Treadmill voice recognition interaction control method based on positioning data and physiological data

Abstract

The application relates to the technical field of voice recognition interaction, in particular to a treadmill voice recognition interaction control method based on positioning data and physiological data, which comprises the following steps: type judgment is performed on control instructions obtained according to voice data, so as to facilitate hierarchical verification of different control instructions; for auxiliary function parameter instructions, position verification is performed; for safe motion parameter instructions, position verification and a first verification operation on physiological data are performed; for high-risk motion parameter instructions, a second verification operation on physiological data, position data and dynamic identity is performed by dynamically binding the positioning data, the physiological data and the voice data, so as to accurately distinguish between a user using the treadmill and a non-using user, realize accurate and safe treadmill voice recognition interaction experience in a multi-person motion scene, and significantly reduce the voice control mis-touch rate, thereby solving the problem of mis-triggering of the treadmill voice control in the multi-person motion scene.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of voice recognition interaction, and particularly relates to a treadmill voice recognition interaction control method based on positioning data and physiological data. BACKGROUND

[0002] With the development of voice recognition interaction, intelligent fitness devices provide more intelligent and thoughtful voice recognition interaction functions, realize voiceprint recognition of users and timely feedback of instructions, and greatly improve the use experience of exercise products. The traditional treadmill voice recognition interaction scheme mainly relies on voiceprint recognition and user position to match users and voice, and realizes voice recognition interaction control of the treadmill. However, in a gym or home exercise environment, there are many people around the treadmill user or other audio interference, such as simultaneous use of multiple treadmills in the gym, external broadcast, and audio. The existing technology has problems of voiceprint recognition confusion and false triggering of treadmill voice control when realizing the function of treadmill voice recognition interaction control.

[0003] The invention patent with publication number CN 118072744B discloses a language recognition method and device based on voiceprint. The patent captures the user position through a visual system and adjusts the pointing direction of the microphone array. Although this method can improve the accuracy of sound pickup, it has the following defects: 1. Body occlusion occurs and positioning signals are lost due to limb swinging during running, and the dynamic scene adaptability is poor; 2. High hardware cost; 3. The acoustic environment of a multi-person exercise scene is complex, and the voiceprint matching accuracy decreases in the exercise state.

[0004] The invention patent with publication number CN 111009242B discloses a false trigger correction of a voice-activated intelligent device. The patent verifies the user's identity through the voiceprint fingerprint detected in the audio input. However, in the exercise scene, the complexity of the acoustic environment leads to a high false trigger rate, mainly including: treadmill motor sound, environmental noise such as conversations between adjacent users, which can increase the azimuth error of the sound source and improve the false trigger rate. Further, when using a treadmill for exercise, the breathing rhythm change of the user during running will cause the fundamental frequency to shift, the MFCC (Mel Frequency Cepstral Coefficient) feature matching degree to decrease, and the voiceprint recognition accuracy to decrease.

[0005] The invention patent with publication number CN118250635A discloses a mine positioning and voice communication system and method based on UWB technology. The patent uses UWB technology for positioning in a complex environment to facilitate voice data transmission between the UWB base station and the miner's lamp. The UWB is only used for basic position sensing, but does not provide a solution to constructing a user holographic image between the user corresponding to the positioning data and the user corresponding to the voice data.

[0006] In order to break through the limitations of the prior art, the present application proposes a treadmill voice recognition interaction control method based on positioning data and physiological data, which dynamically binds positioning data with multi-modal physiological data to solve the problem of false triggering of treadmill voice control in a multi-person exercise scenario. SUMMARY

[0007] The present application aims to propose a treadmill voice recognition interaction control method based on positioning data and physiological data for the above-mentioned technical problems.

[0008] In a first aspect, the present application provides a treadmill voice recognition interaction control method based on positioning data and physiological data, comprising the following steps:

[0009] Real-time acquisition of positioning data, physiological data and voice data of a user using a treadmill, conversion of the positioning data to a position feature in the treadmill coordinate system, feature calculation of the physiological data to obtain physiological features, and voice recognition of the voice data to obtain control instructions;

[0010] A dynamic identity model is constructed, the input of the dynamic identity model being a basic identity at the current time and a dynamic identity at the previous time, and the output being the dynamic identity at the current time, wherein the dynamic identity is initialized according to the position feature and the physiological feature at the initial time when the user is located in a specified area of the treadmill, the basic identity at the current time is generated according to the position feature and the physiological feature at the current time, and the dynamic identity is updated once at each preset update interval based on the dynamic identity model;

[0011] In response to determining that the position feature at the current time is not in the specified area of the treadmill, the previous step is returned, otherwise the type of the control instruction is judged, and the type of the control instruction includes an auxiliary function parameter instruction, a safe exercise parameter instruction and a high-risk exercise parameter instruction;

[0012] In response to determining that the control instruction is an auxiliary function parameter instruction, the auxiliary function parameter instruction is executed;

[0013] In response to determining that the control instruction is a safe exercise parameter instruction, a first verification operation is performed on the physiological data at the current time, if the verification is passed, the safe exercise parameter instruction is executed, otherwise the safe exercise parameter instruction is not executed;

[0014] In response to determining that the control instruction is a high-risk exercise parameter instruction, a second verification operation is performed on the physiological data, the position data and the dynamic identity at the current time, if the verification is passed, the high-risk exercise parameter instruction is executed, otherwise the high-risk exercise parameter instruction is not executed.

[0015] As preferred, the positioning data is converted to a position feature in a treadmill coordinate system, the physiological data is calculated to obtain a physiological feature, and the voice data is recognized to obtain a control instruction, specifically including:

[0016] The positioning data includes UWB positioning coordinates of the user when using the treadmill, the UWB positioning coordinates are converted to the treadmill coordinate system, and normalized to obtain normalized coordinates of the user in the normalized space corresponding to the specified area of the treadmill in the treadmill coordinate system, and the normalized coordinates are taken as the position feature, represented as , wherein and respectively represent the horizontal coordinate and the vertical coordinate of the user when using the treadmill in the treadmill coordinate system after normalization processing;

[0017] The physiological data includes a heart rate signal and an electromyography signal, the standard deviation of adjacent R-wave intervals and the root mean square of adjacent interval differences are calculated according to the heart rate signal, the standard deviation of adjacent R-wave intervals and the root mean square of adjacent interval differences are represented as SDNN and RMSSD respectively, the electromyography root mean square is calculated according to the electromyography signal, the electromyography root mean square is represented as EMG_RMS, and the physiological feature is generated according to the standard deviation of adjacent R-wave intervals, the root mean square of adjacent interval differences and the electromyography root mean square, and the physiological feature is represented as ;

[0018] The voice data is subjected to semantic recognition and voiceprint matching, if the voiceprint matching is successful, the result of the semantic recognition is taken as the control instruction, otherwise no control instruction is output.

[0019] As preferred, a first verification operation is performed on the physiological data at the current time, specifically including:

[0020] It is judged whether the heart rate signal at the current time is located in a first heart rate verification range, if yes, the verification is passed, otherwise the verification is failed.

[0021] As preferred, a basic identity at the current time is generated according to the position feature and the physiological feature at the current time, as shown in the following formula:

[0022] ;

[0023] wherein, represents the basic identity at the current time;

[0024] A dynamic identity model is constructed according to the basic identity at the current time, as shown in the following formula:

[0025] ;

[0026] wherein, represents the dynamic identity at the current time, represents the current time; represents the attenuation coefficient; represents the dynamic identity of the previous time, represents the previous time.

