Low back pain prevention system and program

Abstract

To provide a system and program capable of determining the mass of an object during a motion of lifting the object.SOLUTION: A low back pain prevention system comprises: measurement means worn on the chest of a user who wants to prevent lower back pain in order to measure acceleration during a motion of lifting an object from a forward-leaning posture; and calculation means configured to derive an approximate curve approximating the temporal change in the acceleration based on the measurement results from the measurement means, and to determine the mass of the object based on the derived approximate curve.SELECTED DRAWING: Figure 3

Description

Technical Field

[0001] The present invention relates to a system and a program for preventing low back pain.

Background Art

[0002] Low back pain is a subjective symptom that many adults suffer from. In particular, nurses and caregivers often assume unnatural postures such as forward-leaning postures and twisting postures, such as the movement of transferring a patient from a bed to a wheelchair, which increases the burden on the lumbar spine and leads to a risk of low back pain.

[0003] Therefore, in order to evaluate the burden on the lumbar spine, a technique is known in which the inclination angle of the upper body of a subject is repeatedly detected, compared with an error angle condition, the period during which the error angle condition is continuously satisfied is measured, and a warning is output to the subject when that period reaches a reference value (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] The risk of developing low back pain increases when handling heavy objects. By discriminating the mass, it is possible to self-recognize the risk of low back pain and lead to the prevention of low back pain.

[0006] However, the above prior art has a problem that it can only detect the inclination angle with respect to the forward-leaning posture and output a warning, and cannot discriminate the mass of an object in the operation of lifting an object.

Means for Solving the Problems

[0007] The present invention has been made in view of the above problems, and is a system for preventing low back pain, measurement means that is attached to the chest of a person for whom low back pain is to be prevented and measures the acceleration in the operation of lifting an object from a forward-leaning posture; calculation means that uses the measurement result of the measurement means to derive an approximate curve approximating the temporal change of the acceleration, and discriminates the mass of the object based on the derived approximate curve A low back pain prevention system including the above is provided.

Effects of the Invention

[0008] According to the present invention, it becomes possible to discriminate the mass of an object in the operation of lifting the object.

Brief Description of the Drawings

[0009]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Mode for Carrying Out the Invention

[0010] Hereinafter, the mode for carrying out the present invention will be described in detail. However, the present invention is not limited to the embodiments shown below.

[0011] FIG. 1 is a diagram showing an example of the schematic configuration of a low back pain prevention system. The low back pain prevention system is a system for preventing and improving low back pain and is applicable to all people. The low back pain prevention system is particularly useful for nurses and caregivers who have problems with occupational low back pain. Hereinafter, the target persons for preventing low back pain will be described as nurses and caregivers, but the present invention is not limited thereto.

[0012] The low-back pain prevention system 10 includes a measuring device 11 as a small-sized device that can be worn on the chest, taking into consideration the use by nurses and care workers at the site of nursing and caregiving. The measuring device 11 is equipped with a three-axis acceleration sensor as a single sensor, and measures acceleration as the movement of the upper body when lifting an object 13 during the nursing and caregiving actions of nurses and care workers. The measuring device 11 has a wireless communication function and wirelessly transmits the measurement results.

[0013] The low-back pain prevention system 10 includes an information processing device 12 that performs wireless communication with the measuring device 11. The information processing device 12 discriminates the mass of the lifted object 13 using the measurement results from the measuring device 11. In FIG. 1, for the sake of explanation, an operation of lifting a box-shaped object 13 is shown, but in nursing and caregiving actions, the object 13 is mainly a person such as a patient. The information processing device 12 notifies nurses and care workers who are performing nursing and caregiving actions by, for example, displaying the discriminated mass, making them self-aware of the risk of low-back pain, and promoting the prevention and improvement of low-back pain.

[0014] The measuring device 11 is assumed to house a three-axis acceleration sensor and the like inside a case. The case is formed of a material such as plastic resin and protects the three-axis acceleration sensor and the like. The information processing device 12 is a terminal equipped with a wireless communication function, and a general computer can be used. As the information processing device 12, for example, a notebook PC (Personal Computer), a tablet terminal, a smartphone, a PDA (Personal Digital Assistant), or the like can be used.

[0015] FIG. 2 is a diagram showing an example of the hardware configuration of the measuring device 11 and the information processing device 12. FIG. 2(a) shows the hardware configuration of the measuring device 11, and FIG. 2(b) shows the hardware configuration of the information processing device 12.

[0016] The measuring device 11 includes a storage battery (battery) 20 that supplies power, a power circuit 21 as power control means for controlling the battery 20, a three-axis acceleration sensor 22 as measuring means for measuring acceleration, and a short-distance communication circuit 23 as communication means for performing wireless communication.

[0017] The battery 20 is, for example, a lithium polymer battery, but it is not limited to a lithium polymer battery as long as it is a storage battery. The power circuit 21 controls the power supplied to the three-axis acceleration sensor 22 and the short-distance communication circuit 23.

[0018] The three-axis acceleration sensor 22 is a sensor that measures the acceleration in the three-axis directions of the X-axis, Y-axis, and Z-axis when the lateral direction of the body of a nurse and a caregiver is defined as the X-axis direction, the front-rear direction of the body is defined as the Y-axis direction, and the vertical direction of the body is defined as the Z-axis direction.

[0019] The short-distance communication circuit 23 can perform short-distance wireless communication using Bluetooth (registered trademark) or a wireless LAN (Local Area Network). Note that the measuring device 11 may also include a tact switch for supplying (turning ON) or disconnecting (turning OFF) the power from the battery 20, an LED that lights up when the tact switch is turned ON, and the like. The communication between the measuring device 11 and the information processing device 12 may be vnc (virtual network computing) communication used for remote operation, or ssh (secure shell) communication that performs encrypted communication.

[0020] The information processing device 12 is, for example, a smartphone or a tablet terminal, and includes a CPU (Central Processing Unit) 30, a ROM (Read Only Memory) 31, a RAM (Random Access Memory) 32, an EEPROM (Electrically Erasable Programmable Read-Only Memory) 33, an imaging device 34, an imaging device I / F 35, an acceleration / azimuth sensor 36, a media I / F 37, and a GPS (Global Positioning System) receiver 38.

[0021] The CPU 30 controls the operation of the entire information processing device 12. The ROM 31 stores programs used for starting up the CPU 30. The RAM 32 provides a working area for the CPU 30. The EEPROM 33 stores programs for causing the CPU 30 to execute a process for discriminating the mass of the above object, other programs, various data, and the like.

[0022] The imaging device 34 is a type of imaging means that images a subject such as a CMOS (Complementary Metal Oxide Semiconductor) sensor or a CCD (Charge Coupled Device) sensor and outputs it as image data. The imaging device I / F 35 is a circuit that controls the driving of the imaging device 34. The acceleration / azimuth sensor 36 is a sensor for measuring acceleration and azimuth, and may include a gyro sensor for measuring angular velocity. The media I / F 37 controls the reading and writing of data to and from a recording medium 39 such as a flash memory. The GPS receiver 38 receives GPS signals from GPS satellites.

