CN114587298B

CN114587298B - A method and system for detecting and separating physiological information from multiple human bodies

Info

Publication number: CN114587298B
Application number: CN202210221729.6A
Authority: CN
Inventors: 曹自平; 吴强
Original assignee: Nanjing University of Posts and Telecommunications
Current assignee: Nanjing University of Posts and Telecommunications
Priority date: 2022-03-09
Filing date: 2022-03-09
Publication date: 2025-11-04
Anticipated expiration: 2042-03-09
Also published as: CN114587298A

Abstract

This invention discloses a method and system for detecting and separating physiological information from multiple human bodies, comprising: acquiring mixed signals during human movement; filtering and screening all signals in the acquired signals to separate gait signals from the mixed signals; performing time-domain segmentation on the signals after deducting the gait signals; using the segmented mixed signal data as input to a neural network model, adjusting parameters to train the model, comparing the performance of models with different parameters, and saving the optimal model; using the trained model to separate respiratory signals and heart rate signals, and calculating the frequencies of the three signals. This invention is relatively simple in terms of equipment and imposes less constraint on the human body; the sensors are also characterized by small size, light weight, and low power consumption. The model training complexity of this invention is low; it reduces information loss, increases model robustness, better avoids overfitting, and improves the network's generalization ability.

Description

Multi-human physiological information detection-separation method and system

Technical Field

The invention relates to the technical field of human physiological signal separation, in particular to a method and a system for detecting and separating multiple human physiological information.

Background

Heart rate is an important parameter for observing health status of vital signs, and detection of heart rate in medicine can be carried out through electrocardiograph, electrocardiograph and the like, but the measuring equipment is large in size and has certain binding property to people during use, so that convenience is poor, on the other hand, compared with medical equipment, intelligent wrist rings, wrist bands and the like, although the intelligent wrist rings and wrist bands are very good in use convenience, sensors arranged in the equipment are difficult to stably acquire information at the same position due to large movement displacement of bones and muscles of wrist joints, and therefore distortion degree of output physiological signals of the equipment is high. Compared with the two types of equipment, the intelligent chest belt is equipment which can simultaneously give consideration to operation convenience and data accuracy, the sensing device arranged in the intelligent chest belt can be used for measuring body temperature by a temperature sensor, conducting electrodes and heart parameters and photoelectric or acceleration sensors, so that the intelligent chest belt has a good application prospect.

The intelligent chest belt provided with the acceleration sensor has the characteristics of low cost, small volume and low power consumption, but due to the extremely high measurement sensitivity, the physiological signals obtained by the intelligent chest belt are often mixed with breathing rhythm and gait information, and how to obtain accurate heart rate data through a proper signal processing method is a challenging task.

Disclosure of Invention

This section is intended to outline some aspects of embodiments of the application and to briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section as well as in the description of the application and in the title of the application, which may not be used to limit the scope of the application.

The present invention has been made in view of the above-described problems occurring in the prior art.

The technical problem solved by the invention is therefore how to obtain accurate heart rate data by means of a suitable signal processing method.

The invention provides a technical scheme for solving the technical problems, which comprises the steps of collecting mixed signals during human body movement, screening and filtering all signals in the collected signals, separating gait signals in the mixed signals, carrying out time domain segmentation on the signals after subtraction of the gait signals, taking data of the segmented mixed signals as input of a neural network model, storing an optimal model by adjusting parameters and comparing indexes of different parameter models, separating respiratory signals and heart rate signals by using the model after training, and calculating frequencies of three signals.

As a preferable scheme of the multi-human physiological information detection-separation method, a band-pass filter is utilized to extract gait signals in the mixed signals.

The optimal scheme of the multi-human physiological information detection-separation method comprises the steps of deleting bad points of all acquired data, and filtering the respiration and heart rate signals, wherein a digital filter is adopted in the filtering, weak heart rate signals and body noise in the respiration signals are eliminated through a Butterworth low-pass filter, and baseline drift of the heart rate signals is eliminated through a Butterworth Wo Sigao-pass filter.

The method for detecting and separating the physiological information of the multiple human bodies comprises the steps of adopting a Conv-Tasnet deep learning network model to separate signals, taking data subjected to data preprocessing as training data of the model, dividing the model in a time domain, adjusting parameters of the model, setting learning rate of the model, carrying out counter propagation through a loss value calculated by forward propagation, modifying weights, if the loss value reaches a preset index, storing the model, otherwise, continuously judging whether the maximum iteration number is reached, if the maximum iteration number is reached, storing the model, otherwise, continuously adjusting the parameters.