[0027] As preferred, the second verification operation is performed on the physiological data, the position data and the dynamic identity of the current time, specifically including:

[0028] The heart rate variation of the current time is calculated according to the heart rate signal within a period of time after the initial time when the user is located in the designated area of the treadmill and the heart rate signal of the current time, as shown in the following formula:

[0029] ;

[0030] wherein, represents the heart rate variation of the current time, represents the heart rate signal of the current time, represents the average value of the heart rate within a period of time after the initial time when the user is located in the designated area of the treadmill;

[0031] The first stage verification is entered to determine whether the heart rate variation of the current time is within the second heart rate verification range or the root mean square of the electromyogram of the current time is greater than the first electromyogram root mean square threshold, if yes, the second stage verification is entered, otherwise the verification is not passed;

[0032] The spatial similarity, the physiological similarity and the fusion similarity are calculated according to the position feature, the physiological feature and the dynamic identity of the current time, and the calculation process is as follows:

[0033] ;

[0034] ;

[0035] ;

[0036] wherein, represents the spatial similarity, represents the position feature of the current time, represents the position feature constituted by the first element and the second element in the dynamic identity of the current time, represents the normalized spatial maximum Euclidean distance, represents the vector length symbol; represents the physiological similarity, represents the physiological feature of the current time, represents the physiological feature constituted by the third element to the fifth element in the dynamic identity of the current time; represents the fusion similarity, a weight coefficient representing physiological similarity;

[0037] In the second stage, it is determined whether the fusion similarity is greater than a similarity threshold value. If yes, the third stage is entered. Otherwise, the verification fails.

[0038] In the third stage, it is determined whether the heart rate signal at the current time is greater than a heart rate threshold value, or whether the electromyographic signal is continuously lost at the current time and a period of time before the current time. If yes, the verification fails, and a gradient speed reduction instruction and a voice warning instruction are executed. The speed of the treadmill is reduced to a safe speed within a preset execution time, and a voice prompt of high heart rate or electromyographic signal loss is performed. The relationship between the safe speed and the preset execution time satisfies the following formula:

[0039] ;

[0040] wherein, the safe speed is represented by Vsafe; the minimum speed of the treadmill is represented by Vmin, the speed of the treadmill at the current time is represented by Vt, t represents the execution time of the gradient speed reduction instruction, and e represents a natural constant, the larger value is represented by max,

[0041] As preferred, the weight coefficient of physiological similarity is also dynamically adjusted, specifically including:

[0042] The spatial weight and the physiological weight are constructed and initialized, and the spatial weight and the physiological weight satisfy the following formula:

[0043] ;

[0044] wherein, the spatial weight is represented by Ws, the physiological weight is represented by Wp;

[0045] In response to determining that the electromyographic root mean square is greater than a second electromyographic root mean square threshold value, the weight coefficient of physiological similarity is increased according to the physiological weight and a preset adjustment coefficient. The weight coefficient of physiological similarity is shown in the following formula:

[0046] ;

[0047] wherein, the preset first adjustment coefficient is represented by a1;

[0048] The signal-to-noise ratio is calculated according to the voice data. In response to determining that the signal-to-noise ratio is less than a signal-to-noise ratio threshold value, the weight coefficient of physiological similarity is reduced according to the physiological weight and a preset adjustment coefficient. The weight coefficient of physiological similarity is shown in the following formula:

[0049] ;

[0050] in, This indicates the preset second adjustment coefficient;

[0051] In response to determining that the root mean square of electromyography (RMS) is less than or equal to the second RMS threshold and the signal-to-noise ratio (SNR) is greater than or equal to the SNR threshold, the weighting coefficient of physiological similarity is equal to the physiological weight.

[0052] As a preferred method, determining the type of control command specifically includes:

[0053] The control command includes speed control parameters, gradient control parameters, and other control parameters. In response to determining that the speed control parameters and gradient control parameters are both empty, and the other control parameters are not empty, the control command is an auxiliary function parameter command.

[0054] In response to the determination that the speed control parameter is within the safe speed range and the slope control parameter is less than or equal to the slope threshold, the control command is a safe motion parameter command.

[0055] If the speed control parameter is determined to be outside the safe speed range, or the slope control parameter is greater than the slope threshold, the control command is a high-risk motion parameter command.

[0056] As a preferred method, determining the type of control command specifically includes:

[0057] The control command includes speed control parameters, gradient control parameters, and other control parameters. In response to determining that the speed control parameters and gradient control parameters are both empty, and the other control parameters are not empty, the control command is an auxiliary function parameter command.

[0058] A motion risk quantification model is constructed. The inputs to the motion risk quantification model are speed control parameters and slope control parameters, and the output is the motion risk value. The motion risk quantification model is shown in the following formula:

[0059] ;

[0060] in, Indicates the risk value of exercise. Indicates speed control parameters, This indicates the treadmill's maximum speed. Indicates the weight of the speed control parameters. Indicates the slope control parameters. This indicates the maximum incline of the treadmill. Indicates the weight of the slope control parameters;

[0061] In response to determining that the motion risk value is less than or equal to the motion risk threshold, the control command is a safe motion parameter command;

[0062] In response to the determination that the motion risk value is greater than the motion risk threshold, the control command is a high-risk motion parameter command.

[0063] Secondly, the present invention provides a treadmill voice recognition interactive control device based on positioning data and physiological data, comprising:

[0064] The data acquisition module is configured to acquire the user's positioning data, physiological data and voice data in real time when using the treadmill, convert the positioning data to the treadmill coordinate system to obtain position features, perform feature calculation on the physiological data to obtain physiological features, and perform voice recognition on the voice data to obtain control commands.

[0065] The dynamic identity building module is configured to build a dynamic identity model. The input of the dynamic identity model is the basic identity at the current moment and the dynamic identity at the previous moment, and the output is the dynamic identity at the current moment. Specifically, the dynamic identity is initialized based on the user's initial position characteristics and physiological characteristics in the designated area of ​​the treadmill. The basic identity at the current moment is generated based on the current position characteristics and physiological characteristics. The dynamic identity is updated once at each preset update interval based on the dynamic identity model.

[0066] The instruction judgment module is configured to return to the previous step if it determines that the current position feature is not within the specified area of ​​the treadmill; otherwise, it determines the type of control instruction, which includes auxiliary function parameter instructions, safe exercise parameter instructions, and high-risk exercise parameter instructions.

[0067] The first instruction execution module is configured to execute the auxiliary function parameter instruction in response to determining that the control instruction is an auxiliary function parameter instruction;

[0068] The second instruction execution module is configured to perform a first verification operation on the current physiological data in response to determining that the control instruction is a safe motion parameter instruction. If the verification passes, the safe motion parameter instruction is executed; otherwise, the safe motion parameter instruction is not executed.

[0069] The third instruction execution module is configured to perform a second verification operation on the current physiological data, location data, and dynamic identity identifier in response to determining that the control instruction is a high-risk motion parameter instruction. If the verification passes, the high-risk motion parameter instruction is executed; otherwise, the high-risk motion parameter instruction is not executed.

[0070] Thirdly, the present invention provides an electronic device including one or more processors; and a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any implementation of the first aspect.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] (1) This invention performs graded verification on different types of control commands, especially high-risk exercise parameter commands. By dynamically binding positioning data, physiological data and voice data, it can accurately distinguish between users who are using the treadmill and non-users. In exercise scenarios with multiple people, it can achieve an accurate and safe treadmill voice recognition interaction experience, significantly reduce the false touch rate of voice control, and thus solve the problem of false triggering of treadmill voice control in exercise scenarios with multiple people.

