Intelligent processing method and system for heart rate monitoring data

Abstract

The invention discloses an intelligent processing method and system for heart rate monitoring data, and relates to the technical field of data processing. The method comprises the following steps: S1, collecting heart rate data and motion data; s1, data preprocessing and heart rate risk early warning outputting; s3, based on the received real-time heart rate sequence and the activity state tag associated with the real-time heart rate sequence, packaging the real-time heart rate sequence into a data packet, encrypting the data packet to form a final ciphertext, and sending the final ciphertext to the cloud; s4, performing ciphertext verification based on the received final ciphertext, if verification is passed, storing the ciphertext, otherwise, abandoning the ciphertext; according to the technical scheme, the real-time performance, the accuracy, the robustness, the safety, the auditing performance and the like are comprehensively guaranteed, the requirements for high-precision heart rate monitoring and intelligent early warning are met, and user privacy and system safety are guaranteed.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a method and system for intelligently processing heart rate monitoring data. Background Art

[0002] Heart rate refers to the number of times the human heart beats per minute, and is one of the basic vital signs of a person. Heart rate examination is the most basic item in cardiovascular system examination. Through this examination, we can understand the situation of the heart's beating and reflect the basic life state of the human body. Heart rate is controlled by the autonomic nervous system. When the sympathetic nerve activity is enhanced, the heart rate increases; when the parasympathetic nerve is active, the heart rate slows down. In addition, heart rate is affected by body temperature. When the body temperature rises by 1°C, the heart rate will increase by 12 to 18 times. A large number of clinical and basic studies have shown that a high heart rate state often indicates excessive sympathetic activity and increased cardiovascular load, and is associated with poor prognosis in both epilepsy and cardiovascular disease. Therefore, as one of the basic vital signs of a person, heart rate is a very important monitoring parameter.

[0003] Existing heart rate monitoring data needs to be uploaded to the cloud continuously over a long period of time, which exposes it to risks of data leakage, illegal sharing, and cyberattacks. Current encryption transmission and storage mechanisms vary widely, making it difficult to provide security for heart rate monitoring data.

[0004] In order to solve the above-mentioned defects, a method and system for intelligent processing of heart rate monitoring data are provided. Summary of the Invention

[0005] According to one aspect of the present application, a method for intelligently processing heart rate monitoring data is provided, the method comprising the following steps:

[0006] S1, collects the user's original heart rate data and motion data in real time;

[0007] S2 receives the raw heart rate data and performs a quality check on it to filter out low-quality data. The raw heart rate data that meets the quality requirements is recorded as valid heart rate data. The motion data is then classified into state categories to accurately identify the motion state and assign a motion state label. The valid heart rate data monitored simultaneously is associated with the motion state label. Using the motion data as a reference signal, the RLS adaptive filter is used to remove motion artifacts from the valid heart rate data to output a net PPG signal. A heart rate risk warning is then issued based on the net PPG signal and its corresponding motion state label. The net PPG signal and its corresponding motion state label are encapsulated into a data packet and sent to S3.

[0008] S3: Perform primary encryption processing on the received data packet to form a primary ciphertext. Then, group the values ​​in the primary ciphertext into two-dimensional coordinates, thereby splitting the primary ciphertext into several two-dimensional coordinates. A two-dimensional rectangular coordinate system is established, and coordinate points of each two-dimensional coordinate are plotted on the coordinate axis. The coordinate points are connected in sequence using smooth curves according to the order of each two-dimensional coordinate in the primary ciphertext to form a ciphertext fingerprint. The minimum convex polygon is extracted from the ciphertext fingerprint as a verification image of the ciphertext fingerprint. The ciphertext fingerprint image and its corresponding verification image are packaged into the final ciphertext and uploaded.

[0009] S4, perform ciphertext verification on the uploaded final ciphertext, and store it if the verification passes; otherwise, discard it.

[0010] Optionally, the process of performing quality detection on the raw heart rate data to filter out low-quality data is as follows:

[0011] According to the formula Calculate the signal power P of the PPG waveform signal signal , where n = 1, 2, 3 ... N, N is a positive integer, N represents the total number of signal acquisition points involved in the calculation, and n is the index of any acquisition point; PPG[n] represents the amplitude of the original photoplethysmography signal collected at the nth acquisition point, which reflects the pulse fluctuation caused by blood volume with heartbeat; the Savitzky–Golay filter is used to remove the baseline to obtain the noise component niose[n], the specific formula is: niose[n] = PPG[n] - SG 21,3 PPG[n], where SG 21,3 PPG[n] represents the smoothed value of the original PPG signal at the nth point after applying the Savitzky–Golay filter, which is used to remove the baseline drift; then according to the formula Calculate the noise power P noise ; According to the signal power P signal and noise power P noise By the formula SNR = 10·log 10 (P signal P noise ) Calculate the signal-to-noise ratio (SNR) of the original heart rate data;

