysis means such as short-time Fourier transform are used to extract the time-frequency features of the first heart sound (S1) and the second heart sound (S2), which are used for heart function evaluation. Through the collaborative analysis and feature extraction of the three types of physiological signals, a variety of physiological parameters reflecting the user's cardiovascular status can be calculated, providing a reliable data basis for health monitoring.

[0062] In a possible implementation, the physiological data includes a PPG signal, an ECG signal, and a PCG signal, the physiological parameter includes a user's heart rate, and the step of calculating a physiological parameter based on the physiological data includes: Step G10, respectively calculating the signal quality of the PPG signal, the ECG signal, and the PCG signal; The signal quality of the PPG signal, the ECG signal and the PCG signal is respectively evaluated. The evaluation of the signal quality can be based on the signal-to-noise ratio (SNR) of the signal, the integrity of the signal, the stability of the signal and the like. For example, for the PPG signal, the signal-to-noise ratio is quantified by calculating the ratio of the signal power in a specific frequency band to the noise power in a frequency band, while the waveform continuity, the pulse amplitude stability and the influence degree of motion artifacts are evaluated. The higher the signal-to-noise ratio and the better the waveform continuity, the higher the signal quality score. For the ECG signal, the reliability of QRS complex detection is mainly based on the evaluation, including the consistency of R-wave amplitude, the coefficient of variation of adjacent RR intervals, and the amplitude of baseline drift. The clearer the QRS complex and the more regular the rhythm, the higher the signal quality score. For the PCG signal, the time-frequency characteristics of the heart sound signal are analyzed for evaluation, including the energy concentration of S1 and S2 heart sounds, the background noise level and the consistency of heart sound period. The clearer the heart sound composition and the more stable the periodicity, the higher the signal quality score. The final quality score of each signal is normalized to a value between 0 and 1. Other appropriate signal quality evaluation methods can also be used according to actual application requirements, and the present embodiment does not make specific limitations in this regard.

[0063] In step G20, the fusion weights of the PPG signal, the ECG signal and the PCG signal are respectively set based on the signal quality of each signal, wherein the fusion weight is positively correlated with the signal quality. According to the calculated signal quality, the fusion weights of the PPG signal, the ECG signal and the PCG signal are set respectively. The fusion weight is positively correlated with the signal quality, that is, the higher the signal quality, the greater the corresponding fusion weight. For example, if the signal quality of the PPG signal is high, the signal quality of the ECG signal is medium, and the signal quality of the PCG signal is low, the weight of the PPG signal can be set to 0.6, the weight of the ECG signal can be set to 0.3, and the weight of the PCG signal can be set to 0.1. This weight setting method ensures that high-quality signals occupy a larger proportion in subsequent heart rate calculation, thereby improving the accuracy and reliability of heart rate calculation.

[0064] In step G30, a first heart rate is calculated based on the PPG signal, a second heart rate is calculated based on the ECG signal, and a third heart rate is calculated based on the PCG signal. PPG signal heart rate calculation: the time interval between adjacent peak values is calculated by detecting the pulse wave peak value in the PPG signal, so as to obtain the instantaneous heart rate, i.e. the first heart rate.

[0065] ECG signal heart rate calculation: the time interval between adjacent QRS complexes is calculated by detecting the QRS complex in the ECG signal, so as to obtain the instantaneous heart rate, i.e. the second heart rate.

[0066] By analyzing the interval between the first heart sound (S1) and the second heart sound (S2) in the PCG signal, the heart rate, i.e., the third heart rate, is calculated.

[0067] Step G40, according to the fusion weights corresponding to the PPG signal, the ECG signal and the PCG signal, the first heart rate, the second heart rate and the third heart rate are weighted and summed to obtain the user heart rate.

[0068] User heart rate = fusion weight PPG × first heart rate + fusion weight ECG × second heart rate + fusion weight PCG × third heart rate, for example, assuming the first heart rate is 70 times / minute, the second heart rate is 72 times / minute, and the third heart rate is 71 times / minute, and the corresponding fusion weights are 0.6, 0.3 and 0.1 respectively, then the final user heart rate is: user heart rate = 0.6 × 70 + 0.3 × 72 + 0.1 × 71 = 70.7 times / minute.