[0073] (2) This invention synchronously collects location data, physiological data and voice data in real time, and associates them in the time dimension. It also constructs a dynamic identity identification model to obtain temporary, context-related dynamic identity identification, and associates location data, physiological data and voice data in the spatial dimension. This is to facilitate the verification of control commands when the control command type is a safe movement parameter command or a high-risk movement parameter command, especially a high-risk movement parameter command. Even when the accuracy of voiceprint recognition decreases, it can effectively determine whether the control command comes from a user who is using the treadmill by whether the change in physiological data meets expectations and whether the similarity between the fused location data and physiological data meets expectations, and correctly execute the control command. This overcomes the problem of voice control mis-triggering caused by the distortion of voice data in the movement state and the decrease in the accuracy of voiceprint recognition.

[0074] (3) By dynamically adjusting the weight coefficients of physiological similarity and spatial similarity, this invention avoids the influence of motion state and complex acoustic environment on voiceprint matching from continuing to the verification process of control commands, thus ensuring the accuracy of the second verification operation, thereby achieving more accurate dynamic binding between positioning data, physiological data and voice data, and further improving the accuracy of voice recognition interaction. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0076] Figure 1 This application describes a treadmill voice recognition interaction system as an embodiment of the present application.

[0077] Figure 2 This is a flowchart illustrating a treadmill voice recognition interactive control method based on positioning data and physiological data, as an embodiment of this application.

[0078] Figure 3 This is a schematic diagram of a treadmill voice recognition interactive control device based on positioning data and physiological data, which is an embodiment of this application.

[0079] Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention.

[0080] Reference numerals: 11. UWB positioning device; 12. Voice acquisition device; 13. Physiological sensing device; 14. User registration / configuration device; 15. Treadmill voice recognition interactive control device. Detailed Implementation

[0081] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0082] refer to Figure 1 One specific embodiment of this application provides a treadmill voice recognition interaction system, which includes a UWB positioning device 11, a voice acquisition device 12, a physiological sensing device 13, a user registration / configuration device 14, and a treadmill voice recognition interaction control device 15. The treadmill is 2.8 meters long, 1.2 meters wide, and 2.8 meters high. A rectangular area surrounding the treadmill, centered on its center point and 3 meters long and 2 meters wide, is defined as the designated area of ​​the treadmill. The positioning height of the UWB positioning device 11 is within the range of 0.8-1.8 meters.

[0083] The UWB positioning device 11 includes multiple UWB anchor points and one or more UWB tags. The UWB anchor points are deployed on the treadmill's handrails and / or screen. The UWB tags can be integrated into the user's wearable device or are standalone wearable devices such as wristbands or headphones. The UWB anchor points receive signals emitted by the UWB tags and calculate the tag's position based on the signal's arrival time. The UWB positioning device 11 collects positioning data from the user while using the treadmill, thereby determining the user's position relative to the treadmill in real time with high accuracy, achieving centimeter-level precision at a controllable cost. In one embodiment, the UWB positioning device 11 includes three UWB anchor points (Qorvo DW3110), an independent wristband tag (Decawave DWM3001C), and a positioning controller (NXP SR040), employing a three-sided positioning architecture design with two lateral UWB anchor points and one top UWB anchor point. Positioning is initiated when the signal emitted by the independent wristband tag Decawave DWM3001C is simultaneously captured by three UWB anchor points Qorvo DW3110. Specifically, a TDoA (Time Difference of Arrival) positioning algorithm is used to calculate the user's positioning data by comparing the time difference between the arrival of the UWB tag signal at each UWB anchor point, and a Kalman filter is added to suppress motion jitter errors, achieving a positioning accuracy of ±10 cm. The user's positioning data is acquired at a frequency of 100ms / frame. Specifically, when the user's positioning data is within the designated area of ​​the treadmill for a continuous period of time, the treadmill's voice recognition interactive control function is activated, and dynamic identity updating is performed; when the user's positioning data is outside the designated area of ​​the treadmill for a continuous period of time or the treadmill's voice recognition interactive function is manually turned off, dynamic identity updating is deactivated. In one embodiment, when the user's positioning data is within the designated area of ​​the treadmill for 5 consecutive frames, the treadmill's voice recognition interactive control function is activated, and dynamic identity updating is performed; control commands outside the designated area of ​​the treadmill are automatically blocked.

[0084] The physiological sensing device 13 is integrated into a treadmill handle, wearable device, or clothing to collect at least one physiological data from the user in real time, such as electrocardiogram (ECG), photoplethysmography (PPG), heart rate variability, electrical skin activity (EDA), electromyography (EMG), and respiratory rate, to facilitate subsequent calculation and analysis of physiological characteristics. In one specific embodiment of this application, the physiological sensing device 13 includes a PPG sensor and a flexible electrode. The PPG sensor is integrated into the treadmill handle and has a sampling rate ≥100Hz to acquire heart rate signals; the flexible electrode is integrated into the earphone hook to capture microcurrents in the neck muscles and acquire electromyographic signals.

[0085] The voice acquisition device 12 is integrated into the treadmill or the user's wearable device to collect voice data. The voice acquisition device 12 preferably uses a microphone array with beamforming capabilities. In one specific embodiment of this application, the voice acquisition device 12 includes an Infineon IM69D130 microphone array and an XMOS XVF3610 preprocessor. The sound beam control uses the UWB positioning coordinates collected by the UWB positioning device 11 to calculate the user's azimuth angle θ and distance d, dynamically adjusting the beam direction: beam angle = arctan(vertical offset / horizontal distance). The beamwidth adaptively adjusts with distance: 15° at a distance of 1m, and widens to 25° at a distance of 1.5m.

[0086] The user registration / configuration device 14 is used to record static voiceprints when a user uses the treadmill for the first time, as well as to collect the user's initial voice and physiological data using the voice acquisition device 12 and the physiological sensing device 13.

[0087] The aforementioned treadmill voice recognition interactive control device 15 is deployed on the treadmill local controller or cloud server. Its data acquisition module 1 acquires the positioning data, physiological data and voice data collected by the UWB positioning device 11, physiological sensing device 13 and voice acquisition device 12 in real time, ensuring the spatiotemporal synchronization of the positioning data, physiological data and voice data.

[0088] Figure 2 An embodiment of this application illustrates a voice recognition interactive control method for a treadmill based on positioning data and physiological data, comprising the following steps:

[0089] S1 acquires real-time positioning data, physiological data, and voice data of the user while using the treadmill. It converts the positioning data to the treadmill coordinate system to obtain position features, performs feature calculations on the physiological data to obtain physiological features, and performs voice recognition on the voice data to obtain control commands.

[0090] In a specific embodiment, the positioning data is converted to the treadmill coordinate system to obtain position features, physiological data is processed to obtain physiological features, and voice data is processed to obtain control commands, specifically including:

[0091] Location data includes the user's UWB coordinates while using the treadmill. These UWB coordinates are transformed to the treadmill coordinate system and normalized to obtain the normalized coordinates of the user within a specified area of ​​the treadmill in the treadmill coordinate system. These normalized coordinates are then used as location features, represented as follows: ,in, and These represent the normalized x and y coordinates of the user's coordinates in the treadmill coordinate system when using the treadmill.

