Intelligent rehabilitation pectoral girdle regulation and control method based on real-time data acquisition

Abstract

The invention relates to the technical field of general control systems, in particular to an intelligent rehabilitation pectoral girdle regulation and control method based on real-time data acquisition, which comprises the following steps: determining a heart rate abnormity time period, and determining the heart rate abnormity time period according to the dissimilarity between the respiration abnormity change and the pressure change of the heart rate in the heart rate abnormity time period; screening out a plurality of suspected wound influence sensors from the pressure sensors, and screening out a plurality of wound influence sensors from the suspected wound influence sensors according to the distance between the suspected wound influence sensors and the similarity between the pressure sequential sequences corresponding to the adjacent suspected wound influence sensors; a wound sensor is screened out from the wound surface influence sensors, a trend value is determined, and therefore tightness adjustment is conducted on the adjusting structures corresponding to the wound surface influence sensors. The rehabilitation pectoral girdle at the wound is adjusted in a self-adaptive mode, unnecessary compression on the wound surface is avoided, and therefore recovery of the wound is better promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of general control systems, and particularly to an intelligent rehabilitation chest band regulation method based on real-time data acquisition. Background Art

[0002] An intelligent rehabilitation chest band is a device integrated with multiple sensors for real-time monitoring of a patient's physiological parameters and motion data, which are crucial for personalized treatment and real-time adjustment during the rehabilitation process. The intelligent rehabilitation chest band regulation method based on real-time data acquisition not only improves the efficiency and safety of rehabilitation treatment but also provides a more personalized and flexible rehabilitation plan for patients.

[0003] Existing problems: After breast surgery, patients need to wear a rehabilitation chest band for wound recovery. The existing rehabilitation chest band adjustment method by setting thresholds, although achieving a certain degree of fixation and protection of the chest, fails to fully consider the specific location of the wound on the breast after surgery. When the adjustment is inappropriate, it is very likely that a certain wound site will be overly tightened or the support is insufficient, thus having an adverse impact on the recovery. Summary of the Invention

[0004] The present invention provides an intelligent rehabilitation chest band regulation method based on real-time data acquisition to solve the existing problems.

[0005] The intelligent rehabilitation chest band regulation method based on real-time data acquisition of the present invention adopts the following technical solutions:

[0006] An embodiment of the present invention provides an intelligent rehabilitation chest band regulation method based on real-time data acquisition, which includes the following steps:

[0007] Obtain the heart rate time series of a patient after breast surgery, the pressure time series corresponding to each pressure sensor in the worn intelligent rehabilitation chest band, and the distances between the pressure sensors; each pressure sensor corresponds to an adjustment structure;

[0008] According to the magnitude of the heart rate in the heart rate time series, determine the abnormal heart rate period and the respiratory abnormality of each heart rate within the abnormal heart rate period; according to the dissimilarity between the change in the respiratory abnormality of the heart rate and the pressure change within the abnormal heart rate period, screen out several suspected wound-affecting sensors from the pressure sensors;

[0009] According to the distance between the suspected wound-affecting sensors and the similarity between the pressure time series corresponding to the adjacent suspected wound-affecting sensors, screen out several wound-affecting sensors from the suspected wound-affecting sensors;

[0010] Select wound sensors from the wound-affecting sensors according to the similarity between the pressure time series corresponding to the wound-affecting sensors; determine the trend value according to the change in pressure in the pressure time series corresponding to the wound sensors; adjust the tightness of the adjustment structure corresponding to the wound-affecting sensors according to the magnitude of the trend value.

[0011] Further, the steps for determining the heart rate abnormal period and the respiratory abnormality of each heart rate within the heart rate abnormal period are as follows:

[0012] In the heart rate time series, a heart rate less than the lower limit of the preset normal heart rate is recorded as a low heart rate, and a heart rate greater than the upper limit of the preset normal heart rate is recorded as a high heart rate;

[0013] Calculate the ratio of the difference between each high heart rate and the upper limit of the preset normal heart rate to the upper limit of the preset normal heart rate as the respiratory abnormality of each high heart rate;

[0014] Calculate the ratio of the difference between the lower limit of the preset normal heart rate and each low heart rate to the lower limit of the preset normal heart rate as the respiratory abnormality of each low heart rate;

[0015] The time period where the heart rate time series segment composed of adjacent low heart rates is located is recorded as the low heart rate period, the time period where the heart rate time series segment composed of adjacent high heart rates is located is recorded as the high heart rate period, and the low heart rate period and the high heart rate period are collectively referred to as the heart rate abnormal period.

[0016] Further, the steps for screening out several suspected wound-affecting sensors from the pressure sensors are as follows:

[0017] If the current moment is within the heart rate abnormal period, within the heart rate abnormal period where the current moment is located, obtain the respiratory abnormality time series segment composed of the respiratory abnormalities of all heart rates and the pressure time series segment corresponding to each pressure sensor;

[0018] A pressure sensor with a cosine similarity less than the preset similarity threshold between the respiratory abnormality time series segment and the pressure time series segment is recorded as a suspected wound-affecting sensor.

[0019] Further, the steps for screening out several wound-affecting sensors from the suspected wound-affecting sensors are as follows:

[0020] Assign a label value to each suspected wound-affecting sensor according to the distance between the suspected wound-affecting sensors;

[0021] Select the optimal proximity value B from the preset proximity range sequence according to the label values of the suspected wound-affecting sensors;

[0022] Obtain the top B suspected wound-affecting sensors that are closest to the y-th suspected wound-affecting sensor, denoted as the optimal neighboring sensors. Denote the mean of the cosine similarities between the pressure time series of the y-th suspected wound-affecting sensor and all the optimal neighboring sensors as the pressure distribution uniformity of the area where the y-th suspected wound-affecting sensor is located.

