Intelligent vital sign monitoring method for arc protection clothes

Abstract

The invention relates to the technical field of data processing, in particular to a vital sign monitoring method for intelligent arc protection clothes, which comprises the following steps: according to the change of vital sign data of each dimension when an experimenter wears the arc protection clothes to carry out activities in different states, determining the vital sign data of each dimension; determining the response duration and the response degree of the data of each dimension to the change of the motion state, and further determining the prediction index of each dimension; collecting multi-dimensional real-time vital sign data when the operator wears the arc protection clothes to work; correcting the prediction index of the current dimension according to the change of the real-time vital sign data of each dimension of which the response duration is less than the current dimension; and predicting the vital sign data of each dimension at the future moment according to the corrected prediction index of each dimension, and performing vital sign monitoring according to a prediction result. According to the invention, the prediction and judgment precision of the health condition of the operator can be improved, and the safety of the operator is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, to an intelligent arc protection clothing vital sign monitoring method. Background Art

[0002] Arc protection clothing is an important protective equipment for power workers. It can prevent arc injuries to a certain extent. However, it still has many shortcomings in dynamically monitoring the vital signs of workers and providing real-time warnings.

[0003] With the development of intelligent technology, intelligent protective equipment combined with vital signs monitoring has gradually become an important tool for improving safety. Intelligent arc protection clothing can monitor the vital signs data of operators, such as heart rate, blood oxygen, etc. in real time by integrating sensors, and promptly detect potential health risks. However, simple real-time data monitoring is often difficult to achieve early warning. Therefore, it is particularly important to predict vital signs at future moments based on real-time vital signs data and to provide early warning based on the predicted values.

[0004] Exponential smoothing is a data prediction method that can predict the vital sign values ​​at future moments based on the changing trends of the vital sign data collected by the vital sign monitoring module integrated in the arc protection suit. However, the smoothing coefficient in the exponential smoothing method usually needs to be set manually. If the smoothing coefficient is set too large, the prediction model may pay too much attention to the latest data and ignore the historical trend of the vital sign data, which will lead to inaccurate predictions of vital sign data at future moments and even false alarms; on the contrary, if the smoothing coefficient is set too small, the prediction model will react more slowly to new data, making it difficult to capture sudden changes in vital sign data in a timely manner, thereby reducing the prediction accuracy of vital sign data and possibly causing missed reports. Summary of the invention

[0005] In order to solve the technical problem that the smoothing coefficient of the exponential smoothing method is artificially set, which leads to inaccurate prediction of vital signs data, the present invention proposes an intelligent arc protection vital signs monitoring method, which includes the following steps: Collecting experimental data, the experimental data includes vital sign data of multiple dimensions when the experimenter wears arc protection clothing and performs activities in different states; determining the response time and response degree of the data of each dimension to the change of the motion state according to the changes of the data of each dimension in the experimental data; determining the prediction index of each dimension according to the response time and response degree; collecting real-time vital sign data of multiple dimensions when the operator wears arc protection clothing to work; correcting the prediction index of the current dimension according to the changes of the real-time vital sign data of each dimension whose response time is shorter than that of the current dimension; using the corrected prediction index of each dimension as a smoothing coefficient, and using the exponential smoothing method to predict the vital sign data of each dimension at a future time; and performing vital sign monitoring according to the prediction results of the vital sign data of each dimension at a future time.

[0006] The present invention analyzes the vital signs of each dimension in the experimental data in detail, identifies the response characteristics of each dimension to the change of motion state, determines the response time and response degree of each dimension, and can improve the accuracy of prediction and judgment of the health status of the operators; the working environment and the motion state of the operators are constantly changing. The present invention collects vital signs data of multiple dimensions, and predicts and corrects according to the response time and response degree of each dimension and data changes, so as to ensure that the system effectively copes with the influence of different working states on vital signs in a dynamic environment; the present invention corrects the prediction index of the dimension with a larger response time according to the change of the real-time vital signs data of the dimension with a smaller response time, and can realize dynamic adjustment of the prediction index, so that when the exponential smoothing method is used to predict the vital signs data of each dimension at a future moment, the prediction result can not only reflect the historical trend of the vital signs data, but also timely capture the sudden change of the vital signs data, and make early warning based on the prediction result, which can improve the safety of the operators.