[0092] Physiological data includes heart rate and electromyography (EMG) signals. Based on the heart rate signal, the root mean square (RMS) of the standard deviation of adjacent R-wave intervals and the difference between adjacent intervals is calculated. These RMS values ​​are represented as SDNN and RMSSD, respectively. Based on the EMG signal, the root mean square (RMS) of the electromyography (EMG) signal is calculated and represented as EMG_RMS. Physiological features are generated based on the RMS of the standard deviation of adjacent R-wave intervals, the RMS of the difference between adjacent intervals, and the RMS of the EMG signal. These physiological features are represented as... ;

[0093] The speech data undergoes semantic recognition and voiceprint matching. If the voiceprint matching is successful, the result of semantic recognition is used as a control command; otherwise, no control command is output.

[0094] In one specific embodiment of this application, the treadmill voice recognition interaction system uses a high-precision temperature-compensated crystal oscillator (TCXO) as a unified clock source to ensure that the timing deviation of the positioning data, physiological data and voice data collected by the UWB positioning device 11, physiological sensing device 13 and voice acquisition device 12 is less than 5ms. The system also marks the above data, including UWB positioning coordinates, heart rate signals and voice frames, with timestamps generated by the same clock source at the moment of acquisition, thereby achieving cross-modal data alignment and ensuring the spatiotemporal synchronization of physiological data and voice data.

[0095] Location data includes the user's UWB coordinates while using the treadmill. These UWB coordinates are transformed to the treadmill coordinate system and normalized to obtain the normalized coordinates of the user within a specified area of ​​the treadmill in the treadmill coordinate system. These normalized coordinates are then used as location features, represented as follows: ,in, and These represent the normalized x and y coordinates of the user's coordinates in the treadmill coordinate system when using the treadmill.

[0096] For example, UWB positioning coordinates are The UWB positioning coordinates are determined by the pre-calibrated rotation matrix and translation vector. Transform to the treadmill coordinate system with the treadmill center as the origin to obtain the treadmill's relative coordinates. ,

[0097] in, and These represent the x and y coordinates of the user in the treadmill coordinate system when using the treadmill. The pre-calibrated parameters include the plane angle θ = 12° between the UWB anchor point group and the center point of the treadmill, and the translation vector T = [0.35, 1.2]. The coordinate transformation calibration method is existing technology and will not be elaborated here. The coordinate transformation process is shown in the following formula:

[0098] .

[0099] The relative coordinates of the treadmill are normalized and scaled to the [0,1] interval according to the specified area of ​​the treadmill. and satisfy: , , and These represent the maximum length and maximum width of a specified area on the treadmill, respectively. For example, =3m, ,but .

[0100] Physiological data includes heart rate and electromyography (EMG) signals. R-wave detection and time-domain analysis are performed on the heart rate signal. Specifically, the Christov algorithm is used to identify the peak pulse wave in the PPG signal acquired by the PPG sensor. Then, the root mean square (RMS) of the standard deviation and difference between adjacent R-wave intervals are calculated based on the heart rate signal. These RMS are represented as SDNN and RMSSD, respectively. The root mean square (RMS) of the electromyography (EMG) signal is calculated based on the EMG signal, represented as EMG_RMS. Physiological features are generated based on the standard deviation, RMS, and EMG RMS. In the embodiments of this application, the R-wave detection and time-domain analysis of the heart rate signal are existing modules or networks, and their specific calculation processes will not be described in detail here.

[0101] The speech data undergoes semantic recognition and voiceprint matching. If the voiceprint matching is successful, the result of the semantic recognition is used as a control command; otherwise, no control command is output. The signal-to-noise ratio of the speech data is also calculated. Semantic recognition and voiceprint matching can be processed using existing technologies and will not be elaborated upon here.

[0102] S2, construct a dynamic identity identification model. The input of the dynamic identity identification model is the basic identity identification at the current moment and the dynamic identity identification at the previous moment, and the output is the dynamic identity identification at the current moment. Specifically, the dynamic identity identification is initialized based on the user's initial position characteristics and physiological characteristics within the designated area of ​​the treadmill. The basic identity identification at the current moment is generated based on the current position characteristics and physiological characteristics. The dynamic identity identification is updated once at each preset update interval based on the dynamic identity identification model.

[0103] In a specific embodiment, a basic identity identifier for the current moment is generated based on the current location and physiological characteristics, as shown in the following formula:

[0104] ;

[0105] in, This represents the basic identity identifier at the current moment;

[0106] A dynamic identity model is constructed based on the current basic identity identifier, as shown in the following formula:

[0107] ;

[0108] in, This represents the dynamic identity identifier at the current moment. Indicates the current time; Indicates the attenuation coefficient; This represents the dynamic identity identifier from the previous moment. Indicates the previous moment.

[0109] S3, in response to determining that the current position feature is not within the specified area of ​​the treadmill, return to the previous step; otherwise, determine the type of control instruction. The types of control instructions include assistive function parameter instructions, safe exercise parameter instructions, and high-risk exercise parameter instructions.

[0110] In a specific embodiment, determining the type of control command includes:

[0111] The control command includes speed control parameters, gradient control parameters, and other control parameters. In response to determining that the speed control parameters and gradient control parameters are both empty, and the other control parameters are not empty, the control command is an auxiliary function parameter command.

[0112] In response to the determination that the speed control parameter is within the safe speed range and the slope control parameter is less than or equal to the slope threshold, the control command is a safe motion parameter command.

[0113] If the speed control parameter is determined to be outside the safe speed range, or the slope control parameter is greater than the slope threshold, the control command is a high-risk motion parameter command.

[0114] In a specific embodiment, determining the type of control command includes:

[0115] The control command includes speed control parameters, gradient control parameters, and other control parameters. In response to determining that the speed control parameters and gradient control parameters are both empty, and the other control parameters are not empty, the control command is an auxiliary function parameter command.

[0116] A motion risk quantification model is constructed. The inputs to the motion risk quantification model are speed control parameters and slope control parameters, and the output is the motion risk value. The motion risk quantification model is shown in the following formula:

[0117] ;

[0118] in, Indicates the risk value of exercise. Indicates speed control parameters, This indicates the treadmill's maximum speed. Indicates the weights of the speed control parameters. Indicates the slope control parameters. This indicates the maximum incline of the treadmill. Indicates the weight of the slope control parameters;

[0119] In response to determining that the motion risk value is less than or equal to the motion risk threshold, the control command is a safe motion parameter command;

[0120] In response to the determination that the motion risk value is greater than the motion risk threshold, the control command is a high-risk motion parameter command.

[0121] S4, in response to determining that the control command is an auxiliary function parameter command, execute the auxiliary function parameter command.

[0122] S5, in response to determining that the control command is a safe motion parameter command, performs the first verification operation on the physiological data at the current moment. If the verification passes, the safe motion parameter command is executed; otherwise, the safe motion parameter command is not executed.

[0123] In a specific embodiment, a first verification operation is performed on the physiological data at the current moment, specifically including:

[0124] Determine whether the current heart rate signal is within the first heart rate verification range. If so, the verification passes; otherwise, the verification fails.

[0125] S6, in response to determining that the control command is a high-risk motion parameter command, performs a second verification operation on the current physiological data, location data and dynamic identity identifier. If the verification passes, the high-risk motion parameter command is executed; otherwise, the high-risk motion parameter command is not executed.