[0023] Taking each suspected wound-affecting sensor as a data point, using the label value of each suspected wound-affecting sensor as the horizontal axis and the inverse value of the pressure distribution uniformity of the area where each suspected wound-affecting sensor is located as the vertical axis, construct a sensor scatter plot corresponding to all suspected wound-affecting sensors.

[0024] Use the density-based spatial clustering algorithm to perform clustering operations on all data points in the sensor scatter plot to obtain several clustering clusters.

[0025] In the sensor scatter plot, divide the target area. Denote the ratio of the number of data points in the n-th clustering cluster that are within the target area to the number of data points in the n-th clustering cluster as the proportion of the number of the n-th clustering cluster.

[0026] Denote the suspected wound-affecting sensors in the clustering cluster with the largest proportion of the number as the wound-affecting sensors.

[0027] Further, the specific steps of assigning a label value to each suspected wound-affecting sensor according to the distance between suspected wound-affecting sensors are as follows:

[0028] Denote the sum of the distances between each suspected wound-affecting sensor and all other suspected wound-affecting sensors as the distance threshold of each suspected wound-affecting sensor.

[0029] Denote the suspected wound-affecting sensor with the smallest distance threshold as the target sensor.

[0030] Assign a preset constant as the label value to the target sensor.

[0031] The label value assigned to each suspected wound-affecting sensor other than the target sensor is: the distance between each suspected wound-affecting sensor other than the target sensor and the target sensor.

[0032] Further, the specific steps of screening the optimal neighboring value B from the preset neighboring range sequence according to the label value of the suspected wound-affecting sensor are as follows:

[0033] Select any data value A in the preset adjacent range sequence, and obtain the first A suspected wound-affecting sensors that are closest to any one of the suspected wound-affecting sensors, denoted as adjacent sensors. Combine all the adjacent sensors and the said any one suspected wound-affecting sensor into a sensor group;

[0034] Statistically analyze the variance of the tag values of all suspected wound-affecting sensors in all sensor groups corresponding to A, denoted as the first variance;

[0035] Statistically analyze the variance of the tag values of all suspected wound-affecting sensors in each sensor group corresponding to A, denoted as the second variance. Take the sum value of the second variances of all sensor groups as the third variance;

[0036] Denote the difference obtained by subtracting the third variance from the first variance as the difference threshold corresponding to A;

[0037] Denote the data value in the preset adjacent range sequence corresponding to the minimum difference threshold as the optimal adjacent value B.

[0038] Furthermore, the specific steps for dividing the target area in the sensor scatter plot are as follows:

[0039] In the sensor scatter plot, sort the horizontal axis coordinate values and vertical axis coordinate values of all data points in ascending order respectively to obtain an ascending sequence of horizontal axis coordinate values and an ascending sequence of vertical axis coordinate values;

[0040] Use the APCA segmentation algorithm to perform segmentation operations on the ascending sequence of horizontal axis coordinate values and the ascending sequence of vertical axis coordinate values respectively. Denote the data point corresponding to the last horizontal axis coordinate value in the first segment of the ascending sequence of horizontal axis coordinate values as the first target data point; Denote the data point corresponding to the last vertical axis coordinate value in the first segment of the ascending sequence of vertical axis coordinate values as the second target data point;

[0041] Denote the straight line passing through the first target data point and perpendicular to the horizontal axis as the first straight line;

[0042] Denote the straight line passing through the second target data point and perpendicular to the vertical axis as the second straight line;

[0043] Denote the closed rectangular area enclosed by the first straight line, the second straight line, the horizontal axis, and the vertical axis as the target area.

[0044] Furthermore, the specific steps for screening out wound sensors from the wound-affecting sensors are as follows:

[0045] Obtain the sum value of the cosine similarities between the pressure time series sequences corresponding to the m-th wound-affecting sensor and all other wound-affecting sensors, denoted as the wound possibility of the m-th wound-affecting sensor;

[0046] The wound impact sensor corresponding to the maximum wound possibility is denoted as the wound sensor.

[0047] Furthermore, the determination of the trend value includes the following specific steps:

[0048] In the pressure time series corresponding to the wound sensor, taking each pressure data as a data point, with the sequence number value of each pressure data as the horizontal axis and the pressure data value of each pressure data as the vertical axis, construct a pressure scatter plot of the pressure time series corresponding to the wound sensor;

[0049] In the pressure scatter plot, take all data points as the input of principal component analysis, and output several projection values and the projection vectors corresponding to each projection value. Among them, each projection vector corresponds to a horizontal axis coordinate value and a vertical axis coordinate value;

[0050] Denote the projection vector corresponding to the maximum projection value as the trend vector;

[0051] Denote the ratio of the arctangent angle value corresponding to the ratio of the vertical axis coordinate value to the horizontal axis coordinate value of the trend vector to the preset angle value as the trend value.

[0052] Furthermore, the loosening and tightening adjustment of the adjustment structure corresponding to the wound impact sensor includes the following specific steps:

[0053] Calculate the product of the sum value of 1 and the trend value and the preset reference adjustment amount, and denote it as the adjustment amount for executing the adjustment structure;

[0054] Denote the adjustment structure corresponding to the wound impact sensor as the target adjustment structure;

[0055] If the current moment is within the high heart rate period, then when the trend value is greater than the preset loosening and tightening threshold, use the adjustment amount of the execution adjustment structure to relax the target adjustment structure. When the trend value is less than or equal to the preset loosening and tightening threshold, use the preset reference adjustment amount to relax the target adjustment structure;

[0056] If the current moment is within the low heart rate period, then when the trend value is greater than the preset loosening and tightening threshold, use the adjustment amount of the execution adjustment structure to tighten the target adjustment structure. When the trend value is less than or equal to the preset loosening and tightening threshold, use the preset reference adjustment amount to tighten the target adjustment structure.