[0007] Preferably, determining the response duration and degree of each dimensional data to the change in motion state includes: taking any one dimension as the target dimension; dividing the data of the target dimension into multiple segments according to the moment of the motion state change; determining the degree of change of each data in each segment compared with the corresponding data in each period of the previous segment; forming a sequence with the degrees of change of all data in each segment, performing Otsu threshold segmentation on the sequence, and taking the difference between the moment corresponding to the first data in the second subsequence obtained by the segmentation and the moment corresponding to the first data in the segment as the response duration of the segment; taking the mean value of the degrees of change of all data in the second subsequence as the response degree of the segment; taking the mean value of the response duration of all segments and the mean value of the response degree of all segments as the response duration and degree of the target dimensional data to the change in motion state, respectively.

[0008] The present invention can quantify the dynamic changes of each stage by calculating the degree of change of each segment of data relative to the previous segment. By performing Otsu threshold segmentation on each segment, the response duration of each segment of data can be accurately identified. By calculating the mean of the degree of change of each segment, the response degree of the segment can be obtained, which can provide a quantitative evaluation of the overall reaction level to the change of motion state, and help to accurately evaluate the adaptability and influence of different dimensions on the change of motion state.

[0009] Preferably, determining the degree of change of each data in each segment compared with the corresponding data in each period in the previous segment includes: obtaining the period length of the data of the target dimension, and dividing the data of the target dimension into multiple periods according to the period length; for each data in each segment, taking the position number of the data in its period as the first target number, obtaining the mean of the difference between the data and the data with the position number of the first target number in each period in the previous segment, and taking the ratio of the mean of the difference to the range of all data with the position number of the first target number in each period in the previous segment as the transformation degree of the data.

[0010] The present invention can effectively evaluate the volatility and changing trend of data between cycles by calculating the mean of the difference between each data and the data at the same position in each cycle in the previous section, which helps to identify the potential rules and abnormal fluctuations of the data and helps to more accurately analyze and predict the trend changes of the target dimension.

[0011] Preferably, the method for obtaining the period length is: performing Fourier transform on the data of the target dimension, and obtaining the inverse of the frequency of the component with the largest amplitude after Fourier transform as the period length of the target dimension.

[0012] Preferably, the prediction index of each dimension is determined, including: taking any one dimension as the target dimension, negatively normalizing the response time of the target dimension data to the change in motion state, positively normalizing the response degree of the target dimension data to the change in motion state, and taking the product of the result of the negative correlation normalization of the response time and the result of the positive correlation normalization of the response degree as the prediction index of the target dimension.

[0013] In the present invention, when the data of the target dimension has a short response time and a large response degree to the change of motion state, it indicates that the sensitivity of the dimension is high, and more attention should be paid to the latest data during prediction to reduce the interference of historical data fluctuations on the prediction. By normalizing the response time negatively and the response degree positively, and then taking the product of the two as the prediction index, the data change characteristics of each dimension can be accurately reflected, and it is applied to the exponential smoothing method to adjust the smoothing coefficient. When the prediction index is high, the larger the smoothing coefficient, the more the model will pay attention to the latest data changes, and the accuracy and timeliness of the prediction will be improved.

[0014] Preferably, the prediction index of the current dimension is modified, including: sorting the dimensions in order of response time from small to large, The forecast index after correction of the dimension Satisfies the expression: , Indicates The prediction index of each dimension; , Respectively represent Dimensions, The response time of each dimension to the change of motion state; Indicates dimensions before the current moment In the data The change difference of each data is obtained by: The length of the cycle in the dimension will be The real-time data of the dimension is divided into multiple periods, and the The serial number of each data in its cycle is recorded as , obtain the The data and the data in the previous period The average difference between the data, the normalized result of the average difference is used as the first The difference in data changes.

[0015] The present invention can dynamically adjust the prediction index according to the response time of different dimensions by correcting the prediction index of the current dimension, thereby optimizing the accuracy of data prediction. Dimensions with shorter response times react faster to changes in motion states, and thus can reflect changes in motion states more promptly. For dimensions with longer response times, their prediction indexes are smaller, which helps maintain the regularity of periodic data. When the motion state changes, the prediction index can be corrected by combining the change information of other dimensions, which can more effectively adjust the prediction value of the dimension, so that it can capture sudden changes in vital sign data in a timely manner.

[0016] Preferably, predicting the vital sign data of each dimension at a future moment includes: taking the data sequence collected in real time in each dimension as a data sample, performing exponential smoothing on the data sample according to a smoothing coefficient, and obtaining the prediction result of the vital sign data of each dimension at a future moment.