[0023] The information processing device 12 further includes a communication circuit 40, an imaging device 41, an imaging device I / F 42, a microphone 43, a speaker 44, an audio input / output I / F 45, a display 46, an external device connection I / F 47, and a touch panel 48.

[0024] The communication circuit 40 is a circuit that communicates with other devices via a network and includes a short-range communication circuit such as NFC (Near Field Communication) or Bluetooth (registered trademark). The imaging device 41 and the imaging device I / F 42 have the same functions as the imaging device 34 and the imaging device I / F 35. If the photographer holds the information processing device 12 in hand and the imaging device 34 images a subject on the side beyond the information processing device 12 as viewed from himself / herself, the imaging device 41 images the self-side.

[0025] The microphone 43 is a circuit that converts sound into an electrical signal. The speaker 44 is a circuit that converts an electrical signal into vibration and produces sound. The audio input / output I / F 45 is a circuit that processes the input / output of audio signals between the microphone 43 and the speaker 44. The display 46 is a liquid crystal display, an organic EL (Electro-luminescence) display, etc. that displays an image of a subject or the mass of a discriminated object in numbers or the like.

[0026] The external device connection I / F 47 is an interface for connecting various external devices. The touch panel 48 is a type of input means for operating the information processing device 12 when the user presses the display 46. Each component such as the CPU 30 of the information processing device 12 is connected to the bus 49 and can exchange data and the like with each other via the bus 49.

[0027] Figure 3 is a block diagram showing an example of the functional configuration of the measuring device 11 and the information processing device 12. The information processing device 12 realizes the function of discriminating the mass of the above object by the CPU 30 executing a program. Note that the above function may also be realized using a device such as an ASIC (Application Specific Integrated Circuit), a DSP (Digital Signal Processor), or an FPGA (Field Programmable Gate Array) designed to execute the function and the like.

[0028] The measurement device 11 includes, as functional units, a communication unit 50 constituted by a short-range communication circuit 23 and a measurement unit 51 constituted by a three-axis acceleration sensor 22. The measurement unit 51 measures the acceleration in three axes. The communication unit 50 transmits the acceleration measured by the measurement unit 51 as a measurement result to the information processing device 12.

[0029] The information processing device 12 includes a communication unit 52, an arithmetic unit 53, and a display unit 54 as functional units for realizing the above functions. Note that the information processing device 12 may be provided with other functional units as necessary.

[0030] The communication unit 52 performs wireless communication with the communication unit 50 of the measurement device 11. The communication unit 52 also functions as an acquisition unit and acquires the information on the three-axis acceleration measured by the measurement device 11 as a measurement result.

[0031] The arithmetic unit 53 uses the measurement result acquired by the communication unit 50 to derive an approximate curve approximating the temporal change of the acceleration, and discriminates the mass of the object based on the derived approximate curve.

[0032] The display unit 54 displays the mass of the object discriminated by the arithmetic unit 53. By checking the displayed mass, nurses can self-recognize the risk of low back pain and promote the prevention and improvement of low back pain.

[0033] Here, the risk of developing low back pain increases when handling heavy objects. In order to self-recognize the risk of low back pain and actively work on the prevention and improvement of low back pain, it is important to grasp the mass of the object during the operation of lifting the object. In particular, in the operation of lifting an object of unknown mass, which is considered to impose a large burden on the waist, using the STOOP method, it is important to discriminate the mass of the object of unknown mass. The STOOP method is a method of lifting an object while bending forward with the knees extended.

[0034] Conventionally, a method of determining the mass of an object by spike noise of acceleration during the lifting operation of the object is known. In this method, the spike noise observed from an acceleration sensor attached to the shoulder is used as a feature quantity, and a support vector machine (SVM), which is one of the learned models of machine learning used for predicting the class to which the data belongs, is used to discriminate between an object with a mass of 0 kg and an object with a mass of 25 kg.

[0035] It has been found that the method of discriminating mass by spike noise significantly decreases the discrimination accuracy when the mass difference between objects is reduced. There is a survey result that the risk of developing low back pain differs between a mass of 10 kg and a mass of 23 kg for the object being lifted, and a method of reducing the mass difference between 0 kg and 25 kg above and discriminating accurately is desired.

[0036] The system shown in FIG. 1 focuses on the case where an object of unknown mass, which is particularly considered to impose a large burden on the waist, is lifted by the STOOP method, and can discriminate between the cases where the mass of the object is 10 kg and 20 kg using a single acceleration sensor. Here, although the discrimination between the mass of the object being 10 kg and 20 kg will be described, it is not limited to this.

[0037] When a nurse and a care worker lift an object of unknown mass, a measuring device 11 equipped with a three-axis acceleration sensor is attached to the chests of the nurse and the care worker, and the three-axis acceleration during the lifting operation of the object of unknown mass is measured. The information processing device 12 is equipped with a support vector machine, which is one of the pattern recognition models using supervised learning, and classifies the input data into one of two classes using the support vector machine. Here, the support vector machine is used to discriminate whether the mass of the object of unknown mass is 10 kg or 20 kg.

[0038] First, the acceleration during the lifting operation by the STOOP method will be described. FIG. 4 shows the posture during the lifting operation by the STOOP method. FIG. 4(a) is a view of the posture during the lifting operation from the oblique rear, and FIG. 4(b) is a view of the posture during the lifting operation from the side.

[0039] When the measuring device 11 equipped with the three-axis acceleration sensor 22 is worn on the chests of nurses and care workers, the X-axis direction of the sensor is the lateral direction of the body, the Y-axis direction of the sensor is the front direction (front-back direction) of the body, and the Z-axis direction of the sensor is the vertical direction (upward direction). In Fig. 4(a), α (deg) is the angle around the X-axis and is the forward tilt posture angle.

[0040] Let the acceleration in the Y-axis direction be a y (m / s 2 ). Then, using the downward gravitational acceleration g (m / s 2 ) shown by the arrow line and the forward tilt posture angle α, a y is represented by the following formula (1).

[0041]

Equation

[0042] When the above formula (1) is transformed, the following formula (2) is obtained.

[0043]

Equation

[0044] Referring to the above formula (2), it can be seen that the change in the forward tilt posture can be measured by a single three-axis acceleration sensor 22.

[0045] Fig. 5 is a diagram showing the relationship between the acceleration a y in the Y-axis direction and the forward tilt posture angle α. As shown in Fig. 5, when the object lifting operation is performed at a constant speed, the acceleration a y changes non-linearly. Therefore, it is considered that by obtaining an approximate curve from the acceleration waveform during the lifting operation, the change in the forward tilt posture according to the mass of the object can be discriminated.

[0046] Therefore, the acceleration waveform during the lifting operation is approximated by a quadratic function, and discrimination is performed using the quadratic function approximation formula y = ax 2 + bx + c.