[0069] In this way, the quality of different signals and the calculation results of heart rate are comprehensively considered, and a more accurate and reliable user heart rate value is finally obtained, which provides more accurate data support for health management.

[0070] Exemplarily, in order to help understand the technical concept or technical principle of the device control method combined with the above-mentioned first embodiment, second embodiment and third embodiment, a specific embodiment is listed, in which the smart ring includes the following functional modules: 1. Multi-modal physiological sensing module: PPG sensor: multi-wavelength LED (red, infrared, green) and photodetector are used to measure heart rate, blood oxygen saturation (SpO2) and perfusion index; ECG electrode: dry electrodes are arranged on the outer and inner rings of the ring to form a bipolar or tri-polar lead system to collect high-fidelity electrocardiogram signals; PCG sensor: integrated micro MEMS microphone or vibration sensor, attached to the phalanx to collect heart sound signals, used to assist in judging the status of heart valve and heart rate variability; Each sensor supports synchronous sampling, with millisecond-level time alignment accuracy.

[0071] 2. Inertial Measurement Unit (IMU): Integrated three-axis accelerometer, three-axis gyroscope and optional magnetometer, sampling rate ≥ 200Hz, used to detect finger motion trajectory, angular velocity and spatial attitude; Supports static (such as clenching) and dynamic (such as sliding, clicking, rotating) gesture recognition.

[0072] 3. Low-power processing unit: Equipped with a dedicated co-processor or RISC-V microcontroller, running lightweight AI models (such as TinyML), achieving local gesture classification and physiological signal preprocessing; Supporting edge computing, reducing data upload delay and power consumption.

[0073] 4. Wireless communication module: Support Bluetooth 5.3 or BLE (Bluetooth Low Energy), which can be paired with smartphones, smart glasses, VR headsets, laptops, and other devices; Supporting multiple device concurrent connection and role switching (such as connecting both a smartphone and a VR device simultaneously).

[0074] 5. Power module: Using a miniature solid-state battery or wireless charging coil, with a battery life of ≥7 days; The ring body is made of biocompatible materials (such as titanium alloy, ceramic, or medical-grade silicone), IP68 waterproof, and suitable for different finger sizes.

[0075] 6. Interaction logic engine: Pre-set or user-defined gesture-instruction mapping table (such as "double-click index finger" = answer phone, "clockwise circle" = volume +, "pinch gesture" = PPT next page); Supporting context awareness: automatically switching interaction modes according to the current use device (phone / VR / computer).

[0076] Based on the smart ring with the above functional modules, referring to Figure 2 The device control flow includes: Step S11, synchronous acquisition and preprocessing of multi-modal physiological data and motion data; a) Triggering conditions: Continuous mode: After wearing the ring, it automatically enters a low-power monitoring mode, continuously collecting data at a low sampling rate (such as PPG: 25Hz, IMU: 50Hz).

[0077] Active mode: When the IMU detects a specific wake-up gesture (such as rapid double-clicking fingers) or the user issues a command through the matching APP, the system enters full-function mode, and all sensors are synchronously sampled at the highest accuracy.

[0078] b) Synchronous acquisition mechanism: The system is internally unified by a high-precision clock source (such as a real-time clock RTC) to provide a time reference.

[0079] The low-power processing unit sends a hardware trigger signal to the PPG sensor, ECG electrode, PCG sensor, and IMU simultaneously, ensuring that all data streams are hardware-level time-aligned at the beginning of acquisition, with a time synchronization error of less than 1 millisecond.

[0080] The raw data collected by each sensor is time-stamped uniformly to form a multi-modal data frame.

[0081] c) Signal pre-processing: PPG signal: First pass through a band-pass filter (e.g. 0.5Hz ~ 5.0Hz) to remove baseline drift and high-frequency noise. Then, use an adaptive threshold algorithm to detect pulse wave peaks, and preliminarily calculate the instantaneous heart rate.

[0082] ECG signal: After amplification by a differential amplifier, eliminate power frequency interference through a hardware band-stop filter (50 / 60Hz), and further filter out electromyographic noise through a digital band-pass filter (0.5Hz ~ 40Hz). Use the classic Pan-Tompkins algorithm to detect QRS complexes in real time.