As a preferable scheme of the multi-human physiological information detection-separation method, the Conv-Tasnet deep learning network model is characterized in that a Bahdanau attention mechanism is added in a separator, a random removal unit dropout is added after the separator, and a gradient descent method adopted by the Conv-Tasnet deep learning network model is AdamW +SAM.

The optimal separation model is selected according to the optimal model obtained by comparing the performances of models trained by different parameters, and comprises two indexes of improving and maximizing the signal-to-noise ratio with unchanged scale, and the optimal model with improved or maximized signal-to-noise ratio with unchanged scale is selected as the signal separation model.

The multi-human physiological information detection-separation method is characterized in that for the heart rate signals, an interval maximum value algorithm is adopted to obtain each peak value of the heart rate signals, heart rate is calculated through a signal frequency calculation formula, for the respiratory signals and gait signals, high-frequency signals are filtered through a low-pass filter, the number of peaks in an interval is found through a peak extraction method, and respiratory frequency and step frequency are calculated through a signal frequency calculation formula.

As a preferable scheme of the multi-human physiological information detection-separation method, the signal frequency calculation formula is as follows:

Where f _s is the sampling frequency, X _start is the position of the first peak, X _end is the position of the last peak, and n is the number of peaks in this interval.

The invention further provides a multi-human physiological information detection-separation system which comprises a signal collector, a data preprocessing subsystem and a signal separation subsystem, wherein the signal collector comprises an acceleration sensor, a microcontroller module, an analog-to-digital conversion unit and a wireless transmission module, the acceleration sensor, the microcontroller module, the analog-to-digital conversion unit and the wireless transmission module are connected in a wired or wireless mode and used for collecting mixed signals generated during human body movement, the data preprocessing subsystem is connected with the signal collector and comprises a wireless receiving module, a screening and filtering module and a gait separation module and used for preprocessing collected signals, and the signal separation subsystem is connected with the data preprocessing subsystem and comprises a model training module, a signal separation module and a frequency calculation module and used for constructing a model and separating signals.

The invention has the advantages of simpler equipment, smaller constraint on human body, small volume, light weight, low power consumption and the like of the sensor, lower complexity of model training and smaller model, reduces information loss, increases the robustness of the model, can better avoid overfitting, and improves the generalization capability of the network.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. Wherein:

FIG. 1 is a schematic diagram of a basic flow of a method and a system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention;

FIG. 2 is a schematic flow chart of a signal analysis algorithm of a method and a system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention;

FIG. 3 is a schematic wearing diagram of a signal collector of a method and a system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention;

FIG. 4 is a schematic workflow diagram of a data preprocessing subsystem of a method and system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of a model training process of a method and a system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention;

FIG. 6 is a schematic signal diagram of a signal separation process of a method and a system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention;

Fig. 7 is a schematic block diagram of a method and a system for detecting and separating physiological information of multiple human bodies according to an embodiment of the present invention.

Detailed Description

So that the manner in which the above recited objects, features and advantages of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to the embodiments, some of which are illustrated in the appended drawings. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways other than those described herein, and persons skilled in the art will readily appreciate that the present invention is not limited to the specific embodiments disclosed below.

Further, reference herein to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic can be included in at least one implementation of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments.

While the embodiments of the present invention have been illustrated and described in detail in the drawings, the cross-sectional view of the device structure is not to scale in the general sense for ease of illustration, and the drawings are merely exemplary and should not be construed as limiting the scope of the invention. In addition, the three-dimensional dimensions of length, width and depth should be included in actual fabrication.

Also in the description of the present invention, it should be noted that the orientation or positional relationship indicated by the terms "upper, lower, inner and outer", etc. are based on the orientation or positional relationship shown in the drawings, are merely for convenience of describing the present invention and simplifying the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.

The terms "mounted, connected, and coupled" as used herein, unless otherwise specifically indicated and defined, shall be construed broadly and include, for example, fixed, removable, or integral, as well as mechanical, electrical, or direct, as well as indirect via intermediaries, or communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.

Example 1

Referring to fig. 1 to 6, for an embodiment of the present invention, a method for detecting and separating physiological information of multiple human bodies is provided, including:

s1, collecting a mixed signal during human body movement.

