An energy consumption estimation method and device based on electrocardiosignal and storage medium

Abstract

The present application relates to the technical field of physiological signal processing, and discloses an energy consumption estimation method and device based on electrocardiosignal and a storage medium. The method extracts beat-by-beat heart beat information, heart rate variability features and electrocardio-derived respiration features from an ECG signal, and combines signal quality control, metabolic intensity regression model and start-stop kinetics compensation to realize continuous estimation of instantaneous energy consumption rate and cumulative energy consumption.

Description

Technical Field

[0001] This invention relates to the field of physiological signal processing technology, specifically to a method, apparatus, and storage medium for estimating energy consumption based on electrocardiogram (ECG) signals. Background Technology

[0002] Human energy expenditure reflects the body's metabolic intensity and activity level, and is a key indicator for exercise prescription, training intensity control, obesity intervention, chronic disease management, and long-term health monitoring. Existing energy expenditure measurement and estimation techniques mainly include the following categories:

[0003] (1) Laboratory-level energy consumption measurement method

[0004] Methods such as the indirect calorimetry method obtain oxygen consumption and carbon dioxide emissions through respiratory gas analysis, and then calculate metabolic intensity and energy consumption. This type of method has high accuracy, but the equipment is large, costly, and complex to use, making it difficult to meet the needs of long-term, continuous, and low-intervention monitoring in daily environments.

[0005] (2) Empirical estimation methods based on inertial sensors or motion measurement

[0006] Energy consumption can be estimated by building empirical models using information such as acceleration, step count, and posture. This type of method is simple to implement, but it has a large error in scenarios where the type of activity varies significantly, such as cycling, upper limb-dominant activities, weight-bearing activities, and static exertion activities. At the same time, it is affected by individual gait differences, wearing position, and environmental factors, making it difficult to stably reflect the true metabolic intensity.

[0007] (3) Energy consumption estimation method based on heart rate level

[0008] Heart rate and oxygen consumption are correlated under certain conditions, and some existing technologies estimate energy consumption by using the regression relationship between heart rate level and energy consumption. However, such methods usually rely on the steady-state assumption and are difficult to maintain accuracy under dynamic conditions such as sudden changes in intensity, the start-up phase of exercise, and the recovery phase. At the same time, heart rate is easily affected by non-metabolic factors, such as mental stress, caffeine, changes in body temperature, and changes in posture, which leads to systematic biases in energy consumption estimation. In addition, some solutions require users to perform individual calibration tests, which reduces the convenience of long-term wearable use.

[0009] Therefore, there is an urgent need for a technical solution based on wearable and available ECG signals that can stably estimate metabolic intensity and energy consumption under dynamic conditions, and that can reduce dependence on individual calibration and be robust to noise and artifacts. Summary of the Invention

[0010] To address the aforementioned shortcomings of existing technologies, this invention provides a method, apparatus, and storage medium for estimating energy consumption based on electrocardiogram (ECG) signals. This method extracts beat-by-beat heart rate information, heart rate variability characteristics, and ECG-derived respiratory characteristics from ECG signals, and combines this with signal quality control, a metabolic intensity regression model, and start-stop kinetic compensation to achieve continuous estimation of instantaneous and cumulative energy consumption. This overcomes the deficiencies of existing technologies in addressing dynamic intensity changes, motion artifact interference, and individual calibration dependence.

[0011] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0012] In a first aspect, the present invention proposes an energy consumption estimation method based on electrocardiogram signals, comprising the following steps:

[0013] Continuous ECG signals are acquired and preprocessed to generate signal quality indicators characterizing signal availability.

[0014] Based on the preprocessed ECG signal, the beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features were extracted.

[0015] The beat-by-beat heart rate characteristics, heart rate variability characteristics, ECG-derived respiratory rate, respiratory stability characteristics, individual parameters, and signal quality indicators are input into the trained regression model, which outputs an estimated metabolic intensity.

[0016] The instantaneous energy consumption rate is calculated based on the metabolic intensity, and the instantaneous energy consumption rate is integrated over time to obtain the cumulative energy consumption.

[0017] Preferably, based on the preprocessed ECG signal, beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features are extracted, including:

[0018] R-peak detection is performed on the ECG signal to obtain the R-peak time series, and the beat-by-beat interval and beat-by-beat heart rate are calculated based on the R-peak time series.