[0057] The beneficial effects of the technical solution of the present invention are:

[0058] In an embodiment of the present invention, the heart rate time series of a breast cancer postoperative patient, the pressure time series corresponding to each pressure sensor in the wearable intelligent rehabilitation chest strap, and the distances between the pressure sensors are obtained to determine the heart rate abnormal period, thereby determining whether to perform adjustment and reducing unnecessary regulation. According to the dissimilarity between the abnormal respiratory changes of the heart rate and the pressure changes during the heart rate abnormal period, several suspected wound-affecting sensors are screened out from the pressure sensors, thereby obtaining the pressure sensors at the trauma area with high probability, ensuring the accuracy of subsequent regional adjustment. According to the distances between the suspected wound-affecting sensors and the similarity between the pressure time series corresponding to the adjacent suspected wound-affecting sensors, several wound-affecting sensors are screened out from the suspected wound-affecting sensors, and then wound sensors are screened out from the wound-affecting sensors. According to the pressure changes in the pressure time series corresponding to the wound sensors, a trend value is determined, thereby determining the pressure sensors that have a greater impact on the wound. According to the pressure data collected by them, the adjustment amount is determined, ensuring the accuracy of the adjustment. According to the magnitude of the trend value, the adjustment structure corresponding to the wound-affecting sensor is adjusted in tightness. Thus, the present invention adaptively adjusts the rehabilitation chest strap at the wound, avoids unnecessary compression on the wound surface, and thus better promotes the recovery of the wound. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0060] Figure 1 is a flowchart of the steps of the intelligent rehabilitation chest strap control method based on real-time data collection of the present invention;

[0061] Figure 2 is a schematic diagram of the intelligent rehabilitation chest strap;

[0062] Figure 3 is a schematic diagram of the target area. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the drawings and preferred embodiments, details the specific implementation manners, structures, features, and effects of the intelligent rehabilitation chest strap control method based on real-time data collection proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this invention belongs.

[0065] The following specifically describes the specific solution of the intelligent rehabilitation chest belt regulation method based on real-time data acquisition provided by the present invention in conjunction with the accompanying drawings.

[0066] Please refer to Figure 1 , which shows a flowchart of the steps of the intelligent rehabilitation chest belt regulation method based on real-time data acquisition provided by an embodiment of the present invention. The method includes the following steps:

[0067] Step S001: Obtain the heart rate time series of the patient after breast surgery, the pressure time series corresponding to each pressure sensor in the worn intelligent rehabilitation chest belt, and the distance between the pressure sensors; each pressure sensor corresponds to an adjustment structure.

[0068] It should be noted that: in this embodiment, by analyzing the pressure change situation at the wound position of the patient's wound surface, the corresponding adjustment coefficient is obtained. When the rehabilitation chest belt is too tight or too loose, the rehabilitation chest belt can be better adjusted to avoid unnecessary compression on the wound surface or enhance the supporting effect of the rehabilitation chest belt on the wound surface, thereby better promoting the recovery of the wound.

[0069] Obtain the heart rate time series of the patient after breast surgery, the pressure time series corresponding to each of several pressure sensors in the worn intelligent rehabilitation chest belt, and the distance between the pressure sensors, where each pressure sensor corresponds to an adjustment structure.

[0070] It should be noted that: In this embodiment, the acquisition frequency of the heart rate is once per minute, and the acquisition frequency of the pressure is once per second. The data of the two dimensions are collected simultaneously, and the following description is given by taking this as an example. Each data in the time series corresponds to a timestamp. Inside the intelligent rehabilitation chest strap, a plurality of pressure sensors are distributed along the chest contour for real-time monitoring of the pressure at different positions. The positions of these sensors can be arranged according to the position of the postoperative wound of the patient's breast to ensure that the wound area and its periphery can be covered. Moreover, a plurality of adjustment structures are provided on the chest strap, and the adjustment structures are such as a screw driven by a motor, a gear or an air bag, etc. The adjustment structure is connected to the pressure sensor in the corresponding area, and then automatically adjusts the tightness of the chest strap in this area. For example, when a regulation instruction is received, the motor drives the screw or the gear to tighten or loosen the chest strap. The chest strap is divided into the upper chest, the lower chest and the waist, and each part has its own adjustment structure for independent tightness adjustment of different areas. This design can help the patient obtain better support and comfort in specific body parts, while allowing other parts to maintain appropriate flexibility. The distance between the pressure sensors is the Euclidean distance between the position coordinates of the positions where the pressure sensors are located. The heart rate is measured by using a photoelectric sensor worn on the wrist to detect the blood flow change. This kind of sensor is usually integrated in a smart watch or a wristband, and it should be ensured that the sensor is closely attached to the skin when worn. Thus, the relevant physiological information of the patient is detected and collected by the sensor, and then the intelligent rehabilitation chest strap is intelligently regulated. Schematic diagram of the intelligent rehabilitation chest strap, as Figure 2 shown.

[0071] Step S002: According to the magnitude of the heart rate in the heart rate time series, determine the heart rate abnormal period and the respiratory abnormality of each heart rate within the heart rate abnormal period; according to the dissimilarity between the change of the respiratory abnormality of the heart rate and the pressure change within the heart rate abnormal period, screen out several suspected wound-affecting sensors from the pressure sensors.