It should be noted that, the signal collector is used to collect the mixed signal of human body during exercise, wherein, as shown in fig. 3, the signal collector is fixed on the elastic chest belt, when the physiological signal is collected through the design experiment, the signal collector is placed at the left chest of the human body, and the collected signal is the heart rate, respiration, gait of the subject and the mixed signal of the three signals.

S2, separating gait signals in the mixed signals, and screening and filtering all signals in the mixed signals.

It should be noted that, as shown in fig. 4, all the acquired data are deleted, and the respiration and heart rate signals are filtered, wherein the filtering process adopts a digital filter method, namely, a butterworth low-pass filter is used for eliminating weak heart rate signals and body noise in the respiration signals, a butterworth Wo Sigao-pass filter is used for eliminating baseline drift of the heart rate signals, and a band-pass filter is used for extracting gait signals in the mixed signals.

S3, performing time domain segmentation on the gait signal subtracted signal, taking the data of the segmented mixed signal as the input of a neural network model, and storing an optimal model by adjusting parameters and comparing indexes of different parameter models.

It should be noted that, as shown in fig. 2 and 5, the neural network model preferably adopts a Conv-Tasnet deep learning network model, and the main structure of the neural network model comprises an encoder, a separator and a decoder, the training and generating process comprises the steps of taking the data processed by the data preprocessing subsystem as original training data, dividing the data in the time domain, adjusting the parameters of the model, setting the learning rate of the model, carrying out back propagation through a loss value calculated by forward propagation, modifying the weight, if the loss value reaches an expected index, storing the model, otherwise, continuously judging whether the maximum iteration number is reached, if the maximum iteration number is reached, storing the model, otherwise, continuously adjusting the parameters.

Furthermore, the Conv-Tasnet deep learning network model is characterized in that a Bahdanau attention mechanism is added in a separator to reduce information loss, a random removal unit dropout is added after the separator to avoid overfitting and increase robustness of the model, a gradient descent method adopted by the Conv-Tasnet deep learning network model is AdamW +SAM, adamW is an improved algorithm based on Adam+L2 regularization, and specifically, defects that Adam has larger weight parameters but punishment is not larger are overcome, a sharpness perception minimization (SAM) reduces loss values and loss sharpness, parameters with uniformly lost values are searched in the field of the separator, and generalization capability of the network is improved.

Furthermore, the selection standard of the optimal separation model is an optimal model obtained by comparing the performances of models trained by different parameters, the optimal model comprises two indexes, namely, signal-to-noise ratio improvement (SDRi) and maximization scale (SI-SNR) invariable signal-to-noise ratio, and the optimal model with the signal-to-noise ratio improvement or maximization scale invariable signal-to-noise ratio is selected as the signal separation model.

And S4, separating a respiratory signal and a heart rate signal by using the trained model, and calculating the frequencies of the three signals.

For the heart rate signals, a maximum interval algorithm is adopted to obtain each peak value of the heart rate signals, the heart rate is calculated through a signal frequency calculation formula, for the respiratory signals and the gait signals, the high-frequency signals are filtered through a low-pass filter, the number of the peaks in the interval is found through a peak extraction method, and the respiratory frequency and the step frequency are calculated through the signal frequency calculation formula.

Preferably, the signal frequency calculation formula is:

The physiological signal acquisition device comprises an elastic chest belt and an acceleration sensor, is simple in equipment and has the characteristics of small size, light weight, low power consumption and the like, the sensor also has the characteristics of separating respiratory signals, heart rate signals and gait signals under the dynamic condition, extracting the gait signals firstly, separating the respiratory signals and the heart rate signals through a separation model, reducing the complexity of model training and reducing the size of the model, the deep learning model is Conv-Tasnet, a Bahdanau attention mechanism is added in front of a decoder, the loss of information is reduced, a dropout layer is added behind the separator, the robustness of the model is increased, overfitting can be better avoided, and the gradient descent method is AdamW +SAM which can search parameters of uniformly lost values in the field and improve the generalization capability of the network.

The technical effect adopted in the method is verified and explained, and the embodiment adopts the method for testing, and the actual effect of the method is verified by means of scientific demonstration.

The method comprises the steps that an experimenter needs to collect heart rate and respiratory signals under static conditions, the speed is increased by 2.5km/h step by step on a running machine, the speed is increased by 0.5km/h each time, the respiratory signals, heart rate signals, gait signals and mixed signals of the three signals of the experimenter are collected, the data are sent to a data preprocessing subsystem by a signal collector, and after screening and filtering are carried out on the data, a model with the optimal signal-to-noise ratio and the unchanged maximum scale is obtained through a training module.