[0083] PCG signal: The audio signal collected by the MEMS microphone is first passed through a high-pass filter (>20Hz) to remove low-frequency friction noise conducted by the phalange, and then subjected to time-frequency domain analysis (such as short-time Fourier transform) to extract features of the first heart sound (S1) and the second heart sound (S2).

[0084] IMU signal: Sensor fusion (using complementary filter or Kalman filter algorithm) is performed on three-axis accelerometer and gyroscope data to calculate the absolute attitude angle (pitch, roll, yaw) of the finger in space, and calculate the acceleration and angular velocity of the motion.

[0085] Step S12, signal fusion and physiological parameter calculation; Heart rate variability: Based on the R-R interval sequence extracted from the ECG signal, or the pulse interval sequence extracted from the PPG signal, calculate the HRV through time domain analysis (SDNN), frequency domain analysis (LF / HF power), etc.

[0086] Pulse wave transit time: Use the time difference between the R-wave peak of ECG and the PPG pulse wave foot point to accurately calculate PTT, and estimate blood pressure trend based on pre-established personalized calibration model (user height, weight, static blood pressure calibration value).

[0087] Multi-modal cross-validation: When the quality of a signal is poor due to motion artifacts (such as PPG), prefer to use another signal with higher quality (such as ECG) for heart rate calculation, and output the final result through signal fusion algorithm (such as weighted average) to improve the robustness of measurement.

[0088] Step S13, gesture recognition and context-aware decision-making; a) Gesture feature extraction: From the pre-processed IMU data stream, a time window (e.g. 200ms) is slid to extract the feature vector in real-time. The vector includes but not limited to: mean, variance, peak of 3-axis acceleration, integral of 3-axis angular velocity (change of angle), and FFT spectrum energy of gesture trajectory.

[0089] b) Local AI model inference: The extracted feature vector is input to a lightweight convolutional neural network or recurrent neural network deployed on the RISC-V microcontroller. The TinyML model has been trained in the cloud and quantized before being burned to the device.

[0090] The model outputs a gesture classification probability vector, for example, ["pinch": 0.92, "slide": 0.05, "no gesture": 0.03].

[0091] When the highest probability exceeds a preset threshold (e.g. 0.8), it is determined that the gesture recognition is successful.

[0092] c) Context awareness and instruction mapping: The interaction logic engine maintains a device context state machine in real-time. It determines whether the current interaction subject is "smartphone", "VR glasses" or "computer" through the device ID or service UUID of the BLE connection.

[0093] The engine stores multiple gesture-instruction mapping tables. For example: Smartphone context: "pinch gesture" -> MEDIA_PLAY_PAUSE instruction.

[0094] VR glasses context: the same "pinch gesture" -> VR_TRIGGER_CLICK instruction.

[0095] Computer context: "clockwise circle" -> VOLUME_UP instruction.

[0096] According to the recognized gesture category and the current device context, the corresponding mapping table is dynamically queried to generate the final control instruction code.

[0097] Step S14, encrypted transmission and terminal execution; a) Low-power encrypted transmission: The generated control instruction code (or physiological indicator alarm after fusion calculation) is packaged into a specific data packet protocol.

[0098] The data packet is sent to the paired target terminal device through the BLE 5.3 link after AES-128 encryption. The system supports multiple device concurrency, and can send physiological data to the phone while sending control instructions to the VR glasses.

[0099] b) terminal device executes: The application program or system service running on the terminal device (such as a mobile phone, VR glasses) receives and parses the data packet.

[0100] The application program calls the corresponding API interface provided by the operating system to perform specific operations. For example: On a smart phone, call the MediaSession API to implement music play / pause.

[0101] In the VR glasses, send a "click" event to the 3D engine to drive the virtual cursor interaction.

[0102] On a computer, simulate a keyboard PageDown key press event to implement PPT page turning.

[0103] It should be noted that the above examples are only used to assist in understanding the embodiment and do not constitute a limitation on the device control flow of the embodiment. Further simple transformations based on this technical concept are within the scope of protection of the present application.