[0072] It should be noted that during the postoperative rehabilitation process of the breast, the traditional rehabilitation chest strap adjustment method mainly adjusts the tightness according to the patient's heart rate and respiratory rate to relieve the pressure on the wound, promote the recovery and shaping of chest tissues, but its detection of the wound is not accurate enough. In this embodiment, by setting pressure sensors at multiple positions on the rehabilitation chest strap, the pressure data of each part are monitored in real time, and then these data are analyzed to determine the key sensor position that can best reflect the recovery degree of the wound surface. Based on this position information, the adjustment mechanism is adjusted for tightness. For example, when the value of the pressure sensor at a certain position on the rehabilitation chest strap is large due to reasons such as the patient's nervousness, a larger relaxation coefficient will be automatically assigned to the adjustment mechanism near the wound surface position. This means that the chest strap at this part will be appropriately relaxed to avoid unnecessary pressure on the wound surface due to temporary changes in the patient's physiological parameters, thus ensuring the recovery of the wound surface in a suitable environment. Among them, by considering the adjustment mechanism of the wound surface, the problem of pressure discomfort in the wound area caused by overall adjustment can be effectively avoided, providing a more accurate and comfortable nursing plan for postoperative recovery.

[0073] Preferably, in an embodiment of the present invention, the acquisition method of the suspected wound surface affecting sensor includes:

[0074] Taking the preset lower limit of normal heart rate as 60 beats per minute and the preset upper limit of normal heart rate as 100 beats per minute as an example for description.

[0075] In the heart rate time series, the heart rate less than the preset lower limit of normal heart rate is recorded as low heart rate, and the heart rate greater than the preset upper limit of normal heart rate is recorded as high heart rate.

[0076] Calculate the difference between each high heart rate and the preset upper limit of normal heart rate, and take the ratio of this difference to the preset upper limit of normal heart rate as the respiratory abnormality of each high heart rate.

[0077] Calculate the difference between the preset lower limit of normal heart rate and each low heart rate, and take the ratio of this difference to the preset lower limit of normal heart rate as the respiratory abnormality of each low heart rate.

[0078] In the heart rate time series, the time period where the heart rate time series segment composed of adjacent low heart rates is located is recorded as the low heart rate period, the time period where the heart rate time series segment composed of adjacent high heart rates is located is recorded as the high heart rate period, and the low heart rate period and the high heart rate period are collectively referred to as the heart rate abnormal period.

[0079] It should be noted that after breast surgery, the doctor sets a suitable range for the patient's heart rate according to the specific situation of the patient. When the patient is taking certain drugs, such as beta blockers, etc., the drug effect causes the heart rate to be lower than the set lower limit. At this time, it is necessary to adjust the tightness of the chest strap to ensure proper fixation and support. When the patient exercises, the rehabilitation chest strap may cause additional pressure on the patient at this time, and it is necessary to appropriately loosen the chest strap to reduce the restriction on the patient's movement. Therefore, when the real-time heart rate data value of the patient is not within the normal range, it is necessary to adjust the tightness of the chest strap. For a certain position on the chest, if there is no wound on the chest, the pressure is proportional to the respiratory abnormality. However, due to the existence of the wound, in the chest area affected by the wound, the pressure-bearing capacity at this place is weak, which in turn leads to a larger deformation of the chest at this place compared to other chest areas, and thus the positive correlation between the pressure value collected by the sensor at this place and the respiratory abnormality is weak. Therefore, in this embodiment, first find the chest area affected by the wound, which is obtained through the sensor data in different areas and the respiratory abnormality of the patient.

[0080] If the current moment is not within the heart rate abnormal period, the tightness of the rehabilitation chest strap is not adjusted.

[0081] If the current moment is within the heart rate abnormal period, then within the heart rate abnormal period where the current moment is located, obtain the respiratory abnormality time series segment composed of the respiratory abnormalities of all heart rates and the pressure time series segment corresponding to each pressure sensor.

[0082] The preset similarity threshold is 0.7, and this is used as an example for description.

[0083] Within the heart rate abnormal period where the current moment is located, calculate the cosine similarity between the respiratory abnormality time series segment and the pressure time series segments corresponding to all pressure sensors respectively. Mark the pressure sensors with the cosine similarity between the respiratory abnormality time series segment and the pressure time series segment less than the preset similarity threshold as suspected wound-affected sensors.

[0084] It should be noted that: Since the acquisition frequencies of the heart rate and the pressure are different, first use the interpolation algorithm to make the respiratory abnormality time series segment and the pressure time series segment of equal length, and then calculate the cosine similarity between the two sequences. The value range of the cosine similarity is between -1 and 1. Among them, the calculation of the cosine similarity and the interpolation algorithm are both well-known technologies, and the specific methods are not introduced here. The greater the cosine similarity, the more similar the two sequences are, and the smaller the similarity, the more likely it is to be the sensor at the suspected wound-affected place.

[0085] Step S003: According to the distance between the suspected wound-affected sensors and the similarity between the pressure time series corresponding to the adjacent suspected wound-affected sensors, screen out several wound-affected sensors from the suspected wound-affected sensors.

[0086] It should be noted that due to the influence of the chest shape, in the area where the suspected wound-affecting sensor is located, some areas are close to the bone and some are far from the bone, resulting in an inaccurate area where the wound-affecting sensor is located, which contains many actually non-wound-affecting areas. The truly wound-affecting areas are close in spatial distance, and the pressure change conditions of the corresponding pressure sensors at different times are similar. Since a single sensor can only represent the pressure condition at a single position, it is necessary to combine adjacent sensors to obtain the pressure condition of this sensor. At this time, the obtained pressure condition can represent the pressure conditions of the areas where multiple sensor positions are located.

