Human body posture detection device and detection method based on flexible FBG packaging

Abstract

The invention relates to a human body posture detection device and detection method based on flexible FBG packaging, and belongs to the field of optical fiber sensing, and the method is characterized in that an FBG sensing unit is flexibly packaged by silica gel, such a packaging mode can not only effectively protect the FBG and enable the FBG to stably work in a complex use environment, but also fully utilize the flexible characteristic of the silica gel, so that the FBG sensing unit can stably work in a complex use environment. And various changes of the feet of the human body during movement can be better adapted. The central wavelength of the FBG is drifted by utilizing pressure generated by treading of a human body, a high-precision fiber bragg grating demodulator is used for sampling and repeatedly measuring and recording a drifting value, acquired foot pressure data is imported into a random forest algorithm model subjected to K-fold cross validation and grid search adjustment for classified prediction, no electrical element participates in the whole process, and the accuracy of foot pressure prediction is greatly improved. Accurate detection of the common postures of the human body is achieved, and the method has innovativeness and practicability and has wide application prospects and important practical value in the fields of health monitoring, motion analysis, rehabilitation treatment and the like.

Description

Technical Field

[0001] The invention belongs to the field of optical fiber sensing, and in particular relates to a human body posture detection device and a detection method based on flexible FBG packaging. Background Art

[0002] In recent years, with the rapid development of science and technology and people's increasing attention to fields such as health monitoring and motion analysis, the importance of human posture detection technology has become increasingly significant. In many application scenarios, accurate detection of human posture is of key significance for protecting people's health, improving exercise effects and promoting the development of related industries. For example, in the field of medical rehabilitation, accurate posture detection can provide a scientific basis for patients' rehabilitation treatment and help doctors develop more effective rehabilitation plans; in the field of sports training, it can monitor athletes' movements and postures in real time, help coaches conduct targeted training guidance, improve training effects and prevent sports injuries; in emerging fields such as smart homes and virtual reality, accurate human posture recognition is one of the core technologies for realizing intelligent interaction and immersive experience.

[0003] Traditional human posture detection technologies mainly include inertial sensor technology, pressure sensor technology, and visual recognition technology. Inertial sensor technology uses accelerometers and gyroscopes to measure the motion information of objects and then infer human posture. It is widely used in wearable devices and professional sports monitoring. Pressure sensor technology has many applications in flexible electronics, smart mattresses and other devices that come into contact with the human body. The detection device detects the pressure distribution of the foot through pressure sensors arranged in different positions, thereby judging the posture information such as the center of gravity transfer and footsteps of the human body, and providing data support for rehabilitation training. Visual recognition technology uses a camera to obtain human image information, and then uses image processing and analysis algorithms to identify and judge human posture.

[0004] However, the measurement results of inertial sensors will have the problem of error accumulation over time. The detection accuracy and resolution of pressure sensors have certain limitations, and they may not be able to accurately detect small changes in human foot pressure. Visual recognition technology is extremely sensitive to ambient lighting conditions. In strong or low light environments, the image quality captured by the camera will be seriously affected, resulting in a significant decrease in the accuracy of human posture recognition, and inevitably causing privacy issues.

[0005] Fiber Bragg grating technology has achieved certain application results in other fields such as structural health monitoring and communications. In structural health monitoring, fiber Bragg grating can be pasted or embedded into the materials of large structures such as bridges, buildings, aircraft, satellites, etc., to monitor the changes of physical quantities such as strain and temperature in real time, providing an important basis for timely discovery of structural safety hazards and subsequent maintenance and management.

[0006] When fiber Bragg grating technology is applied to human posture detection, it shows certain potential and advantages. It can effectively overcome the problem that traditional sensors are easily affected by temperature and noise, and has the advantages of light weight, small size, no electromagnetic interference and high sensitivity, providing a new solution for human posture detection. Summary of the invention

[0007] In order to solve the problem that the current visual recognition technology is extremely sensitive to ambient lighting conditions, the image quality captured by the camera will be seriously affected in strong or weak light environments, resulting in a significant decrease in the accuracy of human posture recognition. The present invention provides a human posture detection device and detection method based on flexible FBG packaging, which can accurately detect common human postures and has unique advantages.

[0008] The technical solution of a human posture detection device and detection method based on flexible FBG packaging of the present invention is:

[0009] A human posture detection device based on flexible FBG packaging comprises: a wooden base body, wherein a plurality of FBG sensing units for sensing human foot pressure and causing the central wavelength of the FBG to drift are flexibly packaged in silicone inside the wooden base body, and the FBG sensing units are connected to a fiber grating demodulator for sampling the central wavelength drift of the FBG sensing units and recording the drift value.

[0010] A further improvement of the technical solution of the present invention is that the thickness of the silicone flexible package is 4 mm.