[0047] As a method of approximating with a quadratic function, the least squares method is used. In order to approximate to a quadratic function by the least squares method, in n pieces of measurement data, a set of each coefficient a, b, and c of a quadratic function equation that minimizes L obtained from the following formula (3) is obtained.

[0048]

Number

[0049] Here, the condition for L to be minimized is when the following formulas (4) to (6) are satisfied.

[0050]

Number

[0051]

Number

[0052]

Number

[0053] Using the above formulas (3) to (6), a quadratic function approximation formula of the acceleration waveform obtained from the triaxial acceleration sensor 22 can be obtained.

[0054] For the support vector machine, a non-linear SVM is used. The definition of the non-linear SVM is shown in the following formula (7). Also, the class formula using the learned discriminator is shown in the following formula (8). Note that in formula 8, C 1 represents the class when lifting 10 kg, and C 2 represents the class when lifting 20 kg. w k (k = 1, 2,..., N) are weights, w 0represents bias, N represents the number of learning data, and M represents the Gaussian kernel function. The kernel function is a function for calculating the similarity of data. It is a well-known function, and since the Gaussian kernel function is also well-known as one of the kernel functions, it will not be elaborated here. w k and w 0 are determined by learning from the teacher signal vector as the teacher data composed of the feature data x for learning.

[0055]

Number

[0056]

Number

[0057] When discriminating the mass of an object, first, measure the acceleration of the object during the lifting operation by the STOOP method. The measuring device 11 equipped with the three-axis acceleration sensor 22 can be worn on the chest by putting it in a chest pocket or the like. However, in order to prevent the displacement of the sensor mounting position, as shown in FIG. 6, the measuring device 11 is arranged at the center of the chest, and the position of the sensor can be fixed by winding it with the belt 14 from above. Note that since this example is an example of a method for fixing the position of the sensor, any known method can be adopted as long as the position of the sensor can be fixed.

[0058] Referring to FIG. 7, the lifting operation for discriminating the mass of an object will be described. While placing a hand on the object 60 shown in FIG. 7(a), maintain the posture for several seconds. Since the forward inclination posture angle at the start of lifting is such that the upper body is substantially perpendicular to the vertical direction in which the legs extend, it is approximately 90 deg. The object 60 is, as an example, a semi-transparent storage case for clothing with weights inside. Then, as shown in FIG. 7(b), lift the upper body to lift the object 60, and when a predetermined forward inclination posture angle smaller than the start of lifting is reached, end the lifting operation of the object. After the lifting operation ends, maintain the posture (standing posture) at the end of the lifting operation for several seconds. The several seconds can be, for example, 3 seconds, but it is not limited to this.

[0059] Approximate the acceleration waveform in the Y-axis direction measured from the start to the end of the object lifting operation with a quadratic function to discriminate the mass of the object. Note that at the end of the lifting operation, it is the posture when the upper body is raised to a predetermined forward inclination posture angle, and the predetermined forward inclination posture angle is, for example, 20 deg. Note that the predetermined forward inclination posture angle is not limited to 20 deg.

[0060] Regarding the quadratic function approximation formula y = ax 2 + bx + c obtained by the least squares method, in order to observe the difference in posture change according to the mass of the object, set the intercept c to 0 and eliminate the difference in the forward inclination posture angle at the start of the operation for each measurement data.

[0061] The above quadratic function approximation formula becomes y = ax 2 + bx by c = 0. Use the two coefficients a and b of x in this quadratic function approximation formula as feature quantities for discriminating the mass of the object. Then, based on the feature quantities, discriminate the mass of the object by a support vector machine.

[0062] In order to verify whether this system can accurately discriminate the mass of an object, a test was conducted. At that time, for comparison, a method of using spike noise, which has been conventionally proposed, as a feature quantity and discriminating by a support vector machine was also used.

[0063] The conventional method uses the accelerations in the Y-axis and Z-axis directions measured by the triaxial acceleration sensor of this system, and calculates the maximum value of the data D obtained by the following formula (9) as spike noise, and uses the calculated spike noise as a feature quantity. [k]

[0064]

Equation

[0065] In the above formula (9), k is the signal sample number. d 1[k] (k = 1, 2,...) is the acceleration measurement data every 20 ms which is the sampling period, and d 2[k] is the output obtained by a median filter that outputs the median value from d 1[k] to d 1[k-12] up to. The support vector machine used a non-linear SVM, similar to the support vector machine installed in this system.

[0066] Five subjects were used, and a lifting operation was performed by the STOOP method. The acceleration was measured by the triaxial acceleration sensor 22 of the measurement device 11 worn on the chest of the subject as shown in Fig. 6. Fig. 8 is a diagram showing the acceleration measurement result in the Y-axis direction during the lifting operation. The measurement result is the raw result obtained from the sensor without using a filter. Note that the measurement results were obtained three times each for 10 kg and 20 kg for each of the five subjects, and a total of 30 data were obtained. However, since one data contained a missing value, 29 data were adopted and shown as the acceleration measurement result. The vertical axis indicates the forward tilt posture angle (deg) calculated from the acceleration measurement value in the Y-axis direction, and the horizontal axis indicates the time (s).

[0067] In Fig. 8, the thin solid line indicates the result when the mass of the object is 10 kg, and the thick solid line indicates the result when the mass of the object is 20 kg. Time 0 seconds indicates the start of the operation.

[0068] As shown in Fig. 8, when lifting 20 kg, the speed of posture change from the start of operation was often slower than when lifting 10 kg. This is presumably because the greater the mass, the greater the burden on the waist, and it could not be lifted smoothly.

[0069] Fig. 9 is a diagram showing an approximate curve of a quadratic function obtained by the least squares method. In Fig. 9, the vertical axis represents y of the quadratic function approximation formula y = ax 2 + bx, and the horizontal axis represents x. Similar to Fig. 8, the thin solid line indicates the result when the mass of the object is 10 kg, and the thick solid line indicates the result when the mass of the object is 20 kg. Time 0 seconds indicates the start of operation. As shown in Fig. 9, it can be seen that as the mass of the object to be lifted increases, especially at the moment of lifting, the value tends to become a gentle curve.

[0070] Table 1 below is a table showing the discrimination accuracy (correct rate) of the learning data and the verification data calculated by SVM. Note that approximately 80% of the measurement data measured by the three-axis acceleration sensor is used as the learning data, and approximately 20% is used as the verification data. The number in parentheses next to the accuracy (%) indicates the number of correctly identified data with respect to the total number of data.

[0071]

Table 1

[0072] As shown in Table 1, the discrimination accuracy of this method (quadratic function approximation) by quadratic function approximation was higher than that of the conventional method by spike noise (spike noise). This is presumably because, unlike the values calculated instantaneously such as spike noise, by discriminating from the posture change from the start to the end of the lifting operation, the differences according to the mass are clearer.

[0073] FIG. 10 is a diagram showing the feature value distributions of the learning data and the verification data and the classification boundary obtained from the learning data. The circles indicate the data for lifting 10 kg, and the triangles indicate the data for lifting 20 kg. Those surrounded by a black line, whether circles or triangles, indicate the verification data.