[0087] Preferably, in an embodiment of the present invention, the acquisition method of the wound-affecting sensor includes:

[0088] Preset the adjacent range sequence as {2, 3, 4, 5, 6, 7, 8}, and take this as an example for description.

[0089] Taking any data value A in the preset adjacent range sequence as an example, obtain the first A suspected wound-affecting sensors that are closest to the y-th suspected wound-affecting sensor, denoted as adjacent sensors, and form a sensor group by combining all the adjacent sensors and the y-th suspected wound-affecting sensor.

[0090] According to the above method, obtain the sensor group formed by each suspected wound-affecting sensor and all its adjacent sensors, and obtain several sensor groups corresponding to A. If the number of adjacent sensors is less than A, then take the existing adjacent sensors for subsequent analysis.

[0091] Among all the suspected wound-affecting sensors, record the sum of the distances between each suspected wound-affecting sensor and all other suspected wound-affecting sensors as the distance threshold of each suspected wound-affecting sensor.

[0092] Record the suspected wound-affecting sensor corresponding to the minimum distance threshold as the target sensor.

[0093] Among them, if there are multiple minimum distance thresholds, any one can be taken as an example for analysis.

[0094] Preset the constant as 0, and take this as an example for description.

[0095] Assign the label value of the target sensor as the preset constant, and assign the label value of the z-th suspected wound-affecting sensor other than the target sensor as: the distance between the z-th suspected wound-affecting sensor and the target sensor. Thus, the label value of each suspected wound-affecting sensor is obtained.

[0096] Statistical variance of the label values of all suspected wound-affecting sensors in all sensor groups corresponding to A, denoted as the first variance.

[0097] Statistically calculate the variance of the tag values of all suspected wound-affecting sensors in each sensor group corresponding to A, denoted as the second variance. Take the sum value of the second variances of all sensor groups as the third variance.

[0098] Denote the difference obtained by subtracting the third variance from the first variance as the difference threshold corresponding to A.

[0099] It should be noted that: the smaller the difference threshold, the greater the difference between different combinations of regions corresponding to A, that is, the regions where the sensor combinations are located are affected by external factors. For example, when the patient moves and causes a large pressure on the chest area, the changes are more similar, while the changes of different sensor combinations are more different.

[0100] Obtain the difference threshold corresponding to each data value in the preset adjacent range sequence in the above manner.

[0101] Denote the data value in the preset adjacent range sequence corresponding to the smallest difference threshold as the optimal adjacent value B.

[0102] It should be noted that: if there are multiple smallest difference thresholds, take the data value corresponding to the smallest difference threshold that appears first in the preset adjacent range sequence as the optimal adjacent value B. The optimal adjacent value B means that when the number of suspected wound-affecting sensors adjacent to each suspected wound-affecting sensor is B, the best segmentation area surface can be formed, and the affected degrees of the same segmentation area surface are similar. For example, the distances between all sensors in a certain sensor combination and a certain bone are similar, and thus when affected by external factors, the pressure changes are similar. Otherwise, it is likely to be affected by a wound, resulting in different pressure change situations.

[0103] Obtain the first B suspected wound-affecting sensors that are closest to the y-th suspected wound-affecting sensor, denoted as the optimal adjacent sensors. Obtain the cosine similarity between the pressure time series of the y-th suspected wound-affecting sensor and the pressure time series corresponding to each optimal adjacent sensor, and then obtain the mean value of the cosine similarities between the pressure time series of the y-th suspected wound-affecting sensor and the pressure time series corresponding to all optimal adjacent sensors, denoted as the pressure distribution uniformity of the area where the y-th suspected wound-affecting sensor is located.

[0104] Among them, the pressure distribution uniformity represents the uniformity of the pressure distribution in the area where the y-th suspected wound-affecting sensor and the optimal adjacent sensors are located.

[0105] Obtain the pressure distribution uniformity of the area where each suspected wound-affecting sensor is located in the above manner.

[0106] Taking each suspected wound-affecting sensor as a data point, using the tag value of each suspected wound-affecting sensor as the horizontal axis, and using the inverse value of the pressure distribution uniformity in the area where each suspected wound-affecting sensor is located as the vertical axis, construct a sensor scatter plot corresponding to all suspected wound-affecting sensors.

[0107] Among them, the difference obtained by subtracting the pressure distribution uniformity in the area where each suspected wound-affecting sensor is located from 1 is used as the inverse value of the pressure distribution uniformity in the area where each suspected wound-affecting sensor is located.

[0108] In the sensor scatter plot, use the density-based spatial clustering algorithm to perform clustering operations on all data points to obtain several clustering clusters.

[0109] Among them, the density-based spatial clustering algorithm (DBSCAN density clustering algorithm, Density-Based Spatial Clustering of Applications with Noise) is a well-known technology, and the specific method will not be introduced here.

[0110] In the sensor scatter plot, arrange the horizontal axis coordinate values of all data points in ascending order to obtain an ascending sequence of horizontal axis coordinate values, and arrange the vertical axis coordinate values of all data points in ascending order to obtain an ascending sequence of vertical axis coordinate values.

[0111] Use the APCA segmentation algorithm to perform segmentation operations on the ascending sequence of horizontal axis coordinate values to obtain several ascending sequence segments of horizontal axis coordinate values, and record the data point corresponding to the last horizontal axis coordinate value in the first ascending sequence segment of horizontal axis coordinate values as the first target data point.

[0112] Use the APCA segmentation algorithm to perform segmentation operations on the ascending sequence of vertical axis coordinate values to obtain several ascending sequence segments of vertical axis coordinate values, and record the data point corresponding to the last vertical axis coordinate value in the first ascending sequence segment of vertical axis coordinate values as the second target data point.