[0011] A further improvement of the technical solution of the present invention is that the FBG sensor units are arranged at four positions: the thumb, the first metatarsal, the third metatarsal, and the heel.

[0012] The further improvement of the technical solution of the present invention is that: the sampling frequency of the high-precision fiber Bragg grating demodulator is 100 Hz;

[0013] The accuracy of the high-precision fiber Bragg grating demodulator is ≤3pm.

[0014] A human posture detection method based on flexible FBG packaging, using the above-mentioned human posture detection device based on flexible FBG packaging, comprises the following steps:

[0015] S1. Establish the movement standards for the experiment and conduct theoretical analysis of the force points on the feet of normal people;

[0016] S2, encapsulate the FBG sensor unit and make a detection device, perform repeated experiments and collect data;

[0017] S3, determine the best package size and depth, and perform simulation analysis to determine whether the sensor data is usable;

[0018] S4. Import the data into the random forest algorithm for training, evaluate the accuracy of the random forest algorithm's predictions and improve the algorithm model until the expected results are achieved.

[0019] A further improvement of the above technical solution of the present invention is that: the step S4 specifically includes the following steps:

[0020] S4.1. Score standardization and normalization of experimental data;

[0021] S4.2, perform grid division within the set parameter range and determine multiple parameter combinations; at the same time, randomly divide the experimental data into K groups;

[0022] S4.3, select different parameter combinations from the divided grids in turn to train the model;

[0023] S4.4. According to the test results of each parameter combination in S4.3, select the parameter combination with better performance and retrain the model;

[0024] S4.5. Carefully determine whether the current parameter combination is optimal; if the parameters are not optimal, return to step S4.2, readjust the parameter range and grid division, and continue experimental exploration.

[0025] The present invention provides a human body posture detection device and detection method based on flexible FBG packaging, which has the following beneficial effects:

[0026] The present invention flexibly encapsulates the FBG sensor unit with silicone and then embeds it into the wooden base body. When the human foot generates pressure, the central wavelength of the FBG sensor unit drifts. The wavelength drift is sampled and recorded by a high-precision fiber grating demodulator, and then the data is imported into a random forest algorithm model optimized by K-fold cross validation and grid search for classification prediction.

[0027] Combining fiber grating technology with machine learning algorithms, and taking advantage of the comfort and stability of the wooden substrate, accurate detection of common human postures is achieved. Silicone encapsulation ensures the safety and stable performance of the FBG, and the layout in specific positions improves the accuracy of pressure perception. The high-precision sampling of the high-precision fiber grating demodulator and the optimization of the random forest algorithm model make the invention have higher detection accuracy and reliability. The FBG sensing unit in the present invention does not require the use of a power supply, thereby avoiding problems such as electromagnetic interference, and improving the stability and safety of the present invention. The present invention can be widely used in fields such as health monitoring and motion analysis, providing strong support for people's lives and health. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1A schematic diagram of the structure of a human posture detection device based on a flexible packaged FBG provided by the present invention;

[0029] Figure 2 It is a flow chart of a human posture detection method based on flexible FBG packaging of the present invention;

[0030] Figure 3 The invention discloses a classification prediction algorithm based on an improved random forest algorithm for a human posture detection method based on a flexible packaged FBG.

[0031] In the attached drawings: 1. wooden base body; 2. first FBG sensor unit; 3. second FBG sensor unit; 4. third FBG sensor unit; 5. fourth FBG sensor unit. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all 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.

[0033] The present invention provides a human posture detection device based on flexible packaged FBG, referring to Figure 1 It can be seen that it includes a wooden base body 1; four FBG sensor units flexibly encapsulated in the wooden base body 1 using silicone, the FBG sensor unit is a fiber Bragg grating commonly used in the prior art, and the four FBG sensor units with different central wavelengths are cascaded, and the central wavelengths of the four FBG sensor units are 1540nm, 1550nm, 1560nm, and 1570nm respectively; the four FBG sensor units are respectively a first FBG sensor unit 2, a second FBG sensor unit 3, a third FBG sensor unit 4, and a fourth FBG sensor unit 5; wherein the first FBG sensor unit 2 corresponds to the big toe on the sole, the second FBG sensor unit 3 corresponds to the first metatarsal on the sole, the third FBG sensor unit 4 corresponds to the third metatarsal on the sole, and the fourth FBG sensor unit 5 corresponds to the heel on the sole. In the embodiment of the present invention, the FBG sensor unit is encapsulated in the wooden base body on a mold, and the dimensions of the mold are 4mm in thickness, 30mm in length, and 15mm in width. The packaging mold was 3D printed using white resin, and the embedding depth of the FBG sensor unit was set to half the thickness of the mold, i.e., 2 mm.