[0012] Detect the time corresponding to all peaks of the PPG waveform signal and form a time series. Calculate the intervals between adjacent peaks based on the time series. Then, calculate the standard deviation of the peak intervals based on the intervals between adjacent peaks using the standard deviation formula, which is recorded as σ. PPI , the signal-to-noise ratio SNR and the peak-to-peak standard deviation σ PPI Through the weighted formula A signal quality index SQI is obtained by performing fusion calculation; if the signal quality index SQI is greater than a preset quality threshold, it is marked as valid heart rate data; otherwise, it is marked as invalid heart rate data and discarded.

[0013] Optionally, the process of classifying the motion data to accurately identify the motion state and assign a motion state label is as follows:

[0014] Extract motion data, which includes three-axis accelerometer data recorded as {(x[n], y[n], z[n])}, where (x[n], y[n], z[n]) represents the acceleration of the three axes at the nth acquisition point; according to the formula Calculate the Euclidean norm of the three-axis acceleration vector amplitude to obtain the vector amplitude VM[n], which is used to measure the instantaneous exercise intensity; use a moving average filter with a certain window width to smooth the VM curve to suppress short-term jitter and obtain the sliding amplitude; set intensity threshold 1, intensity threshold 2, and intensity threshold 3; if the sliding amplitude is less than intensity threshold 1, the running state is marked as static; if intensity threshold 1 ≤ sliding amplitude < intensity threshold 2, the running state is marked as low-intensity activity; if intensity threshold 2 ≤ sliding amplitude < intensity threshold 3, the running state is marked as medium-intensity activity; if sliding amplitude ≥ intensity threshold 3, the exercise state is marked as high-intensity activity; associate the original heart rate data with the activity state label corresponding to the exercise data collected in the same time period, thereby obtaining the original heart rate data and its associated activity state label.

[0015] Optionally, the motion artifact removal process is:

[0016] The improved RLS adaptive filtering is adopted, and acceleration is used as the reference signal to construct the reference signal equation: r[n]=0.4×x[n]+0.3×y[n]+0.3×z[n], where r[n] is the acceleration reference signal, which is determined by the linear combination of the weighted three-axis components; then the equation is used to calculate the reference signal. The PPG signal and the reference signal are normalized to accelerate the convergence of the RLS algorithm and ensure the consistency of the variable dimensions, where μ PPG and σ PPG are the mean and standard deviation of the PPG signal, μ r and σ r are the mean and standard deviation of the reference signal r[n] respectively; and {r norm [n],......r norm [n-M+1]} Τ As the input vector of RLS filter, PPG norm [n] is used as the expected output vector, and the relevant parameters of the RLS adaptive filter are set by the user: the forgetting factor and the filter order.

[0017] Optionally, the heart rate risk warning process is:

[0018] Peak detection is performed on the clean PPG signal to extract peak indices, forming a peak index sequence. The minimum and maximum allowable intervals between adjacent peaks are set, and the impurity peaks outside the minimum sample interval to the maximum sample interval are filtered out. A sliding window is used to calculate the real-time heart rate sequence. The real-time heart rate sequence and its associated activity state label are obtained and sent to step S3.

[0019] Set each activity status label to correspond to a heart rate limit interval, extract the activity status label of each heart rate sequence, and compare it with all the set activity status labels to match the corresponding heart rate limit. Perform distance analysis on the heart rate sequence and the corresponding heart rate limit interval. If all heart rate sequences belong to the corresponding heart rate limit interval, the output is that the heart rate is normal and no warning is needed; if the heart rate of the heart rate sequence deviates from the upper and lower limits of the heart rate limit interval by 5-10bpm, a slight warning risk is output and the user is reminded to pay attention to heart rate changes; if the heart rate sequence deviates from the upper and lower limits of the heart rate limit interval by more than 10bpm and continues for 60 seconds, a serious warning risk is output and the user is reminded to rest.

[0020] Optionally, the primary encryption process is:

[0021] Call the JSON serialization function to convert the data packet into JSON text, flattening the data packet into a continuous string of characters. Then convert the JSON text into a byte stream to obtain a UTF-8 byte sequence. The UTF-8 byte sequence is a sequence of unsigned integers between 0 and 255.

[0022] The UTF-8 byte sequence is padded so that the length of the padded UTF-8 byte sequence is an integer multiple of 16 bytes. After the padding is completed, the padded UTF-8 byte sequence is encrypted using the AES-CBC algorithm to obtain a pseudo-random byte stream with the same length as the original byte, which is recorded as the primary ciphertext.