[0074] FIG. 10(a) is a diagram showing the distribution diagram and the classification boundary of the conventional spike noise method, and FIG. 10(b) is a diagram showing the distribution diagram and the classification boundary of this method. The vertical axis of FIG. 10(a) indicates the value of spike noise calculated from the acceleration measurement value in the Z-axis direction, and the horizontal axis indicates the value of spike noise calculated from the acceleration measurement value in the Y-axis direction. The vertical axis of FIG. 10(b) indicates the coefficient a of the quadratic function approximation formula, and the horizontal axis indicates the coefficient b.

[0075] Referring to FIG. 10(a), it can be seen that as the mass of the object decreases, the values of the spike noise calculated from the acceleration measurement values in the Y-axis and Z-axis directions both tend to decrease. Therefore, as shown in FIG. 10(a), the classification boundary is set in the part that separates the lower left region where the spike noise values calculated from the acceleration measurement values in the Y-axis and Z-axis directions are both small and the other regions. However, since cases where the value of the spike noise when lifting 20 kg is smaller than the value of the spike noise when lifting 10 kg are occasionally seen, it is hard to say that the results classified by this classification boundary are classified with high accuracy.

[0076] Referring to Fig. 10(b), it can be seen that as the mass of the object increases, although there is no obvious difference in the value of a, the value of b tends to increase. This is because when lifting 10 kg, the forward tilt posture angle changes more rapidly, and the value of b depends on the coordinates of the vertex of the curve depicted by the quadratic function approximation formula and the slope of the tangent line at the intercept. Therefore, the classification boundary will be set at the part that separates the area below a certain value of b from the area above it. The results when lifting 10 kg are almost all classified into the area below the classification boundary, and the results when lifting 20 kg are almost all classified into the area above the classification boundary, indicating that classification can be achieved with high accuracy.

[0077] From this, it can be understood that by approximately fitting the waveform obtained from the acceleration from the start to the end of the lifting operation with a quadratic function and using the values of a and b, which are the coefficients of x in the approximation formula, as characteristic quantities, it is possible to accurately determine whether the mass of the object is 10 kg or 20 kg. Based on this, by setting thresholds for the values of a and b based on the classification boundary obtained by SVM, it becomes possible to distinguish finer differences in the mass of an object using only a single sensor.

[0078] Incidentally, for preventing low back pain, in addition to discriminating the mass of the object, it is desirable to detect unnatural postures that impose a large burden on the waist and output a warning sound to prompt improvement of the posture. Unnatural postures that impose a large burden on the waist include, in addition to the forward tilt posture in the above-mentioned STOOP method, twisting postures that are often taken in nursing and caregiving operations.

[0079] FIG. 11 is a diagram showing another example of the hardware configuration of the measuring device 11. The measuring device 11 includes a storage battery (battery) 20 that supplies power, a power circuit 21 as power control means for controlling the battery 20, a 9-axis sensor 24 that measures angular velocity and azimuth in addition to acceleration as measuring means, an electronic buzzer 25 as output means for outputting a warning as a warning sound, and a short-range communication circuit 23 as communication means for performing wireless communication. Since the battery 20, the power circuit 21, and the short-range communication circuit 23 have already been described, only the 9-axis sensor 24 and the electronic buzzer 25 will be described here.

[0080] The 9-axis sensor 24 is a sensor having a function of measuring acceleration in three-axis directions of the X-axis, Y-axis, and Z-axis, angular velocity in three-axis directions, and azimuth in three-axis directions when the lateral direction of the body of a nurse and a caregiver is defined as the X-axis direction, the front-back direction of the body is defined as the Y-axis direction, and the up-down direction of the body is defined as the Z-axis direction. In this example, the 9-axis sensor 24 is used. However, since the postures of the nurse and the caregiver can be measured by the acceleration in the three-axis directions and the angular velocity in the three-axis directions, a 6-axis sensor having a function of measuring the acceleration in the three-axis directions and the angular velocity in the three-axis directions may also be used. Further, in addition to the 3-axis acceleration sensor 22 shown in FIG. 2, a 3-axis gyro sensor for measuring the angular velocity in the three-axis directions may be provided separately.

[0081] The electronic buzzer 25 may output one type of warning sound, but in order to distinguish which posture the unnatural posture is due to, it may be capable of outputting two or more types of warning sounds such as by changing the pitch of the sound. Note that when outputting two or more types of warning sounds, it is not limited to changing the pitch of the sound, and it may be, for example, changing the period when turning the sound on / off.

[0082] The information processing device 12 acquires the measurement result of the 9-axis sensor 24, detects the posture from the measurement result, determines whether the detected posture is an unnatural posture, and when it is determined that the posture is an unnatural posture, executes a process of instructing the measuring device 11 to output a warning. This process can be realized with a hardware configuration similar to that shown in FIG. 2(b).

[0083] FIG. 12 is a block diagram showing another example of the functional configuration of the measuring device 11 and the information processing device 12. In order to implement the process of outputting the above warning, the measuring device 11 includes an output unit 55 constituted by an electronic buzzer 25 in addition to the communication unit 50 and the measuring unit 51.

[0084] As a functional unit for determining whether or not the above unnatural posture exists and implementing the process of instructing to output a warning, the information processing device 12 includes an instruction unit 56 in addition to the communication unit 52, the calculation unit 53, and the display unit 54. Note that the information processing device 12 may include other functional units as necessary.

[0085] The measuring unit 51 of the measuring device 11 measures the three-axis acceleration and the three-axis angular velocity. The communication unit 50 transmits the acceleration and the angular velocity measured by the measuring unit 51 to the information processing device 12 as measurement results. The output unit 55 outputs a warning in accordance with an instruction to output a warning received from the information processing device 12.

[0086] The communication unit 52 of the information processing device 12 acquires the information on the acceleration and the angular velocity transmitted from the measuring device 11 as measurement results, and transmits an instruction to output a warning to the measuring device 11.

[0087] The calculation unit 53 calculates a forward inclination posture angle, which is the angle of the upper body in the forward inclination posture state with respect to the upright state of a nurse or a caregiver, and a twisting posture angle, which is the rotation angle around the upper body during the forward inclination posture, using the measurement results acquired by the communication unit 50.

[0088] When the forward inclination posture angle calculated by the calculation unit 53 is equal to or greater than a first threshold value, or the twisting posture angle calculated by the calculation unit 53 is equal to or greater than a second threshold value, or both, the instruction unit 56 instructs to output a warning. The instruction unit 56 transmits an instruction to output a warning to the measuring device 11 via the communication unit 52.

[0089] When the instruction unit 56 determines that a warning is necessary, the display unit 54 can display a warning message such as "The forward leaning posture angle exceeds 20 deg, which is a dangerous posture." In this way, it is not possible to immediately understand the reason why the warning sound is sounding only with the electronic buzzer 25, but by displaying it as a warning message, it becomes easier to immediately understand the reason for the warning.

