Sweat amount monitoring system and method based on RFID passive multi-frequency perception

Abstract

The invention discloses a sweat amount monitoring system and method based on RFID passive multi-frequency sensing, and the system comprises a passive sweat amount sensor which is used for receiving a wireless radio frequency signal transmitted by a sweat amount sensing all-in-one machine, modulating the received wireless radio frequency signal, and transmitting the modulated wireless radio frequency signal to the sweat amount sensing all-in-one machine; the sweat amount sensing all-in-one machine is used for generating a multi-frequency wireless radio frequency signal to activate the passive sweat amount sensor, collecting a wireless radio frequency signal backscattered by the passive sweat amount sensor in real time and sending the wireless radio frequency signal to the data processing platform; and the data processing platform is used for acquiring the phase data and the RSSI feature data sent by the sweat amount sensing all-in-one machine in real time, extracting features related to the sweat amount by using a multi-frequency signal difference method, and predicting the sweat amount to obtain the sweat amount. According to the method, RFID passive sensing is adopted, a cheap RFID tag is adopted, a battery is not needed for energy supply, and accurate sweat amount monitoring is achieved by analyzing changes of RFID signals.

Description

Technical Field

[0001] The present invention belongs to the technical field of RFID passive sensing, and in particular relates to a sweat volume monitoring system and method based on RFID passive multi-frequency sensing. Background Art

[0002] With the development of health monitoring technology, real-time monitoring systems based on biological signals are increasingly used in personal health management and medical diagnosis. As an important physiological indicator, human sweat volume can reflect an individual's metabolic level, fluid balance, and potential health problems, especially in high temperature environments, during exercise, and in the monitoring of specific diseases. However, current sweat-related research and technological applications are mainly focused on the detection of sweat components, such as the analysis of electrolytes, glucose, lactate, and pH in sweat. Although this information can reveal a specific metabolic state, it often ignores the basic but key health indicator of sweat volume.

[0003] Existing research and technology for sweat volume detection is relatively limited. Even though a small number of studies have attempted to develop solutions for monitoring sweat volume, there are still significant deficiencies:

[0004] (1) Reliance on carefully designed dedicated sensors: These sensors usually require complex manufacturing processes and precise calibration, which results in high equipment costs and is not suitable for large-scale applications.

[0005] (2) Energy limitation: Most existing work relies on active devices that require regular charging or battery replacement. Especially when used for a long time or monitored at a high frequency, the energy supply problem becomes a major bottleneck.

[0006] Therefore, based on the above considerations, there is an urgent need for a low-cost, simple and passive technology that can overcome the limitations of traditional methods and provide a new means of sweat monitoring for daily health management. Summary of the invention

[0007] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a sweat volume monitoring system and method based on RFID passive multi-frequency sensing, so as to solve the problems of short battery life and high cost of relying on active dedicated sensors in the existing sweat volume detection technology; the present invention adopts RFID passive sensing, uses cheap RFID tags, and does not require battery power, and realizes accurate sweat volume monitoring by analyzing the changes in RFID signals.

[0008] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0009] The sweat volume monitoring system based on RFID passive multi-frequency sensing of the present invention comprises: a passive sweat volume sensor, a sweat volume sensing integrated machine and a data processing platform; wherein:

[0010] A passive sweat volume sensor, used to receive the wireless radio frequency signal transmitted by the sweat volume sensing integrated machine, and modulate the received wireless radio frequency signal and send it to the sweat volume sensing integrated machine;

[0011] The sweat volume sensing integrated machine is used to generate a multi-frequency wireless radio frequency signal to activate the passive sweat volume sensor, and collect the wireless radio frequency signal sent by the passive sweat volume sensor in real time and send it to the data processing platform;

[0012] The data processing platform obtains the phase data and RSSI characteristic data in the wireless radio frequency signal sent by the sweat volume sensing integrated device in real time, and uses the multi-frequency signal difference method to extract the features related to the sweat volume, predict the sweat volume, and obtain the sweat volume.

[0013] Furthermore, the passive sweat amount sensor includes: an RFID tag pair and a sponge block for storing sweat, the RFID tag pair includes a sensing tag and a reference tag, wherein the sensing tag is used to sense the sweat amount, and the reference tag is placed 1 cm near the sensing tag to eliminate dynamic interference; the sensing tag and the reference tag are respectively adhered to the front side of the sponge block with double-sided tape, and the back side of the sponge block is adhered to the sweat monitoring target position.

[0014] Furthermore, the sweat volume sensing integrated machine includes: an RFID antenna and an RFID reader; the RFID reader continuously transmits multi-frequency wireless radio frequency signals through the RFID antenna, and collects the wireless radio frequency signals backscattered by the passive sweat volume sensor in real time, and sends them to the data processing platform.

[0015] Furthermore, the RFID reader and RFID tag are of UHF specification with a frequency of 860-960 MHz, and the RFID reader reads and writes the RFID tag through the EPC Global C1 G2 protocol.

[0016] Furthermore, the multi-frequency wireless radio frequency signal is an OFDM symbol, and the specific generation steps are as follows:

[0017] Generate a random sequence S of length n, where each element has a probability of half being +1 or -1, and n is the custom OFDM symbol length;

[0018] Perform an inverse Fourier transform (IFFT) on the random sequence S to obtain the OFDM symbol x in the time domain. The OFDM symbol consists of n complex numbers, denoted by x = {c1, c2, …, c i ,…,c n}, where each complex number c i Indicates a specific frequency f i Signal;

[0019] A cyclic suffix CS of length Q is added to the end of the generated OFDM symbol x, where CS is the first Q elements of x, CS = {c1, c2, ..., c i ,…,c Q};

[0020] The OFDM symbol after adding CS is multiplied by the amplification factor α. The amplitude of the OFDM symbol is proportional to the energy, ensuring that the OFDM symbol has enough energy to activate the RFID tag. The value of α is 6.

[0021] Furthermore, during the EPC stage when the RFID signal is sent, the sweat volume sensing integrated machine replaces a single-frequency continuous wave (CW) with an OFDM symbol having multi-frequency information to activate the passive sweat volume sensor.