[0023] Optionally, the process of extracting the minimum convex polygon from the ciphertext fingerprint is:

[0024] Remove duplicate coordinate points from the ciphertext fingerprint. Find the point with the smallest ordinate among all coordinate points. If there are two or more, select the point with the smallest abscissa as the vertex of the convex hull. Calculate the polar angle and distance from the convex hull vertex to the coordinate points other than the vertex. Use the convex hull vertex as the starting point and sort the other coordinate points in descending order of their corresponding polar angles. If two points have the same polar angle, sort the one with the larger distance first. The resulting coordinate sequence is recorded as {V0, V1, V2, V3...}.

[0025] Set up an empty "stack" to store the coordinate sequence being constructed. Output three points V0, V1, and V2 in sequence and push them into the stack. Starting from the third point V3 after sorting, and continuing to the last point, loop through the following process:

[0026] 1-1: The top of the stack is recorded as B, and the next one is recorded as A;

[0027] 1-2: Try a new point, let the current point to be processed be C=V k , where k is the index of any point in the coordinate sequence except the vertex of the convex hull;

[0028] 1-3: Points A, B, and C form a vector and And use the cross product formula to calculate the vector and The steering amount Δ;

[0029] 1-4: When Δ≤0, keep popping the top of the stack, that is, discarding B, and recalculating the cross product with the next point C using the new top of the stack; when Δ>0, stop popping and push C into the stack, that is, connecting C to the current convex hull boundary;

[0030] When all V k After all processing is completed, the remaining points in the stack are the convex hull vertex sequence arranged in a counterclockwise direction. Connecting these points in the stack end to end in sequence will obtain the minimum convex polygon containing all the original points, that is, the convex hull.

[0031] According to one aspect of the present application, a system for intelligent processing of heart rate monitoring data is provided, comprising: a portable heart rate monitoring device, a mobile phone, and a cloud; the mobile phone has a built-in preprocessing module, a data encryption module, and a first transceiver module; the cloud includes a second transceiver module and a database;

[0032] The portable heart rate monitoring device is used to collect the user's original heart rate data and exercise data in real time;

[0033] The preprocessing module receives raw heart rate data and performs quality checks on it to filter out low-quality data. Raw heart rate data that meets quality requirements is recorded as valid heart rate data. The motion data is then classified to accurately identify the motion state and assign a motion state label. The simultaneously monitored valid heart rate data is associated with the motion state label. Using the motion data as a reference signal, the RLS adaptive filter is used to remove motion artifacts from the valid heart rate data and output a net PPG signal. A heart rate risk warning is then issued based on the net PPG signal and its corresponding motion state label. The net PPG signal and its corresponding motion state label are encapsulated into a data packet and sent to the data encryption module.

[0034] The data encryption module performs primary encryption processing on the received data packet to form a primary ciphertext, and then groups the values ​​in the primary ciphertext into two-dimensional coordinates, thereby splitting the primary ciphertext into several two-dimensional coordinates, establishing a two-dimensional rectangular coordinate system, and plotting coordinate points of each two-dimensional coordinate on the coordinate axis. The coordinate points are then connected in sequence using smooth curves according to the order of each two-dimensional coordinate in the primary ciphertext to form a ciphertext fingerprint graph, and the minimum convex polygon is extracted from it as a verification graph of the ciphertext fingerprint graph; the ciphertext fingerprint graph and its corresponding verification graph are encapsulated into a final ciphertext, and sent to the first transceiver module, which is then uploaded to the second transceiver module in the cloud via the first transceiver module;

[0035] The second transceiver module performs ciphertext verification on the uploaded final ciphertext, and stores it if the verification passes; otherwise, it is discarded.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] (1) This application evaluates the original heart rate data to eliminate low-quality data, and then combines it with real-time three-axis acceleration to match the heart rate with the current exercise intensity, effectively distinguishing physiological fluctuations from abnormal fluctuations, and classifying it into four levels: static / low / medium / high. It can adjust the parameters of users in different usage scenarios to improve monitoring accuracy; finally, based on the heart rate range of different exercise states, it outputs three levels of risk in real time: "normal / mild warning / serious warning", provides timely feedback and guides users to adjust or rest; this not only improves the reliability and accuracy of heart rate data, but also gives the system scene perception and intelligent warning capabilities;

[0038] (2) This application converts the data packet into a pseudo-random byte stream with the same length as the original byte through JSON serialization and AES-CBC encryption to obtain the primary ciphertext; then the primary ciphertext is converted into a coordinate point set, and a ciphertext fingerprint is drawn based on it. The convex hull algorithm is used to extract the minimum convex deformation from it as a verification graph. The geometric features of the fingerprint graph are unique, and any bit-level modification can be detected by convex hull comparison; only the final ciphertext that passes the verification is written into the database to ensure the confidentiality of the data and protect the data security. The ciphertext fingerprint graph and convex hull verification are combined to achieve efficient detection and audit of data tampering.

