Fall judgment system and method based on genetic training

Abstract

The invention provides a fall judgment system based on genetic training, which comprises a protective garment, a microprocessor device and a power supply device are arranged in the protective garment, air bag devices are arranged at the neckline and the waist of the protective garment, and air inlets of the air bag devices are connected with an inflation device in the protective garment. The protective clothing is characterized in that magic tape devices are arranged on the neck and the waist of the protective clothing, N pairs of sensor modules are fixed in the magic tape devices, and each sensor module is provided with an acceleration sensor and a gyroscope sensor. The anti-falling device has the advantages of being simple in structure, capable of accurately judging whether a wearer is about to fall down or not and providing safety protection for the wearer about to fall down in time, low in use cost, suitable for most families and capable of being produced universally.

Description

Technical Field

[0001] This invention relates to the field of protective wearable devices, and in particular to a fall detection system and method based on genetic training. Background Technology

[0002] With the rapid aging of society and the increasing number of elderly people, their health is receiving more attention, especially regarding the issue of falls among the elderly, leading to the development of numerous protective products. However, currently available protective products often suffer from inaccurate fall detection and untimely protection. Furthermore, these products are often heavy, inconvenient to carry, and expensive, making them unsuitable for ordinary households.

[0003] The drawbacks of existing technology are: it is not accurate enough in judging falls, it is not timely in providing protection, it is not convenient to carry, and it is relatively expensive. Summary of the Invention

[0004] The main objective of this invention is to provide a fall detection system and method based on genetic training, which can accurately determine whether the wearer is about to fall and provide timely safety protection for the wearer who is about to fall.

[0005] To achieve the above objectives, the present invention provides a fall detection system based on genetic training, comprising a protective suit, which contains a microprocessor and a power supply. Airbags are located at the collar and waist of the protective suit, with their air inlets connected to an inflation device inside the suit. Crucially, the protective suit has Velcro fasteners at the neck and waist, each fastener housing N pairs of sensor modules, each equipped with an accelerometer and a gyroscope.

[0006] The N pairs of sensor modules are used to collect acceleration data α and angular velocity data ω of the wearer's neck and waist in real time from various directions, such as front and back, left and right, left front and right back, and right front and left back, to ensure that the wearer can be detected by the corresponding sensor module no matter which direction they fall from. The Velcro device's design for fixing the sensor modules ensures that each pair of sensor modules always collects acceleration data α and angular velocity data ω from the corresponding direction.

[0007] Preferably, the outputs of all the sensor modules are connected to the input of the same microprocessor device, the output of which is connected to the inflation device, and the power supply device is used to power the other modules.

[0008] Through the above design, each pair of sensor modules collects acceleration data α and angular velocity data ω from two different parts of the protective suit, namely the neck and waist, and transmits them to a microprocessor device. The microprocessor device performs neural network calculations on the acceleration data α and angular velocity data ω collected by each pair of sensor modules, and then compares the calculated acceleration weight X3 and angular velocity weight X4 with the fall threshold weight parameters C9 and C1, respectively. 10 The system compares the data to determine if the wearer is about to fall. If the wearer is not about to fall, it continues to monitor the acceleration data α and angular velocity data ω of the neck and waist of the protective suit. If the wearer is about to fall, the microprocessor immediately outputs a control command to the inflation device, which instantly inflates the airbag, providing timely protection for the body and preventing serious injury from a fall.

[0009] Preferably, the microprocessor device is also wirelessly connected to a user mobile terminal. The user mobile terminal is used to input the wearer's age (Y), gender (S), height (H), and weight (W) data into the microprocessor device.

[0010] A fall detection method based on genetic training, the key of which includes the following steps:

[0011] Step 1: This step involves inputting the wearer's age Y, gender S, height H, and weight W into the microprocessor device via the user's mobile terminal, completing the initialization of the microprocessor device, and determining the influence weights X1 and X2 of acceleration and angular velocity determined by the wearer's age Y, gender S, height H, and weight W data.