[0019] Heart rate variability characteristics are calculated based on the beat-by-beat interval within a sliding window;

[0020] The ECG-derived respiratory signal is constructed by extracting the amplitude of the R peak in the ECG signal and passing it through a low-pass filter. The ECG-derived respiratory signal is then subjected to spectral analysis to obtain the ECG-derived respiratory frequency, and respiratory stability characteristics are calculated.

[0021] Preferably, the individual parameters include at least one or any combination of age, gender, height, and weight.

[0022] Preferably, the estimation of metabolic intensity also includes start-stop kinetic compensation, specifically:

[0023] Obtain the initial value of metabolic intensity output by the regression model;

[0024] The exercise is determined to be in the starting or stopping / recovering state based on whether the current rate of change of heart rate and rate of change of respiratory rate both exceed a preset threshold.

[0025] Based on the discrimination results, the corresponding adaptive coefficient is selected, and the metabolic intensity is dynamically updated according to the following formula:

[0026] ;

[0027] in, The metabolic intensity is updated at time t. For adaptive coefficients, This is the initial value for metabolic intensity. This is the sampling update interval.

[0028] Preferably, the estimation of metabolic intensity also includes performing extended Kalman filtering, specifically:

[0029] The output of the regression model is used as an observation, and based on the state equation... and observation equations Construct an extended Kalman filter, where As a state variable of metabolic intensity, for, For process noise, for, for, To observe noise;

[0030] The metabolic intensity is estimated by fusion using the extended Kalman filter, and the filtered output is used as the final metabolic intensity.

[0031] Preferably, the weight or magnitude of the observation noise in the extended Kalman filter is adjusted according to the signal quality index. When the signal quality index is lower than a preset quality threshold, the observation noise is increased or its weight is decreased.

[0032] Preferably, the calculation of instantaneous energy consumption rate based on the metabolic intensity further includes signal quality control, specifically:

[0033] If the signal quality index at the current moment is lower than the preset quality threshold, the metabolic intensity used to calculate the current instantaneous energy consumption rate will be frozen to the metabolic intensity of the previous effective moment.

[0034] Preferably, calculating the instantaneous energy consumption rate based on the metabolic intensity includes:

[0035] Based on the metabolic intensity and oxygen consumption calorific value, the instantaneous energy consumption rate is calculated as follows:

[0036] ;

[0037] in, Let be the instantaneous energy consumption rate at time t. Metabolic intensity, It is the oxygen consumption heat equivalent.

[0038] Secondly, this invention proposes an energy consumption estimation device based on electrocardiogram (ECG) signals, which applies the energy consumption estimation method based on ECG signals as described above, including:

[0039] The preprocessing module is used to acquire continuous ECG signals and preprocess the ECG signals to generate signal quality indicators that characterize the usability of the signals.

[0040] The feature extraction module is used to extract beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features based on the preprocessed ECG signal.

[0041] The metabolic intensity estimation module is used to input the beat-by-beat heart rate characteristics, heart rate variability characteristics, ECG-derived respiratory rate, respiratory stability characteristics, individual parameters, and signal quality indicators into the trained regression model and output the estimated metabolic intensity.

[0042] The energy calculation module is used to calculate the instantaneous energy consumption rate based on the metabolic intensity, and to integrate the instantaneous energy consumption rate over time to obtain the cumulative energy consumption.

[0043] Thirdly, the present invention proposes a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the energy consumption estimation method based on electrocardiogram signals as described above.

[0044] The present invention has the following beneficial effects:

[0045] This invention extracts beat-by-beat heart rate information, heart rate variability features, and ECG-derived respiratory features from ECG signals, and combines these with signal quality control, metabolic intensity regression models, and start-stop kinetic compensation to achieve continuous estimation of instantaneous energy consumption rate and cumulative energy consumption. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the energy consumption estimation method based on electrocardiogram signals according to the present invention;

[0047] Figure 2 Flowchart for ECG preprocessing, R peak detection, and RR / HR calculation;

[0048] Figure 3Flowchart for calculating EDR and respiratory rate RespR;

[0049] Figure 4 Here is a flowchart of metabolic intensity regression estimation and start-stop kinetics compensation.