[0039] To sum up, the entire technical solution of this application is fully guaranteed in terms of real-time, accuracy, robustness, security, auditability, etc., which not only meets the needs of high-precision heart rate monitoring and intelligent early warning, but also ensures user privacy and system security. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0041] Figure 1 Schematic diagram of the method flow of the present invention;

[0042] Figure 2 It is a system connection block diagram of the present invention;

[0043] Figure 3 Schematic diagram of the data preprocessing process of the present invention. DETAILED DESCRIPTION

[0044] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0045] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0046] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0047] like Figure 1-2As shown, an embodiment of the present application provides an intelligent processing method for heart rate monitoring data, which is implemented based on an intelligent processing system for heart rate monitoring data, the system comprising: a portable heart rate monitoring device (heart rate monitoring watch), a mobile phone and a cloud; the portable heart rate monitoring device collects the user's original heart rate data, motion data and location in real time and uploads them to the mobile phone; the mobile phone preprocesses the original heart rate data and motion data to obtain preprocessed effective heart rate data, and judges the user's heart rate status based on the preprocessed data to output a corresponding warning, encrypts the preprocessed heart rate data to form a final ciphertext, and uploads it to the second transceiver module in the cloud via the first transceiver module. The second transceiver module performs ciphertext verification on the received final ciphertext, and sends it to the database for storage after successful verification, so that the original data does not leave the mobile phone, the uploaded information is available but not decipherable, and the user's heart rate data can be effectively prevented from being abused;

[0048] The method comprises the following steps:

[0049] S1: The portable heart rate monitoring device collects the user's original heart rate data and exercise data in real time and transmits it to the mobile phone;

[0050] S2: The mobile phone terminal includes a pre-processing module, a data encryption module and a first transceiver module, such as Figure 3 As shown in the figure, the pre-processing module determines the user's heart rate status based on the real-time collected raw heart rate data and motion data, and outputs the corresponding heart rate risk, specifically:

[0051] Step 1: Receive the original heart rate data, which is a PPG waveform signal:

[0052] According to the formula Calculate the signal power P of the PPG waveform signal signal , where n = 1, 2, 3 ... N, N is a positive integer, N represents the total number of signal acquisition points involved in the calculation, and n is the index of any acquisition point; PPG[n] represents the amplitude of the original photoplethysmography (PPG) signal collected at the nth acquisition point (time), which reflects the pulse fluctuation caused by blood volume with heartbeat; the Savitzky–Golay filter is used to remove the baseline to obtain the noise component niose[n], the specific formula is: niose[n] = PPG[n] - SG 21,3 PPG[n], where SG 21,3 PPG[n] represents the smoothed value of the original PPG signal at the nth point after applying the Savitzky–Golay filter (window length 21 points, polynomial fitting order 3) to remove baseline drift; then according to the formula Calculate the noise power P noise ; According to the signal power P signal and noise power P noiseBy the formula SNR = 10·log 10 (P signal P noise ) Calculate the signal-to-noise ratio (SNR) of the original heart rate data (PPG waveform signal);

[0053] The time corresponding to all peaks of the PPG waveform signal is detected and a time series is formed. The interval between adjacent peaks is calculated based on the time series (interval represents the time interval). It should be noted that the interval between two adjacent PPG peaks reflects the length of a complete heartbeat cycle. The standard deviation formula is then used to calculate the standard deviation of the peak interval based on the interval between each adjacent peak, which is recorded as σ PPI , which indicates the stability of the heartbeat interval. The smaller the value, the more regular the heartbeat rhythm. The signal-to-noise ratio SNR and the peak-to-peak standard deviation σ PPI Through the weighted formula A signal quality index (SQI) is obtained by fusion calculation. If the SQI is greater than a preset quality threshold (those skilled in the art use a value of 5.0, which can be manually adjusted according to actual conditions), the original heart rate data is usable and marked as valid. Otherwise, it is marked as invalid and discarded.