[0022] Furthermore, the method for the data processing platform to calculate the sweat volume result data is:

[0023] (1) Obtain the phase data of the reference tag and the sensing tag at different RF signal frequencies, which are recorded as in is the reference label at frequency f i The phase data below, is the perceived label at frequency f i Phase data below;

[0024] (2) Calculate the phase data W of the sensing tag s The phase data W of the reference tag r The difference W d , denoted as in is the frequency f i The phase data difference between the sensing tag and the reference tag;

[0025] (3) Calculate the difference W d The adjacent frequency f i With f i+1 The phase data difference between the two is expressed as the sweat volume feature X,

[0026] (4) Calculate all sweat volume characteristics within the measurement range;

[0027] (5) Train a regressor using all the sweat volume features in step (4);

[0028] (6) Use the trained regressor to predict sweat volume.

[0029] Furthermore, the step (1) specifically includes:

[0030] (11) According to the wireless radio frequency signal amplitude threshold A thresh and duration threshold T thresh , separate the EPC phase signal from the received wireless RF signal, where A thresh is 0.05, T thresh is 3;

[0031] (12) Subcarrier alignment is performed on the OFDM symbol x in the EPC phase signal, specifically including:

[0032] (121) Perform a fast Fourier transform (FFT) on x to obtain the frequency domain representation as follows:

[0033]

[0034] Where E is the subcarrier offset of the OFDM symbol, X n is the random sequence used to generate OFDM symbols, j is the imaginary unit, and N is the length of the OFDM symbol;

[0035] (122)Compared with the initial X n and The phase offset of the subcarrier is obtained by:

[0036] (123) Shift x by E units to obtain the OFDM symbol after subcarrier offset compensation alignment, denoted as x align ;

[0037] (13) x align The amplitude A x The OFDM symbol amplitude A in the T1 phase when the RFID signal is transmitted T For comparison, if A x =A T , then x align It is in OFF state, recorded as x off , otherwise it is in the ON state, recorded as x on ;

[0038] (14) According to the obtained x off and x on Extract the phase data w, the expression is as follows:

[0039]

[0040] Here, arg represents the angle at which the complex number is calculated.

[0041] Furthermore, the step (2) specifically includes:

[0042] Set the phase data of the perception tag to: (t s1 ,W s1 ),(t s2 ,W s2 ),…,(t sm ,W sm ), where t s1 ,t s2 ,…,t sm represents the time point of perceiving the label, W s1 ,W s2 ,…,W sm Indicates the corresponding phase value; the phase data of the reference label is:

[0043] (t r1 ,W r1 ),(t r2 ,W r2 ),…,(t rm ,W rm ), where t r1 ,t r2 ,…,t rm represents the time point of perceiving the label, W r1 ,W r2 ,…,W rm Represents the corresponding phase value; two time series t r and t s If they are inconsistent, the phase data is interpolated so that the phase data of the perception tag is mapped to the time point of the reference tag; linear interpolation is used to estimate the phase value of the perception tag at the time point of the reference tag, as follows:

[0044] (21) Determine the time point of the reference tag: The time point of the reference tag is t r1 ,t r2 ,…,t rm , indicating the time point you wish to align;

[0045] (22) Find the time interval of the perception label: For each reference time point t rk , find the perception label time series t s Two time points that satisfy the following relationship:

[0046] t sj ≤t rk ≤t sj+1 ,j∈[1,m-1]

[0047] Among them, t sj and t sj+1 For rk The two closest time points;

[0048] (23) According to the definition of linear interpolation, the perceived tag phase value W s (t rk ) at time point t rk The estimated values ​​at are as follows:

[0049]

[0050] Among them, W sj , W sj+1 are time points t sj and t sj+1 The perceived tag phase value at t sj ,t sj+1 are time points t rk The nearest time point;

[0051] (24) Loop through all reference time points: For each reference label time point t rk , both find the two most recent time points t in the time series of the perception label sj and t sj+1 , use the linear interpolation formula to calculate the corresponding perception tag phase value W s (t rk );

[0052] (25) After aligning the time points of the perception label and the reference label by interpolation, for each time point t i , calculate W d (t i )=W s (t i )-W r (t i ), all W d (t i ) value average, and get the difference W d , where W s (t i ) is the perception label at time point t i The phase value, W r (t i ) is the reference label at time point t i The phase value, W d (t i ) is the perceived label and the reference label at time point t i The phase value difference.

[0053] Furthermore, the step (5) specifically includes:

[0054] (51) The sweat volume feature X and the corresponding sweat volume Y are standardized to zero mean and unit variance, respectively, and are denoted as X std and Y std ,as follows:

[0055]

[0056] in, and δX are the mean and standard deviation of X, respectively. and δY are the mean and standard deviation of Y, respectively. std is the standardized sweat volume characteristic, Y std is the standardized sweat volume;

[0057] (52) Use the cross validation method to compare the performance of models with different numbers of components p and select the p with the smallest mean square error;

[0058] (53) Establish a PLSR model to project X and Y into a low-dimensional space and maximize the correlation between them. The specific steps are as follows:

[0059] (531) For each PLSR component t h (h-th latent variable) is extracted and modeled to obtain the weight matrix W, score matrix T, loading matrix P and regression coefficient matrix c, W = [w1, w2, …, w k ],T=[t1,t2,…,t k ],P=[p1,p2,…,p k ],c=[c1,c2,…,c k ]; specifically:

[0060] (5311) Calculate the weight vector w of the input sweat volume feature X h , so that the component t h The variance of is the largest, and the expression is as follows:

[0061]

[0062] (5312) Calculate the hth component t h (score vector), the expression is as follows:

[0063] t h =Xw h

[0064] (5313) The ingredients t h Establish a regression relationship with the sweat volume Y and calculate the loading vector c h , the expression is as follows:

[0065]

[0066] (5314) Calculate the residual matrix of X and Y, removing the influence of the current component, the expression is as follows:

[0067]

[0068] Among them, p h is the loading matrix of X, X new Y is the residual matrix of the sweat volume feature after removing the influence of the current component, new is the residual matrix of sweat volume after removing the influence of current components;

[0069]

[0070] (5315) Repeat steps (5311) to (5314) until p principal components are extracted;

[0071] (532) Calculate the regression coefficient β, connect the sweat volume feature X with the sweat volume Y, and the expression is as follows:

[0072] β=W(P T W) -1 c

[0073] The PLSR model at this time is:

[0074]

[0075] Among them, ∈ is the error term, which establishes the sweat volume feature X and the sweat volume prediction value The mapping relationship between them.

