Data acquisition and identification method of sensor

Abstract

The invention discloses a data acquisition and identification method of a sensor, which relates to the technical field of sensors, and comprises the following steps: sensor installation, data acquisition, data processing, processing of converted data by using a data weighted filtering algorithm, gait identification, and acquisition of gait data after data processing. The gait recognition algorithm is used for carrying out recognition and exercise risk assessment on the exercise mode of the current user, the health risk assessment algorithm is used for assessing the current risk state of the user, the health risk score and data storage at the current time are obtained, the weight is dynamically adjusted through the weighted filtering algorithm, the influence of noise and abnormal fluctuation is reduced, and the user experience is improved. The gait recognition algorithm is combined with acceleration and gait frequency, the motion state is accurately judged, errors are reduced, health risk assessment is combined with gait, heart rate and body temperature changes, the health condition of a user is comprehensively reflected, and reminding is provided when the motion intensity is too high.

Description

Technical Field

[0001] The present invention relates to the technical field of sensors, and in particular to a data collection and identification method for a sensor. Background Art

[0002] In recent years, due to the innovation of sensor technology, especially the technological progress of micro-electromechanical system sensors and biosensors, wearable smart devices containing a variety of microsensors have continued to emerge and are used to collect various data related to the human body, and then derive corresponding applications and services, such as smart homes in the field of automated control, positioning applications in security, remote monitoring of human health, etc. Wearable devices are not just a kind of hardware device, but also realize powerful functions through software support, data interaction, and cloud interaction.

[0003] In the process of sensor data collection in wearable smart devices, there is often noise interference, signal instability or mutation. Especially in the actual application environment, the device may be affected by vibration, environmental noise, etc., which reduces the quality of collected data. In addition, due to the complex changes in human gait, traditional methods usually only judge based on movement speed, resulting in increased recognition errors and difficulty in comprehensively and accurately evaluating the individual's healthy exercise status. Summary of the invention

[0004] The purpose of the present invention is to provide a data collection and identification method for a sensor, which solves the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a data collection and identification method of a sensor, comprising the following steps: Sensor installation: select a wearable device, embed a wireless communication module, a data standardization module and a sensor module inside the wearable device, and establish a connection between the wearable device and the data receiving end through the wireless communication module; Data collection, formulate data collection strategy, the sensor module collects data according to the data collection strategy, and sends the data to the data receiving end, and converts the data format through the data standardization module into data in a unified data format; Data processing, using data weighted filtering algorithm to process the converted data, to obtain the processed data E at the current time t , to ensure data quality; Gait recognition: After the data is processed, the gait recognition algorithm is used to identify the current user's movement pattern and obtain the total acceleration intensity A at the current time. total (t) and the current time gait frequency f gait (t); Movement risk assessment, combined with the current time gait frequency f gait(t) and the data collected by the sensor module, and use the health risk assessment algorithm to evaluate the user's current risk status, obtain the current time health risk score R(t), and set the risk threshold to Y3. When R(t) is greater than Y3, it means that the current user's exercise intensity is too high and will cause damage to the body. At this time, the risk information is sent to the data receiving end through the wireless communication module to remind the user; Data storage: The data collected by the sensor module is stored in the cloud database via Wi-Fi, and the user's health trends and exercise trajectories are displayed using data visualization tools so that users can view historical data and understand their exercise status. Optionally, the data receiving end includes a smart phone, and the wireless communication module is a Bluetooth module.

[0006] Optionally, the weighted filtering algorithm process is as follows: ; Where E t It is the data processed at the current time; S t It is the original data of the current time; W t is the dynamic weighting coefficient; E t－1 is the data processed at time t-1; Dynamic weighting coefficient W t It is expressed as: ; Where α is the adjustment constant, which is used to adjust the sensitivity of the dynamic enhancement coefficient to data changes, and its value is 0.8; |S t -S t－1 | is the absolute difference between the current time and the original data at time t-1; |S t -S －1 The larger the difference, the greater the change in the sensor data at the current time compared with the previous moment. At this time, the dynamic weighting coefficient W t The smaller the difference, the less dependence on historical data. On the contrary, if the difference is smaller, it means that the data changes steadily, and the dynamic weighting coefficient W t The larger the value, the more historical data will be affected by E t The impact is greater, thus being able to adjust its reliance on current and historical data based on fluctuations in sensor data.