[0050] Figure 5 This is a flowchart for calculating instantaneous energy consumption rate and cumulative energy consumption integration. Detailed Implementation

[0051] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0052] Example 1

[0053] Reference Figure 1 The present invention provides an energy consumption estimation method based on electrocardiogram signals, comprising the following steps S1 to S4:

[0054] S1. Acquire continuous ECG signals and preprocess the ECG signals to generate signal quality indicators characterizing signal availability;

[0055] In an optional embodiment of the present invention, step S1 involves using a single-lead ECG sensor or a multi-lead ECG sensor on a wearable terminal to acquire continuous ECG signals and obtain a digital ECG sequence ECG[n] through analog-to-digital conversion; at the same time, individual parameters are input, including one or any combination of age, gender, height or weight, for individualized correction of metabolic intensity estimation.

[0056] In this embodiment, the digital ECG sequence is preprocessed to suppress noise and drift. The preprocessing uses a bandpass filter to remove baseline drift and high-frequency noise, with the passband set to 0.5Hz to 40Hz. At the same time, a notch filter is used to suppress power frequency interference, with the notch center frequency set to 50Hz or 60Hz.

[0057] This embodiment further employs wavelet transform for denoising, with Symlet5 as the wavelet mother function and a wavelet decomposition level of 6. The threshold for denoising satisfies the following:

[0058] ;

[0059] in, The standard deviation of noise. The signal length is used to improve the robustness of R-peak detection and preserve the dynamic characteristics of the heartbeat.

[0060] Motion artifact segments generate signal quality indicators through evaluation of local energy abrupt changes in ECG, peak shape consistency, and R-peak detectability. and signal quality indicators The signal quality index (SQI) is compared with a preset threshold to distinguish between valid and low-quality fragments. This SQI characterizes the availability of the current ECG fragment, providing a basis for subsequent R-peak detection reliability control, metabolic intensity output control, and cumulative energy consumption accumulation strategies.

[0061] S2. Based on the preprocessed ECG signal, extract the beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features.

[0062] In an optional embodiment of the present invention, step S2 extracts beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features based on the preprocessed ECG signal, including:

[0063] R-peak detection is performed on the ECG signal to obtain the R-peak time series, and the beat-by-beat interval and beat-by-beat heart rate are calculated based on the R-peak time series.

[0064] Heart rate variability characteristics are calculated based on the beat-by-beat interval within a sliding window;

[0065] The ECG-derived respiratory signal is constructed by extracting the amplitude of the R peak in the ECG signal and passing it through a low-pass filter. The ECG-derived respiratory signal is then subjected to spectral analysis to obtain the ECG-derived respiratory frequency, and respiratory stability characteristics are calculated.

[0066] This embodiment performs R-peak detection on the preprocessed ECG sequence to obtain beat-by-beat heart rate information. R-peak detection is achieved using a combination of threshold detection and template matching, or it can be implemented using a Pan-Tompkins-based detection method. The R-peak time series {t} is obtained from the R-peak detection. i} Calculate the beat interval and heart rate:

[0067] ;

[0068] ;

[0069] in, The interval between heartbeats is measured in seconds. For the i-th peak; Heart rate, measured in bpm. Outlier correction is performed on each beat interval. When the preset physiological constraint threshold is not met, the corresponding heartbeat is marked as abnormal and replaced with the interpolation result of the interval between adjacent normal heartbeats to avoid the accumulation of errors caused by false positives and false negatives.

[0070] This embodiment extracts a heart rate variability (HRV) feature set within a sliding window to characterize autonomic nervous system regulation and dynamic heart rate fluctuations, thereby improving the identifiability of metabolic state changes. The HRV feature set is calculated from beat-by-beat intervals within the sliding window. The HRV feature set includes RMSSD and SDNN. RMSSD is calculated from the root mean square difference of adjacent RR intervals, and SDNN is calculated from the standard deviation of RR intervals within the window.

[0071] This embodiment improves the ability to identify changes in metabolic state and reduces errors caused by relying solely on heart rate levels, based on the characteristics of ECG beat-by-beat intervals and heart rate variability.

[0072] In this embodiment, an ECG-derived respiratory sequence (EDR) is constructed, and the respiratory rate (RespR) and respiratory stability feature set are calculated. Respiratory features are obtained through the ECG-derived respiratory EDR, which is constructed using R-peak amplitude modulation: the R-peak amplitude of each heartbeat is extracted to form an R-peak amplitude sequence, and this sequence is low-pass filtered to obtain the ECG-derived respiratory sequence EDR(t). Spectral analysis of the ECG-derived respiratory sequence EDR(t) is performed within a fixed time window to obtain the main peak frequency. Calculate respiratory rate:

[0073] ;

[0074] It also calculates the respiratory stability feature set, which includes the respiratory cycle variation coefficient. Respiratory stability characteristics can be characterized by calculating the variability of respiratory cycles or amplitudes (such as standard deviation, coefficient of variation, frequency domain indices, etc.), and are calculated based on the EDR of the ECG-derived respiratory sequence.