[0012] Step 2: The step of using N pairs of sensor modules to collect the acceleration data α and angular velocity data ω of the wearer's neck and waist in real time and transmit them to the microprocessor device;

[0013] Step 3: The microprocessor device receives and saves the acceleration data α and angular velocity data ω, then performs neural network calculations, and compares the calculated acceleration weights X3 and angular velocity weights X4 with the fall threshold weight parameters C9 and C1, respectively. 10 The steps for comparison;

[0014] When two or more pairs of acceleration weights X3 > C9 and angular velocity weights X4 > C 10 If the wearer is about to fall, proceed to step 4; otherwise, proceed to step 2.

[0015] Step 4: This step involves using a microprocessor device to send control commands to an inflation device, which instantly inflates the airbag to provide fall protection for the wearer.

[0016] Through the above design, the present invention realizes the detection of the wearer's behavior and posture during normal walking and the accurate judgment of the wearer's fall behavior. It also implements fall protection for the wearer who is about to fall through the airbag device, effectively avoiding serious physical injury caused by the fall.

[0017] Preferably, in step 1, the weighting formula for the influence of the wearer's age Y, gender S, height H, and weight W on the wearer's acceleration during walking or falling is as follows:

[0018] X1 = C1×Y + C2×S + C3×H + C4×W

[0019] Where C1 is the acceleration weight parameter for age Y, C2 is the acceleration weight parameter for gender S, C3 is the acceleration weight parameter for height H, and C4 is the acceleration weight parameter for weight W.

[0020] The formula for calculating the weighted average (X2) of the influence of the wearer's age (Y), gender (S), height (H), and weight (W) on the wearer's angular velocity during walking or falling is as follows:

[0021] X2 = C5×Y + C6×S + C7×H + C8×W

[0022] Wherein, C5 is the angular velocity weight parameter for age Y, C6 is the angular velocity weight parameter for gender S, C7 is the angular velocity weight parameter for height H, and C8 is the angular velocity weight parameter for weight W.

[0023] Preferably, in step 3, the formula for calculating the acceleration weight X3 is as follows:

[0024]

[0025] Among them, a x a y a z These represent the accelerations of the human body in the X, Y, and Z directions, respectively.

[0026] The formula for calculating the angular velocity weight X4 is as follows:

[0027]

[0028] Wherein, where ω x ω y ω z These are the angular velocities of the human body in the X, Y, and Z directions, respectively.

[0029] The weighting parameter C1-C 10 The following training process was used to obtain:

[0030] Step S1: The product developer uses sensors to collect a large amount of acceleration and angular velocity data of people of different ages Y, genders S, heights H and weights W during normal walking and falling, to obtain the initial feature dataset.

[0031] Step S2: Construct a neural network model, input the initial feature dataset for network algorithm training, continuously optimize the neural network model and its weight parameters, and finally obtain and save the optimal fall judgment model and its weight parameters.

[0032] The network algorithm training formula in step S2 is consistent with the calculation formulas in steps 1 and 3. A large amount of data from the initial feature dataset is used to train the network algorithm, continuously optimizing the weight parameters C1-C. 10 The optimal fall judgment model is obtained by finding the value of , which can accurately judge the fall behavior of the vast majority of people.

[0033] The optimal fall detection model is stored in the microprocessor device. This optimal fall detection model is periodically trained using a genetic algorithm to optimize the acceleration and angular velocity data of the wearer during normal walking and during falls, which are stored in the microprocessor device, and to complete the optimization of the weight parameters C1-C1. 10 Minor optimizations are made and saved, and the data is continuously optimized and updated as the wearer ages, making it more in line with the wearer's daily habits and making fall judgment results more accurate.

[0034] The genetic algorithm optimization training process is as follows:

[0035] Step 1: Generate the initial population through hybrid encoding: The encoding formula is as follows

[0036] x i =l i +β(u i -l i )

[0037] Among them, u i l is the upper limit of the gene solution space. i β is the lower bound of the gene solution space, and β is a random number between 0 and 1, generating an initial population with M chromosomes.