[0054] Step 2: Classify the user's exercise status based on exercise data:

[0055] Extract motion data, which includes three-axis accelerometer data recorded as {(x[n], y[n], z[n])}, where (x[n], y[n], z[n]) represents the acceleration of the three axes at the nth acquisition point; according to the formula Calculate the Euclidean norm of the three-axis acceleration vector amplitude to obtain the vector amplitude VM[n], which is used to measure the instantaneous motion intensity; use a moving average filter with a certain window width (the technician in this field sets the window width to 50 samples) to smooth the VM curve to suppress short-term jitter and obtain the sliding amplitude; set intensity threshold 1, intensity threshold 2, and intensity threshold 3, which are set by technicians in this field to 0.1g, 0.3g, and 0.6g respectively, where g is the unit of acceleration; if the sliding amplitude is less than the intensity threshold 1, the running state is marked as static; if the intensity threshold 1 ≤ the sliding amplitude < the intensity threshold 2, the running state is marked as low-intensity activity; if the intensity threshold 2 ≤ the sliding amplitude < the intensity threshold 3, the running state is marked as medium-intensity activity; if the sliding amplitude ≥ the intensity threshold 3, the motion state is marked as high-intensity activity; associate the original heart rate data with the activity state labels (static, low-intensity activity, medium-intensity activity, and high-intensity activity) corresponding to the motion data collected in the same time period, thereby obtaining the original heart rate data and its associated activity state labels;

[0056] Step 3: Motion artifact removal:

[0057] The improved RLS adaptive filtering is adopted, and acceleration is used as the reference signal to construct the reference signal equation: r[n]=0.4×x[n]+0.3×y[n]+0.3×z[n], where r[n] is the acceleration reference signal, which is determined by the linear combination of the weighted three-axis components; then the equation is used to calculate the reference signal. The PPG signal and the reference signal are normalized to accelerate the convergence of the RLS algorithm and ensure the consistency of the variable dimensions, where μ PPG and σ PPG are the mean and standard deviation of the PPG signal, μ r and σ r are the mean and standard deviation of the reference signal r[n] respectively; and {r norm [n],......r norm [n-M+1]} Τ As the input vector of RLS filter, PPG norm [n] is used as the desired output vector (net PPG signal), and the relevant parameters of the RLS adaptive filter are set by the user: forgetting factor and filter order. The technicians set the forgetting factor to 0.99 and the filter order to 10.

[0058] Step 4: Extract the net PPG signal and its associated activity status label, perform dynamic heart rate calculation and status classification, and judge the heart rate risk accordingly:

[0059] Peak detection is performed on the clean PPG signal to extract peak indices to form a peak index sequence. The minimum and maximum allowable intervals between adjacent peaks are set, and impurity peaks outside the minimum sample interval to the maximum sample interval are filtered out. Those skilled in the art set the minimum sample interval to 30 sample points and the maximum sample interval to 200 sample points. A sliding window is used to calculate the real-time heart rate sequence, thereby obtaining the real-time heart rate sequence and its associated activity state label, which are then sent to the data encryption module in step S3.

[0060] Set each activity status label to correspond to a heart rate limit interval. Specifically, the heart rate upper limit for stillness is 40-100bpm, the heart rate limit interval for low intensity is 50-130bpm, and the heart rate limit interval for medium intensity is 70-170bpm. Extract the activity status label of each heart rate sequence and compare it with all the set activity status labels to match the corresponding heart rate limit. Perform distance analysis on the heart rate sequence and the corresponding heart rate limit interval. If all heart rate sequences belong to the corresponding heart rate limit interval, the heart rate is output as normal and no warning is required. If the heart rate of the heart rate sequence deviates from the upper and lower limits of the heart rate limit interval by 5-10bpm, a slight warning risk is output and the user is reminded to pay attention to heart rate changes. If the heart rate sequence deviates from the upper and lower limits of the heart rate limit interval by more than 10bpm and continues for 60 seconds, a serious warning risk is output and the user is reminded to rest.

[0061] By performing signal evaluation on the raw heart rate data to eliminate low-quality data, and combining it with real-time three-axis acceleration, the heart rate is matched with the current exercise intensity, effectively distinguishing physiological fluctuations from abnormal fluctuations. The four-level classification of static / low / medium / high can be used to personalize parameter adjustments for users in different usage scenarios to improve monitoring accuracy. Finally, based on the heart rate range of different exercise states, the three-level risk of "normal / mild warning / serious warning" is output in real time, and timely feedback is given to guide users to adjust or rest. This not only improves the reliability and accuracy of heart rate data, but also gives the system scene perception and intelligent warning capabilities.

[0062] S3: The data encryption module appends the continuously received real-time heart rate sequence and its associated activity status tags to the buffer in chronological order. When the number of buffered data items or the time span reaches the preset conditions, it encapsulates them into a data packet and then encrypts the data packet. Specifically:

[0063] Step 1: Call the JSON serialization function to convert the data packet into JSON text, flattening the data packet into a continuous string of characters. Then convert the JSON text into a byte stream to obtain a UTF-8 byte sequence. The UTF-8 byte sequence is a sequence of unsigned integers between 0 and 255.