[0113] Among them, the APCA segmentation algorithm (Adaptive Piecewise Constant Approximation) is a well-known technology, and the specific method will not be introduced here.

[0114] In the sensor scatter plot, record the straight line passing through the first target data point and perpendicular to the horizontal axis as the first straight line. Record the straight line passing through the second target data point and perpendicular to the vertical axis as the second straight line, and record the closed rectangular area enclosed by the first straight line, the second straight line, the horizontal axis, and the vertical axis as the target area.

[0115] Among them, the schematic diagram of the target area is as Figure 3 shown.Figure 3 The horizontal axis (X-axis) and the vertical axis (Y-axis) respectively represent the tag value of each suspected wound-affecting sensor and the inverse proportional value of the pressure distribution uniformity in the area where each suspected wound-affecting sensor is located. Each circle represents a suspected wound-affecting sensor. The dashed line perpendicular to the horizontal axis and the dashed line perpendicular to the vertical axis are the first straight line and the second straight line respectively.

[0116] The ratio of the number of data points within the target area in the nth clustering cluster to the number of data points in the nth clustering cluster is denoted as the proportion of the number of the nth clustering cluster.

[0117] Among all the clustering clusters, the suspected wound-affecting sensors in the clustering cluster corresponding to the largest proportion of the number are denoted as wound-affecting sensors.

[0118] It should be noted that: if there are multiple largest proportions of the number, the suspected wound-affecting sensors in the clustering clusters corresponding to these largest proportions of the number are all denoted as wound-affecting sensors. When the patient's heart rate is abnormal, the data collected by the wound-affecting sensors can represent the information that the wound is affected, and then the position of the wound can be indicated, that is, the sensor that best matches the wound position can achieve precise positioning and pressure monitoring of the wound position.

[0119] Step S004: Screen out the wound sensors from the wound-affecting sensors according to the similarity between the pressure time series corresponding to the wound-affecting sensors; determine the trend value according to the change of the pressure in the pressure time series corresponding to the wound sensors; adjust the tightness of the adjustment structure corresponding to the wound-affecting sensors according to the magnitude of the trend value.

[0120] Preferably, in an embodiment of the present invention, the method for adjusting the tightness of the adjustment structure includes:

[0121] Obtain the cosine similarity between the pressure time series corresponding to the mth wound-affecting sensor and each other wound-affecting sensor, and then obtain the sum value of the cosine similarities between the pressure time series corresponding to the mth wound-affecting sensor and all other wound-affecting sensors, which is denoted as the wound possibility of the mth wound-affecting sensor.

[0122] Among all the wound-affecting sensors, the wound-affecting sensor corresponding to the largest wound possibility is denoted as the wound sensor.

[0123] Among them, if there are multiple largest wound possibilities, any one is taken as an example for subsequent analysis. The pressure data corresponding to the wound sensor can best reflect the influence of the patient's current heart rate change on the wound. If the increase degree of the pressure value at the wound position is large, and the overall current requirement is to loosen, a larger relaxation coefficient is assigned to the wound-affecting sensor. If the overall current requirement is to tighten, a larger tightening coefficient is assigned to the wound-affecting sensor.

[0124] In the pressure time series corresponding to the wound sensor, with each pressure data as a data point, the ordinal value of each pressure data as the horizontal axis, and the value of each pressure data as the vertical axis, a pressure scatter plot of the pressure time series corresponding to the wound sensor is constructed.

[0125] In the pressure scatter plot, all data points are used as the input of principal component analysis, and several projection values and the projection vectors corresponding to each projection value are output. Each projection vector corresponds to a horizontal axis coordinate value and a vertical axis coordinate value.

[0126] Among them, principal component analysis (PCA, Principal Component Analysis) is a well-known technology, and the specific method will not be introduced here. Each projection vector (i.e., the principal component vector) output by principal component analysis is a unit vector, and the horizontal axis coordinate value and the vertical axis coordinate value of each projection vector are the components of the projection vector on the horizontal axis and the vertical axis, respectively.

[0127] The preset angle value is 90 degrees, and this is used as an example for description.

[0128] The projection vector corresponding to the maximum projection value is denoted as the trend vector, and the arctangent angle value corresponding to the ratio of the vertical axis coordinate value to the horizontal axis coordinate value of the trend vector is obtained. The ratio of the arctangent angle value to the preset angle value is denoted as the trend value.

[0129] Among them, the arctangent angle value refers to the angle value calculated by the arctangent function (arctan), which is a well-known technology, and the specific method will not be introduced here. The larger the trend value, the greater the increase degree of the pressure value at the wound position. If the current overall situation requires relaxation, a larger relaxation coefficient is assigned to the wound-affecting sensor to ensure that the patient's current behavior will not cause too much pressure on the wound and affect recovery. If the current overall situation requires tightening, a larger tightening coefficient is assigned to the wound-affecting sensor to ensure fixation and support.

[0130] The adjustment structure corresponding to the wound-affecting sensor is denoted as the target adjustment structure.

[0131] The preset tightness threshold is 0.5, and the preset reference adjustment amount is 3 cm, and this is used as an example for description.

[0132] If the current moment is within a high heart rate period, when the trend value is greater than the preset tightness threshold, calculate the sum value of 1 and the trend value, and denote the product of the sum value and the preset reference adjustment amount as the adjustment amount of the execution adjustment structure. Use the adjustment amount of the execution adjustment structure to perform a relaxation adjustment on the target adjustment structure. When the trend value is less than or equal to the preset tightness threshold, use the preset reference adjustment amount as the adjustment amount of the execution adjustment structure to perform a relaxation adjustment on the target adjustment structure.