[0017] Preferably, predicting the vital signs data of each dimension at a future moment includes: taking any one dimension as a target dimension, and taking a data sequence collected in real time from the target dimension as a first data sample; dividing the real-time data of the target dimension into multiple periods according to the period length of the target dimension, if the position number of the data at the current moment in its period is less than the period length of the target dimension, adding 1 to the position number of the data at the current moment in its period as the second target number, otherwise, taking 1 as the second target number, and forming a sequence of data with the position number of the second target number in all periods as a second data sample; performing exponential smoothing on the first data sample and the second data sample according to the smoothing coefficient to obtain a first prediction value and a second prediction value; and performing weighted summation on the first prediction value and the second prediction value to obtain a prediction result of the vital signs data of the target dimension at a future moment.

[0018] Preferably, when the first prediction value and the second prediction value are weighted and summed, the weight of the first prediction value is , the weight of the second prediction value is ,in is the corrected prediction index of the target dimension.

[0019] The present invention effectively combines the periodic characteristics and real-time changes of vital signs data, and improves the accuracy and reliability of vital signs data prediction. When the corrected prediction index of the target dimension is small, it indicates that the state of the operator is stable, and the prediction result depends more on the periodic changes of the target dimension, thereby enhancing the periodic characteristics of the prediction by increasing the weight of the second prediction value; and when the corrected prediction index is large, it indicates that the state of the operator has changed, and the prediction result focuses more on real-time data changes, so by increasing the weight of the first prediction value, the prediction can better reflect the impact of state changes.

[0020] Preferably, vital signs monitoring is performed based on the predicted results of vital signs data of each dimension at a future moment, including: in response to the predicted result of one dimension exceeding the normal numerical range, or the actual value of the dimensional data exceeding the normal numerical range, or the difference between the predicted result and the actual value of the dimensional data exceeding a preset difference threshold, vital signs become abnormal.

[0021] The present invention has the following beneficial effects: The prediction results of the vital signs data of various dimensions at future moments in the present invention can not only reflect the historical trend of the vital signs data, but also timely capture the sudden changes of the vital signs data, improve the accuracy of prediction and judgment of the health status of operators, and provide early warning based on the prediction results, which can effectively improve the safety of operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1This is a flow chart of the steps of an intelligent arc protection vital signs monitoring method of the present invention. Figure 2 It is a flow chart of step S2 of a method for monitoring vital signs of intelligent arc protection clothing according to the present invention; Figure 3 This is a schematic diagram of the carbon dioxide concentration curve. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0024] In order to monitor the vital signs of the operators, carbon dioxide sensors and oxygen sensors are installed on the inside of the arc protection suit, and heart rate sensors and blood oxygen sensors are installed on the inside of the cuffs of the arc protection suit. After the experimenters or operators wear the arc protection suit, the heart rate sensors and blood oxygen sensors can be close to the wrists of the experimenters or operators to collect their heart rate and blood oxygen in real time.

[0025] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] See also Figure 1 , which shows a flow chart of the steps of an intelligent arc protection vital signs monitoring method provided by the present invention, the method comprising the following steps: S1. Collecting experimental data, wherein the experimental data includes multiple dimensions of vital sign data when the experimenter wears arc protection clothing and performs activities in different states.

[0027] The experimenters wear arc protection suits to carry out activities in different motion states, such as standing, sitting, walking, running, etc., and record the moment when the motion state changes. Collect the vital signs data of the experimenters during the experiment, such as the carbon dioxide concentration, oxygen concentration, heart rate, blood oxygen, etc. in the arc protection suit. The carbon dioxide concentration and oxygen concentration in the arc protection suit are collected by the carbon dioxide sensor and oxygen sensor on the inside of the arc protection suit, and the heart rate and blood oxygen of the experimenters are collected by the heartbeat sensor and blood oxygen sensor on the inside of the cuff of the arc protection suit. The collection frequency is set by the implementer according to the actual implementation situation, for example, 1 second / time.

[0028] S2. According to the changes of each dimension data in the experimental data, determine the response time and response degree of each dimension data to the change of motion state.