FIG. 1 is a flow chart of a multi-body signal separation work, as shown in FIG. 6, wherein the signal represents an original mixed signal acquired by a signal acquisition device, the original acceleration mixed signal is filtered by a digital band-pass filter to obtain a gait signal, as shown in FIG. 6, the mixed signal after subtraction of the gait signal is transmitted to a signal separation module, as shown in FIG. 6, the heart rate and the respiratory signal are separated by a Conv-Tasnet model, and finally the frequencies of three signals are calculated by adopting a method and a formula of the following method, namely, for the heart rate signal, an algorithm of a maximum value of a section is adopted to find each peak value of the heart rate signal, and then the heart rate is calculated by the formula, for the respiratory and the gait signals, a high-frequency signal is filtered by a low-pass filter, the number of peaks in the section is found by a peak extraction method, and then the respiratory frequency and the step frequency are calculated by the formula. For the calculation of the three signal frequencies, the following formula is used:

According to the human body multi-physiological signal separation method and system based on acceleration, which are provided by the invention, the flow of collecting physiological signals is relatively accurate, the realization is relatively simple, the constraint with wearable equipment is small and convenient, and the performance evaluation of the signal separation result is more accurate.

Example 2

Referring to fig. 7, another embodiment of the present invention, which is different from the first embodiment, provides a multi-human physiological information detection-separation system, and the multi-human physiological information detection-separation method is implemented by the system, and specifically includes:

The signal collector comprises an acceleration sensor, a microcontroller module, an analog-to-digital conversion unit and a wireless transmission module, wherein the signal collector is connected in a wired or wireless mode and is used for collecting mixed signals generated during human body movement, preferably, the acceleration sensor is used for collecting human body physiological information, the microcontroller module is used for filtering high-frequency noise and controlling the transmission of signals, the digital-to-analog conversion module is used for converting analog signals into digital signals, and the wireless transmission module is used for transmitting the digital signals to the data preprocessing subsystem.

The data preprocessing subsystem is connected with the signal acquisition device and comprises a wireless receiving module, a screening and filtering module and a gait separation module, wherein the wireless receiving module is used for preprocessing acquired signals, preferably, the wireless receiving module is used for receiving data transmitted by the signal acquisition device, the screening and filtering module is used for screening bad points of the acquired multiple physiological signals and filtering interference noise in the signals, and the gait separation module is used for separating gait signals from mixed signals of heart rate, respiration and gait.

The signal separation subsystem is connected with the data preprocessing subsystem and comprises a model training module, a signal separation module and a frequency calculation module, wherein the model training module is used for constructing a model and separating signals, and preferably, the model training module is used for obtaining an optimal signal separation model through training, the signal separation module is used for separating heart rate and respiratory signals from a mixed signal with gait signals subtracted, and the frequency calculation module is used for calculating step frequency, heart rate and respiratory frequency through the separated signals.

It should be appreciated that embodiments of the invention may be implemented or realized by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer readable memory. The methods may be implemented in a computer program using standard programming techniques, including a non-transitory computer readable storage medium configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner, in accordance with the methods and drawings described in the specific embodiments. Each program may be implemented in a high level procedural or object oriented programming language to communicate with a computer system. However, the program(s) can be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language. Furthermore, the program can be run on a programmed application specific integrated circuit for this purpose.

Furthermore, the operations of the processes described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes (or variations and/or combinations thereof) described herein may be performed under control of one or more computer systems configured with executable instructions, and may be implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications), by hardware, or combinations thereof, collectively executing on one or more processors. The computer program includes a plurality of instructions executable by one or more processors.

Further, the method may be implemented in any type of computing platform operatively connected to a suitable computing platform, including, but not limited to, a personal computer, mini-computer, mainframe, workstation, network or distributed computing environment, separate or integrated computer platform, or in communication with a charged particle tool or other imaging device, and so forth. Aspects of the invention may be implemented in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and/or write storage medium, RAM, ROM, etc., such that it is readable by a programmable computer, which when read by a computer, is operable to configure and operate the computer to perform the processes described herein. Further, the machine readable code, or portions thereof, may be transmitted over a wired or wireless network. When such media includes instructions or programs that, in conjunction with a microprocessor or other data processor, implement the steps described above, the invention described herein includes these and other different types of non-transitory computer-readable storage media. The invention also includes the computer itself when programmed according to the methods and techniques of the present invention. The computer program can be applied to the input data to perform the functions described herein, thereby converting the input data to generate output data that is stored to the non-volatile memory. The output information may also be applied to one or more output devices such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects produced on a display.