[0126] In a specific embodiment, a second verification operation is performed on the current physiological data, location data, and dynamic identity identifier, specifically including:

[0127] The change in heart rate at the current moment is calculated based on the heart rate signal of the user within a certain period after the initial moment when the user is in the designated area of ​​the treadmill, and the heart rate signal at the current moment, as shown in the following formula:

[0128] ;

[0129] in, This indicates the change in heart rate at the current moment. This indicates the heart rate signal at the current moment. This represents the average heart rate over a period of time after the initial moment when the user is positioned within the designated area of ​​the treadmill.

[0130] If the heart rate change at the current moment is determined to be within the range of the second heart rate verification or the root mean square of electromyography at the current moment is greater than the first root mean square threshold of electromyography, then the second stage of verification will be entered; otherwise, the verification will fail.

[0131] Spatial similarity, physiological similarity, and fusion similarity are calculated based on the current location features, physiological features, and dynamic identity identifiers. The calculation process is as follows:

[0132] ;

[0133] ;

[0134] ;

[0135] in, Representing spatial similarity, Indicates the location characteristics at the current moment. This represents the positional feature formed by the first and second elements of the dynamic identity identifier at the current moment. This represents the maximum Euclidean distance in the normalized space. Indicates the sign of the vector magnitude; Indicating physiological similarity, Indicates the physiological characteristics at the current moment. This refers to the physiological characteristics formed by the third to fifth elements of the dynamic identity identifier at the current moment; Indicates fusion similarity, Weighting coefficients representing physiological similarity;

[0136] If the fusion similarity is greater than the similarity threshold, the second stage of verification is initiated; otherwise, the verification fails.

[0137] The third stage of verification begins. If the heart rate signal at the current moment is greater than the heart rate threshold, or if the electromyography (EMG) signal has been continuously lost at the current moment or for a period of time prior to the current moment, the verification fails. A gradient deceleration command and a voice warning command are then executed to reduce the treadmill speed to a safe speed within a preset execution time. Simultaneously, a voice prompt indicating a high heart rate or lost EMG signal is provided. The relationship between the safe speed and the preset execution time satisfies the following equation:

[0138] ;

[0139] in, Indicates a safe speed; This indicates the minimum speed of the treadmill. Let represent the current speed of the treadmill, t represent the execution time of the gradient deceleration command, and e represent the natural constant. If the value is larger, the check will pass; otherwise, the check will pass.

[0140] In a specific embodiment, the method further includes dynamically adjusting the weighting coefficients of physiological similarity, specifically including:

[0141] Construct and initialize spatial and physiological weights, which satisfy the following equation:

[0142] ;

[0143] in, Indicates spatial weights, Indicates physiological weight;

[0144] In response to determining that the root mean square of electromyography (EMG) is greater than the second EMG threshold, the weighting coefficient of physiological similarity is increased according to the physiological weight and a preset adjustment coefficient. The weighting coefficient of physiological similarity is shown in the following formula:

[0145] ;

[0146] in, This represents the preset first adjustment factor;

[0147] The signal-to-noise ratio (SNR) is calculated based on the speech data. In response to a SNR threshold, the physiological similarity weighting coefficient is reduced based on physiological weights and a preset adjustment coefficient. The physiological similarity weighting coefficient is shown in the following formula:

[0148] ;

[0149] in, This indicates the preset second adjustment coefficient;

[0150] In response to determining that the root mean square of electromyography (RMS) is less than or equal to the second RMS threshold and the signal-to-noise ratio (SNR) is greater than or equal to the SNR threshold, the weighting coefficient of physiological similarity is equal to the physiological weight.

[0151] In one specific embodiment of this application, the update interval of the dynamic identity is set to 200ms, and the decay coefficient is... The decay rate is 5% every 200ms. User location data is acquired at a frequency of 100ms / frame. The dynamic identity is updated every 200ms when the user's location data is within the designated area of ​​the treadmill for 5 consecutive frames, the signal-to-noise ratio of the voice data is greater than 15dB, and the electromyography (EMG) signal is continuously valid for 5 seconds. The first heart rate verification range is set to [60, 140], in bpm; the second heart rate verification range is set to [40, 120], in bpm; the first EMG root mean square threshold is set to 0.3mV; the second EMG root mean square threshold is set to 0.2mV; the similarity threshold is set to 0.75; the heart rate threshold is set to 180bpm; and the EMG signal loss time threshold is set to 5 seconds. The gradient deceleration command execution time is 5 seconds. The minimum speed of the treadmill is 5km / h, the maximum speed is 12km / h, the maximum incline is 8%, the exercise risk threshold is 0.7, and the speed control parameter weights are... Slope control parameter weights The safe speed range is (0, 12], in km / h. Spatial weights are initialized to... Physiological weights are initialized to First adjustment coefficient Set to 0.2, the second adjustment factor. Set to 0.2, and the signal-to-noise ratio threshold to 20dB.

[0152] The following section provides a further explanation of the treadmill voice recognition interactive control method based on positioning data and physiological data proposed in this application, using specific scenarios as examples.

[0153] When the treadmill is activated, the voice data collected by the Infineon IM69D130 microphone array is matched with the static voiceprint of the user using the treadmill. The semantic recognition result is "Start running, speed 6km / h". The UWB anchor point detects the UWB positioning coordinates as (1.2, 0.3, 1.5), which is within the designated area of ​​the treadmill. The PPG sensor detects a heart rate signal of 72bpm, which is within the first heart rate verification range. This control command is a safe exercise parameter command. The treadmill executes the "Start running" safe exercise parameter command, controlling the motor to accelerate to the speed control parameter of 6km / h. At the same time, the Infineon IM69D130 microphone array is pointed at the user's head, and the dynamic identity is updated every 200ms based on the real-time acquired positioning data and physiological data.

[0154] In the scenario of executing assistive function parameter commands, semantic recognition is performed based on the voice data collected by the Infineon IM69D130 microphone array. If the user's voiceprint is successfully matched, the semantic recognition result is only "play music," and the treadmill executes the "play music" assistive function parameter command. In a specific embodiment of this application, the treadmill voice recognition interaction system also includes an air outlet device, and the assistive function parameter commands also include querying data and starting / stopping the air outlet device.

[0155] In the dynamic identity update scenario, the user's current location features [0.72, 0.85] and physiological features are calculated based on the real-time collected user location data and physiological data. SDNN is measured in milliseconds (ms), RMSSD in milliseconds (ms), and EMG_RMS in mV. If the user remains within the designated area of ​​the treadmill and the dynamic identity identifier of the previous moment is [0.70, 0.80, 60, 40, 0.25], then the dynamic identity identifier of the current moment is updated based on the dynamic identity identifier model, as shown in the following formula:

[0156]

[0157] .

[0158] In a scenario where a safe exercise parameter command is executed, the treadmill's current speed is 6 km / h and the incline is 0°. Semantic recognition is performed based on voice data collected by the Infineon IM69D130 microphone array. The user's voiceprint is successfully matched, and the control command is "Adjust speed to 7 km / h," without any incline control parameter. This control command only includes the speed control parameter, specifically 7 km / h. A threshold judgment method for control parameters is used to determine the type of this control command. Since the speed control parameter 7 km / h is less than the treadmill's maximum speed of 12 km / h, the control command falls under the category of safe exercise parameter commands. In a specific embodiment of this application, the type of control command can also be determined based on a motion risk quantification model, and a motion risk value can be calculated. If the value is less than the motion risk threshold of 0.7, the control command is considered a safe motion parameter command. Preferably, both the control parameter threshold judgment method and the motion risk quantification model method can be used simultaneously to determine the type of the control command. If both methods meet the requirements for determining a safe motion parameter command, then the control command is a safe motion parameter command.