[0038] Step 2: Calculate the population fitness: calculate the difference between the model's predicted data and the actual data.

[0039] Step 3: Selection operation: If the fitness is greater than the set threshold a, a feedback mutation based on the dynamic coding idea is performed to recalculate the fitness of individuals in the population and sort the fitness values ​​of individuals in the entire population; if the fitness is less than the set threshold a, a roulette wheel selection method is used to select two chromosomes for crossover operation, and then an orthogonal matrix experiment is performed to produce a better chromosome.

[0040] Step 4: Repeat steps 2-3 until the algorithm terminates when the termination condition is met.

[0041] The genetic algorithm does not run during normal use. It is only optimized based on historical data stored in the microprocessor device when charging. The model parameters are updated only after the algorithm obtains better parameters.

[0042] The significant advantages of this invention are: simple structure, accurate judgment of whether the wearer is about to fall, timely safety protection for the wearer who is about to fall, low cost of use, suitable for most families, and can be widely produced. Attached Figure Description

[0043] Figure 1 This is a schematic diagram of the structure of an embodiment. Detailed Implementation

[0044] The present invention will be further described in detail below with reference to the accompanying drawings and specific examples. The following embodiments or drawings are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0045] like Figure 1 As shown: A fall detection system based on genetic training includes a protective suit. The protective suit is equipped with a microprocessor device 1 and a power supply device 2. An airbag device 3 is provided at the collar and waist of the protective suit. The air inlet of the airbag device 3 is connected to an inflation device 4 inside the protective suit. The key feature is that the neck and waist of the protective suit are provided with Velcro devices 5. N pairs of sensor modules are fixed inside the Velcro devices 5. Each sensor module is equipped with an accelerometer sensor and a gyroscope sensor.

[0046] The outputs of all the sensor modules are connected to the input of the same microprocessor device 1, the output of which is connected to the inflation device 4. The power supply device 2 is used to supply power to the other modules.

[0047] The microprocessor device 1 is also wirelessly connected to the user's mobile terminal.

[0048] The genetically trained fall detection system specifically achieves accurate judgment and protection against falls by using a genetically trained fall detection method, including the following steps:

[0049] Step 1: This step involves inputting the wearer's age Y, gender S, height H, and weight W into the microprocessor device 1 via a user mobile terminal, completing the initialization of the microprocessor device 1, and determining the influence weights X1 and X2 of acceleration and angular velocity determined by the wearer's age Y, gender S, height H, and weight W data.

[0050] Step 2: A step for using N pairs of sensor modules to collect in real time the acceleration data α and angular velocity data ω of the wearer's neck and waist, and to transmit them to the microprocessor device 1;

[0051] Step 3: The microprocessor device 1 receives and saves the acceleration data α and angular velocity data ω, then performs neural network calculations, and compares the calculated acceleration weights X3 and angular velocity weights X4 with the fall threshold weight parameters C9 and C1, respectively. 10 The steps for comparison;

[0052] When two or more pairs of acceleration weights X3 > C9 and angular velocity weights X4 > C 10 If the wearer is about to fall, proceed to step 4; otherwise, proceed to step 2.

[0053] Step 4: The step of sending control commands from the microprocessor device 1 to the inflation device 4, which inflates the airbag device 3 instantly to provide fall protection for the wearer.

[0054] Specifically, in step 1, the formula for calculating the weight X1 of the influence of the wearer's age Y, gender S, height H, and weight W on the wearer's acceleration when walking or falling is as follows:

[0055] X1 = C1×Y + C2×S + C3×H + C4×W

[0056] Where C1 is the acceleration weight parameter for age Y, C2 is the acceleration weight parameter for gender S, C3 is the acceleration weight parameter for height H, and C4 is the acceleration weight parameter for weight W.

[0057] The formula for calculating the weighted average (X2) of the influence of the wearer's age (Y), gender (S), height (H), and weight (W) on the wearer's angular velocity during walking or falling is as follows:

[0058] X2 = C5×Y + C6×S + C7×H + C8×W

[0059] Wherein, C5 is the angular velocity weight parameter for age Y, C6 is the angular velocity weight parameter for gender S, C7 is the angular velocity weight parameter for height H, and C8 is the angular velocity weight parameter for weight W.