[0075] The construction of ECG-derived respiratory sequence (EDR) utilizes the modulation effect of respiration on ECG waveform amplitude, morphology, and baseline changes to form a proxy signal related to respiratory activity. The respiratory frequency (RespR) is obtained through spectral analysis, and respiratory rhythm changes are quantified through respiratory stability characteristics, providing additional physiological constraint information for metabolic intensity estimation.

[0076] This embodiment introduces ECG-derived respiratory sequence EDR and respiratory rate RespR to enhance the physiological constraints on metabolic intensity estimation and improve stability under dynamic activity and intensity abrupt changes.

[0077] S3. Input the beat-by-beat heart rate characteristics, heart rate variability characteristics, ECG-derived respiratory rate, respiratory stability characteristics, individual parameters, and signal quality indicators into the trained regression model, and output the estimated metabolic intensity.

[0078] In an optional embodiment of the present invention, step S3 takes the heart rate, HRV characteristics, RespR, individual parameters, and SQI as inputs, and outputs metabolic intensity through a regression model. Metabolic intensity is defined as oxygen consumption (VO2) or metabolic equivalent (MET(t)).

[0079] Metabolic intensity estimation is achieved using a neural network regression model, with the model input vector defined as:

[0080] ;

[0081] The regression model outputs an initial value of metabolic intensity M0(t), in the form of metabolic equivalent MET0(t) or oxygen consumption VO2(t). When the output is a metabolic equivalent, the metabolic equivalent and oxygen consumption satisfy the following:

[0082] ;

[0083] in, The unit is ml / kg / min, and 3.5 is the baseline value for resting oxygen consumption, in ml / kg / min.

[0084] In an optional embodiment of the present invention, to accommodate the differences in metabolic kinetics between the start-up and recovery phases of exercise, step S3, when estimating metabolic intensity, further includes start-stop kinetic compensation, specifically:

[0085] Obtain the initial value of metabolic intensity output by the regression model;

[0086] The exercise is determined to be in the starting or stopping / recovering state based on whether the current rate of change of heart rate and rate of change of respiratory rate both exceed a preset threshold.

[0087] Based on the discrimination results, the corresponding adaptive coefficient is selected, and the metabolic intensity is dynamically updated according to the following formula:

[0088] ;

[0089] in, The metabolic intensity is updated at time t. For adaptive coefficients, This is the initial value for metabolic intensity. This is the sampling update interval. When the state is determined to be in motion start-up, When it is determined that the movement has stopped or resumed, command ,and The start-up and stop / recovery states of exercise are determined based on the rate of change of heart rate and the rate of change of respiratory rate: when both the rate of change of heart rate and the rate of change of respiratory rate are greater than a preset threshold, it is determined to be the start-up state of exercise; otherwise, it is determined to be the stop / recovery state of exercise.

[0090] This embodiment uses a start-stop dynamics compensation mechanism to provide differentiated responses during the motion start-up and recovery phases, thereby improving the dynamic deviations caused by the steady-state assumption.

[0091] In an optional embodiment of the present invention, step S3, when estimating metabolic intensity, further includes performing extended Kalman filtering, specifically:

[0092] The output of the regression model is used as an observation, and based on the state equation... and observation equations Construct an extended Kalman filter, where As a state variable of metabolic intensity, This is the state transition function. For process noise, For observation purposes, For the observation function, To observe noise;

[0093] The metabolic intensity is fused and estimated using the extended Kalman filter, and the filtered output is taken as the final metabolic intensity. The function is used to describe the state evolution of metabolic intensity in adjacent time steps. This describes the observation generation mechanism. The extended Kalman filter output is used as the final metabolic intensity. It participates in energy consumption calculation and cumulative integration.

[0094] To enable the filtering process to adapt to signal quality, this embodiment utilizes signal quality indicators. Adjusting observation noise: When signal quality indicators When the signal quality index is less than a preset threshold, increase the observation noise or decrease the observation weight to weaken the impact of low-quality observations on state estimation; when the signal quality index is less than a preset threshold, increase the observation noise or decrease the observation weight to weaken the impact of low-quality observations on state estimation. When the noise level is not less than the preset threshold, the default observation noise setting is restored, thereby improving output stability while ensuring response speed.