[0090] Hereinafter, using the measuring device 11 shown in FIG. 11 and the information processing device 12 having the functions shown in FIG. 12, the process of determining whether or not the above-mentioned unnatural posture exists and outputting a warning will be described in detail.

[0091] Referring to FIG. 13, the nursing and caregiving operations during the transfer assistance from the bed to the wheelchair will be described. Here, it will be described as being performed by the caregiver 70. As shown in FIG. 13(a), the caregiver 70 moves the wheelchair 73 and places it beside the bed 71 of the patient 72 lying on the bed 71. Next, the caregiver 70 puts hands under the neck and feet of the patient 72 to hold them, moves to the side where the wheelchair 73 is placed on the bed 71, raises the upper body, and sits on the edge of the bed 71. If the patient 72 can move to the edge of the bed 71 by himself / herself, he / she may move by himself / herself. Also, if the patient 72 can sit on the edge of the bed 71 by himself / herself, he / she may sit on the edge of the bed 71 by himself / herself.

[0092] As shown in FIG. 13(b), the caregiver 70 asks the patient 72 to raise both hands, and while the patient 72 is in a forward leaning posture, inserts both hands of the caregiver 70 into both sides of the patient 72. Then, as shown in FIG. 13(c), the caregiver 70 raises his / her upper body to lift the patient 72 from the bed 71.

[0093] As shown in FIG. 13(d), the caregiver 70 twists his / her body horizontally to move the back of the patient 72 so that it faces the side of the wheelchair 73. Then, as shown in FIG. 13(e), while supporting the patient 72 in a forward leaning posture, the caregiver 70 sits on the wheelchair 73 behind the patient 72.

[0094] The above is the operation during the transfer assistance from the bed 71 to the wheelchair 73. However, there are postures that impose a burden on the waist of the caregiver 70 during this operation. Referring to FIG. 14, the forward inclination posture and the twisting posture will be described as postures that impose a burden on the waist. FIG. 14(a) is a diagram for explaining the forward inclination posture, and FIG. 14(b) is a diagram for explaining the twisting posture.

[0095] The forward inclination posture is the posture when only the upper body is tilted forward from the upright state and the upper body is rotated forward. Note that the upright state is a state where the body extends linearly from the head to the feet. However, if it is approximately linear, it may be a state where the feet, waist, back, etc. are slightly bent. As shown in FIG. 14(a), the direction of gravity is indicated by the dashed arrow line. When in the forward inclination posture, the upper body is inclined forward at an arbitrary angle with respect to the dashed line. Note that in FIG. 14(a), the upward, forward, and lateral directions during the forward inclination posture are indicated by solid arrow lines. Specifically, the posture of the caregiver 70 in the forward bent state shown in FIG. 13(b) is the forward inclination posture.

[0096] The twisting posture is the posture when the upper body is rotated about the axis during the forward inclination posture. As shown in FIG. 14(b), in the twisting posture, with the upper body in the forward inclination state, the upper body is twisted left and right, and both the lateral and forward directions of the upper body are inclined at arbitrary angles from the forward inclination posture. For example, when the lateral direction of the body is inclined by 10 deg about the axis from the forward inclination posture, the forward direction is also inclined by 10 deg. The twisting posture is one of the special postures in which the angles are dependent on two axes. The posture of the caregiver 70 when the body is twisted horizontally as shown in FIG. 13(d) is the twisting posture.

[0097] By the way, the measurement method using Euler angles, which has been conventionally used for posture measurement, represents the posture of an object by three angles when the object is rotated three times in a specific order along the X-axis, Y-axis, and Z-axis of the object, and is a commonly used method as a method for representing a posture. However, in nursing and caregiving operations, special postures in which the angles are dependent on two axes, such as the above-mentioned twisting posture, are likely to occur, and it becomes difficult to accurately represent the posture.

[0098] Also, when measuring the torsional posture angle of the upper body with a single 9-axis sensor 24, it is necessary to distinguish between the rotational motion due to a change in direction during walking and the rotational motion about the upper body with the forward tilt posture angle during the forward tilt posture.

[0099] Therefore, in this method, attention is paid to limiting it to the torsional motion during the forward tilt posture, and by using the integration by the part (gyro sensor) that measures the three-axis angular velocity of the 9-axis sensor 24 from the time point of the forward tilt posture and the decomposition of the gravitational acceleration component by the part (acceleration sensor) that measures the three-axis acceleration of the 9-axis sensor 24, each torsional posture angle is calculated. These calculated torsional posture angles include errors as noise, and a Kalman filter is used to synthesize these torsional posture angles including the noise to calculate the accurate torsional posture angle.

[0100] A specific calculation method will be described. When taking the X-axis in the lateral direction of the body, the Y-axis in the front-rear direction of the body, and the Z-axis in the vertical direction of the body, the acceleration of each axis can be represented by A x 、A y 、A z The forward tilt posture angle θ x 、A y 、A z can be calculated by the following formula (10) using the accelerations A facc measured by the acceleration sensor.

[0101]

Equation

[0102] Also, when setting the angular velocity of each axis as G x 、G y 、G z the forward tilt posture angle θ x 、G y 、G z can be calculated by the following formula (11) using the angular velocities G fgyro measured by the gyro sensor. In the following formula 2, δ t represents the sampling time.

[0103]

Number

[0104] The torsional attitude angle θ obtained using acceleration tacc can be expressed as in the following formula (12) using the above formula (10) and the gravitational acceleration g. Also, the torsional attitude angle θ obtained using the angular velocity tgyro can be expressed as in the following formula (13) using the above formula (11) and the gravitational acceleration g.

[0105]

Number

[0106]

Number

[0107] Here, establish the persistent prediction models shown in the following formulas (14) and (15). In formulas (14) and (15), v and w represent normal white noise, and x represents the noise of the gyro sensor. θ f represents the forward tilt attitude angle, and θ t represents the true value of the torsional attitude angle.

[0108]

Number

[0109]

Number

[0110] Here, the observations by the acceleration sensor (observation equations) can be expressed by the following formulas (16) and (17), and the observations by the gyro sensor (observation equations) can be expressed by the following formulas (18) and (19). Also, the noise models can be expressed by the following formulas (20) and (21).

[0111]

Number

[0112]

Number

[0113]

Number

[0114]

Number

[0115]

Number

[0116]

Number

[0117] The model combined by these can be expressed by the following formula (22) and the following formula (23).

[0118]

Number

[0119]

Number

[0120] By implementing the Kalman filter in this way, it becomes possible to analyze the twisting motion during the forward-leaning posture, which is a problem specific to nursing operations. By measuring the twisting posture angle after the forward-leaning posture, it is possible to substantially reduce the non-computable region that becomes a constraint in gimbal lock due to the relationship of the joint range of motion. In addition, by using the Kalman filter in combination, it is possible to eliminate the accumulation of integration noise caused by the gyro sensor, so that the twisting posture angle during the forward-leaning posture during nursing operations can be uniquely measured by a single sensor. And it is also possible to distinguish it from rotational motions such as turning during walking.