[0029] It should be noted that changes in motion state will cause changes in data in each dimension. For example, when the experimenter changes from a standing posture to a walking posture, the experimenter's breathing speeds up, the heart rate increases, the body temperature increases, and the oxygen consumption increases, which reduces the blood oxygen, reduces the oxygen concentration in the arc protection suit, and increases the exhaled carbon dioxide, which increases the carbon dioxide concentration in the arc protection suit. However, the response time and degree of the data in each dimension to the change in motion state are different. For example, when the experimenter changes from a standing posture to a walking posture, the muscles continue to move, consume oxygen and energy, metabolize to produce carbon dioxide, and expel carbon dioxide from the body. Therefore, the oxygen concentration in the arc protection suit is first affected by the change in motion state, and the carbon dioxide concentration is later affected by the change in motion state. Therefore, the response time of oxygen concentration to the change in motion state is shorter than that of carbon dioxide concentration to the change in motion state. When the experimenter changes from a standing posture to a walking posture, the heart rate increases to 1.5 to 2 times the previous one, while the degree of reduction in blood oxygen is limited, usually not exceeding 5%. Therefore, the response degree of heart rate to the change in motion state is greater than that of blood oxygen to the change in motion state. The present invention determines the response time and degree of each dimensional data to the change of motion state according to the change of each dimensional data in the experimental data.

[0030] See also Figure 2 , which shows a flowchart of step S2 of a smart arc protection clothing vital signs monitoring method of the present invention, including steps S201-step S202, specifically: S201, dividing the data of each dimension into multiple segments according to the moment when the motion state changes, and determining the degree of change of each data in each segment compared with the corresponding data in each cycle in the previous segment.

[0031] It should be noted that when inhaling, the expansion and pressure changes of the chest cavity affect the blood return of the heart, which accelerates the heart rate, oxygen enters the lungs, and diffuses into the blood through the alveoli, and blood oxygen usually rises, while when exhaling, the heart rate will slow down slightly, the carbon dioxide in the body is discharged, and the blood oxygen will drop slightly, so the heart rate and blood oxygen change with the breathing cycle. During one breathing process of the experimenter, the experimenter inhales oxygen and exhales carbon dioxide, so that the carbon dioxide concentration in the arc protection suit increases and the oxygen concentration decreases. Since the arc protection suit is designed with vents, the outside air enters the arc protection suit through the vents, and the gas in the arc protection suit leaves through the vents, so that the carbon dioxide concentration in the arc protection suit decreases and the oxygen concentration increases. As the experimenter continues to breathe, the above process is repeated, so that the carbon dioxide concentration and oxygen concentration also show the characteristics of periodic fluctuations. Therefore, the present invention obtains the transformation degree of each data according to the law of periodic changes of each dimension data, and uses the degree of change to reflect the degree of influence of the data on the change of motion state.

[0032] Specifically, any dimension is used as the target dimension, the data of the target dimension is Fourier transformed, and the inverse of the frequency of the component with the largest amplitude after the Fourier transform is obtained as the period length of the target dimension. The data of the target dimension is divided into multiple periods according to the period length.

[0033] The data of the target dimension is divided into multiple segments according to the moment when the motion state changes. The first data in each segment is the data of the target dimension corresponding to the moment when the motion state changes, and the last data in each segment is the data of the target dimension at the moment before the next moment when the motion state changes.

[0034] For each data in each segment, the position number of the data in its cycle is taken as the first target number, the mean of the difference between the data and the data whose position number is the first target number in each cycle in the previous segment is obtained, and the ratio of the mean of the difference to the range of all data whose position number is the first target number in each cycle in the previous segment is taken as the transformation degree of the data.

[0035] It should be noted that the data with the position number of the first target number in each period in the previous segment are the data at the same position in each period under the previous motion state. Under the periodic law, these data are the same or similar, and the current data are also the data at the same position in the period. If the difference between the current data and the data with the position number of the first target number in each period in the previous segment is greater, it means that the change of the current data compared with the previous data is greater, and the current data is more affected by the change of the motion state. Conversely, if the difference between the current data and the data with the position number of the first target number in each period in the previous segment is smaller, it means that the change of the current data compared with the previous data is smaller, and the current data is less affected by the change of the motion state.

[0036] For example Figure 3 is a schematic diagram of the carbon dioxide concentration curve. Figure 3 A in the equation is the two moments when the motion state changes. Figure 3 Among them, B1 and B2 are two segments, data point D is a data point in segment B2, data points C1, C2, C3, C4, C5 are data in segment B1 that are in the same periodic position as data point D, data points C1, C2, C3, C4, C5 are basically similar, and there is a certain difference between data point D and data points C1, C2, C3, C4, C5, indicating that data point D is greatly affected by the change of motion state.