[0034] Before using molds and silicone to flexibly encapsulate the FBG sensor unit, a series of preparations need to be done. First, the FBG sensor unit must be cleaned to remove impurities and contaminants on the surface to ensure that its surface is smooth for subsequent encapsulation operations. At the same time, the Shore hardness of the silicone used is 40 degrees.

[0035] When the preparation work is completed, the flexible packaging operation of the FBG sensor unit is started. Pour the prepared silicone material into the mold to form a silicone layer of a certain thickness. Then, place the FBG sensor unit flatly in the middle of the silicone layer to ensure that the position of the FBG sensor unit is accurate and remains in a straight line. In the process of placing the FBG sensor unit, avoid contact between the FBG sensor unit and the edge of the mold or other objects to avoid damage to the FBG sensor unit. Then, use special tools or equipment to heat the silicone to solidify it and encapsulate the gate area of ​​the FBG sensor unit inside the silicone. During the curing process, control the heating temperature to 80°C and the time to 3 hours to ensure the performance and packaging quality of the silicone. After the silicone is cured, check whether the encapsulated FBG sensor unit has bubbles, wrinkles or other defects.

[0036] The present invention also provides a human posture detection method based on flexible package FBG, referring to Figure 2 and Figure 3 It can be seen that the following steps are included:

[0037] Step S1, formulate the experimental action (posture) standard and conduct theoretical analysis of the force points of normal human feet.

[0038] First, an in-depth theoretical analysis of the force points of the normal human foot is conducted. By consulting literature, referring to previous research results, and applying relevant theoretical knowledge such as biomechanics, the main force areas and force characteristics of the foot under different movements are determined.

[0039] Step S2, encapsulating the FBG sensor unit and making the above detection device, performing repeated experiments and collecting data.

[0040] Select appropriate FBG sensor units and materials for packaging. Pre-treat the FBG sensor units to remove impurities on their surfaces. Determine the packaging size and method, and perform packaging operations according to the preset plan. Embed the packaged FBG sensor units in specific locations. Repeat the experiment to simulate different human body movements and postures, and use the fiber Bragg grating demodulator to collect data from the sensor device under different movements, including but not limited to posture data, wavelength changes and other information.

[0041] Step S3: determine the best package size and depth, and perform simulation analysis to determine whether the sensing data is usable.

[0042] By changing the package size and depth, multiple groups of experiments are performed. The collected data is analyzed and processed, and the data under different packaging conditions are simulated and analyzed with simulation software to evaluate the accuracy and reliability of the sensor data and determine the best package size and depth combination. If the sensor data is not available, return to step S2 to adjust the package parameters and conduct experiments and analysis again.

[0043] Step S4, importing the sensor data in the above step S3 into the random forest algorithm for training, evaluating the accuracy of the random forest algorithm prediction and improving the algorithm model until the expected result is achieved.

[0044] Import the collected valid sensor data into the pre-designed random forest algorithm model, train the posture data, and enable it to accurately predict human motion posture based on the sensor data. Evaluate the accuracy of the random forest algorithm prediction by comparing it with known standard motion posture data. If the accuracy of the random forest algorithm prediction does not meet the expected requirements, analyze the reasons and adjust and optimize the random forest algorithm model, such as increasing the amount of training data, adjusting algorithm parameters, and improving the algorithm structure. Repeat steps S3 and S4 until the accuracy of the algorithm prediction reaches the expected result.

[0045] When the above step S4 is implemented, refer to Figure 3 It can be seen that the following steps are included:

[0046] Step S4.1: Data collation.

[0047] The sensor data in step S3 is score-standardized and normalized to lay the foundation for subsequent analysis. In addition, the value range of key parameters is reasonably set according to experience and experimental requirements to make full preparations for subsequent parameter optimization.

[0048] Step S4.2: detailed planning and grouping.

[0049] Grid division is performed within the set parameter range to determine multiple possible parameter combinations. At the same time, the experimental data is randomly divided into K groups to provide a data basis for cross-validation. This grouping method can more comprehensively evaluate the performance of the model on different data subsets and effectively avoid overfitting.

[0050] Step S4.3, cyclic exploration and testing.

[0051] Different parameter combinations are selected from the divided grids to train the model. During the training process, the accuracy and generalization ability of the model are improved according to the characteristics of the model and the characteristics of the data. After the training is completed, the model is simulated and tested using other groups of data that did not participate in the training to evaluate the performance of the model. By continuously looping this process, each parameter combination is fully tested and evaluated.

[0052] Step S4.4, optimization training and statistical analysis.

[0053] According to the test results of the previous parameter combinations, select the parameter combinations with better performance and retrain the model. Through repeated training, statistical analysis is performed on the experimental results under different parameter combinations, and the mean is taken as the final performance indicator. This can reduce the impact of random factors and more accurately evaluate the performance of the model.