[0159] In the first active protection scenario, the user, while panting, says "Stop." According to the UWB positioning coordinates collected by UWB positioning device 11, the user is located within the designated area of ​​the treadmill. Based on the voice data collected by the Infineon IM69D130 microphone array, signal-to-noise ratio calculation, semantic recognition, and voiceprint matching are performed. The signal-to-noise ratio is 25dB, and the user's voiceprint matching is successful. The control command is "Stop," which is a high-risk exercise parameter command, requiring a second verification operation. The voiceprint matching accuracy is 62%. The user's average heart rate within the designated treadmill area for the previous 30 seconds was 150 bpm, and the current heart rate signal is 198 bpm. The current heart rate change... Within the second heart rate verification range [40, 120], EMG_RMS = 0.38mV, which is greater than the first EMG root mean square threshold of 0.3mV. The physiological mutation is consistent with the expected sudden arrest, and the first verification stage passes, proceeding to the second verification stage. Spatial similarity is calculated based on the user's current location characteristics, physiological characteristics, and dynamic identity identifier. Physiological similarity Since EMG_RMS is greater than the second electromyographic root mean square threshold of 0.2mV, the weighting coefficient of physiological similarity is calculated according to the dynamic adjustment rule of the weighting coefficient of physiological similarity, and the fusion similarity is calculated based on spatial similarity and physiological similarity. The calculation process is as follows:

[0160] ;

[0161] .

[0162] The fusion similarity of 0.886 is greater than the similarity threshold of 0.75, so the second verification stage passes and the system proceeds to the third verification stage. The current heart rate signal is 198 bpm, which is greater than the heart rate threshold of 180 bpm, so the third verification stage fails. A gradient deceleration command and a voice warning command are then executed. Within 5 seconds of the gradient deceleration command's execution time, the treadmill speed is reduced to a safe speed, and an alarm light and voice prompt are triggered to provide exercise risk warnings, achieving proactive protection and ensuring user safety. In a specific embodiment of this application, the physiological weight ranges from [0.4, 0.8], with its lower limit representing the safe boundary of the physiological weight and its upper limit used to prevent spatial weight failure. Clinical studies have shown that when the sternocleidomastoid muscle EMG_RMS > 0.2 mV, the standard deviation of the vocal cord fundamental frequency increases by 120 ± 15 Hz, and the voiceprint recognition accuracy decreases to ≤ 50%, where 0.2 mV is the exercise breathing threshold. In the embodiments of this application, existing voiceprint recognition technology is used to match voice data with the user's voiceprint already stored in the user registration / configuration device 14. When the intensity of movement increases, the accuracy of recognition based solely on voiceprint decreases, and the location data changes rapidly. At this time, it is necessary to increase the user's physiological weight during the dynamic binding process of location data, physiological data, and voice data. Specifically, it is necessary to increase the weight coefficient of physiological similarity to compensate for voiceprint failure. In a specific embodiment of this application, based on the fitting curve of the sports medicine database, a compensation rate of 20% is set and used as the first adjustment coefficient. In practical applications, the first adjustment coefficient Adjustments can be made based on experiments and different equipment. In the first active protection scenario, the voice data is in a state of motion distortion due to breathing, resulting in a voiceprint matching accuracy of 62%, which is lower than the voiceprint matching threshold of 70%. By adjusting the weight coefficients of physiological features, the problem of decreased voiceprint recognition accuracy caused by motion distortion of the voice data can be compensated.

[0163] The second active protection scenario differs from the first active protection scenario in that, up to the current moment, the electromyographic signal has been continuously lost for more than 5 seconds, the third verification stage fails the verification, and the gradient deceleration command and voice warning command are executed.

[0164] The difference between the emergency stop scenario and the first active protection scenario lies in the user's average heart rate during the first 30 seconds within the designated treadmill area: 130 bpm, and the current heart rate signal is 175 bpm. (Current heart rate change...) Within the second heart rate verification range [40, 120], the first verification stage passed. The verification process for the second verification stage is the same as that for the first active protection scenario, and the second verification stage passed. The current heart rate signal is 175 bpm, which is less than the heart rate threshold of 180 bpm, and the electromyographic signal has been continuously valid from the initial moment when the user was in the designated area of ​​the treadmill until the current moment. The third verification stage passed, and the treadmill executed the "stop" command, triggering the emergency stop mechanism within 800ms.

[0165] The third active protection scenario differs from the first in that the EMG_RMS is less than 0.2mV, and the user is in a high-noise environment, such as a neighboring treadmill user shouting loudly. The signal-to-noise ratio (SNR) calculated from the real-time collected voice data is 18dB. Acoustic experiments (ICASSP 2022 dataset) show that acoustic reliability decreases by 20% when SNR < 20dB. Moreover, in emergency situations, the user may have lost physiological data or exceeded safety limits. In this case, increasing spatial weights during the dynamic binding of location data, physiological data, and voice data is more reliable. Increasing the weight coefficient of spatial similarity and decreasing the weight coefficient of physiological similarity reduces the impact of the noisy environment on physiological characteristics. In a specific embodiment of this application, based on the noise robustness test dataset, an attenuation rate of 20% is set and used as the second adjustment coefficient. In practical applications, the second adjustment coefficient Adjustments can be made based on experiments and different equipment. The weighting coefficients of physiological similarity are calculated according to the dynamic adjustment rules of the physiological similarity weighting coefficients, and the fusion similarity is calculated based on spatial similarity and physiological similarity. The calculation process is as follows:

[0166] ;

[0167] .

[0168] Fusion similarity If the similarity score is greater than the threshold of 0.75, the second verification stage passes and the third verification stage begins. If the current heart rate is 198 bpm, which is greater than the heart rate threshold of 180 bpm, the third verification stage fails, and the gradient deceleration command and voice warning command are executed.

[0169] In the first and third active protection scenarios, to address the issue of complex acoustic environments and decreased accuracy of voiceprint matching in motion scenarios with multiple participants, dynamic adjustments are made to the weighting coefficients of physiological similarity and spatial similarity. This prevents the impact of motion and complex acoustic environments on voiceprint matching from extending into the verification process of control commands, ensuring the accuracy of the second verification operation. Consequently, more precise dynamic binding between positioning data, physiological data, and voice data is achieved, further improving the accuracy of voice recognition interaction.

[0170] The fourth active protection scenario differs from the first and third active protection scenarios in that: EMG_RMS and signal-to-noise ratio do not conform to the two cases in the first and third active protection scenarios; the weight coefficient of physiological similarity is equal to the physiological weight; and the fusion similarity... If the similarity score is greater than the threshold of 0.75, the second verification stage passes and the third verification stage begins. If the current heart rate is 198 bpm, which is greater than the heart rate threshold of 180 bpm, the third verification stage fails, and the gradient deceleration command and voice warning command are executed.

[0171] The treadmill voice recognition interactive control method proposed in this application integrates location data and physiological data to construct a dynamic identity identifier. It dynamically binds synchronously collected voice data, location data, and physiological data, resulting in strong dynamic adaptability. It overcomes the problems of signal loss when using vision systems for location and sound pickup in existing technologies, the complex acoustic environment of multi-person exercise scenarios, the decrease in voiceprint matching accuracy during exercise, the high rate of false touches in voice recognition interaction based solely on voiceprint recognition, and the inability to construct a holographic user profile between the user corresponding to the location data and the user corresponding to the voice data. It is not only applicable to treadmills but can also be extended to smart exercise bikes, rowing machines, and other smart gym equipment in all scenarios, laying the technical foundation for the next generation of fitness human-computer interaction systems.