[0095] This embodiment improves robustness under conditions of motion artifacts, poor contact, and abrupt changes in intensity by fusing the regression model output with the dynamic update results through extended Kalman filtering.

[0096] S4. Calculate the instantaneous energy consumption rate based on the metabolic intensity, and integrate the instantaneous energy consumption rate over time to obtain the cumulative energy consumption.

[0097] In an optional embodiment of the present invention, step S4, calculating the instantaneous energy consumption rate based on the metabolic intensity, includes:

[0098] Based on the metabolic intensity and oxygen consumption calorific value, the instantaneous energy consumption rate is calculated as follows:

[0099] ;

[0100] in, Let be the instantaneous energy consumption rate at time t. Metabolic intensity, It is the oxygen consumption heat equivalent.

[0101] This embodiment obtains cumulative energy consumption by integrating instantaneous energy consumption rate, which is suitable for training evaluation, weight management and long-term health monitoring, and has continuous output capability.

[0102] This embodiment is based on signal quality indicators. Perform energy consumption output control when When the intensity of frozen metabolism is less than a preset threshold, Metabolic intensity at the previous moment Furthermore, the metabolic intensity after freezing is used in the calculation of cumulative energy consumption, thereby reducing the impact of low-quality fragments on the cumulative results.

[0103] This embodiment reduces the impact of noise segments such as motion artifacts and poor contact on the output through SQI quality control and freezing strategies, thereby improving robustness in wearable scenarios.

[0104] This invention utilizes continuous ECG signals acquired by wearable devices to generate a Signal Quality Index (SQI) through bandpass filtering, power frequency suppression, and motion artifact detection. Based on this, R-peak detection is performed, and beat-by-beat RR intervals and heart rate are calculated. Heart rate variability features are further extracted within a sliding window. Simultaneously, an ECG-derived respiratory EDR sequence is constructed, and respiratory rate (RespR) and respiratory stability features are obtained through spectral analysis. Heart rate, HRV features, RespR, individual parameters, and SQI are input into a regression model to output metabolic intensity. Start-stop kinetic compensation is used to improve the dynamic response capability during exercise initiation and recovery phases, and extended Kalman filtering is incorporated to enhance robustness when necessary. Finally, instantaneous energy consumption rate is calculated based on oxygen consumption caloric equivalent and integrated to obtain cumulative energy consumption. An SQI freezing strategy is used to reduce the impact of low-quality segments on the cumulative results. This invention has advantages such as continuous estimation, resistance to artifact interference, and strong dynamic adaptability, and can be applied to wearable health monitoring, exercise training assessment, weight management, and long-term chronic disease management.

[0105] Example 2

[0106] This invention also provides an energy consumption estimation device based on electrocardiogram (ECG) signals, which applies an energy consumption estimation method based on ECG signals as described in Embodiment 1, including:

[0107] The preprocessing module is used to acquire continuous ECG signals and preprocess the ECG signals to generate signal quality indicators that characterize the usability of the signals.

[0108] The feature extraction module is used to extract beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features based on the preprocessed ECG signal.

[0109] The metabolic intensity estimation module is used to input the beat-by-beat heart rate characteristics, heart rate variability characteristics, ECG-derived respiratory rate, respiratory stability characteristics, individual parameters, and signal quality indicators into the trained regression model and output the estimated metabolic intensity.

[0110] The energy calculation module is used to calculate the instantaneous energy consumption rate based on the metabolic intensity, and to integrate the instantaneous energy consumption rate over time to obtain the cumulative energy consumption.

[0111] This embodiment may also include a wireless communication module, which sends the energy consumption estimation results to a mobile terminal or cloud server via Bluetooth or Wi-Fi for visualization and long-term data management.

[0112] Example 3

[0113] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements an energy consumption estimation method based on electrocardiogram signals as described in Embodiment 1.

[0114] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0115] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0116] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0117] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

[0118] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A method for estimating energy consumption based on electrocardiogram (ECG) signals, characterized in that, Includes the following steps: Continuous ECG signals are acquired and preprocessed to generate signal quality indicators characterizing signal availability. Based on the preprocessed ECG signal, the beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features were extracted. The beat-by-beat heart rate characteristics, heart rate variability characteristics, ECG-derived respiratory rate, respiratory stability characteristics, individual parameters, and signal quality indicators are input into the trained regression model, which outputs an estimated metabolic intensity. The instantaneous energy consumption rate is calculated based on the metabolic intensity, and the instantaneous energy consumption rate is integrated over time to obtain the cumulative energy consumption.