[0060] In step 3, the formula for calculating the acceleration weight X3 is as follows:

[0061]

[0062] Among them, a x a y a zThese represent the accelerations of the human body in the X, Y, and Z directions, respectively.

[0063] The formula for calculating the angular velocity weight X4 is as follows:

[0064]

[0065] Wherein, where ω x ω y ω z These are the angular velocities of the human body in the X, Y, and Z directions, respectively.

[0066] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A fall detection system based on genetic training, comprising a protective suit, wherein a microprocessor device (1) and a power supply device (2) are disposed inside the protective suit, and an airbag device (3) is disposed at the collar and waist of the protective suit, the air inlet of the airbag device (3) being connected to an inflation device (4) inside the protective suit, characterized in that: The protective suit is provided with Velcro devices (5) at the neck and waist. N pairs of sensor modules are fixed inside the Velcro devices (5), and each sensor module is provided with an accelerometer and a gyroscope.

2. The fall detection system based on genetic training according to claim 1, characterized in that: The outputs of all the sensor modules are connected to the input of the same microprocessor device (1), the output of which is connected to the inflation device (4), and the power supply device (2) is used to power the other modules.

3. The fall detection system based on genetic training according to claim 1, characterized in that: The microprocessor device (1) is also wirelessly connected to the user's mobile terminal.

4. A fall detection method based on genetic training, the fall detection system based on genetic training according to claims 1-3, characterized in that: Includes the following steps: Step 1: This step involves inputting the wearer's age Y, gender S, height H, and weight W into the microprocessor device (1) via the user's mobile terminal, completing the initialization of the microprocessor device (1), and determining the influence weights X1 and X2 of acceleration and angular velocity determined by the wearer's age Y, gender S, height H, and weight W data. Step 2: The step of using N pairs of sensor modules to collect the acceleration data α and angular velocity data ω of the wearer's neck and waist in real time and transmit them to the microprocessor device (1); Step 3: The microprocessor device (1) receives and saves the acceleration data α and angular velocity data ω, then performs neural network calculations, and compares the calculated acceleration weights X3 and angular velocity weights X4 with the fall threshold weight parameters C9 and C1, respectively. 10 The steps for comparison; When two or more pairs of acceleration weights X3 > C9 and angular velocity weights X4 > C 10 If the wearer is about to fall, proceed to step 4; otherwise, proceed to step 2. Step 4: The step of sending control commands to the inflation device (4) using the microprocessor device (1) to inflate the airbag device (3) instantly to protect the wearer from falls.

5. The fall detection method based on genetic training according to claim 4, characterized in that: In step 1, the weighting formula for the influence of the wearer's age Y, gender S, height H, and weight W on the wearer's acceleration when walking or falling is as follows: X1 = C1×Y + C2×S + C3×H + C4×W Where C1 is the acceleration weight parameter for age Y, C2 is the acceleration weight parameter for gender S, C3 is the acceleration weight parameter for height H, and C4 is the acceleration weight parameter for weight W. The formula for calculating the weighted average (X2) of the influence of the wearer's age (Y), gender (S), height (H), and weight (W) on the wearer's angular velocity during walking or falling is as follows: X2 = C5×Y + C6×S + C7×H + C8×W Wherein, C5 is the angular velocity weight parameter for age Y, C6 is the angular velocity weight parameter for gender S, C7 is the angular velocity weight parameter for height H, and C8 is the angular velocity weight parameter for weight W.

6. The fall detection method based on genetic training according to claim 4, characterized in that: In step 3, the formula for calculating the acceleration weight X3 is as follows: Among them, a x a y a z These represent the accelerations of the human body in the X, Y, and Z directions, respectively. The formula for calculating the angular velocity weight X4 is as follows: Wherein, where ω x ω y ω z These are the angular velocities of the human body in the X, Y, and Z directions, respectively.

Images