[0076] Furthermore, the step (6) specifically includes:

[0077] (61) For the test data X test Perform standardization to obtain the standardized sweat volume characteristic test data X test,std , the expression is as follows:

[0078]

[0079] (62) Use the regression coefficient β to make predictions and get the prediction results The expression is as follows:

[0080]

[0081] (63) The prediction results are denormalized to obtain the final prediction value of sweat volume The expression is as follows:

[0082]

[0083] A sweat volume monitoring method based on RFID passive multi-frequency sensing of the present invention, based on the above system, includes the following steps:

[0084] 1) Paste the passive sweat volume sensor at the target position of the human body;

[0085] 2) The sweat volume sensing integrated device continuously transmits multi-frequency wireless radio frequency signals, and collects the wireless radio frequency signals backscattered by the passive sweat volume sensor in real time, and sends them to the data processing platform;

[0086] 3) The data processing platform calculates the above wireless radio frequency signal to obtain the phase information of the multi-frequency signal;

[0087] 4) Eliminate the influence of human motion on the sensing results through dual RFID tags and multi-frequency signal phase difference method to obtain sweat volume characteristics;

[0088] 5) Calculate all sweat volume features within the measurement range and train the sweat volume regressor;

[0089] 6) Input the sweat volume feature test data into the sweat volume regressor for prediction to obtain the sweat volume.

[0090] Furthermore, the multi-frequency wireless radio frequency signal in step 2) is an OFDM symbol, and the specific generation steps are as follows:

[0091] 21) Generate a random sequence S of length n, where each element has a probability of being +1 or -1 half, and n is the custom OFDM symbol length;

[0092] 22) Perform inverse Fourier transform on the random sequence S to obtain the OFDM symbol x in the time domain. The OFDM symbol consists of n complex numbers, denoted as x = {c1, c2, …, c i ,…,c n}, where each complex number c i Indicates a specific frequency f i Signal;

[0093] 23) Add a cyclic suffix CS of length Q to the end of the generated OFDM symbol x, where CS is the first Q elements of x, CS = {c1, c2, ..., c i ,…,c Q};

[0094] 24) Multiply the OFDM symbol after adding CS by the amplification factor α. The amplitude of the OFDM symbol is proportional to the energy, ensuring that the OFDM symbol has enough energy to activate the RFID tag. The value of α is 6;

[0095] 25) During the EPC stage when the RFID signal is sent, the sweat volume sensing integrated machine replaces the single-frequency continuous wave with an OFDM symbol with multi-frequency information to activate the passive sweat volume sensor.

[0096] Furthermore, the method for calculating the phase of the multi-frequency signal in step 3) is:

[0097] 31) According to the wireless radio frequency signal amplitude threshold A thresh and duration threshold T thresh , separate the EPC phase signal from the received wireless RF signal, where A thresh is 0.05, T thresh is 3;

[0098] 32) Perform subcarrier alignment on the OFDM symbol x in the EPC phase signal, specifically including:

[0099] 321) Perform a fast Fourier transform on x to get the frequency domain representation as follows:

[0100]

[0101] Where E is the subcarrier offset of the OFDM symbol, X n is the random sequence used to generate OFDM symbols, j is the imaginary unit, and N is the length of the OFDM symbol;

[0102] 322)Compare with initial X n and The phase offset of the subcarrier is obtained by:

[0103] 323) Shift x by E units to obtain the OFDM symbol after subcarrier offset compensation alignment, denoted as x align ;

[0104] 33) x align The amplitude A x The OFDM symbol amplitude A in the T1 phase when the RFID signal is transmitted T For comparison, if A x =A T , then x align It is in OFF state, recorded as x off , otherwise it is in the ON state, recorded as x on ;

[0105] 34) According to the obtained x off and x on Extract the phase data w, the expression is as follows:

[0106]

[0107] Here, arg represents the angle at which the complex number is calculated.

[0108] Furthermore, in step 4), the step of obtaining the sweat volume characteristics by dual RFID tags and multi-frequency signal phase difference is as follows:

[0109] Set the phase data of the perception tag to: (t s1 ,W s1 ),(t s2 ,W s2 ),…,(t sm ,W sm ), where t s1 ,t s2 ,…,t sm represents the time point of perceiving the label, W s1 ,W s2 ,…,W sm Indicates the corresponding phase value; the phase data of the reference label is:

[0110] (t r1 ,W r1 ),(t r2 ,W r2 ),…,(t rm ,W rm ), where t r1 ,t r2 ,…,t rm represents the time point of perceiving the label, W r1 ,W r2 ,…,W rm Represents the corresponding phase value; two time series t r and t s If they are inconsistent, the phase data is interpolated so that the phase data of the perception tag is mapped to the time point of the reference tag; linear interpolation is used to estimate the phase value of the perception tag at the time point of the reference tag, as follows:

[0111] 41) Determine the time point of the reference tag: The time point of the reference tag is t r1 ,t r2 ,…,t rm , indicating the time point you wish to align;

[0112] 42) Find the time interval of the perception label: For each reference time point t rk , find the perception label time series t s Two time points that satisfy the following relationship:

[0113] t sj ≤t rk ≤t sj+1 ,j∈[1,m-1]

[0114] Among them, t sj and t sj+1 Forrk The two closest time points;

[0115] 43) According to the definition of linear interpolation, the perceived tag phase value W s (t rk ) at time point t rk The estimated values ​​at are as follows:

[0116]

[0117] Among them, W sj , W sj+1 are time points t sj and t sj+1 The perceived tag phase value at t sj ,t sj+1 are time points t rk The nearest time point;

[0118] 44) Loop through all reference time points: For each reference label time point t rk , both find the two most recent time points t in the time series of the perception label sj and t sj+1 , use the linear interpolation formula to calculate the corresponding perception tag phase value W s (t rk );

[0119] 45) After aligning the time points of the perception label and the reference label by interpolation, for each time point t i , calculate W d (t i )=W s (t i )-W r (t i ), all W d (t i ) value average, and get the difference W d , where W s (t i ) is the perception label at time point t i The phase value, W r (t i ) is the reference label at time point t i The phase value, W d (t i ) is the perceived label and the reference label at time point t i The difference in phase values;

[0120] 46) Calculate the phase data W of the sensing tag s The phase data W of the reference tag r The difference W d , denoted as in is the frequency f i The phase data difference between the sensing tag and the reference tag;

[0121] 47) Calculate the difference W d The adjacent frequency f i With f i+1 The phase data difference between the two is expressed as the sweat volume feature X,

[0122] Furthermore, the steps of training the regressor in step 5) are:

[0123] 51) The sweat volume feature X and the corresponding sweat volume Y are standardized to zero mean and unit variance, respectively, and are denoted as X std and Y std ,as follows:

[0124]

[0125] in, and δX are the mean and standard deviation of X, respectively. and δY are the mean and standard deviation of Y, respectively. std is the standardized sweat volume characteristic, Y std is the standardized sweat volume;

[0126] 52) Use the cross validation method to compare the performance of models with different numbers of components p and select the p with the smallest mean square error;

[0127] 53) Establish a PLSR model to project X and Y into a low-dimensional space while maximizing the correlation between the two. The specific steps are as follows:

[0128] 531) For each PLSR component t h (h-th latent variable) is extracted and modeled to obtain the weight matrix W, score matrix T, loading matrix P and regression coefficient matrix c, W = [w1, w2, …, w k ],T=[t1,t2,…,t k ],P=[p1,p2,…,p k ],c=[c1,c2,…,c k ]; specifically:

[0129] 5311) Calculate the weight vector w of the input sweat volume feature X h , so that the component t h The variance of is the largest, and the expression is as follows:

[0130]

[0131] 5312) Calculate the hth component t h (score vector), the expression is as follows:

[0132] t h =Xw h

[0133] 5313) the ingredients t h Establish a regression relationship with the sweat volume Y and calculate the loading vector c h , the expression is as follows:

[0134]

[0135] 5314) Calculate the residual matrix of X and Y to remove the influence of the current component. The expression is as follows:

[0136]

[0137] Among them, p h is the loading matrix of X, X new Y is the residual matrix of the sweat volume feature after removing the influence of the current component, new is the residual matrix of sweat volume after removing the influence of current components;

[0138]

[0139] 5315) Repeat step 5311)-step 5314) until p principal components are extracted;

[0140] 532) Calculate the regression coefficient β, connect the sweat volume feature X with the sweat volume Y, and the expression is as follows:

[0141] β=W(P T W) -1 c

[0142] The PLSR model at this time is:

[0143]

[0144] Among them, ∈ is the error term, which establishes the sweat volume feature X and the sweat volume prediction value The mapping relationship between them.

[0145] Furthermore, the step 6) specifically includes:

[0146] 61) For the test data X test Perform standardization to obtain the standardized sweat volume characteristic test data X test,std, the expression is as follows:

[0147]

[0148] 62) Use the regression coefficient β to make predictions and get the prediction results The expression is as follows:

[0149]

[0150] 63) Denormalize the prediction results to obtain the final prediction value of sweat volume The expression is as follows:

[0151]

[0152] Beneficial effects of the present invention:

[0153] 1. Dynamic environment robustness: The present invention utilizes dual tags and multi-frequency signal phase differential method to effectively eliminate the interference of human body movement on RFID wireless radio frequency signals.

[0154] 2. Multi-dimensional feature extraction: The present invention uses OFDM technology to provide multi-frequency signals, which facilitates the extraction of multi-frequency and multi-dimensional features, thereby increasing the accuracy and robustness of sweat volume perception.

[0155] 3. Lightweight sweat volume prediction: The present invention uses the PLSR lightweight regression predictor to ensure the easy training and real-time performance of sweat volume prediction reasoning.

[0156] 4. Easy deployment and low cost: The present invention uses commercial RFID tags to complete sensing, which is easy to deploy at any location and low cost, and is convenient for large-scale deployment. BRIEF DESCRIPTION OF THE DRAWINGS

[0157] Figure 1 is an architecture diagram of the system of the present invention;

[0158] Figure 2 Schematic diagram of the sweat volume sensing principle. DETAILED DESCRIPTION

[0159] In order to facilitate the understanding of those skilled in the art, the present invention is further described below in conjunction with embodiments and drawings. The contents mentioned in the implementation modes are not intended to limit the present invention.

[0160] Reference Figure 1 , Figure 2 As shown, a sweat volume monitoring system based on RFID passive multi-frequency sensing of the present invention comprises: a passive sweat volume sensor, a sweat volume sensing integrated machine and a data processing platform; wherein,

[0161] A passive sweat volume sensor, used to receive the wireless radio frequency signal transmitted by the sweat volume sensing integrated machine, and modulate the received wireless radio frequency signal and send it to the sweat volume sensing integrated machine;

[0162] Among them, the passive sweat volume sensor includes: an RFID tag pair and a sponge block for storing sweat, the RFID tag pair includes a sensing tag and a reference tag, wherein the sensing tag is used to sense the sweat volume, and the reference tag is placed 1 cm near the sensing tag to eliminate dynamic interference; the sensing tag and the reference tag are respectively attached to the front side of the sponge block with double-sided tape, and the back side of the sponge block is attached to the sweat monitoring target position (such as the armpit or the back).

[0163] The sweat volume sensing integrated machine is used to generate a multi-frequency wireless radio frequency signal to activate the passive sweat volume sensor, and collect the wireless radio frequency signal backscattered by the passive sweat volume sensor in real time and send it to the data processing platform;

[0164] Among them, the sweat volume sensing integrated machine includes: an RFID antenna and an RFID reader; the RFID reader continuously transmits multi-frequency wireless radio frequency signals through the RFID antenna, and collects the wireless radio frequency signals backscattered by the passive sweat volume sensor in real time, and sends them to the data processing platform.

[0165] Specifically, the RFID reader and RFID tag are of UHF specifications with a frequency of 860-960 MHz, and the RFID reader reads and writes the RFID tag through the EPC Global C1 G2 protocol.

[0166] Specifically, the multi-frequency wireless radio frequency signal is an OFDM symbol, and the specific generation steps are as follows:

[0167] Generate a random sequence S of length n, where each element has a probability of half being +1 or -1, and n is the custom OFDM symbol length;

[0168] Perform an inverse Fourier transform (IFFT) on the random sequence S to obtain the OFDM symbol x in the time domain. The OFDM symbol consists of n complex numbers, denoted by x = {c1, c2, …, c i ,…,c n}, where each complex number c i Indicates a specific frequency f i Signal;

[0169] A cyclic suffix CS of length Q is added to the end of the generated OFDM symbol x, where CS is the first Q elements of x, CS = {c1, c2, ..., c i ,…,c Q};

[0170] The OFDM symbol after adding CS is multiplied by the amplification factor α. The amplitude of the OFDM symbol is proportional to the energy, ensuring that the OFDM symbol has enough energy to activate the RFID tag. The value of α is 6.