[0037] S202: Determine the response time and degree of each dimension to the change of the motion state according to the degree of change of each data in each segment.

[0038] Specifically, the transformation degrees of all data in each segment are formed into a sequence, and the sequence is segmented by Otsu threshold to be divided into two subsequences.

[0039] It should be noted that the variation of each data in the first subsequence obtained by segmentation is relatively small, and the influence of the change of motion state on the data of this dimension has not yet appeared. The variation of each data in the second subsequence obtained by segmentation is relatively large, that is, the data of this dimension begins to be affected by the change of motion state. The moment corresponding to the first data in the segment is the moment of the motion state change. Therefore, the difference between the moment corresponding to the first data in the second subsequence and the moment corresponding to the first data in the segment reflects the delay time of the influence of the motion state change on the segment, and the variation of each data in the second subsequence reflects the degree of influence of the motion state change on the segment. Therefore, the present invention obtains the response duration and response degree of the segment according to the delay time of the influence of the motion state change on the segment and the degree of influence of the motion state change on the segment.

[0040] Specifically, the difference between the time corresponding to the first data in the second subsequence and the time corresponding to the first data in the segment is taken as the response duration of the segment, and the average of the change degrees of all data in the second subsequence is taken as the response degree of the segment.

[0041] The average response duration of all segments of the target dimension is taken as the response duration of the target dimension data to the change of motion state, and the average response degree of all segments of the target dimension is taken as the response degree of the target dimension data to the change of motion state.

[0042] Similarly, the response time and degree of each dimension of data to the change in motion state are obtained.

[0043] S3. Determine the prediction index of each dimension according to the response time and degree of each dimension data to the change of motion state.

[0044] It should be noted that for any dimension, when the response time of the dimension data to the change of motion state is shorter and the response degree to the change of motion state is greater, the data of the dimension is more sensitive to the change of motion state, and when the subsequent data prediction is performed, the less attention is paid to the previous data, so as to avoid the fluctuation law of the previous data affecting the accuracy of the predicted data, and the more attention is paid to the latest data, so that the predicted data can reflect the data changes caused by the change of motion state. The present invention uses the exponential smoothing method to predict data later. The larger the smoothing coefficient of the exponential smoothing method is, the more attention is paid to the latest data, and the smaller the smoothing coefficient is, the more attention is paid to the previous data. Therefore, the present invention determines the prediction index of each dimension according to the response time and response degree of each dimension data to the change of motion state, and obtains the smoothing coefficient using the prediction index. When the response time of each dimension data to the change of motion state is shorter and the response degree to the change of motion state is greater, the prediction index of the dimension is greater, and when the subsequent data prediction is performed, the more attention is paid to the latest data.

[0045] Specifically, the response time of the target dimension data to the change of motion state is negatively normalized, and the response degree of the target dimension data to the change of motion state is positively normalized. The product of the negative correlation normalization result of the response time and the positive correlation normalization result of the response degree is used as the prediction index of the target dimension.

[0046] Furthermore, the method used in the present invention to normalize the negative correlation of the response time is: ,in It is the result of negative correlation normalization of the response time of the target dimension data to the change of motion state. is the response time of the target dimension data to the change of motion state, is the period length of the target dimension, is an exponential function with a natural constant as its base.

[0047] The function used in the present invention for normalizing the response degree by positive correlation is a hyperbolic tangent function. In other embodiments, the implementer can select the method of negative correlation normalization and positive correlation normalization according to the actual implementation situation.

[0048] Similarly, obtain the prediction index of each dimension.

[0049] S4. Collect real-time vital sign data in multiple dimensions when workers wear arc protection clothing to work.

[0050] Specifically, when the operator wears the arc protection suit to work, the vital signs data of the operator during the work process are collected in real time, such as the carbon dioxide concentration, oxygen concentration, heart rate, blood oxygen, etc. in the arc protection suit. The carbon dioxide concentration and oxygen concentration in the arc protection suit are collected by the carbon dioxide sensor and oxygen sensor inside the arc protection suit, and the heart rate and blood oxygen of the operator are collected by the heartbeat sensor and blood oxygen sensor inside the cuff of the arc protection suit. The collection frequency is the same as the collection frequency of the experimental data.