2. The energy consumption estimation method based on electrocardiogram signals according to claim 1, characterized in that, Based on the preprocessed ECG signal, beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features were extracted, including: R-peak detection is performed on the ECG signal to obtain the R-peak time series, and the beat-by-beat interval and beat-by-beat heart rate are calculated based on the R-peak time series. Heart rate variability characteristics are calculated based on the beat-by-beat interval within a sliding window; The ECG-derived respiratory signal is constructed by extracting the amplitude of the R peak in the ECG signal and passing it through a low-pass filter. The ECG-derived respiratory signal is then subjected to spectral analysis to obtain the ECG-derived respiratory frequency, and respiratory stability characteristics are calculated.

3. The energy consumption estimation method based on electrocardiogram signals according to claim 1, characterized in that, The individual parameters include at least one or any combination of age, gender, height, and weight.

4. The energy consumption estimation method based on electrocardiogram signals according to claim 1, characterized in that, Estimating metabolic intensity also includes compensation for start-stop kinetics, specifically: Obtain the initial value of metabolic intensity output by the regression model; The exercise is determined to be in the starting or stopping / recovering state based on whether the current rate of change of heart rate and rate of change of respiratory rate both exceed a preset threshold. Based on the discrimination results, the corresponding adaptive coefficient is selected, and the metabolic intensity is dynamically updated according to the following formula: ； in, The metabolic intensity is updated at time t. For adaptive coefficients, This is the initial value for metabolic intensity. This is the sampling update interval.

5. The energy consumption estimation method based on electrocardiogram signals according to claim 1, characterized in that, Estimating metabolic intensity also includes performing extended Kalman filtering, specifically: The output of the regression model is used as an observation, and based on the state equation... and observation equations Construct an extended Kalman filter, where As a state variable of metabolic intensity, This is the state transition function. For process noise, For observation purposes, For the observation function, To observe noise; The metabolic intensity is estimated by fusion using the extended Kalman filter, and the filtered output is used as the final metabolic intensity.

6. The energy consumption estimation method based on electrocardiogram signals according to claim 5, characterized in that, The weight or magnitude of the observation noise in the extended Kalman filter is adjusted according to the signal quality index. When the signal quality index is lower than a preset quality threshold, the observation noise is increased or its weight is decreased.

7. The energy consumption estimation method based on electrocardiogram signals according to claim 1, characterized in that, The calculation of instantaneous energy consumption rate based on the metabolic intensity also includes signal quality control, specifically: If the signal quality index at the current moment is lower than the preset quality threshold, the metabolic intensity used to calculate the current instantaneous energy consumption rate will be frozen to the metabolic intensity of the previous effective moment.

8. The energy consumption estimation method based on electrocardiogram signals according to claim 1, characterized in that, The instantaneous energy consumption rate is calculated based on the metabolic intensity, including: Based on the metabolic intensity and oxygen consumption calorific value, the instantaneous energy consumption rate is calculated as follows: ； in, Let be the instantaneous energy consumption rate at time t. Metabolic intensity, It is the oxygen consumption heat equivalent.

9. An energy consumption estimation device based on electrocardiogram (ECG) signals, employing an energy consumption estimation method based on ECG signals as described in any one of claims 1 to 8, characterized in that, include: The preprocessing module is used to acquire continuous ECG signals and preprocess the ECG signals to generate signal quality indicators that characterize the usability of the signals. The feature extraction module is used to extract beat-by-beat heart rate features, heart rate variability features, ECG-derived respiratory rate, and respiratory stability features based on the preprocessed ECG signal. The metabolic intensity estimation module is used to input the beat-by-beat heart rate characteristics, heart rate variability characteristics, ECG-derived respiratory rate, respiratory stability characteristics, individual parameters, and signal quality indicators into the trained regression model and output the estimated metabolic intensity. The energy calculation module is used to calculate the instantaneous energy consumption rate based on the metabolic intensity, and to integrate the instantaneous energy consumption rate over time to obtain the cumulative energy consumption.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements an energy consumption estimation method based on electrocardiogram signals as described in any one of claims 1 to 8.

Images