[0133] If the current moment is within the low heart rate period, when the trend value is greater than the preset tightness threshold, calculate the sum value of 1 and the trend value, and record the product of this sum value and the preset reference adjustment amount as the adjustment amount of the execution adjustment structure. Use the adjustment amount of the execution adjustment structure to perform a tightening adjustment on the target adjustment structure. When the trend value is less than or equal to the preset tightness threshold, use the preset reference adjustment amount as the adjustment amount of the execution adjustment structure to perform a tightening adjustment on the target adjustment structure.

[0134] It should be noted that: the preset reference adjustment amount is the change amount of the chest strap length. If there are multiple target adjustment structures, the multiple target adjustment structures are adjusted simultaneously. Therefore, when a patient takes certain drugs, the drug effect causes the heart rate to be lower than the set lower limit. At this time, it is necessary to adjust the tightness of the chest strap to ensure proper fixation and support. When the patient exercises, the heart rate may be too high. At this time, the rehabilitation chest strap may cause additional pressure on the patient, and it is necessary to appropriately loosen the chest strap to reduce the restriction on the patient's movement.

[0135] So far, the present invention is completed.

[0136] In summary, in the embodiment of the present invention, the heart rate time series of the breast cancer postoperative patient, the pressure time series corresponding to each pressure sensor in the worn intelligent rehabilitation chest strap, and the distances between the pressure sensors are obtained, the heart rate abnormal period is determined, and according to the dissimilarity between the respiratory abnormal change of the heart rate and the pressure change within the heart rate abnormal period, several suspected wound-affecting sensors are screened out from the pressure sensors. According to the distance between the suspected wound-affecting sensors and the similarity between the pressure time series corresponding to the adjacent suspected wound-affecting sensors, several wound-affecting sensors are screened out from the suspected wound-affecting sensors, and then the wound sensors are screened out from the wound-affecting sensors. According to the change of the pressure in the pressure time series corresponding to the wound sensor, the trend value is determined, and according to the size of the trend value, the tightness of the adjustment structure corresponding to the wound-affecting sensor is adjusted. The present invention can avoid unnecessary compression of the wound surface by adaptively adjusting the rehabilitation chest strap at the wound, so as to better promote the recovery of the wound.

[0137] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An intelligent rehabilitation chest strap regulation method based on real-time data acquisition, characterized in that, The method includes the following steps: Obtain the heart rate time series of the patient after breast surgery, the pressure time series corresponding to each pressure sensor in the worn intelligent rehabilitation chest band, and the distances between the pressure sensors; an adjustment structure corresponding to each pressure sensor; According to the magnitude of the heart rate in the heart rate time series, determine the heart rate abnormal period and the respiratory abnormality of each heart rate within the heart rate abnormal period; according to the dissimilarity between the change in the respiratory abnormality of the heart rate and the pressure change within the heart rate abnormal period, screen out several suspected wound-affecting sensors from the pressure sensors; According to the distances between the suspected wound-affecting sensors and the similarity between the pressure time series corresponding to the adjacent suspected wound-affecting sensors, screen out several wound-affecting sensors from the suspected wound-affecting sensors; According to the similarity between the pressure time series corresponding to the wound-affecting sensors, screen out the wound sensors from the wound-affecting sensors; determine the trend value according to the change in pressure in the pressure time series corresponding to the wound sensors; adjust the tightness of the adjustment structure corresponding to the wound-affecting sensors according to the magnitude of the trend value.

2. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 1, characterized in that, The specific steps included in determining the heart rate abnormal period and the respiratory abnormality of each heart rate within the heart rate abnormal period are as follows: In the heart rate time series, record the heart rate less than the lower limit of the preset normal heart rate as the low heart rate, and record the heart rate greater than the upper limit of the preset normal heart rate as the high heart rate; Calculate the ratio of the difference between each high heart rate and the upper limit of the preset normal heart rate to the upper limit of the preset normal heart rate as the respiratory abnormality of each high heart rate; Calculate the ratio of the difference between the lower limit of the preset normal heart rate and each low heart rate to the lower limit of the preset normal heart rate as the respiratory abnormality of each low heart rate; Record the time period where the heart rate time series segment composed of adjacent low heart rates is located as the low heart rate period, record the time period where the heart rate time series segment composed of adjacent high heart rates is located as the high heart rate period, and collectively refer to the low heart rate period and the high heart rate period as the heart rate abnormal period.

3. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 1, wherein, The specific steps included in screening out several suspected wound-affecting sensors from the pressure sensors are as follows: If the current moment is within the heart rate abnormal period, within the heart rate abnormal period where the current moment is located, obtain the respiratory abnormality time series segment composed of the respiratory abnormalities of all heart rates and the pressure time series segment corresponding to each pressure sensor; Record the pressure sensor with the cosine similarity between the respiratory abnormality time series segment and the pressure time series segment less than the preset similarity threshold as the suspected wound-affecting sensor.