[0051] S5. Correct the prediction index of the current dimension according to the changes in the real-time vital sign data of each dimension whose response time is shorter than the current dimension.

[0052] It should be noted that the data of the dimension with a long response time responds slowly to the change of the motion state, and its prediction index is small. If the motion state has not changed, the smaller prediction index can make the prediction value of this dimension better reflect the periodic change law of the data. If the motion state changes, it is necessary to increase its prediction index so that the prediction value of this dimension can reflect the changes brought about by the change of the motion state. Therefore, the present invention determines whether the motion state has changed according to the changes in the real-time vital sign data of the dimension with a small response time, thereby adjusting the prediction index of the dimension with a large response time.

[0053] Specifically, the dimensions are sorted in ascending order of response time.

[0054] After sorting The forecast index after correction of the dimension Satisfies the expression: ; in, Indicates The prediction index of each dimension; , Respectively represent Dimensions, The response time of each dimension to the change of motion state; Indicates dimensions before the current moment The data The change difference of each data is obtained by: The length of the cycle in the dimension will be The real-time data of the dimension is divided into multiple periods, and the The period where the data is located is recorded as ,when When Otherwise, the The serial number of each data in its cycle is recorded as , obtain the Data and previous In the cycle The average difference between the data, the normalized result of the average difference is used as the The difference in data changes.

[0055] Jordi dimensions before the current moment The greater the difference in the change of data, the greater the difference between the current moment and the Within a moment, the motion state of the operator may change. The prediction index of each dimension is increased, so that the subsequent prediction When the data of the first dimension is collected, the more attention is paid to the latest data, so that the predicted data can reflect the data changes caused by the change of motion state. dimensions before the current moment The greater the difference in the change of data, the greater the difference between the current moment and the Within a moment, the motion state of the operator changes. The prediction index of each dimension remains unchanged as much as possible, so that the subsequent prediction When predicting data of a dimension, we should pay more attention to the previous data so that the predicted data can reflect the overall rules of the data of that dimension.

[0056] It should be noted that when At that time, There is no other dimension before the dimension, so it is stipulated that: when At that time, The forecast index after correction of the dimension Compared with the forecast index before revision equal.

[0057] At this point, the correction of the prediction index of each dimension has been achieved.

[0058] S6. Use the corrected prediction index of each dimension as a smoothing coefficient to predict the vital sign data of each dimension at future moments.

[0059] In the first embodiment, the data sequence collected in real time in the target dimension is used as the data sample, and the prediction index after correction of the target dimension is used as the smoothing coefficient. The data sample is exponentially smoothed to obtain the prediction result of the vital sign data of the target dimension at the future moment.

[0060] In the second embodiment, the data sequence collected in real time in the target dimension is used as the first data sample, and the corrected prediction index of the target dimension is used as the smoothing coefficient. The first data sample is exponentially smoothed to obtain the first predicted value of the vital sign data of the target dimension at the future moment.

[0061] The real-time data of the target dimension is divided into multiple periods according to the cycle length of the target dimension. If the position number of the data at the current moment in its cycle is less than the cycle length of the target dimension, the position number of the data at the current moment in its cycle is increased by 1 as the second target number, otherwise, 1 is used as the second target number. The data with the position number of all cycles as the second target number are formed into a sequence as the second data sample, and the corrected prediction index of the target dimension is used as the smoothing coefficient. The second data sample is exponentially smoothed to obtain the second predicted value of the vital sign data of the target dimension at the future moment.

[0062] The first predicted value and the second predicted value of the vital sign data of the target dimension at the future moment are weighted and summed to obtain the prediction result of the vital sign data of the target dimension at the future moment. When the first predicted value and the second predicted value of the vital sign data of the target dimension at the future moment are weighted and summed, the weights of the first predicted value and the second predicted value are both 0.5.

[0063] It should be noted that since the vital signs data of each dimension changes periodically, the results may have certain deviations if predictions are made only based on the real-time data series. Therefore, in Example 2, the periodic variation characteristics of the data are also taken into account. The data at future times are predicted based on the data at the same position in the cycle as the data at future times. Compared with predictions based only on the real-time data series, the results are more accurate.

[0064] In the third embodiment, the data sequence collected in real time in the target dimension is used as the first data sample, and the corrected prediction index of the target dimension is used as the smoothing coefficient. The first data sample is exponentially smoothed to obtain the first predicted value of the vital sign data of the target dimension at the future moment.