[0121] By using the above formula (22) and the above formula (23), the true values θ f 、θ t of the forward-leaning posture angle and the twisting posture angle can be calculated from the acceleration and angular velocity measured by the acceleration sensor and the gyro sensor, respectively, as the forward-leaning posture angle and the twisting posture angle. The arithmetic unit 53 can calculate the forward-leaning posture angle and the twisting posture angle by this calculation method, and a test was conducted to verify its effectiveness.

[0122] The test included this system including the measuring device 11 shown in FIG. 11 and the information processing device 12 having the functional configuration shown in FIG. 12, and for comparison, a DSP (Digital Signal Processor) wireless 9-axis motion sensor (manufactured by Sports Sensing Co., Ltd.), which is generally used as a motion sensor, and a motion capture system (manufactured by OptiTrack Japan Co., Ltd.) that captures the movement of an object, converts it into digital data, and outputs it were used. The motion capture system (motion capture) includes markers and six cameras, and based on the images captured by the six cameras, three-dimensional motion analysis is performed to calculate substantially accurate forward-leaning posture angles and twisting posture angles. The markers were used by attaching them to the upper, lower, left, and right ends after expanding the upper, lower, left, and right of the substantially rectangular parallelepiped measuring device 11. In addition, a DSP wireless 9-axis motion sensor (motion sensor) was fixed so as to overlap the measuring device 11.

[0123] The sampling period of this system is 240 ms, the sampling period of motion capture is 1 ms, and the sampling period of the motion sensor is 1 ms.

[0124] In motion capture, the forward tilt angle and the twist angle are defined as the angles formed by the direction vectors of the Z-axis and X-axis at the start of measurement and the direction vectors of the Z-axis and X-axis after tilting the measuring device 11, with the positive direction of the Y-axis being the front direction of the measuring device 11 of this system, the positive direction of the Z-axis being vertically upward, and the positive direction of the X-axis being to the right. The motion sensor outputs a quaternion used to represent the rotation and posture of an object, calculates the roll angle, pitch angle, and yaw angle from the output quaternion, and calculates the forward tilt angle and the twist angle based on these roll angle, pitch angle, and yaw angle.

[0125] In the test, referring to the forward tilt angle calculated in real time by the motion sensor, the object was tilted forward over about 3 seconds until the forward tilt angle reached about 60 deg. Also, referring to the forward tilt angle calculated in real time by the motion sensor, a twisting motion of about 30 - 45 deg was performed while maintaining the forward tilt postures at 20 deg, 40 deg, and 60 deg respectively.

[0126] Figure 15 shows the measurement results of the forward tilt angle when each system and device is tilted forward up to 60 deg. The vertical axis represents the forward tilt angle (deg), and the horizontal axis represents time (s). The time (s) is the measurement time since the start of measurement. Since the sampling periods vary depending on the system and device, the measurement results every 240 ms of the sampling period of this system were used after unifying them to the sampling period of 240 ms of this system.

[0127] Since it is difficult to start measurements simultaneously with three systems or devices, the angle at the start of measurement of the motion sensor is approximately 28 deg. However, as the forward tilt posture angle increases, the difference decreases. Also, in this system, similar to motion capture, the angle at the start of measurement is almost 0 deg, and the change in angle after the start of measurement also progressed in almost the same manner. From this, the effectiveness of the measurement accuracy of this system was confirmed.

[0128] Note that the measured value of acceleration changes non-linearly as the forward tilt increases. The smaller the forward tilt posture angle, the larger the error, and the larger the forward tilt posture angle, the smaller the error. For this reason, the motion sensor has a large error in the range where the forward tilt posture angle is small. However, for example, in the range where the measurement angle exceeds 50 deg (the range where the forward tilt posture angle is large), the value approximates the angle measured by the motion capture.

[0129] FIG. 16 is a diagram showing the results of measuring the torsion posture angle when a torsion operation is performed when each system or device is tilted forward at three angles of 20 deg, 40 deg, and 60 deg. FIG. 16(a) shows the results when tilted forward by 20 deg, FIG. 16(b) shows the results when tilted forward by 40 deg, and FIG. 16(c) shows the results when tilted forward by 60 deg.

[0130] The motion sensor always measures the angle by integrating the angular velocity and correcting with the gravitational acceleration. For this reason, there is a possibility that small errors accumulate. Also, the motion sensor calculates the changes in roll angle, pitch angle, and yaw angle from the changes in the X-axis, Y-axis, and Z-axis during calibration performed before the start of measurement. If the X-axis, Y-axis, and Z-axis are tilted during calibration, there is a possibility that a large difference will occur in the calculated value of the torsion posture angle.

[0131] Due to such factors, the torsion posture angle calculated from the motion sensor is significantly different from the torsion posture angle calculated from this system and the motion capture system, and the difference did not decrease even as the angle increased.

[0132] On the one hand, in this system, for any twisting motion starting from the forward leaning posture angles shown in Figs. 16(a) to (c), results that were almost identical to those obtained by motion capture were obtained. Note that although the sampling period of this system is 240 ms, which is a longer time interval than that of other devices, it can reduce the integration error. Thus, it can be seen that it is possible to practically measure the twisting posture angle as a system for preventing low back pain.

[0133] Next, criteria for determining whether the measured forward leaning posture angle and twisting posture angle are unnatural postures that impose a large burden on the waist, i.e., dangerous postures, will be explained. The dangerous forward leaning posture angle can be set with reference to REBA (Rapid Entire Body Assessment), which is adopted as an international occupational low back pain risk assessment method. REBA is a posture evaluation index for the whole body, and it performs a risk assessment by adding the posture evaluation score of the whole body in the sagittal plane and the activity evaluation score regarding the duration. Regarding the forward leaning posture angle, it is subdivided into 20 deg, 20 - 60 deg, and 60 deg or more. Therefore, as the forward leaning posture angle, 20 deg and 60 deg will be set as the reference angles for dangerous postures.

[0134] Regarding the twisting posture angle, conventionally, only the presence or absence of twisting was visually judged, and the degree of twisting was not considered. In previous studies, in a twisting posture of 30 deg, an increase has been confirmed in the erector spinae muscle, external oblique abdominal muscle, and internal oblique abdominal muscle significantly compared to the initial posture. Therefore, 30 deg or more will be set as the reference angle for dangerous postures.

[0135] Actually, a test was conducted to see if dangerous postures would decrease during the nursing operation by setting to output a warning sound every 1.1 seconds at a forward leaning posture angle of 20 deg, every 0.6 seconds at 60 deg, and every 0.4 seconds at a twisting posture angle of 30 deg. The sampling period of this system is 240 ms. The moving average every 4 samples was used as the measured value, and the determination of dangerous postures was performed every 1 second.