4. The intelligent rehabilitation chest band regulation method based on real-time data acquisition according to claim 1, wherein, The specific steps included in screening out several wound-affecting sensors from the suspected wound-affecting sensors are as follows: According to the distances between the suspected wound-affecting sensors, assign a label value to each suspected wound-affecting sensor; According to the label values of the suspected wound-affecting sensors, screen out the optimal adjacent value B from the preset adjacent range sequence; Obtain the top B suspected wound-affecting sensors that are closest to the y-th suspected wound-affecting sensor, denoted as the optimal neighboring sensors. Denote the mean of the cosine similarities between the pressure time series of the y-th suspected wound-affecting sensor and all the optimal neighboring sensors as the pressure distribution uniformity of the area where the y-th suspected wound-affecting sensor is located. Take each suspected wound-affecting sensor as a data point, use the label value of each suspected wound-affecting sensor as the horizontal axis, and use the inverse value of the pressure distribution uniformity of the area where each suspected wound-affecting sensor is located as the vertical axis to construct a sensor scatter plot corresponding to all suspected wound-affecting sensors. Use the density-based spatial clustering algorithm to perform clustering operations on all data points in the sensor scatter plot to obtain several clustering clusters. In the sensor scatter plot, divide the target area. Denote the ratio of the number of data points in the n-th clustering cluster that are within the target area to the number of data points in the n-th clustering cluster as the proportion of the number of the n-th clustering cluster. Denote the suspected wound-affecting sensors in the clustering cluster with the largest proportion of the number as the wound-affecting sensors.

5. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 4, characterized in that The step of assigning a label value to each suspected wound-affecting sensor according to the distance between the suspected wound-affecting sensors is as follows: Denote the sum of the distances between each suspected wound-affecting sensor and all other suspected wound-affecting sensors as the distance threshold of each suspected wound-affecting sensor. Denote the suspected wound-affecting sensor with the smallest distance threshold as the target sensor. Assign a label value of a preset constant to the target sensor. The label value assigned to each suspected wound-affecting sensor other than the target sensor is: the distance between each suspected wound-affecting sensor other than the target sensor and the target sensor.

6. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 4, characterized in that, The step of screening out the optimal neighboring value B from the preset neighboring range sequence according to the label value of the suspected wound-affecting sensor is as follows: Select any data value A in the preset neighboring range sequence, obtain the top A suspected wound-affecting sensors that are closest to any suspected wound-affecting sensor, denoted as neighboring sensors, and form a sensor group with all the neighboring sensors and the any suspected wound-affecting sensor. Statistically calculate the variance of the label values of all suspected wound-affecting sensors in all sensor groups corresponding to A, denoted as the first variance. Statistically calculate the variance of the label values of all suspected wound-affecting sensors in each sensor group corresponding to A, denoted as the second variance, and take the sum of the second variances of all sensor groups as the third variance. Denote the difference between the first variance and the third variance as the difference threshold corresponding to A. Denote the data value in the preset neighboring range sequence corresponding to the smallest difference threshold as the optimal neighboring value B.

7. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 4, wherein, The step of dividing the target area in the sensor scatter plot is as follows: In the sensor scatter plot, sort the horizontal axis coordinate values and vertical axis coordinate values of all data points in ascending order respectively to obtain an ascending sequence of horizontal axis coordinate values and an ascending sequence of vertical axis coordinate values. Use the APCA segmentation algorithm to perform segmentation operations on the ascending sequence of horizontal axis coordinate values and the ascending sequence of vertical axis coordinate values respectively. Denote the data point corresponding to the last horizontal axis coordinate value in the first segment of the ascending sequence of horizontal axis coordinate values as the first target data point; Denote the data point corresponding to the last vertical axis coordinate value in the first segment of the ascending sequence of vertical axis coordinate values as the second target data point; Denote the straight line passing through the first target data point and perpendicular to the horizontal axis as the first straight line; Denote the straight line passing through the second target data point and perpendicular to the vertical axis as the second straight line; Denote the closed rectangular area enclosed by the first straight line, the second straight line, the horizontal axis, and the vertical axis as the target area.

8. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 1, characterized in that The steps of screening out the wound sensor from the wound surface impact sensors are as follows: Obtain the sum value of the cosine similarities between the pressure time series corresponding to the m-th wound surface impact sensor and all other wound surface impact sensors, and denote it as the wound possibility of the m-th wound surface impact sensor; Denote the wound surface impact sensor corresponding to the maximum wound possibility as the wound sensor.

9. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 1, characterized in that, The steps of determining the trend value are as follows: In the pressure time series corresponding to the wound sensor, with each pressure data as a data point, use the ordinal value of each pressure data as the horizontal axis and the value of each pressure data as the vertical axis to construct a pressure scatter plot of the pressure time series corresponding to the wound sensor; In the pressure scatter plot, use all data points as the input of principal component analysis, and output several projection values and the projection vectors corresponding to each projection value. Among them, each projection vector corresponds to a horizontal axis coordinate value and a vertical axis coordinate value; Denote the projection vector corresponding to the maximum projection value as the trend vector; Denote the ratio of the arctangent angle value corresponding to the ratio of the vertical axis coordinate value to the horizontal axis coordinate value of the trend vector to the preset angle value as the trend value.

10. The intelligent rehabilitation chest strap regulation method based on real-time data acquisition according to claim 2, characterized in that, The steps of adjusting the tightness of the adjustment structure corresponding to the wound surface impact sensor are as follows: Calculate the product of the sum value of 1 and the trend value and the preset reference adjustment amount, and denote it as the adjustment amount for executing the adjustment structure; Denote the adjustment structure corresponding to the wound surface impact sensor as the target adjustment structure; If the current moment is within the high heart rate period, then when the trend value is greater than the preset tightness threshold, use the adjustment amount of the execution adjustment structure to perform a relaxation adjustment on the target adjustment structure. When the trend value is less than or equal to the preset tightness threshold, perform a relaxation adjustment on the target adjustment structure with the preset reference adjustment amount; If the current moment is within the low heart rate period, then when the trend value is greater than the preset tightness threshold, use the adjustment amount of the execution adjustment structure to perform a tightening adjustment on the target adjustment structure. When the trend value is less than or equal to the preset tightness threshold, perform a tightening adjustment on the target adjustment structure with the preset reference adjustment amount.

Images