[0064] Step 2: Pad the UTF-8 byte sequence so that the length of the padded UTF-8 byte sequence is an integer multiple of 16 bytes. After padding, use the AES-CBC algorithm to perform primary encryption on the padded UTF-8 byte sequence to obtain a pseudo-random byte stream with the same length as the original byte, and record it as the primary ciphertext;

[0065] Step 3: Group the values ​​in the primary ciphertext into two-dimensional coordinates, thereby splitting the primary ciphertext into several two-dimensional coordinates, establishing a two-dimensional rectangular coordinate system, and plotting coordinate points on the coordinate axis. The coordinate points are then connected in sequence using smooth curves according to the order of the two-dimensional coordinates in the primary ciphertext to form a ciphertext fingerprint.

[0066] Step 4: Extract the coordinate point set of the ciphertext fingerprint image, use the convex hull algorithm to calculate the minimum convex polygon of all points to output the minimum convex polygon, use it as the verification image of the ciphertext fingerprint image, and package the ciphertext fingerprint image and its corresponding verification image into the final ciphertext; specifically, the process of the convex hull algorithm outputting the minimum convex polygon is as follows:

[0067] Remove duplicate coordinate points from the ciphertext fingerprint. Find the point with the smallest ordinate among all coordinate points. If there are multiple points, select the point with the smallest abscissa as the vertex of the convex hull. Calculate the polar angle and distance from the convex hull vertex to the coordinate points other than the vertex. Use the convex hull vertex as the starting point and sort the other coordinate points in descending order of their corresponding polar angles. If two points have the same polar angle, sort the one with the larger distance first. The resulting coordinate sequence is recorded as {V0, V1, V2, V3...}.

[0068] Set up an empty "stack" (last-in, first-out structure) to store the coordinate sequence being constructed. Output three points V0, V1, and V2 in sequence and push them into the stack. Starting from the third point V3 after sorting, and continuing to the last point, loop through the following process:

[0069] 1-1: The top of the stack is recorded as B, and the next one is recorded as A;

[0070] 1-2: Try a new point, let the current point to be processed be C=V k , where k is the index of any point in the coordinate sequence except the vertex of the convex hull;

[0071] 1-3: Points A, B, and C form a vector and And use the cross product formula to calculate the vector and The turning amount Δ; if Δ>0, it means that from A to C via B is a left turn, which is consistent with the direction of the convex hull; if Δ≤0, it is a "straight line" or "right turn", which means that B cannot become a vertex of the convex hull and needs to "return";

[0072] 1-4: When Δ≤0, keep popping the top of the stack (i.e. discarding B) and recalculating the cross product with the next point C using the new top of the stack; when Δ>0 (turn left), stop popping and push C onto the stack, indicating that C is connected to the current convex hull boundary;

[0073] When all V k After all processing is completed, the remaining points in the stack are the convex hull vertex sequence arranged in a counterclockwise direction. Connecting these points in the stack end to end will obtain the minimum convex polygon containing all the original points - the convex hull;

[0074] The final ciphertext is sent to the first transceiver module, and uploaded to the second transceiver module in the cloud via the first transceiver module;

[0075] Through JSON serialization and AES-CBC encryption, the data packet is converted into a pseudo-random byte stream with the same length as the original bytes to obtain the primary ciphertext; the primary ciphertext is then converted into a set of coordinate points, and a ciphertext fingerprint is drawn based on it. The convex hull algorithm is used to extract the minimum convex deformation from it as the verification image. The fingerprint image has unique geometric features, and any bit-level modification can be detected by convex hull comparison; only the final ciphertext that passes the verification is written to the database to ensure data confidentiality and security. The ciphertext fingerprint image and convex hull verification are combined to achieve efficient detection and auditing of data tampering.

[0076] S4: The second transceiver module performs identity authentication based on the received final ciphertext: extracts the ciphertext fingerprint image and its corresponding verification image in the final ciphertext, and uses the convex hull algorithm to output the minimum convex hull of the ciphertext fingerprint image. If it is consistent with the verification image, Figure 1 If the result is the same, the verification is successful and the final ciphertext is sent to the database for storage; otherwise, the verification is judged to have failed and the final ciphertext is discarded.