[0171] Among them, in the EPC stage when the RFID signal is sent, the sweat amount sensing integrated machine replaces the single-frequency continuous wave (CW) with an OFDM symbol with multi-frequency information to activate the passive sweat amount sensor.

[0172] The data processing platform obtains the phase data and RSSI characteristic data in the wireless radio frequency signal sent by the sweat volume sensing integrated device in real time, and uses the multi-frequency signal difference method to extract the features related to the sweat volume, perform sweat volume prediction, and obtain the sweat volume;

[0173] The method for the data processing platform to calculate the sweat volume result data is:

[0174] (1) Obtain the phase data of the reference tag and the sensing tag at different RF signal frequencies, which are recorded as in is the reference label at frequency f i The phase data below, is the perceived label at frequency f i Phase data below;

[0175] (2) Calculate the phase data W of the sensing tag s The phase data W of the reference tag r The difference W d , denoted as in is the frequency f i The phase data difference between the sensing tag and the reference tag;

[0176] (3) Calculate the difference W d The adjacent frequency f i With f i+1 The phase data difference between the two is expressed as the sweat volume feature X,

[0177] (4) Calculate all sweat volume characteristics within the measurement range;

[0178] (5) Train a regressor using all the sweat volume features in step (4);

[0179] (6) Use the trained regressor to predict sweat volume.

[0180] Specifically, the step (1) comprises:

[0181] (11) According to the wireless radio frequency signal amplitude threshold A thresh and duration threshold T thresh , separate the EPC phase signal from the received wireless RF signal, where A thresh is 0.05, T thresh is 3;

[0182] (12) Subcarrier alignment is performed on the OFDM symbol x in the EPC phase signal, specifically including:

[0183] (121) Perform a fast Fourier transform (FFT) on x to obtain the frequency domain representation as follows:

[0184]

[0185] Where E is the subcarrier offset of the OFDM symbol, X n is the random sequence used to generate OFDM symbols, j is the imaginary unit, and N is the length of the OFDM symbol;

[0186] (122)Compared with the initial X n and The phase offset of the subcarrier is obtained by:

[0187] (123) Shift x by E units to obtain the OFDM symbol after subcarrier offset compensation alignment, denoted as x align ;

[0188] (13) x align The amplitude A x The OFDM symbol amplitude A in the T1 phase when the RFID signal is transmitted T For comparison, if A x =A T , then x align It is in OFF state, recorded as x off , otherwise it is in the ON state, recorded as x on ;

[0189] (14) According to the obtained x off and x on Extract the phase data w, the expression is as follows:

[0190]

[0191] Here, arg represents the angle at which the complex number is calculated.

[0192] Specifically, the step (2) comprises:

[0193] Set the phase data of the perception tag to: (t s1 ,W s1 ),(t s2 ,W s2 ),…,(t sm ,W sm ), where t s1 ,t s2 ,…,t sm represents the time point of perceiving the label, W s1 ,W s2 ,…,W sm Indicates the corresponding phase value; the phase data of the reference label is:

[0194] (t r1 ,W r1 ),(t r2 ,W r2 ),…,(t rm ,W rm ), where t r1 ,t r2 ,…,t rm represents the time point of perceiving the label, W r1 ,W r2 ,…,W rm Represents the corresponding phase value; two time series t r and t s If they are inconsistent, the phase data is interpolated so that the phase data of the perception tag is mapped to the time point of the reference tag; linear interpolation is used to estimate the phase value of the perception tag at the time point of the reference tag, as follows:

[0195] (21) Determine the time point of the reference tag: The time point of the reference tag is t r1 ,t r2 ,…,t rm , indicating the time point you wish to align;

[0196] (22) Find the time interval of the perception label: For each reference time point t rk , find the perception label time series t s There are two time points in which the following relationship is satisfied:

[0197] t sj ≤t rk ≤t sj+1 ,j∈[1,m-1]

[0198] Among them, t sj and t sj+1 For rk The two closest time points;

[0199] (23) According to the definition of linear interpolation, the perceived tag phase value Ws (t rk ) at time point t rk The estimated values ​​at are as follows:

[0200]

[0201] Among them, W sj , W sj+1 are time points t sj and t sj+1 The perceived tag phase value at t sj ,t sj+1 are time points t rk The nearest time point;

[0202] (24) Loop through all reference time points: For each reference label time point t rk , both find the two most recent time points t in the time series of the perception label sj and t sj+1 , use the linear interpolation formula to calculate the corresponding perception tag phase value W s (t rk );

[0203] (25) After aligning the time points of the perception label and the reference label by interpolation, for each time point t i , calculate W d (t i )=W s (t i )-W r (t i ), all W d (t i ) value average, and get the difference W d , where W s (t i ) is the perception label at time point t i The phase value, W r (t i ) is the reference label at time point t i The phase value, W d (t i ) is the perceived label and the reference label at time point t i The phase value difference.

[0204] Specifically, the step (5) comprises:

[0205] (51) The sweat volume feature X and the corresponding sweat volume Y are standardized to zero mean and unit variance, respectively, and are denoted as X std and Y std ,as follows:

[0206]

[0207] in, and δX are the mean and standard deviation of X, respectively. and δY are the mean and standard deviation of Y, respectively. std is the standardized sweat volume characteristic, Y std is the standardized sweat volume;

[0208] (52) Use the cross validation method to compare the performance of models with different numbers of components p and select the p with the smallest mean square error;

[0209] (53) Establish a PLSR model to project X and Y into a low-dimensional space and maximize the correlation between them. The specific steps are as follows:

[0210] (531) For each PLSR component t h (h-th latent variable) is extracted and modeled to obtain the weight matrix W, score matrix T, loading matrix P and regression coefficient matrix c, W = [w1, w2, …, w k ],T=[t1,t2,…,t k ],P=[p1,p2,…,p k ],c=[c1,c2,…,c k ]; specifically:

[0211] (5311) Calculate the weight vector w of the input sweat volume feature X h , so that the component t h The variance of is the largest, and the expression is as follows:

[0212]

[0213] (5312) Calculate the hth component t h (score vector), the expression is as follows:

[0214] t h =Xw h

[0215] (5313) The ingredients t h Establish a regression relationship with the sweat volume Y and calculate the loading vector c h , the expression is as follows:

[0216]

[0217] (5314) Calculate the residual matrix of X and Y, removing the influence of the current component, the expression is as follows:

[0218] X Ynew =Yt h c h

[0219] Among them, p h is the loading matrix of X, X new Y is the residual matrix of the sweat volume feature after removing the influence of the current component, new is the residual matrix of sweat volume after removing the influence of current components;

[0220]

[0221] (5315) Repeat steps (5311) to (5314) until p principal components are extracted;

[0222] (532) Calculate the regression coefficient β, connect the sweat volume feature X with the sweat volume Y, and the expression is as follows:

[0223] β=W(P T W) -1 c

[0224] The PLSR model at this time is:

[0225]

[0226] Among them, ∈ is the error term, which establishes the sweat volume feature X and the sweat volume prediction value The mapping relationship between them.