[0104] In addition, the present application also provides a device control system, as shown in Figure 5 The device control system comprises a smart ring and a terminal device in communication connection, the smart ring comprises a motion sensor; The smart ring is used to: Obtain motion data collected by the motion sensor, and perform feature extraction on the motion data to obtain feature data; Input the feature data into a pre-trained gesture classification model to obtain a gesture classification result; Obtain a control instruction corresponding to the gesture classification result, and send the control instruction to the terminal device; The terminal device is used to: Execute the control instruction.

[0105] In addition, the present application also provides a smart ring, which comprises a motion sensor and a processor, the motion sensor is used to collect motion data, and the processor is used to execute the steps of the device control method as described above.

[0106] The smart ring provided by the present application adopts the device control method in the above embodiment, which can solve the technical problem of how to improve the practicability of the smart ring. Compared with the prior art, the smart ring provided by the present application has the same beneficial effects as the device control method provided by the above embodiment, and the other technical features in the smart ring are the same as the features disclosed in the previous embodiment method, which will not be repeated here.

[0107] It should be understood that portions of the application disclosed can be implemented in hardware, software, firmware, or combinations thereof. In the description of the embodiments above, specific features, structures, materials or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0108] The above description is merely illustrative of the application and is not intended to limit the scope of the application. Any changes and modifications that can be made to the application in accordance with the principles of the application would be readily apparent to those skilled in the art and the present application is therefore not limited to the description and examples contained herein but is only limited by the claims.

[0109] In addition, to achieve the above object, the embodiment of the present application further provides a readable storage medium having computer readable program instructions (i.e. computer programs) stored thereon, the computer readable program instructions being used to execute the device control method in the above embodiment.

[0110] The computer readable storage medium provided by the embodiment of the present application may, for example, be a U disk, but is not limited to an electric, magnetic, optical, electromagnetic, infrared, or semiconductor system, system or device, or any combination of the above. More specific examples of the computer readable storage medium can include, but are not limited to, an electric connection with one or more conductive wires, a portable computer disk, a hard disk, a random access memory (RAM), a read only memory (ROM), an erasable programmable read only memory (EPROM or flash memory), an optical fiber, a portable compact disk read only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the embodiment, the computer readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system or device. The program code contained on the computer readable storage medium can be transmitted by any suitable medium, including but not limited to an electric wire, an optical cable, an RF (Radio Frequency: radio frequency), etc., or any suitable combination of the above.

[0111] The above computer readable storage medium can be contained in the smart ring; or can exist separately and not be assembled into the smart ring.

[0112] The above computer readable storage medium carries one or more programs, which, when executed by the smart ring, cause the smart ring to perform the steps of any of the above embodiments.

[0113] Computer program code for carrying out operations of the present application can be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider).

[0114] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0115] The modules involved in the embodiments of the present application can be implemented in software or hardware. In some cases, the name of the module does not constitute a limitation on the module itself.

[0116] The readable storage medium provided by the present application is a computer readable storage medium, which stores computer readable program instructions (i.e. computer programs) for executing the device control method described above, and can solve the technical problem of how to improve the practicability of the intelligent ring. Compared with the prior art, the computer readable storage medium provided by the present application has the same beneficial effects as the device control method provided by the above-mentioned embodiments, which will not be described here.

[0117] In addition, the embodiment of the present application further provides a computer program product comprising a computer program, which, when executed by a processor, implements the steps of the device control method as described above.

[0118] The computer program product embodiment of the present application is basically the same as the above-mentioned device control method embodiments, and will not be described here.

[0119] It should be noted that in this paper, the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or system. Without more limitations, the element defined by the statement "comprises a" does not exclude the presence of another identical element in the process, method, article or system including the element.

[0120] The above-mentioned serial numbers of the embodiments of the present application are only for description, not representing the advantages and disadvantages of the embodiments.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be realized by software and necessary general hardware platform, of course, also can be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application essentially or say the part that contributes to the prior art can be embodied in the form of software sensor, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, including a number of instructions to make a smart ring (which can be a mobile phone, computer, server or network equipment, etc.) execute the method described in each embodiment of the present application.

[0122] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation using the content of the present application specification and drawings, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A device control method characterized by, The device control method is applied to a smart ring, the smart ring comprising a motion sensor, the smart ring being connected with a terminal device, and the device control method comprising the following steps: acquiring motion data collected by the motion sensor, and performing feature extraction on the motion data to obtain feature data; inputting the feature data into a pre-trained gesture classification model to obtain a gesture classification result; acquiring a control instruction corresponding to the gesture classification result to control the terminal device based on the control instruction.