[0007] Optionally, the gait recognition algorithm includes an acceleration unit and a gait frequency unit, and the calculation process of the acceleration unit is as follows: ; Among them A total(t) is the total acceleration intensity at the current time; A x (t) is the acceleration in the x-axis direction at the current time; A y (t) is the acceleration in the y-axis direction at the current time; A z (t) is the acceleration in the z-axis direction at the current time; The gait frequency unit calculation process is as follows: ; where f gait (t) represents the gait frequency at the current time; N is the length of the time window, which represents the time range used to calculate the acceleration; A total (i) is the total acceleration intensity at time i; Set the activity state threshold Y1 and gait classification threshold Y2, when A total (t)>Y1, it means the user is in motion. total When (t)≤Y1, it means the user is in a stationary state. total (t)>Y1 when the gait frequency f gait (t) Determine the user's motion state. gait (t)>Y2, it is judged as running. gait When (t)≤Y2, it is judged as walking.

[0008] Optionally, the health risk assessment algorithm process is as follows: ; Where R(t) is the health risk score at the current time; f gait (t) represents the gait frequency at the current time; β is f gait (t) influence coefficient, ranging from 0 to 1; L(t) is the heart rate at the current time; L avg It is the average heart rate over a period of time; γ is the heart rate influence coefficient, ranging from 0 to 1; T(t) is the body temperature at the current time; T avg It is the average temperature over a period of time; δ is the body temperature influence coefficient, ranging from 0 to 1; The larger the current health risk score R(t), the higher the current user's exercise intensity, and vice versa. The risk threshold of R(t) is set to Y3. When R(t) is greater than Y3, it means that the current user's exercise intensity is too high. At this time, the user is reminded through the data receiving end.

[0009] Optionally, when R(t) is greater than Y3, it indicates that the current user's exercise intensity is too high, and the heart rate and body temperature change range is too large. At this time, in order to ensure the accuracy of the data, the adjustment constant α in the weighted filtering algorithm is reduced from the original 0.8 to 0.6, which will increase the dynamic weighting coefficient W t , increasing reliance on historical data.

[0010] Optionally, in the data collection step, the data collection strategy is a combination of a change amplitude trigger and a time interval trigger, and the absolute difference between the original data at the current time and the time t-1 is set |S t -S t－1 The change threshold of | is D, that is, when |S t -S t－1 When | is greater than D, the sensor module will trigger data collection, and the time interval trigger is to automatically collect data intermittently according to the set time.

[0011] Optionally, in the data collection step, SSL encryption technology is used to encrypt and transmit the data when the data is sent to the data receiving end.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention processes data through a weighted filtering algorithm, dynamically adjusts weights according to the degree of data fluctuation, and can assign different weights to data at different times in the presence of noise or abnormal fluctuations, effectively reducing noise problems caused by environmental or equipment errors and improving data acquisition quality. The current motion module is then identified through a gait recognition algorithm, which makes a comprehensive judgment based on the acceleration of different axes, determines whether it is in a motion state based on the total acceleration intensity at the current time, and then determines what kind of motion state it is in based on the gait frequency at the current time, thereby improving the sensitivity of gait recognition, achieving accurate distinction of motion patterns, and reducing recognition errors. The health risk assessment algorithm comprehensively evaluates users by combining multiple factors such as gait frequency, heart rate changes, and body temperature changes, which can comprehensively reflect the user's health status. When the exercise intensity is too high, the user will be reminded to avoid excessive exercise and prevent damage to the body.

[0013] 2. When the current health risk score is greater than Y3, the present invention reduces the adjustment constant α to increase the reliance on historical data, thereby effectively reducing the impact of instantaneous data fluctuations on the system, allowing the method to be dynamically adjusted according to actual data, preventing the algorithm from making incorrect health risk assessments in a short period of time, and improving the stability and accuracy of the data. Through this dynamic adjustment, the method can better adapt to the drastic physiological changes under high-intensity exercise conditions, ensuring that the health risk score results are more reliable. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0015] 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 only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0016] For examples, see Figure 1 , this embodiment provides a sensor data collection and identification method, comprising the following steps: Sensor installation: select a wearable device, embed a wireless communication module, a data standardization module and a sensor module inside the wearable device, and establish a connection between the wearable device and the data receiving end through the wireless communication module; Data collection, formulate data collection strategy, the sensor module collects data according to the data collection strategy, and sends the data to the data receiving end, and converts the data format through the data standardization module into data in a unified data format; When data is sent to the data receiving end, SSL encryption technology is used to encrypt the data for transmission to ensure the security of data transmission; Data processing, using data weighted filtering algorithm to process the converted data, to obtain the processed data E at the current time t , ensure data quality; Gait recognition: After the data is processed, the gait recognition algorithm is used to identify the current user's movement pattern and obtain the total acceleration intensity A at the current time. total (t) and the current time gait frequency f gait (t); Movement risk assessment, combined with the current time gait frequency f gait(t) and the data collected by the sensor module, and use the health risk assessment algorithm to evaluate the user's current risk status, obtain the current time health risk score R(t), and set the risk threshold to Y3. When R(t) is greater than Y3, it means that the current user's exercise intensity is too high and will cause damage to the body. At this time, the risk information is sent to the data receiving end through the wireless communication module to remind the user; Data storage: The data collected by the sensor module is stored in the cloud database via Wi-Fi, and the user's health trends and exercise trajectories are displayed using data visualization tools so that users can view historical data and understand their exercise status.