[0227] Specifically, the step (6) comprises:

[0228] (61) For the test data X test Perform standardization to obtain the standardized sweat volume characteristic test data X test,std , the expression is as follows:

[0229]

[0230] (62) Use the regression coefficient β to make predictions and get the prediction results The expression is as follows:

[0231]

[0232] (63) The prediction results are denormalized to obtain the final prediction value of sweat volume The expression is as follows:

[0233]

[0234] A sweat volume monitoring method based on RFID passive multi-frequency sensing of the present invention, based on the above system, includes the following steps:

[0235] 1) Stick the passive sweat volume sensor to the target position of the human body, such as the armpit or the back;

[0236] 2) The sweat volume sensing integrated device continuously transmits multi-frequency wireless radio frequency signals, and collects the wireless radio frequency signals backscattered by the passive sweat volume sensor in real time, and sends them to the data processing platform;

[0237] Wherein, the multi-frequency wireless radio frequency signal in step 2) is an OFDM symbol, and the specific generation steps are as follows:

[0238] 21) Generate a random sequence S of length n, where each element has a probability of being +1 or -1 half, and n is the custom OFDM symbol length;

[0239] 22) Perform inverse Fourier transform on the random sequence S to obtain the OFDM symbol x in the time domain. The OFDM symbol consists of n complex numbers, denoted as x = {c1, c2, …, c i ,…,c n}, where each complex number c i Indicates a specific frequency f i Signal;

[0240] 23) Add a cyclic suffix CS of length Q to the end of the generated OFDM symbol x, where CS is the first Q elements of x, CS = {c1, c2, ..., c i ,…,c Q};

[0241] 24) Multiply the OFDM symbol after adding CS by the amplification factor α. The amplitude of the OFDM symbol is proportional to the energy, ensuring that the OFDM symbol has enough energy to activate the RFID tag. The value of α is 6;

[0242] 25) During the EPC stage when the RFID signal is sent, the sweat volume sensing integrated machine replaces the single-frequency continuous wave with an OFDM symbol with multi-frequency information to activate the passive sweat volume sensor.

[0243] 3) The data processing platform calculates the collected wireless radio frequency signals to obtain phase information of the multi-frequency signals;

[0244] Among them, the method for calculating the phase of the multi-frequency signal is:

[0245] 31) According to the wireless radio frequency signal amplitude threshold A thresh and duration threshold T thresh , separate the EPC phase signal from the received wireless RF signal, where A thresh is 0.05, T thresh is 3;

[0246] 32) Perform subcarrier alignment on the OFDM symbol x in the EPC phase signal, specifically including:

[0247] 321) Perform a fast Fourier transform on x to get the frequency domain representation as follows:

[0248]

[0249] Where E is the subcarrier offset of the OFDM symbol, X n is the random sequence used to generate OFDM symbols, j is the imaginary unit, and N is the length of the OFDM symbol;

[0250] 322)Compare with initial X n and The phase offset of the subcarrier is obtained by:

[0251] 323) Shift x by E units to obtain the OFDM symbol after subcarrier offset compensation alignment, denoted as x align ;

[0252] 33) x align The amplitude A x The OFDM symbol amplitude A in the T1 phase when the RFID signal is transmitted T For comparison, if A x =A T , then x align It is in OFF state, recorded as x off , otherwise it is in the ON state, recorded as x on ;

[0253] 34) According to the obtained x off and x on Extract the phase data w, the expression is as follows:

[0254]

[0255] Here, arg represents the angle at which the complex number is calculated.

[0256] 4) Eliminate the influence of human motion on the sensing results through dual RFID tags and multi-frequency signal phase difference method to obtain sweat volume characteristics;

[0257] Wherein, the step of obtaining the sweat volume characteristics by using dual RFID tags and multi-frequency signal phase difference in step 4) is:

[0258] Set the phase data of the perception tag to: (t s1 ,W s1 ),(t s2 ,W s2),…,(t sm ,W sm ), where t s1 ,t s2 ,…,t sm represents the time point of perceiving the label, W s1 ,W s2 ,…,W sm Indicates the corresponding phase value; the phase data of the reference tag is: (t r1 ,W r1 ),(t r2 ,W r2 ),…,(t rm ,W rm ), where t r1 ,t r2 ,…,t rm represents the time point of perceiving the label, W r1 ,W r2 ,…,W rm Represents the corresponding phase value; two time series t r and t s If they are inconsistent, the phase data is interpolated so that the phase data of the perception tag is mapped to the time point of the reference tag; linear interpolation is used to estimate the phase value of the perception tag at the time point of the reference tag, as follows:

[0259] 41) Determine the time point of the reference tag: The time point of the reference tag is t r1 ,t r2 ,…,t rm , indicating the time point you wish to align;

[0260] 42) Find the time interval of the perception label: For each reference time point t rk , find the perception label time series t s Two time points that satisfy the following relationship:

[0261] t sj ≤t rk ≤t sj+1 ,j∈[1,m-1]

[0262] Among them, t sj and t sj+1 For rk The two closest time points;

[0263] 43) According to the definition of linear interpolation, the perceived tag phase value W s (t rk ) at time point t rk The estimated values ​​at are as follows:

[0264]

[0265] Among them, W sj , W sj+1 are time points t sj and t sj+1 The perceived tag phase value at t sj ,t sj+1 are time points t rk The nearest time point;

[0266] 44) Loop through all reference time points: For each reference label time point t rk , both find the two most recent time points t in the time series of the perception label sj and t sj+1 , use the linear interpolation formula to calculate the corresponding perception tag phase value W s (t rk );

[0267] 45) After aligning the time points of the perception label and the reference label by interpolation, for each time point t i , calculate W d (t i )=W s (t i )-W r (t i ), all W d (t i ) value average, and get the difference W d , where W s (t i ) is the perception label at time point t i The phase value, W r (t i ) is the reference label at time point t i The phase value, W d (t i ) is the perceived label and the reference label at time point t i The difference in phase values;