[0172] The steps S1-S6 above do not necessarily represent the order of the steps, but are represented by step symbols. The order of the steps can be adjusted.

[0173] Further reference Figure 3 As an implementation of the methods shown in the above figures, this application provides an embodiment of a treadmill voice recognition interactive control device based on positioning data and physiological data. This device embodiment is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0174] This application provides a treadmill voice recognition interactive control device based on positioning data and physiological data, including:

[0175] Data acquisition module 1 is configured to acquire in real time the user's positioning data, physiological data and voice data when using the treadmill, convert the positioning data to the treadmill coordinate system to obtain position features, perform feature calculation on the physiological data to obtain physiological features, and perform voice recognition on the voice data to obtain control commands.

[0176] The dynamic identity construction module 2 is configured to build a dynamic identity model. The input of the dynamic identity model is the basic identity at the current moment and the dynamic identity at the previous moment, and the output is the dynamic identity at the current moment. Specifically, the dynamic identity is initialized based on the user's initial position characteristics and physiological characteristics in the designated area of ​​the treadmill. The basic identity at the current moment is generated based on the current position characteristics and physiological characteristics. The dynamic identity is updated once at each preset update interval based on the dynamic identity model.

[0177] The instruction judgment module 3 is configured to return to the previous step if it determines that the current position feature is not within the specified area of ​​the treadmill; otherwise, it determines the type of control instruction, which includes auxiliary function parameter instructions, safe exercise parameter instructions, and high-risk exercise parameter instructions.

[0178] The first instruction execution module 4 is configured to execute the auxiliary function parameter instruction in response to determining that the control instruction is an auxiliary function parameter instruction;

[0179] The second instruction execution module 5 is configured to perform a first verification operation on the physiological data at the current moment in response to determining that the control instruction is a safe motion parameter instruction. If the verification passes, the safe motion parameter instruction is executed; otherwise, the safe motion parameter instruction is not executed.

[0180] The third instruction execution module 6 is configured to perform a second verification operation on the current physiological data, location data and dynamic identity identifier in response to determining that the control instruction is a high-risk motion parameter instruction. If the verification passes, the high-risk motion parameter instruction is executed; otherwise, the high-risk motion parameter instruction is not executed.

[0181] Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. For example... Figure 4 As shown, the electronic device in this embodiment includes a processor 401 and a memory 402; wherein the memory 402 is used to store computer execution instructions; and the processor 401 is used to execute the computer execution instructions stored in the memory to implement the various steps performed by the electronic device in the above embodiment. For details, please refer to the relevant descriptions in the foregoing method embodiments.

[0182] Alternatively, the memory 402 can be either standalone or integrated with the processor 401.

[0183] When the memory 402 is set up independently, the electronic device also includes a bus 403 for connecting the memory 402 and the processor 401.

[0184] This invention also provides a computer storage medium storing computer execution instructions, which, when executed by processor 401, implement the above method.

[0185] This invention also provides a computer program product, including a computer program, which, when executed by a processor 401, implements the above-described method.

[0186] In the embodiments provided by this invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0187] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.

[0188] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.

[0189] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor 401 to execute some steps of the methods of the various embodiments of this application.

[0190] It should be understood that the processor 401 described above can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor, or the processor 401 can be any conventional processor 401. The steps of the method disclosed in this invention can be directly manifested as the hardware processor 401 executing the steps, or as a combination of hardware and software modules within the processor 401 executing the steps.

[0191] The memory 402 may include high-speed RAM memory, and may also include non-volatile memory NVM, such as at least one disk storage device, and may also be a USB flash drive, portable hard drive, read-only memory, disk or optical disc, etc.

[0192] Bus 403 can be an Industry Standard Architecture (ISA), a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Bus 403 can be divided into address bus, data bus, control bus, etc. For ease of illustration, the bus 403 in the accompanying drawings of this application is not limited to only one bus 403 or one type of bus 403.

[0193] The aforementioned storage medium can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0194] An exemplary storage medium is coupled to processor 401, enabling processor 401 to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of processor 401. Processor 401 and storage medium can reside in application-specific integrated circuits (ASICs). Alternatively, processor 401 and storage medium can exist as discrete components in an electronic device or host device.

[0195] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0196] 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 or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A voice recognition interactive control method for a treadmill based on positioning data and physiological data, characterized in that, Includes the following steps: The system acquires real-time positioning data, physiological data, and voice data of the user while using the treadmill. It converts the positioning data to the treadmill coordinate system to obtain position features, performs feature calculation on the physiological data to obtain physiological features, and performs voice recognition on the voice data to obtain control commands. A dynamic identity identification model is constructed. The input of the dynamic identity identification model is the basic identity identification at the current moment and the dynamic identity identification at the previous moment, and the output is the dynamic identity identification at the current moment. The dynamic identity identification is initialized according to the position characteristics and physiological characteristics of the user at the initial moment when the user is located in the designated area of ​​the treadmill. The basic identity identification at the current moment is generated according to the position characteristics and physiological characteristics at the current moment. The dynamic identity identification is updated once at each preset update interval based on the dynamic identity identification model. If it is determined that the current position feature is not within the designated area of ​​the treadmill, the previous step is returned; otherwise, the type of the control instruction is determined. The types of control instructions include assistive function parameter instructions, safe exercise parameter instructions, and high-risk exercise parameter instructions. In response to determining that the control command is an auxiliary function parameter command, the auxiliary function parameter command is executed; In response to determining that the control command is a safe motion parameter command, a first verification operation is performed on the physiological data at the current moment. If the verification passes, the safe motion parameter command is executed; otherwise, the safe motion parameter command is not executed. In response to determining that the control command is a high-risk motion parameter command, a second verification operation is performed on the current physiological data, location data, and dynamic identity identifier. If the verification passes, the high-risk motion parameter command is executed; otherwise, the high-risk motion parameter command is not executed.

2. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 1, characterized in that, The process of converting the positioning data to the treadmill coordinate system to obtain position features, performing feature calculations on the physiological data to obtain physiological features, and performing voice recognition on the voice data to obtain control commands specifically includes: The positioning data includes the user's UWB positioning coordinates when using the treadmill. These UWB coordinates are converted to the treadmill coordinate system and normalized to obtain the normalized coordinates of the user within the normalized space corresponding to a specified area on the treadmill. These normalized coordinates are then used as position features, represented as follows: ,in, and These represent the normalized x and y coordinates of the user's coordinates in the treadmill coordinate system when using the treadmill. The physiological data includes heart rate signals and electromyography (EMG) signals. Based on the heart rate signals, the root mean square (RMS) of the standard deviation of adjacent R-wave intervals and the root mean square of the difference between adjacent intervals are calculated. These RMS are represented as SDNN and RMSSD, respectively. Based on the EMG signals, the root mean square (RMS) of the electromyography (EMG) signals is calculated. This RMS is represented as EMG_RMS. Physiological features are generated based on the standard deviation of adjacent R-wave intervals, the RMS of the difference between adjacent intervals, and the RMS of the electromyography (EMG). These physiological features are represented as... ; The speech data is subjected to semantic recognition and voiceprint matching. If the voiceprint matching is successful, the result of semantic recognition is used as a control command; otherwise, no control command is output.

3. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 2, characterized in that, The first verification operation performed on the physiological data at the current moment specifically includes: Determine whether the current heart rate signal is within the first heart rate verification range. If so, the verification passes; otherwise, the verification fails.

4. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 2, characterized in that, Based on the location and physiological characteristics at the current moment, a basic identity identifier for the current moment is generated, as shown in the following formula: ； in, This represents the basic identity identifier at the current moment; A dynamic identity model is constructed based on the current basic identity identifier, as shown in the following formula: ； in, This represents the dynamic identity identifier at the current moment. Indicates the current time; Indicates the attenuation coefficient; This represents the dynamic identity identifier from the previous moment. Indicates the previous moment.

5. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 4, characterized in that, The second verification operation performed on the current physiological data, location data, and dynamic identity identifier specifically includes: The change in heart rate at the current moment is calculated based on the heart rate signal over a period of time after the initial moment when the user is in the designated area of ​​the treadmill, and the heart rate signal at the current moment, as shown in the following formula: ； in, This indicates the change in heart rate at the current moment. This indicates the heart rate signal at the current moment. This represents the average heart rate over a period of time after the initial moment when the user is positioned within the designated area of ​​the treadmill. Enter the first stage of verification, and determine whether the current heart rate change is within the second heart rate verification range or whether the current electromyography root mean square is greater than the first electromyography root mean square threshold. If so, enter the second stage of verification; otherwise, the verification fails. Based on the current location features, physiological features, and dynamic identity identifier, spatial similarity, physiological similarity, and fusion similarity are calculated. The calculation process is as follows: ； ； ； in, Representing spatial similarity, Indicates the location characteristics at the current moment. This represents the positional feature formed by the first and second elements of the dynamic identity identifier at the current moment. This represents the maximum Euclidean distance in the normalized space. Indicates the sign of the vector magnitude; Indicating physiological similarity, Indicates the physiological characteristics at the current moment. This refers to the physiological characteristics formed by the third to fifth elements of the dynamic identity identifier at the current moment; Indicates fusion similarity, Weighting coefficients representing physiological similarity; The second stage of verification is initiated to determine whether the fusion similarity is greater than the similarity threshold. If so, the third stage of verification is initiated; otherwise, the verification fails. The third stage of verification involves determining whether the current heart rate signal exceeds a heart rate threshold, or whether the electromyographic signal has been continuously lost at the current moment or for a period prior to the current moment. If so, the verification fails, and a gradient deceleration command and a voice warning command are executed. Within a preset execution time, the treadmill speed is reduced to a safe speed, while a voice prompt indicating a high heart rate or lost electromyographic signal is provided. The relationship between the safe speed and the preset execution time satisfies the following formula: ； in, Indicates a safe speed; This indicates the minimum speed of the treadmill. Let represent the current speed of the treadmill, t represent the execution time of the gradient deceleration command, and e represent the natural constant. If the value is larger, the check will pass; otherwise, the check will pass.

6. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 5, characterized in that, It also includes dynamically adjusting the weighting coefficients of physiological similarity, specifically including: Construct and initialize spatial weights and physiological weights, which satisfy the following equation: ； in, Indicates spatial weights, Indicates physiological weight; In response to determining that the root mean square of electromyography (EMG) is greater than a second EMG threshold, the weighting coefficient of physiological similarity is increased according to the physiological weight and a preset adjustment coefficient, wherein the weighting coefficient of physiological similarity is as follows: ； in, This represents the preset first adjustment factor; The signal-to-noise ratio (SNR) is calculated based on the speech data. In response to determining that the SNR is less than a SNR threshold, the weight coefficient of physiological similarity is reduced according to the physiological weight and a preset adjustment coefficient. The weight coefficient of physiological similarity is shown in the following formula: ； in, This indicates the preset second adjustment coefficient; In response to determining that the root mean square of electromyography (RMS) is less than or equal to a second RMS threshold and the signal-to-noise ratio (SNR) is greater than or equal to a SNR threshold, the weighting coefficient of the physiological similarity is equal to the physiological weight.

7. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 1, characterized in that, The determination of the type of the control command specifically includes: The control command includes speed control parameters, slope control parameters, and other control parameters. In response to determining that the speed control parameters and slope control parameters are both empty, and the other control parameters are not empty, the control command is an auxiliary function parameter command. In response to determining that the speed control parameter is within a safe speed range and the slope control parameter is less than or equal to a slope threshold, the control command is a safe motion parameter command; In response to determining that the speed control parameter is not within the safe speed range, or that the slope control parameter is greater than the slope threshold, the control command is a high-risk motion parameter command.

8. The treadmill voice recognition interactive control method based on positioning data and physiological data according to claim 1, characterized in that, The determination of the type of the control command specifically includes: The control command includes speed control parameters, slope control parameters, and other control parameters. In response to determining that the speed control parameters and slope control parameters are both empty, and the other control parameters are not empty, the control command is an auxiliary function parameter command. A motion risk quantification model is constructed. The inputs of the motion risk quantification model are speed control parameters and slope control parameters, and the output is the motion risk value. The motion risk quantification model is shown in the following formula: ； in, Indicates the risk value of exercise. Indicates speed control parameters, This indicates the treadmill's maximum speed. Indicates the weights of the speed control parameters. Indicates the slope control parameters. This indicates the maximum incline of the treadmill. Indicates the weight of the slope control parameters; In response to determining that the motion risk value is less than or equal to the motion risk threshold, the control command is a safe motion parameter command; In response to determining that the exercise risk value is greater than the exercise risk threshold, the control command is a high-risk exercise parameter command.

9. A treadmill voice recognition interactive control device based on positioning data and physiological data, characterized in that, include: The data acquisition module is configured to acquire the user's positioning data, physiological data and voice data in real time when using the treadmill, convert the positioning data to the treadmill coordinate system to obtain position features, perform feature calculation on the physiological data to obtain physiological features, and perform voice recognition on the voice data to obtain control commands. The dynamic identity construction module is configured to construct a dynamic identity model. The input of the dynamic identity model is the basic identity at the current moment and the dynamic identity at the previous moment, and the output is the dynamic identity at the current moment. Specifically, the dynamic identity is initialized based on the location features and physiological features of the user at the initial moment when the user is located in the designated area of ​​the treadmill. The basic identity at the current moment is generated based on the location features and physiological features at the current moment. The dynamic identity is updated once at each preset update interval based on the dynamic identity model. The instruction judgment module is configured to return to the previous step in response to determining that the current position feature is not within the designated area of ​​the treadmill; otherwise, it determines the type of the control instruction, which includes auxiliary function parameter instructions, safe exercise parameter instructions, and high-risk exercise parameter instructions. The first instruction execution module is configured to execute the auxiliary function parameter instruction in response to determining that the control instruction is an auxiliary function parameter instruction; The second instruction execution module is configured to perform a first verification operation on the physiological data at the current moment in response to determining that the control instruction is a safe motion parameter instruction. If the verification passes, the safe motion parameter instruction is executed; otherwise, the safe motion parameter instruction is not executed. The third instruction execution module is configured to, in response to determining that the control instruction is a high-risk motion parameter instruction, perform a second verification operation on the physiological data, location data and dynamic identity identifier at the current moment. If the verification passes, the high-risk motion parameter instruction is executed; otherwise, the high-risk motion parameter instruction is not executed.

10. An electronic device, comprising: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-8.

Images