[0065] The real-time data of the target dimension is divided into multiple periods according to the cycle length of the target dimension. If the position number of the data at the current moment in its cycle is less than the cycle length of the target dimension, the position number of the data at the current moment in its cycle is increased by 1 as the second target number, otherwise, 1 is used as the second target number. The data with the position number of all cycles as the second target number are formed into a sequence as the second data sample, and the corrected prediction index of the target dimension is used as the smoothing coefficient. The second data sample is exponentially smoothed to obtain the second predicted value of the vital sign data of the target dimension at the future moment.

[0066] The first predicted value and the second predicted value of the vital sign data of the target dimension at the future moment are weighted and summed to obtain the prediction result of the vital sign data of the target dimension at the future moment. Among them, when the first predicted value and the second predicted value of the vital sign data of the target dimension at the future moment are weighted and summed, the weight of the first predicted value is , the weight of the second prediction value is ,in is the corrected prediction index of the target dimension.

[0067] It should be noted that in Example 3, when the prediction index after the target dimension is corrected is smaller, it means that the motion state of the operator remains unchanged. At this time, a smaller weight is given to the first prediction value, and a larger weight is given to the second prediction value, so that the result of the final prediction value depends more on the periodic change characteristics, so that the final prediction value can better reflect the periodic change law of the target dimension. On the contrary, when the prediction index after the target dimension is corrected is larger, it means that the motion state of the operator has changed. At this time, a larger weight is given to the first prediction value, and a smaller weight is given to the second prediction value, so that the result of the final prediction value pays more attention to the real-time changes of the sequence, so that the final prediction value can better reflect the changes brought about by the change of the motion state.

[0068] S7. Monitor vital signs based on the predicted results of vital sign data of various dimensions at future moments.

[0069] It should be noted that when the predicted result exceeds the normal numerical range, it means that the change trend of the operator's vital signs data is abnormal, and arc flash, high temperature, high pressure and other abnormalities may have occurred. The operator's vital signs will be abnormal in the future. When the actual value exceeds the normal numerical range, it means that the operator's vital signs have become abnormal. When the difference between the predicted result and the actual value exceeds the difference threshold, it means that the change of the operator's vital signs does not conform to the change pattern of the previous data. At this time, arc flash, high temperature, high pressure and other abnormalities may have occurred.

[0070] Specifically, in response to the prediction result of a dimension exceeding the normal numerical range, or the actual value of the dimension data exceeding the normal numerical range, or the difference between the prediction result and the actual value exceeds the preset difference threshold, it is considered that an abnormality has occurred. At this time, the monitoring personnel should immediately instruct the operating personnel to stop working and check the physical condition of the operating personnel.

[0071] Among them, the normal data range and difference threshold of each dimension are set by the implementers according to the actual implementation situation. For example, the normal value range of heart rate is 60 times / minute-110 times / minute, and the difference threshold is 5 times / minute. The normal value range of blood oxygen is 95%-100%, and the difference threshold is 1%. The normal value range of carbon dioxide concentration is 0.04%-0.1%, and the difference threshold is 0.005%. The normal value range of oxygen concentration is 19.5% to 23.5%, and the difference threshold is 0.01%.

[0072] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent arc protection clothing vital signs monitoring method, characterized in that: include: Collecting experimental data, wherein the experimental data includes multiple dimensions of vital sign data of the experimenter when wearing arc protection clothing and performing activities in different states; Determine the response time and degree of each dimension data to the change of motion state according to the change of each dimension data in the experimental data; determine the prediction index of each dimension according to the response time and degree; Collect real-time vital sign data of multiple dimensions when workers wear arc protection clothing to work; modify the prediction index of the current dimension according to the changes of real-time vital sign data of each dimension whose response time is shorter than the current dimension; use the modified prediction index of each dimension as the smoothing coefficient, and use the exponential smoothing method to predict the vital sign data of each dimension in the future; Vital signs monitoring is performed based on the predicted results of vital signs data in various dimensions at future moments.