[0017] More specifically, in this embodiment: after the data is collected, the data format is converted through the data standardization module to a unified data format, so that the sensor data can be analyzed later, and then the data is processed through the weighted filtering algorithm, and the weight is dynamically adjusted according to the degree of data fluctuation. In the case of noise or abnormal fluctuation, the data of the current time is given a lower weight, and the weight of the historical data is enhanced, so that the noise problem caused by the environment or equipment error can be effectively reduced, and the stability of the data is guaranteed. Then, the current motion module is identified through the gait recognition algorithm, and the gait recognition algorithm is combined with the acceleration of different axes for comprehensive judgment to obtain the total acceleration intensity A of the current time. total (t), and calculate the current time gait frequency f gait (t), through the total acceleration intensity A at the current time total (t) Determine whether it is in motion, and then calculate the gait frequency f at the current time. gait (t) Determine the state of motion, improve the sensitivity of gait recognition, accurately distinguish motion patterns with large gait frequency changes, and reduce recognition errors.

[0018] Finally, the user's current exercise status is scored through the health risk assessment algorithm to obtain the current time health risk score R(t). The health risk assessment algorithm conducts a comprehensive assessment of the user by combining multiple factors such as gait frequency, heart rate changes, and body temperature changes. It can comprehensively reflect the user's health status and set the risk threshold to Y3. When the exercise intensity is too high, the user will be reminded to avoid excessive exercise and prevent damage to the body.

[0019] Furthermore, the data receiving end includes a smart phone, and the wireless communication module is a Bluetooth module.

[0020] Specifically, the penetration rate of smartphones is extremely high, which avoids users from purchasing additional dedicated equipment, reducing hardware costs and thresholds. By setting the wireless communication module to a Bluetooth module, Bluetooth consumes extremely low power during communication, can achieve long-term connection, and can ensure that the device continues to run without frequent charging. It also has strong short-range communication capabilities and fast data transmission speeds, and can transmit sensor data to smartphones in real time, achieving rapid response and real-time monitoring, and improving user experience.

[0021] Furthermore, the weighted filtering algorithm process is as follows: ; Where E t It is the data processed at the current time; S t It is the original data of the current time; W t is the dynamic weighting coefficient; E t－1 is the data processed at time t-1; Dynamic weighting coefficient W t It is expressed as: ; Where α is the adjustment constant, which is used to adjust the sensitivity of the dynamic enhancement coefficient to data changes, and its value is 0.8; |S t -S t－1 | is the absolute difference between the current time and the original data at time t-1; Specifically, |S t -S －1 The larger the difference, the greater the change in the sensor data at the current time compared with the previous moment. At this time, the dynamic weighting coefficient W t The smaller the difference, the less dependence on historical data. On the contrary, if the difference is smaller, it means that the data changes steadily, and the dynamic weighting coefficient W t The larger the value, the more historical data will be affected by E t The impact is greater, and it can adjust its dependence on current data and historical data according to the fluctuation of sensor data, thereby effectively suppressing noise, maintaining the smoothness of data, and improving the quality of data collected by the sensor module.

[0022] Furthermore, the gait recognition algorithm includes an acceleration unit and a gait frequency unit, and the calculation process of the acceleration unit is as follows: ; Among them A total (t) is the total acceleration intensity at the current time; A x (t) is the acceleration in the x-axis direction at the current time; A y (t) is the acceleration in the y-axis direction at the current time; A z (t) is the acceleration in the z-axis direction at the current time; The gait frequency unit calculation process is as follows: ; where f gait (t) represents the gait frequency at the current time; N is the length of the time window, which represents the time range used to calculate the acceleration; A total (i) is the total acceleration intensity at time i; Specifically, the activity state threshold Y1 and gait classification threshold Y2 are set. total (t)>Y1, it means the user is in motion. total When (t)≤Y1, it means the user is in a stationary state. total (t)>Y1 when the gait frequency f gait (t) Determine the user's motion state. gait (t)>Y2, it is judged as running. gait When (t)≤Y2, it is judged as walking. By combining the accelerations of different axes, a comprehensive judgment is made to obtain the total acceleration intensity A at the current time. total (t), and calculate the current time gait frequency f gait (t), through the total acceleration intensity A at the current time total (t) Determine whether it is in motion, and then calculate the gait frequency f at the current time. gait (t) Determine the state of motion, improve the sensitivity of gait recognition, accurately distinguish the motion patterns, and reduce recognition errors.