[0136] Table 2 and Table 3 below are tables showing the average values of the forward tilt posture angles and the twist posture angles for each subject. There were 8 subjects, and these average values are the average values of the forward tilt posture angles and the twist posture angles measured when the measuring device 11 of this system was worn in the chest pocket and the transfer assistance operation from the bed to the wheelchair was performed.

[0137] Table 2 shows the average values of the forward tilt posture angles for each subject, and Table 3 shows the average values of the twist posture angles for each subject. For noise removal, a moving average for every 4 samples was applied to all the data, and a twist posture angle of 65 deg or more was regarded as an outlier and processed as 65 deg.

[0138]

Table 2

[0139]

Table 3

[0140] Table 2 includes the subjects, the transfer assistance time without a warning sound and the average value of the forward tilt posture angle during assistance, and the transfer assistance time with a warning sound and the average value of the twist posture angle during assistance. Table 3 includes the subjects, the transfer assistance time without a warning sound and the average value of the twist posture angle during assistance, the transfer assistance time with a warning sound and the average value of the twist posture angle during assistance, and the ratio of outliers of the twist posture angle. In Table 3, the ratio of outliers is the ratio of the number of samples in which a dangerous posture was detected to the total number of samples during assistance.

[0141] Regarding the forward tilt posture, it was confirmed that the forward tilt posture angle during assistance decreased for all the subjects when there was a warning sound compared to the case without a warning sound. Regarding the twist posture, it was confirmed that for 6 out of 8 subjects, the twist posture angle during assistance decreased when there was a warning sound compared to the case without a warning sound.

[0142] A t-test was conducted to determine whether the difference in the average value of the torsion posture angle was statistically significant under two conditions: without a warning sound and with a warning sound. For the t-test, a one-sided test was adopted, and the significance level was set at α = 0.05.

[0143] As a result of the t-test, the p-value, which is the probability that the value of the average torsion posture angle with a warning sound is equal to or greater than the value without a warning sound, was 0.002, which was significantly lower than the significance level β = 0.05. Therefore, it was confirmed that the average torsion posture angle was significantly lower with a warning sound.

[0144] Similarly, when a t-test was conducted for the average value of the forward tilt posture angle, the p-value was 0.003, which was significantly lower than the significance level β = 0.05, and it was confirmed that the forward tilt posture angle was significantly lower with a warning sound.

[0145] Figure 17 shows the temporal change in the forward tilt posture angle of the subject (Sub.1) shown in Table 2, and Figure 18 shows the temporal change in the torsion posture angle of the subject (Sub.1) shown in Table 3. In Figure 17, the vertical axis represents the forward tilt posture angle (deg), and the horizontal axis represents the measurement time (s). In Figure 18, the vertical axis represents the torsion posture angle (deg), and the horizontal axis represents the measurement time (s). In Figure 18, the timing when the warning sound sounded is indicated as 1, and the others are indicated as 0, showing three warning sounds at intervals of 1.1 seconds, 0.6 seconds, and 0.4 seconds.

[0146] Referring to Figure 18, it is estimated that the movement posture was corrected after the warning sound sounded, and the subsequent forward tilt posture angle and torsion posture angle became smaller. Comparing Figure 17 and Figure 18, the period during which the forward tilt posture angle and torsion posture angle were larger was shorter in the case with a warning sound than in the case without a warning sound. From these facts, it can be seen that the forward tilt posture angle and torsion posture angle are improved in the case with a warning sound. This was the same for the other subjects except for the results of two subjects among the eight subjects, for whom the torsion posture angle was smaller without a warning sound.

[0147] FIG. 19 is a flowchart showing an example of a process for discriminating the mass of an object. By turning on the power of the measuring device 11 and starting up the information processing device 12, the process starts from step 100. The subject holds the object in a forward-leaning posture and lifts the upper body to lift the object. In step 101, the measuring device 11 measures the three-axis acceleration in the operation of lifting the object. The measured acceleration information is transmitted to the information processing device 12 as a measurement result.

[0148] In step 102, the information processing device 12 derives a quadratic function approximation formula of the acceleration waveform from the measurement result based on the above formulas 3 to 6. In step 103, the information processing device 12 obtains the coefficients a and b of the quadratic function approximation formula as feature quantities. In step 104, the information processing device 12 discriminates the mass of the object by a support vector machine based on the feature quantities. Then, in step 105, the information processing device 12 displays the discrimination result and ends the process in step 106.

[0149] FIG. 20 is a flowchart showing an example of a process for outputting a warning for a dangerous posture. The process shown in FIG. 20 can be executed in parallel with the process shown in FIG. 19. By turning on the power of the measuring device 11 and starting up the information processing device 12, the process starts from step 200.

[0150] In step 201, the measuring device 11 measures the three-axis acceleration and three-axis angular velocity in the operation of the subject. The measured three-axis acceleration and three-axis angular velocity information is transmitted to the information processing device 12 as a measurement result.

[0151] In step 202, the information processing device 12 obtains the measurement result and calculates the forward-leaning posture angle and the twisting posture angle using the above formulas 10 to 23. The information processing device 12 applies a moving average to the forward-leaning posture angle or the twisting posture angle calculated from a plurality of data sampled in order at a sampling period for noise removal with respect to the calculated forward-leaning posture angle and twisting posture angle, and the average value can be used for determining a dangerous posture.

[0152] In step 203, the information processing apparatus 12 determines whether the forward tilt posture angle is greater than or equal to a preset first threshold value. When using the above average value, it determines whether the average value is greater than or equal to the first threshold value. The first threshold value is, for example, 20 deg as described above. If the forward tilt posture angle is greater than or equal to the first threshold value, the process proceeds to step 205. If it is less than the first threshold value, the process proceeds to step 204.

[0153] In step 204, the information processing apparatus 12 determines whether the twist posture angle is greater than or equal to a preset second threshold value. When using the above average value, it determines whether the average value is greater than or equal to the second threshold value. The second threshold value is, for example, 30 deg as described above. If the twist posture angle is greater than or equal to the second threshold value, the process proceeds to step 205. If it is less than the second threshold value, the process proceeds to step 207 and the process ends.

[0154] In step 205, the information processing apparatus 12 instructs the measurement apparatus 11 as to which posture should be regarded as a dangerous posture and a warning should be output according to which posture is greater than or equal to the threshold value. Specifically, it instructs how many seconds the buzzer should output a warning sound.

[0155] In step 206, the measurement apparatus 11 outputs a warning sound in response to the instruction from the information processing apparatus 12, and in step 207, the process ends. Note that the measurement apparatus 11 and the information processing apparatus 12 can repeat the processes of steps 200 to 207 while the power is on and they are communicating with each other.

[0156] Also, for the forward tilt posture angle and the twist posture angle, a plurality of threshold values may be provided, it may be determined whether each is greater than or equal to the threshold value, and an instruction may be given to output a warning corresponding to the angle in addition to the posture, and a warning sound may be output. Specifically, it can be instructed to output a warning when the above forward tilt posture angle is 20 deg and 60 deg, and different warning sounds can be output according to the angle.