[0268] 46) Calculate the phase data W of the sensing tag s The phase data W of the reference tag r The difference W d , denoted as in is the frequency f i The phase data difference between the sensing tag and the reference tag;

[0269] 47) Calculate the difference W d The adjacent frequency f i With fi+1 The phase data difference between the two is expressed as the sweat volume feature X,

[0270] 5) Calculate all sweat volume features within the measurement range and train the sweat volume regressor;

[0271] The steps of training the regressor in step 5) are:

[0272] 51) The sweat volume feature X and the corresponding sweat volume Y are standardized to zero mean and unit variance, respectively, and are denoted as X std and Y std ,as follows:

[0273]

[0274] in, and δX are the mean and standard deviation of X, respectively. and δY are the mean and standard deviation of Y, respectively. std is the standardized sweat volume characteristic, Y std is the standardized sweat volume;

[0275] 52) Use the cross validation method to compare the performance of models with different numbers of components p and select the p with the smallest mean square error;

[0276] 53) Establish a PLSR model to project X and Y into a low-dimensional space while maximizing the correlation between the two. The specific steps are as follows:

[0277] 531) For each PLSR component t h (h-th latent variable) is extracted and modeled to obtain the weight matrix W, score matrix T, loading matrix P and regression coefficient matrix c, W = [w1, w2, …, w k ],T=[t1,t2,…,t k ],P=[p1,p2,…,p k ],c=[c1,c2,…,c k ]; specifically:

[0278] 5311) Calculate the weight vector w of the input sweat volume feature X h , so that the component t h The variance of is the largest, and the expression is as follows:

[0279]

[0280] 5312) Calculate the hth component t h (score vector), the expression is as follows:

[0281] t h =Xw h

[0282] 5313) the ingredients t h Establish a regression relationship with the sweat volume Y and calculate the loading vector c h , the expression is as follows:

[0283]

[0284] 5314) Calculate the residual matrix of X and Y to remove the influence of the current component. The expression is as follows:

[0285]

[0286] Among them, p h is the loading matrix of X, X new Y is the residual matrix of the sweat volume feature after removing the influence of the current component, new is the residual matrix of sweat volume after removing the influence of current components;

[0287]

[0288] 5315) Repeat step 5311)-step 5314) until p principal components are extracted;

[0289] 532) Calculate the regression coefficient β, connect the sweat volume feature X with the sweat volume Y, and the expression is as follows:

[0290] β=W(P T W) -1 c

[0291] The PLSR model at this time is:

[0292]

[0293] Among them, ∈ is the error term, which establishes the sweat volume feature X and the sweat volume prediction value The mapping relationship between them.

[0294] 6) Inputting the sweat volume feature test data into the sweat volume regressor for prediction to obtain the sweat volume; specifically including:

[0295] 61) For the test data X test Perform standardization to obtain the standardized sweat volume characteristic test data X test,std , the expression is as follows:

[0296]

[0297] 62) Use the regression coefficient β to make predictions and get the prediction results The expression is as follows:

[0298]

[0299] 63) Denormalize the prediction results to obtain the final prediction value of sweat volume The expression is as follows:

[0300]

[0301] The present invention has many specific application paths. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements can be made without departing from the principle of the present invention. These improvements should also be regarded as the protection scope of the present invention.

Claims

1. A sweat volume monitoring system based on RFID passive multi-frequency sensing, characterized in that: include: Passive sweat volume sensor, sweat volume sensing integrated machine and data processing platform; A passive sweat volume sensor, used to receive the wireless radio frequency signal transmitted by the sweat volume sensing integrated machine, and modulate the received wireless radio frequency signal and send it to the sweat volume sensing integrated machine; The sweat volume sensing integrated machine is used to generate a multi-frequency wireless radio frequency signal to activate the passive sweat volume sensor, and collect the wireless radio frequency signal sent by the passive sweat volume sensor in real time and send it to the data processing platform; The data processing platform obtains the phase data and RSSI characteristic data in the wireless radio frequency signal sent by the sweat volume sensing integrated device in real time, extracts the features related to the sweat volume, predicts the sweat volume, and obtains the sweat volume.

2. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 1 is characterized in that: The passive sweat volume sensor includes: an RFID tag pair and a sponge block for storing sweat, wherein the RFID tag pair includes a sensing tag and a reference tag, wherein the sensing tag is used to sense the sweat volume, and the reference tag is placed 1 cm away from the sensing tag to eliminate dynamic interference; the sensing tag and the reference tag are respectively attached to the front side of the sponge block by double-sided tape, and the back side of the sponge block is attached to the sweat monitoring target position.

3. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 1 is characterized in that: The multi-frequency wireless radio frequency signal is an OFDM symbol, and the specific generation steps are as follows: Generate a random sequence S of length n, where each element has a probability of half being +1 or -1, and n is the custom OFDM symbol length; Perform inverse Fourier transform on the random sequence S to obtain the OFDM symbol x in the time domain. The OFDM symbol consists of n complex numbers, denoted as x = {c1, c2, ..., c i , ..., c n }, where each complex number c i Indicates a specific frequency f i Signal; A cyclic suffix CS of length Q is added to the end of the generated OFDM symbol x, where CS is the first Q elements of x, CS = {c1, c2, ..., c i , ..., c Q }; The OFDM symbol after adding CS is multiplied by the amplification factor α, and the amplitude of the OFDM symbol is proportional to the energy, ensuring that the OFDM symbol has enough energy to activate the RFID tag.

4. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 1 is characterized in that: During the EPC phase when the RFID signal is sent, the sweat volume sensing integrated machine replaces the single-frequency continuous wave with an OFDM symbol having multi-frequency information to activate the passive sweat volume sensor.

5. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 1 is characterized in that: The method for the data processing platform to calculate the sweat volume result data is: (1) Obtain the phase data of the reference tag and the sensing tag at different RF signal frequencies, which are recorded as and in is the reference label at frequency f i The phase data below, is the perceived label at frequency f i Phase data below; (2) Calculate the phase data W of the sensing tag s The phase data W of the reference tag r The difference W d , denoted as in is the frequency f i The phase data difference between the sensing tag and the reference tag; (3) Calculate the difference W d The adjacent frequency f i With f i+1 The phase data difference between the two is expressed as the sweat volume feature X, (4) Calculate all sweat volume characteristics within the measurement range; (5) Train a regressor using all the sweat volume features in step (4); (6) Use the trained regressor to predict sweat volume.

6. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 5 is characterized in that: The step (1) specifically includes: (11) According to the wireless radio frequency signal amplitude threshold A thresh and duration threshold T thresh , separating the EPC phase signal from the received wireless radio frequency signal; (12) Subcarrier alignment is performed on the OFDM symbol x in the EPC phase signal, specifically including: (121) Perform a fast Fourier transform on x and get the frequency domain representation as follows: Where E is the subcarrier offset of the OFDM symbol, X n is the random sequence used to generate OFDM symbols, j is the imaginary unit, and N is the length of the OFDM symbol; (122)Compared with the initial X n and The phase offset of the subcarrier is obtained by: (123) Shift x by E units to obtain the OFDM symbol after subcarrier offset compensation alignment, denoted as x align ; (13) x align The amplitude A x The OFDM symbol amplitude A in the T1 phase when the RFID signal is transmitted T For comparison, if A x =A T , then x align It is in OFF state, recorded as x off , otherwise it is in the ON state, recorded as x on ; (14) According to the obtained x off and x on Extract the phase data w, the expression is as follows: Here, arg represents the angle at which the complex number is calculated.

7. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 5 is characterized in that: The step (2) specifically comprises: Set the phase data of the perception tag to: (t s1 , W s1 ), (t s2 , W s2 ),...,(t sm , W sm ), where t s1 , t s2 , ..., t sm represents the time point of perceiving the label, W s1 , W s2 , ..., W sm Indicates the corresponding phase value; the phase data of the reference tag is: (t r1 , W r1 ), (t r2 , W r2 ),...,(t rm , W rm ), where t r1 , t r2 , ..., t rm represents the time point of perceiving the label, W r1 , W r2 , ..., W rm Represents the corresponding phase value; two time series t r and t s If they are inconsistent, the phase data is interpolated so that the phase data of the perception tag is mapped to the time point of the reference tag; linear interpolation is used to estimate the phase value of the perception tag at the time point of the reference tag, as follows: (21) Determine the time point of the reference tag: The time point of the reference tag is t r1 , t r2 , ..., t rm , indicating the time point you wish to align; (22) Find the time interval of the perception label: For each reference time point t rk , find the perception label time series t s There are two time points in which the following relationship is satisfied: t sj ≤t rk ≤t sj+1 ,j∈[1,m-1] Among them, t sj and t sj+1 For rk The two closest time points; (23) According to the definition of linear interpolation, the perceived tag phase value W s (t rk ) at time point t rk The estimated values ​​at are as follows: Among them, W sj , W sj+1 are time points t sj and t sj+1 The perceived tag phase value at t sj ,t sj+1 are time points t rk The nearest time point; (24) Loop through all reference time points: For each reference label time point t rk , both find the two most recent time points t in the time series of the perception label sj and t sj+1 , use the linear interpolation formula to calculate the corresponding perception tag phase value W s (t rk ); (25) After aligning the time points of the perception label and the reference label by interpolation, for each time point t i , calculate W d (t i )=W s (t i )-W r (t i ), all W d (t i ) value average, and get the difference W d , where W s (t i ) is the perception label at time point t i The phase value, W r (t i ) is the reference label at time point t i The phase value, W d (t i ) is the perceived label and the reference label at time point t i The phase value difference.

8. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 5 is characterized in that: The step (5) specifically comprises: (51) The sweat volume feature X and the corresponding sweat volume Y are standardized to zero mean and unit variance, respectively, and are denoted as X std and Y std ,as follows: in, and δX are the mean and standard deviation of X, respectively. and δY are the mean and standard deviation of Y, respectively. std is the standardized sweat volume characteristic, Y std is the standardized sweat volume; (52) Use the cross-validation method to compare the performance of models with different numbers of components p and select the p with the smallest mean square error; (53) Establish a PLSR model to project X and Y into a low-dimensional space and maximize the correlation between them. The specific steps are as follows: (531) For each PLSR component t h Extraction and modeling are performed to obtain the weight matrix W, score matrix T, loading matrix P and regression coefficient matrix c, W = [w1, w2, ..., w k ], T = [t1, t2, ..., t k ], P = [p1, p2, ..., p k ],c=[c1,c2,...,c k ]; specifically: (5311) Calculate the weight vector w of the input sweat volume feature X h , so that the component t h The variance of is the largest, and the expression is as follows: (5312) Calculate the hth component t h , the expression is as follows: t h =Xw h (5313) The ingredients t h Establish a regression relationship with the sweat volume Y and calculate the loading vector c h , the expression is as follows: (5314) Calculate the residual matrix of x and Y to remove the influence of the current component. The expression is as follows: Among them, p h is the loading matrix of X, X new Y is the residual matrix of the sweat volume feature after removing the influence of the current component, new is the residual matrix of sweat volume after removing the influence of current components; (5315) Repeat steps (5311) to (5314) until p principal components are extracted; (532) Calculate the regression coefficient β, connect the sweat volume feature X with the sweat volume Y, and the expression is as follows: β=W(P T W) -1 c The PLSR model at this time is: Among them, ∈ is the error term, which establishes the sweat volume feature X and the sweat volume prediction value The mapping relationship between them.

9. The sweat volume monitoring system based on RFID passive multi-frequency sensing according to claim 5, characterized in that: The step (6) specifically comprises: (61) For the test data X test Perform standardization to obtain the standardized sweat volume characteristic test data X test，std , the expression is as follows: (62) Use the regression coefficient β to make predictions and get the prediction results The expression is as follows: (63) The prediction results are denormalized to obtain the final prediction value of sweat volume The expression is as follows:

10. A sweat volume monitoring method based on RFID passive multi-frequency sensing, based on the system described in any one of claims 1-9, characterized in that: The steps are as follows: 1) Paste the passive sweat volume sensor at the target position of the human body; 2) The sweat volume sensing integrated device continuously transmits multi-frequency wireless radio frequency signals, and collects the wireless radio frequency signals backscattered by the passive sweat volume sensor in real time, and sends them to the data processing platform; 3) The data processing platform calculates the above wireless radio frequency signal to obtain the phase information of the multi-frequency signal; 4) Eliminate the influence of human motion on the sensing results through dual RFID tags and multi-frequency signal phase difference method to obtain sweat volume characteristics; 5) Calculate all sweat volume features within the measurement range and train the sweat volume regressor; 6) Input the sweat volume feature test data into the sweat volume regressor for prediction to obtain the sweat volume.

Images