[0023] Furthermore, the health risk assessment algorithm process is as follows: ; Where R(t) is the health risk score at the current time; f gait (t) represents the gait frequency at the current time; β is f gait (t) influence coefficient, ranging from 0 to 1; L(t) is the heart rate at the current time; L avg It is the average heart rate over a period of time; γ is the heart rate influence coefficient, ranging from 0 to 1; T(t) is the body temperature at the current time; Tavg It is the average temperature over a period of time; δ is the body temperature influence coefficient, ranging from 0 to 1; Specifically, the larger the current health risk score R(t), the higher the current user's exercise intensity, and vice versa. The risk threshold of R(t) is set to Y3. When R(t) is greater than Y3, it means that the current user's exercise intensity is too high. At this time, the user is reminded through the data receiving end. In actual applications, the risk threshold Y3 can be adjusted according to the user's gender, age, etc. to suit different groups of people. The health risk assessment algorithm comprehensively evaluates the user's current exercise status by combining multiple factors, which can avoid errors caused by single data and more comprehensively reflect the user's exercise status. By setting a threshold reminder mechanism, the user can be reminded in time when the exercise intensity is too high to avoid excessive damage to the body and achieve the effect of healthy exercise.

[0024] Furthermore, when R(t) is greater than Y3, it means that the current user's exercise intensity is too high, and the heart rate and body temperature change too much. At this time, in order to ensure the accuracy of the data, the adjustment constant α in the weighted filtering algorithm is reduced from the original 0.8 to 0.6, which will increase the dynamic weighting coefficient W. t , increasing reliance on historical data.

[0025] Specifically, when R(t) is greater than Y3, by reducing the adjustment constant α to increase reliance on historical data, the impact of instantaneous data fluctuations on the system can be effectively reduced, so that the method can be dynamically adjusted according to actual data. For example, excessive exercise in a short period of time can cause a sharp increase in heart rate or body temperature, but the rapid changes in heart rate and body temperature are temporary. By increasing reliance on historical data, the impact of current data on the algorithm output is reduced, preventing the algorithm from making incorrect health risk assessments in a short period of time, and improving data stability and accuracy. Through this dynamic adjustment, the method can better adapt to the drastic physiological changes under high-intensity exercise conditions, ensuring that the health risk scoring results are more reliable.

[0026] Furthermore, in the data collection step, the data collection strategy is a combination of change amplitude trigger and time interval trigger, setting the absolute difference between the current time and the original data at time t-1 | S t -S t－1 The change threshold of | is D, that is, when |S t -S t－1 When | is greater than D, the sensor module will trigger data collection, and the time interval trigger is to automatically collect data intermittently according to the set time.

[0027] Specifically, the data collection strategy based on conditional triggering can help the system collect data more efficiently, reduce unnecessary resource consumption, and collect data at critical moments according to actual needs. In actual applications, the time interval trigger can be changed according to the specific data. For example, if the time interval is set to M seconds, the sensor module will collect data every M seconds. When R(t) is greater than Y3, it means that the exercise intensity is too high at this time. In order to ensure the real-time nature of the data, the data collection interval can be shortened. For example, setting it to 2 / M seconds can ensure that this method collects data accurately and timely, meeting the needs of monitoring and analysis, while reducing power consumption.

[0028] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sensor data collection and identification method, characterized in that: The following steps are involved: Step S1: sensor installation, selecting a wearable device, embedding a wireless communication module, a data standardization module and a sensor module inside the wearable device, and establishing a connection between the wearable device and a data receiving end through the wireless communication module; Step S2: Data collection, formulate a data collection strategy, the sensor module collects data according to the data collection strategy, and sends the data to the data receiving end, and converts the data format through the data standardization module into data in a unified data format; Step S3: Data processing: using the data weighted filtering algorithm to process the converted data to obtain the processed data E at the current time t , to ensure data quality; Step S4: Gait recognition: after the data processing is completed, the gait recognition algorithm is used to identify the current user's movement pattern to obtain the total acceleration intensity A at the current time. total (t) and the current time gait frequency f gait (t); Step S5: Movement risk assessment, combined with the current time gait frequency f gait (t) and the data collected by the sensor module, and use the health risk assessment algorithm to evaluate the user's current risk status, obtain the current time health risk score R(t), and set the risk threshold to Y3. When R(t) is greater than Y3, it means that the current user's exercise intensity is too high and will cause damage to the body. At this time, the risk information is sent to the data receiving end through the wireless communication module to remind the user; Step S6: Data storage: The data collected by the sensor module is stored in a cloud database via Wi-Fi, and the user's health trends and exercise trajectories are displayed using data visualization tools so that the user can view historical data and understand the exercise status.