As used in this disclosure, the terms "component," "module," "system," and the like are intended to refer to a computer-related entity, either hardware, firmware, a combination of hardware and software, or software in execution. For example, a component may be, but is not limited to being, a process running on a processor, an object, an executable, a thread of execution, a program, and/or a computer. By way of example, both an application running on a computing device and the computing device can be a component. One or more components may reside within a process and/or thread of execution and a component may be localized on one computer and/or distributed between two or more computers. Furthermore, these components can execute from various computer readable media having various data structures thereon. The components may communicate by way of local and/or remote processes such as in accordance with a signal having one or more data packets (e.g., data from one component interacting with another component in a local system, distributed system, and/or across a network such as the internet with other systems by way of the signal).

It should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention may be modified or substituted without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered in the scope of the claims of the present invention.

Claims

1. A multi-human physiological information detection-separation method, comprising:

collecting a mixed signal when a human body moves;

screening and filtering all signals in the acquired signals, and separating gait signals in the mixed signals;

Performing time domain segmentation on the gait signal subtracted signal, taking the data of the segmented mixed signal as the input of a neural network model, and storing an optimal model by adjusting parameters and comparing indexes of different parameter models;

separating a respiratory signal and a heart rate signal by using the model after training, and calculating the frequencies of the three signals;

extracting gait signals in the mixed signal by using a band-pass filter;

The method for separating the signals by adopting Conv-Tasnet deep learning network model comprises the following steps:

Taking the data after data preprocessing as training data of a model, and carrying out time domain segmentation on the training data;

parameters of the model are adjusted, learning rate of the model is set, back propagation is carried out through a loss value calculated by forward propagation, and weight is modified;

If the loss value reaches the preset index, saving the model, otherwise, continuing to judge whether the maximum iteration number is reached, if the maximum iteration number is reached, saving the model, otherwise, continuing to adjust the parameters;

The Conv-Tasnet deep learning network model is characterized in that a Bahdanau attention mechanism is added in a separator, a random removal unit dropout is added after the separator, and a gradient descent method adopted by the Conv-Tasnet deep learning network model is AdamW +SAM.

2. The method for detecting and separating physiological information of multiple human bodies according to claim 1, further comprising the steps of deleting dead spots from all acquired data and filtering the respiration and heart rate signals, wherein a digital filter is adopted in the filtering process, weak heart rate signals and body noise in the respiration signals are eliminated through a Butterworth low-pass filter, and baseline drift of the heart rate signals is eliminated through a Butterworth Wo Sigao-pass filter.

3. The multi-human physiological information detection-separation method according to claim 2, wherein the selection criteria of the optimal separation model is an optimal model obtained by comparing performance between models trained by different parameters, the optimal model comprises two indexes of improving signal to noise ratio and maximizing scale-invariant signal to noise ratio, and the optimal model of improving signal to noise ratio or maximizing scale-invariant signal to noise ratio is selected as the signal separation model.

4. The method for detecting and separating physiological information of multiple human bodies according to claim 3, further comprising obtaining each peak value of the heart rate signal by using a interval maximum algorithm for the heart rate signal, calculating the heart rate by using a signal frequency calculation formula, filtering high-frequency signals by using a low-pass filter for the respiratory signal and the gait signal, finding the number of peaks in an interval by using a peak extraction method, and calculating the respiratory frequency and the step frequency by using a signal frequency calculation formula.

5. The multi-body physiological information detection-separation method according to claim 4, wherein the signal frequency calculation formula is:

wherein, the Is the sampling frequency at which the sample is to be taken,Is the location of the first peak,Is the location of the last peak value,Is the number of peaks in this interval.

6. A multi-human physiological information detection-separation system, applying the multi-human physiological information detection-separation method according to any one of claims 1 to 5, comprising:

The signal collector comprises an acceleration sensor, a microcontroller module, an analog-to-digital conversion unit and a wireless transmission module, and is connected in a wireless mode and used for collecting mixed signals during human body movement;

the data preprocessing subsystem is connected with the signal collector and comprises a wireless receiving module, a screening and filtering module and a gait separation module, and is used for preprocessing the collected signals;

The signal separation subsystem is connected with the data preprocessing subsystem and comprises a model training module, a signal separation module and a frequency calculation module, and is used for constructing a model and separating signals.