[0157] From the above, by providing the low-back pain prevention system and program of the present invention, it becomes possible to determine the mass of an object in the operation of lifting the object. In addition, it becomes possible to output a warning against a twisting posture that increases the lumbar burden, promote posture improvement, and reduce the lumbar burden.

[0158] So far, the embodiments of the low-back pain prevention system and program of the present invention have been described in detail. However, the present invention is not limited to the above-described embodiments, and can be changed within the scope that those skilled in the art can conceive, such as other embodiments, addition, modification, deletion, etc., and is included in the scope of the present invention as long as the functions and effects of the present invention are exhibited in any aspect.

[0159] Therefore, this system is not limited to having both the function of determining the mass of an object and the function of determining whether it is a dangerous posture and outputting a warning, and may have only one of the functions. The program can be provided as a program that is installed in the information processing device 12 and can execute only the processing to be executed by the information processing device 12. In addition, in the present invention, in addition to the system and program, it is also possible to provide the method of determining the mass of the above object, the method of determining whether it is the above dangerous posture, and outputting a warning, etc.

Explanation of Signs

[0160] 10…Low-back pain prevention system 11…Measuring device 12…Information processing device 13…Object 14…Belt 20…Battery 21…Power supply circuit 22…Three-axis acceleration sensor 23…Short-range communication circuit 24…Nine-axis sensor 25…Electronic buzzer 30…CPU 31…ROM 32…RAM 33…EEPROM 34…Imaging element 35…Image sensor I / F 36…Acceleration and azimuth sensor 37…Media I / F 38…GPS receiver 39…Recording medium 40…Communication circuit 41…Image sensor 42…Image sensor I / F 43…Microphone 44…Speaker 45…Audio input / output I / F 46…Display 47…External device connection I / F 48…Touch panel 49…Bus 50…Communication unit 51…Measurement unit 52…Communication unit 53…Arithmetic unit 54…Display unit 55…Output unit 56…Instruction unit 60…Object 70…Nurse 71…Bed 72…Patient 73…Wheelchair

Claims

1. A system for preventing low back pain, comprising: measuring means that is attached to the chest of a person for whom low back pain is to be prevented and measures the acceleration in the operation of lifting an object from a forward-leaning posture; calculating means that uses the measurement result of the measuring means to derive an approximate curve approximating the temporal change of the acceleration, and discriminates the mass of the object based on the derived approximate curve A low back pain prevention system.

2. The calculation means derives a quadratic function approximation formula y = ax ２ + bx + c of the approximate curve, uses the coefficients a and b of the derived quadratic function approximation formula as feature quantities, uses the feature quantities as inputs, and discriminates the mass of the object using a learned model of machine learning. The low-back pain prevention system according to claim 1.

3. Taking the forward direction of the person as the Y-axis direction, The low back pain prevention system according to claim 2, wherein the calculating means derives an approximate curve approximating the temporal change of the acceleration in the Y-axis direction.

4. The low back pain prevention system according to any one of claims 1 to 3, wherein the calculating means discriminates an object with a mass of 10 kg and an object with a mass of 20 kg.

5. The low back pain prevention system according to claim 4, wherein the measuring means measures the acceleration during the operation of lifting the object by the STOOP method of lifting the object with the knees extended and bending forward.

6. The measuring means is mounted on a measuring device attached to the chest of the person, The measuring device includes communication means for performing wireless communication with an information processing device including the calculating means. The low back pain prevention system according to any one of claims 1 to 3.

7. The measuring means measures the acceleration and angular velocity in the operation of lifting the object and moving it laterally, The calculating means uses the measurement result of the measuring means to calculate a forward-leaning posture angle, which is the angle of the upper body in the forward-leaning posture state with respect to the upright state of the person, and a twisting posture angle, which is the rotation angle around the upper body during the forward-leaning posture, The low back pain prevention system further includes output means for outputting a warning when the forward-leaning posture angle is greater than or equal to a first threshold value, or when the twisting posture angle is greater than or equal to a second threshold value. The low back pain prevention system according to claim 1.

8. The calculating means calculates a first forward-leaning posture angle and a first twisting posture angle using the acceleration result among the measurement results, calculates a second forward-leaning posture angle and a second twisting posture angle using the angular velocity result among the measurement results, and calculates the forward-leaning posture angle and the twisting posture angle using a Kalman filter from the calculated first forward-leaning posture angle, the first twisting posture angle, the second forward-leaning posture angle, and the second twisting posture angle. The low back pain prevention system according to claim 7.

9. The low-back pain prevention system according to claim 7 or 8, wherein the output means outputs a warning sound with a changed tone according to whether the forward leaning posture angle is equal to or greater than the first threshold value or whether the twisting posture angle is equal to or greater than the second threshold value.

10. The warning sound is a sound output at a set time interval, and the time interval of the warning sound is set to different intervals when the forward leaning posture angle is equal to or greater than the first threshold value and when the twisting posture angle is equal to or greater than the second threshold value. The low-back pain prevention system according to claim 9.

11. The measurement means and the output means are mounted in a measuring device. The measuring device has a size that can be worn in the chest pocket of the subject. The low-back pain prevention system according to claim 7 or 8.

12. A program for causing a computer to execute processes related to the prevention of low-back pain, acquiring, as a measurement result, the acceleration in the operation of lifting an object from a forward leaning posture measured by measurement means mounted on the chest of the subject for whom the low-back pain is to be prevented; deriving an approximate curve approximating the temporal change of the acceleration using the measurement result, and determining the mass of the object based on the derived approximate curve; A program for causing the above to be executed.

13. In the step of determination, a quadratic function approximation formula y = ax ２ + bx + c of the approximate curve is derived, the coefficients a and b of the derived quadratic function approximation formula are used as feature quantities, the feature quantities are used as inputs, and a learned model of machine learning is used to determine the mass of the object. The program according to claim 12.

14. The measurement means can measure the acceleration and angular velocity in the operation of lifting and laterally moving an object, and is mounted in a measuring device together with an output means for outputting a warning. The program acquires the acceleration and the angular velocity measured by the measurement means as a second measurement result; calculates a forward leaning posture angle, which is the angle of the upper body in a forward leaning posture state with respect to the upright state of the subject for whom low-back pain is to be prevented, and a twisting posture angle, which is the rotation angle around the upper body during the forward leaning posture, using the second measurement result; when the forward leaning posture angle is equal to or greater than a first threshold value or the twisting posture angle is equal to or greater than a second threshold value, instructs the measuring device to output a warning; The program according to claim 12, which further causes the above to be executed.

15. In the calculating step, the first forward tilt attitude angle and the first twist attitude angle are calculated using the acceleration result among the second measurement results, and the second forward tilt attitude angle and the second twist attitude angle are calculated using the angular velocity result among the second measurement results. The forward tilt attitude angle and the twist attitude angle are calculated using a Kalman filter from the calculated first forward tilt attitude angle, the first twist attitude angle, the second forward tilt attitude angle, and the second twist attitude angle. The program according to claim 14.

Images