2. The sensor data collection and identification method according to claim 1, characterized in that: The data receiving end includes a smart phone, and the wireless communication module is a Bluetooth module.

3. The sensor data collection and identification method according to claim 2, characterized in that: The weighted filtering algorithm process is as follows: ； Where E t It is the data processed at the current time; S t It is the original data of the current time; W t is the dynamic weighting coefficient; E t－1 is the data processed at time t-1; Dynamic weighting coefficient W t It is expressed as: ； Where α is the adjustment constant, which is used to adjust the sensitivity of the dynamic enhancement coefficient to data changes, and its value is 0.8; |S t -S t－1 | is the absolute difference between the current time and the original data at time t-1; |S t -S －1 The larger the difference, the greater the change in the sensor data at the current time compared with the previous moment. At this time, the dynamic weighting coefficient W t The smaller the difference, the less reliance on historical data. On the contrary, if the difference is smaller, it means that the data changes steadily, and the dynamic weighting coefficient W t The larger the value, the more historical data will be affected by E t The impact is greater, thus being able to adjust its reliance on current and historical data based on fluctuations in sensor data.

4. The sensor data collection and identification method according to claim 3, characterized in that: The gait recognition algorithm includes an acceleration unit and a gait frequency unit. The calculation process of the acceleration unit is as follows: ； Among them A total (t) is the total acceleration intensity at the current time; A x (t) is the acceleration in the x-axis direction at the current time; A y (t) is the acceleration in the y-axis direction at the current time; A z (t) is the acceleration in the z-axis direction at the current time; The gait frequency unit calculation process is as follows: ； where f gait (t) represents the gait frequency at the current time; N is the length of the time window, which represents the time range used to calculate the acceleration; A total (i) is the total acceleration intensity at time i; Set the activity state threshold Y1 and gait classification threshold Y2, when A total (t)>Y1, it means the user is in motion. total When (t)≤Y1, it means the user is in a stationary state. total (t)>Y1 when the gait frequency f gait (t) Determine the user's motion state. gait (t)>Y2, it is judged as running. gait When (t)≤Y2, it is judged as walking.

5. The sensor data collection and identification method according to claim 4, characterized in that: The health risk assessment algorithm process is as follows: ； Where R(t) is the health risk score at the current time; f gait (t) represents the gait frequency at the current time; β is f gait (t) influence coefficient, ranging from 0 to 1; L(t) is the heart rate at the current time; L avg It is the average heart rate over a period of time; γ is the heart rate influence coefficient, ranging from 0 to 1; T(t) is the body temperature at the current time; T avg It is the average body temperature over a period of time; δ is the body temperature influence coefficient, ranging from 0 to 1; The larger the current health risk score R(t), the higher the current user's exercise intensity, and vice versa. The risk threshold of R(t) is set to Y3. When R(t) is greater than Y3, it means that the current user's exercise intensity is too high. At this time, the user is reminded through the data receiving end.

6. The sensor data collection and identification method according to claim 5, characterized in that: When R(t) is greater than Y3, it means that the current user's exercise intensity is too high, and the heart rate and body temperature change too much. In order to ensure the accuracy of the data, the adjustment constant α in the weighted filtering algorithm is reduced from the original 0.8 to 0.6, which will increase the dynamic weighting coefficient W. t , increasing reliance on historical data.

7. The sensor data collection and identification method according to claim 1, characterized in that: The data acquisition strategy in the data acquisition step is a combination of change amplitude triggering and time interval triggering, setting the absolute difference between the current time and the original data at time t-1 | S t -S t－1 The change threshold of | is D, that is, when |S t -S t－1 When | is greater than D, the sensor module will trigger data collection, and the time interval trigger is to automatically collect data intermittently according to the set time.

8. The sensor data collection and identification method according to claim 1, characterized in that: In the data collection step, when the data is sent to the data receiving end, the SSL encryption technology is used to encrypt the data for transmission.

Images