2. The device control method according to Claim 1, wherein The step of acquiring the control instruction corresponding to the gesture classification result comprises: acquiring a device identifier of the terminal device; searching for the control instruction corresponding to the gesture classification result in a preset instruction library according to the device identifier.

3. The device control method according to Claim 1, wherein The motion data comprises acceleration data and angular velocity data, and the step of performing feature extraction on the motion data to obtain feature data comprises: acquiring an acceleration mean value, an acceleration variance and an acceleration peak value based on the acceleration data; performing integral calculation on the angular velocity data to obtain an angle change amount; acquiring a gesture trajectory based on the acceleration data, and calculating signal energy of the gesture trajectory in a frequency domain to obtain frequency spectrum energy; combining the acceleration mean value, the acceleration variance, the acceleration peak value, the angle change amount and the frequency spectrum energy to obtain feature data.

4. The device control method of claim 1, wherein, The step of inputting the feature data into the pre-trained gesture classification model to obtain the gesture classification result comprises: inputting the feature data into the pre-trained gesture classification model to output a gesture classification probability vector; acquiring a preset probability threshold, and searching for a maximum classification probability in the gesture classification probability vector; if the maximum classification probability is greater than or equal to the probability threshold, a gesture category indicated by the maximum classification probability is determined as the gesture classification result; if the maximum classification probability is less than the probability threshold, it is determined that the gesture recognition fails this time.

5. The device control method according to Claim 4, wherein The feature data comprises an acceleration mean value, an acceleration variance, an angle change amount and frequency spectrum energy of a gesture trajectory, and the step of acquiring a preset probability threshold comprises: selecting at least one parameter from the acceleration mean value, the acceleration variance and the angle change amount as time domain reference data; performing weighted summation calculation on the time domain reference data and the frequency spectrum energy to obtain gesture energy; setting a probability threshold based on the gesture energy, wherein the probability threshold and the gesture energy have a positive correlation trend.

6. The device control method according to Claim 5, wherein The step of setting the probability threshold based on the gesture energy comprises: if the motion data indicates that the smart ring is in a stationary state, calculating noise energy based on the acceleration variance, wherein the noise energy and the acceleration variance are positively correlated; setting a probability threshold based on the gesture energy and the noise energy, wherein the probability threshold and the noise energy have a negative correlation trend.

7. The device control method according to any one of claims 1 to 6, wherein, The smart ring further comprises a physiological sensor, and the device control method further comprises: acquiring physiological data collected by the physiological sensor, and calculating a physiological parameter based on the physiological data; If the physiological parameter matches a preset health abnormality parameter threshold, the terminal device is controlled to perform health abnormality reminding.

8. The device control method according to Claim 7, wherein The physiological data includes PPG signals, ECG signals and PCG signals, and the physiological parameter includes a user heart rate. The step of calculating a physiological parameter based on the physiological data includes: calculating signal quality of the PPG signals, the ECG signals and the PCG signals, respectively; setting fusion weights of the PPG signals, the ECG signals and the PCG signals based on the signal quality, respectively, wherein the fusion weights are positively correlated with the signal quality; calculating a first heart rate based on the PPG signals, a second heart rate based on the ECG signals and a third heart rate based on the PCG signals; performing weighted sum calculation on the first heart rate, the second heart rate and the third heart rate according to the fusion weights corresponding to the PPG signals, the ECG signals and the PCG signals, to obtain the user heart rate.

9. An apparatus control system characterized by comprising: The device control system includes a smart ring and a terminal device in communication connection, and the smart ring includes a motion sensor. The smart ring is configured to: acquire motion data collected by the motion sensor, and perform feature extraction on the motion data to obtain feature data; input the feature data into a pre-trained gesture classification model to obtain a gesture classification result; acquire a control instruction corresponding to the gesture classification result, and send the control instruction to the terminal device; The terminal device is configured to: execute the control instruction.

10. A smart ring, characterized by The smart ring includes a motion sensor and a processor, the motion sensor is configured to collect motion data, and the processor is configured to execute the device control method according to any one of claims 1 to 8.

Images