[0077] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0078] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A method for intelligently processing heart rate monitoring data, comprising: S1, collects the user's original heart rate data and motion data in real time; It is characterized by further comprising: S2 receives the raw heart rate data and performs a quality check on it to filter out low-quality data. The raw heart rate data that meets the quality requirements is recorded as valid heart rate data. The motion data is then classified into state categories to accurately identify the motion state and assign a motion state label. The valid heart rate data monitored simultaneously is associated with the motion state label. Using the motion data as a reference signal, the RLS adaptive filter is used to remove motion artifacts from the valid heart rate data to output a net PPG signal. A heart rate risk warning is then issued based on the net PPG signal and its corresponding motion state label. The net PPG signal and its corresponding motion state label are encapsulated into a data packet and sent to S3. S3, based on the received data packet, primary encryption processing is performed to form a primary ciphertext, and then the numerical values ​​in the primary ciphertext are grouped into two groups to form a two-dimensional coordinate. In this way, the primary ciphertext can be split into several two-dimensional coordinates, and a two-dimensional rectangular coordinate system is established. Coordinate points of each two-dimensional coordinate are plotted on the coordinate axis, and the coordinate points are connected in sequence with smooth curves according to the order of each two-dimensional coordinate in the primary ciphertext to form a ciphertext fingerprint map, and the minimum convex polygon is extracted from it as the verification map of the ciphertext fingerprint map; the ciphertext fingerprint map and its corresponding verification map are encapsulated into the final ciphertext and uploaded.

2. The intelligent processing method for heart rate monitoring data according to claim 1, characterized in that: The process of quality checking the raw heart rate data to filter out low-quality data is as follows: The signal power of the PPG waveform signal is calculated based on the original heart rate data and recorded as P signal , noise power is recorded as P noise , according to the signal power P signal and noise power P noise By formula Calculate the signal-to-noise ratio (SNR) of the original heart rate data; The times corresponding to all peaks of the PPG waveform signal are detected and formed into a time series. The intervals between adjacent peaks are calculated based on the time series. The standard deviation formula is then used to calculate the standard deviation of the peak intervals based on the intervals between adjacent peaks. The signal-to-noise ratio and the standard deviation of the peak intervals are fused using a weighted formula to calculate the signal quality index (SQI). If the signal quality index (SQI) is greater than the preset quality threshold, it is marked as valid heart rate data; otherwise, it is marked as invalid heart rate data and discarded.

3. The intelligent processing method for heart rate monitoring data according to claim 1, characterized in that: The process of classifying motion data to accurately identify the motion state and assign motion state labels is as follows: Extract motion data, which includes three-axis accelerometer data recorded as {(x[n], y[n], z[n])}, where (x[n], y[n], z[n]) represents the acceleration of the three axes at the nth acquisition point; according to the formula Calculate the Euclidean norm of the three-axis acceleration vector amplitude to obtain the vector amplitude VM[n], which is used to measure the instantaneous motion intensity; use a moving average filter with a certain window width to smooth the VM curve to suppress short-term jitter and obtain the sliding amplitude; set the intensity threshold 1, intensity threshold 2, and intensity threshold 3; if the sliding amplitude is less than the intensity threshold 1, the running state is marked as static; if the intensity threshold 1 ≤ sliding amplitude < intensity threshold 2, the running state is marked as low-intensity activity; if the intensity threshold 2 ≤ sliding amplitude < intensity threshold 3, the running state is marked as medium-intensity activity; if the sliding amplitude is greater than or equal to the intensity threshold 3, the motion state is marked as high-intensity activity; The original heart rate data is associated with the activity state labels corresponding to the exercise data collected in the same time period, thereby obtaining the original heart rate data and its associated activity state labels.

4. The intelligent processing method for heart rate monitoring data according to claim 1, characterized in that: The motion artifact removal process is: The improved RLS adaptive filtering is adopted, and acceleration is used as the reference signal to construct the reference signal equation: r[n]=0.4×x[n]+0.3×y[n]+0.3×z[n], where r[n] is the acceleration reference signal, which is determined by the linear combination of the weighted three-axis components; then the equation is used to calculate the reference signal. The PPG signal and the reference signal are normalized to accelerate the convergence of the RLS algorithm and ensure the consistency of the variable dimensions, where μ PPG and σ PPG are the mean and standard deviation of the PPG signal, μ r and σ r are the mean and standard deviation of the reference signal r[n] respectively; and {r norm [n],......r norm [n-M+1]} Τ As the input vector of RLS filter, PPG norm [n] is used as the expected output vector, and the relevant parameters of the RLS adaptive filter are set by the user: the forgetting factor and the filter order.

5. The intelligent processing method for heart rate monitoring data according to claim 1, characterized in that: The process of heart rate risk warning is as follows: Peak detection is performed on the clean PPG signal to extract peak indices, forming a peak index sequence. The minimum and maximum allowable intervals between adjacent peaks are set, and the impurity peaks outside the minimum sample interval to the maximum sample interval are filtered out. A sliding window is used to calculate the real-time heart rate sequence. The real-time heart rate sequence and its associated activity state label are obtained and sent to step S3. Set each activity status label to correspond to a heart rate limit interval, extract the activity status label of each heart rate sequence, and compare it with all the set activity status labels to match the corresponding heart rate limit. Perform distance analysis on the heart rate sequence and the corresponding heart rate limit interval. If all heart rate sequences belong to the corresponding heart rate limit interval, the output is that the heart rate is normal and no warning is needed; if the heart rate of the heart rate sequence deviates from the upper and lower limits of the heart rate limit interval by 5-10bpm, a slight warning risk is output and the user is reminded to pay attention to heart rate changes; if the heart rate sequence deviates from the upper and lower limits of the heart rate limit interval by more than 10bpm and continues for 60 seconds, a serious warning risk is output and the user is reminded to rest.