2. The intelligent arc protection clothing vital signs monitoring method according to claim 1 is characterized in that: Determine the response time and degree of each dimension data to the change of motion state, including: taking any one dimension as the target dimension; dividing the data of the target dimension into multiple segments according to the moment of motion state change; determining the degree of change of each data in each segment compared with the corresponding data in each cycle of the previous segment; forming a sequence with the degree of change of all data in each segment, performing Otsu threshold segmentation on the sequence, taking the difference between the moment corresponding to the first data in the second subsequence obtained by segmentation and the moment corresponding to the first data in the segment as the response time of the segment; taking the average of the degree of change of all data in the second subsequence as the response degree of the segment; The average value of the response duration of all segments and the average value of the response degree of all segments are respectively used as the response duration and response degree of the target dimension data to the change of motion state.

3. The intelligent arc protection clothing vital signs monitoring method according to claim 2 is characterized in that: Determine the degree of change of each data in each segment compared with the corresponding data in each cycle in the previous segment, including: Obtain the cycle length of the data of the target dimension, and divide the data of the target dimension into multiple cycles according to the cycle length; for each data in each segment, use the position number of the data in its cycle as the first target number, obtain the mean of the difference between the data and the data with the position number of the first target number in each cycle of the previous segment, and use the ratio of the mean of the difference to the range of all the data with the position number of the first target number in each cycle of the previous segment as the transformation degree of the data.

4. The intelligent arc protection clothing vital signs monitoring method according to claim 3 is characterized in that: The method for obtaining the cycle length is: Perform Fourier transform on the data of the target dimension, and obtain the inverse of the frequency of the component with the largest amplitude after Fourier transform as the period length of the target dimension.

5. The intelligent arc protection clothing vital signs monitoring method according to claim 1 is characterized in that: Determine the prediction index for each dimension, including: Take any dimension as the target dimension, perform negative correlation normalization on the response time of the target dimension data to the change of motion state, and perform positive correlation normalization on the response degree of the target dimension data to the change of motion state. The product of the negative correlation normalization result of the response time and the positive correlation normalization result of the response degree is used as the prediction index of the target dimension.

6. The intelligent arc protection clothing vital signs monitoring method according to claim 1 is characterized in that: Modify the prediction index of the current dimension, including: Sort each dimension in order of response time from small to large. The forecast index after correction of the dimension Satisfies the expression: , Indicates The prediction index of each dimension; , Respectively represent Dimensions, The response time of each dimension to the change of motion state; Indicates dimensions before the current moment In the data The change difference of each data is obtained by: The length of the cycle in the dimension will be The real-time data of the dimension is divided into multiple periods, and the The serial number of each data in its cycle is recorded as , obtain the The data and the data in the previous period The average difference between the data, the normalized result of the average difference is used as the first The difference in data changes.

7. The intelligent arc protection clothing vital signs monitoring method according to claim 1 is characterized in that: Predict vital signs data in various dimensions at future moments, including: The data sequences collected in real time in each dimension are taken as data samples, and the data samples are exponentially smoothed according to the smoothing coefficient to obtain the prediction results of the vital signs data of each dimension at the future moment.

8. The intelligent arc protection clothing vital signs monitoring method according to claim 1 is characterized in that: Predicting vital sign data of each dimension at a future moment, including: taking any one dimension as a target dimension, and taking a data sequence collected in real time from the target dimension as a first data sample; dividing the real-time data of the target dimension into multiple periods according to the period length of the target dimension, if the position number of the data at the current moment in its period is less than the period length of the target dimension, adding 1 to the position number of the data at the current moment in its period as a second target number, otherwise, taking 1 as the second target number, and forming a sequence of data with the position number of the second target number in all periods as a second data sample; performing exponential smoothing on the first data sample and the second data sample according to the smoothing coefficient to obtain a first prediction value and a second prediction value; and performing weighted summation on the first prediction value and the second prediction value to obtain a prediction result of the vital sign data of the target dimension at a future moment.

9. The intelligent arc protection clothing vital signs monitoring method according to claim 8, characterized in that: When the first predicted value and the second predicted value are weighted and summed, the weight of the first predicted value is , the weight of the second prediction value is ,in is the corrected prediction index of the target dimension.

10. The intelligent arc protection clothing vital signs monitoring method according to claim 1, 7 or 8, characterized in that: Vital sign monitoring is performed based on the predicted results of vital sign data in various dimensions at future moments, including: In response to a prediction result of a dimension exceeding a normal value range, or an actual value of the dimension data exceeding a normal value range, or a difference between the prediction result and the actual value of the dimension data exceeding a preset difference threshold, an abnormal vital sign occurs.

Images