[0054] Step S4.5, accurate judgment and adjustment.

[0055] Carefully determine whether the current parameter combination is optimal. This can be determined based on a preset performance indicator threshold or a comparison with other methods. If the parameters are not optimal, return to step S4.2, readjust the parameter range and grid division, and continue experimental exploration. Through continuous adjustment and optimization, gradually approach the optimal parameter combination.

[0056] Theoretical and experimental results show that when the human body performs a variety of common posture changes, such as standing on both feet, sitting, standing on one foot, leaning forward, leaning backward and folding forward, etc. The experiment collected data from 10 volunteers aged 22-25 (5 males and 5 females). Each person repeated each group of actions five times, and each sampling lasted for 10s with a sampling interval of 20s. Finally, 300 sets of data were obtained, with a prediction accuracy of 97%-99%. The detection device can accurately detect and distinguish different postures. Among them, the thickness of the silicone package is 4mm, the FBG sensor unit is arranged at four positions: the thumb, the first metatarsal, the third metatarsal and the heel, the sampling frequency of the high-precision fiber grating demodulator is 100Hz, the accuracy is ≤3pm, and the number of decision trees and the maximum depth in the random forest algorithm model are 100 and 5. In addition, the invention is not affected by factors such as age, gender, and weight, and has good universality. It can be seen that the detection device has the advantages of structural innovation, stable performance, and high detection accuracy, and can be widely used in health monitoring, motion analysis, rehabilitation treatment and other fields.

[0057] In the above embodiment, the present invention proposes a human posture detection device and detection method based on flexible FBG packaging. The present invention flexibly packages the FBG sensor unit with silicone and then embeds it into the wooden base body. When the human foot generates pressure, the central wavelength of the FBG sensor unit drifts. The wavelength drift is sampled and recorded by a high-precision fiber grating demodulator, and then the data is imported into a random forest algorithm model optimized by K-fold cross validation and grid search for classification prediction. The entire invention does not involve electrical components, avoids electromagnetic interference and other problems, and improves the stability and safety of the invention.

[0058] The above-described embodiments are merely descriptions of preferred implementations of the present invention, and are not intended to limit the concept and scope of the present invention. Without departing from the design concept of the present invention, various modifications and improvements made by ordinary persons in the art to the technical solution of the present invention should fall within the protection scope of the present invention, and the technical contents for which protection is sought in the present invention have been fully recorded in the claims.

Claims

1. A human posture detection device based on flexible FBG packaging, characterized in that: include: A wooden base body (1), wherein a plurality of FBG sensor units for sensing human foot pressure and causing the central wavelength of the FBG to drift are flexibly encapsulated in silicone inside the wooden base body (1), and the FBG sensor units are connected to a fiber optic Bragg grating demodulator for sampling the central wavelength drift of the FBG sensor units and recording the drift value via a single-mode optical fiber jumper.

2. A human posture detection device based on flexible FBG packaging according to claim 1, characterized in that: The thickness of the silicone flexible package is 4 mm.

3. A human posture detection device based on flexible FBG packaging according to claim 1, characterized in that: The FBG sensing units are arranged at four locations: the thumb, the first metatarsal bone, the third metatarsal bone, and the heel.

4. A human posture detection device based on flexible FBG packaging according to claim 1, characterized in that: The sampling frequency of the high-precision fiber Bragg grating demodulator is 100 Hz; The accuracy of the high-precision fiber Bragg grating demodulator is ≤3pm.

5. A human posture detection method based on flexible FBG packaging, characterized in that: The human posture detection device using the flexible FBG package of any one of claims 1 to 4 comprises the following steps: S1. Establish the movement standards for the experiment and conduct theoretical analysis of the force points on the feet of normal people; S2, encapsulate the FBG sensor unit and make a detection device, perform repeated experiments and collect data; S3, determine the best package size and depth, and perform simulation analysis to determine whether the sensor data is usable; S4. Import the data into the random forest algorithm for training, evaluate the accuracy of the random forest algorithm's predictions and improve the algorithm model until the expected results are achieved.

6. A human posture detection method based on flexible FBG packaging according to claim 5, characterized in that: The step S4 specifically includes the following steps: S4.

1. Score standardization and normalization of experimental data; S4.2, meshing is performed within the set parameter range to determine multiple parameter combinations; at the same time, the experimental data are randomly divided into K groups; S4.3, select different parameter combinations from the divided grids in turn to train the model; S4.

4. According to the test results of each parameter combination in S4.3, select the parameter combination with better performance and re-train the model; S4.

5. Carefully determine whether the current parameter combination is optimal; if the parameters are not optimal, return to step S4.2, readjust the parameter range and grid division, and continue experimental exploration.

Images