6. The intelligent processing method for heart rate monitoring data according to claim 1, characterized in that: The primary encryption process is: Call the JSON serialization function to convert the data packet into JSON text, flattening the data packet into a continuous string of characters. Then convert the JSON text into a byte stream to obtain a UTF-8 byte sequence. The UTF-8 byte sequence is a sequence of unsigned integers between 0 and 255. The UTF-8 byte sequence is padded so that the length of the padded UTF-8 byte sequence is an integer multiple of 16 bytes. After the padding is completed, the padded UTF-8 byte sequence is encrypted using the AES-CBC algorithm to obtain a pseudo-random byte stream with the same length as the original byte, which is recorded as the primary ciphertext.

7. According to the method for intelligent processing of heart rate monitoring data of claim 6, the process of extracting the minimum convex polygon from the ciphertext fingerprint image is as follows: Remove duplicate coordinate points from the ciphertext fingerprint. Find the point with the smallest ordinate among all coordinate points. If there are two or more, select the point with the smallest abscissa as the vertex of the convex hull. Calculate the polar angle and distance from the convex hull vertex to the coordinate points other than the vertex. Use the convex hull vertex as the starting point and sort the other coordinate points in descending order of their corresponding polar angles. If two points have the same polar angle, sort the one with the larger distance first. The resulting coordinate sequence is recorded as {V0, V1, V2, V3...}. Set up an empty "stack" to store the coordinate sequence being constructed. Output three points V0, V1, and V2 in sequence and push them into the stack. Starting from the third point V3 after sorting, and continuing to the last point, loop through the following process: 1-1: The top of the stack is recorded as B, and the next one is recorded as A; 1-2: Try a new point, let the current point to be processed be C=V k , where k is the index of any point in the coordinate sequence except the vertex of the convex hull; 1-3: Points A, B, and C form a vector and And use the cross product formula to calculate the vector and The steering amount Δ; 1-4: When Δ≤0, keep popping the top of the stack, that is, discarding B, and recalculating the cross product with the next point C using the new top of the stack; when Δ>0, stop popping and push C into the stack, that is, connecting C to the current convex hull boundary; When all V k After all processing is completed, the remaining points in the stack are the convex hull vertex sequence arranged in a counterclockwise direction. Connecting these points in the stack end to end will obtain the minimum convex polygon containing all the original points, that is, the convex hull.

8. An intelligent processing system for heart rate monitoring data, characterized in that An intelligent processing method for heart rate monitoring data according to any one of claims 1 to 7, the system comprising: a portable heart rate monitoring device, a mobile phone, and a cloud; the mobile phone having a built-in preprocessing module, a data encryption module, and a first transceiver module; and the cloud comprising a second transceiver module and a database; The portable heart rate monitoring device is used to collect the user's original heart rate data and exercise data in real time; The preprocessing module receives raw heart rate data and performs quality checks on it to filter out low-quality data. Raw heart rate data that meets quality requirements is recorded as valid heart rate data. The motion data is then classified to accurately identify the motion state and assign a motion state label. The simultaneously monitored valid heart rate data is associated with the motion state label. Using the motion data as a reference signal, the RLS adaptive filter is used to remove motion artifacts from the valid heart rate data and output a net PPG signal. A heart rate risk warning is then issued based on the net PPG signal and its corresponding motion state label. The net PPG signal and its corresponding motion state label are encapsulated into a data packet and sent to the data encryption module. The data encryption module performs primary encryption processing on the received data packet to form a primary ciphertext, and then groups the values ​​in the primary ciphertext into two-dimensional coordinates, thereby splitting the primary ciphertext into several two-dimensional coordinates, establishing a two-dimensional rectangular coordinate system, and plotting coordinate points of each two-dimensional coordinate on the coordinate axis. The coordinate points are then connected in sequence using smooth curves according to the order of each two-dimensional coordinate in the primary ciphertext to form a ciphertext fingerprint graph, and the minimum convex polygon is extracted from it as a verification graph of the ciphertext fingerprint graph; the ciphertext fingerprint graph and its corresponding verification graph are encapsulated into a final ciphertext, and sent to the first transceiver module, which is then uploaded to the second transceiver module in the cloud via the first transceiver module; The second transceiver module performs ciphertext verification on the uploaded final ciphertext, and stores it if the verification passes; otherwise, it is discarded.

Images