JP2007160076A - Human posture and motion discrimination apparatus and energy consumption calculation apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a human posture and motion discrimination apparatus capable of precisely discriminating that a person is in either state of standing, sitting, being recumbent, running, walking, ascending the stairs, descending the stairs, or tumbling using a triaxial acceleration sensor worn on a human body. <P>SOLUTION: This human posture and motion discrimination apparatus for discriminating the posture and motion states of a person using an acceleration detecting means worn on the human body placing its front/rear in the X direction, its right/left in the Y direction and its vertical direction in the Z direction, is provided with: the acceleration detecting means detecting the X-directional acceleration, the Y-directional acceleration and the Z-directional acceleration which orthogonally cross one another; a means for calculating a Z-directional posture component from the detected Z-directional acceleration; a means for calculating a mean value of motor components of resultant composite accelerations of three-directional accelerations from the detected three-directional accelerations; and a first discrimination means discriminating that the person is in either state of the posture states or the motion states based on the calculated Z-directional posture component and the mean value of the motion components of the resultant composite accelerations of the tree-directional accelerations. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

The present invention relates to a human posture motion discriminating apparatus, and in particular, using a three-axis acceleration sensor mounted on a human body, the person is in a standing, sitting, lying, running, walking, stair climbing, stair descent, or falling state The present invention relates to a posture motion discrimination device for a person who discriminates whether or not there is any.
Furthermore, the present invention relates to an energy consumption calculation device, and in particular, using a three-axis acceleration sensor attached to a human body, the person can be in a standing position, sitting position, standing position, running, walking, stair climbing, or stair descent. It is related with the energy consumption calculation apparatus which discriminate | determines whether it exists and calculates the energy consumption according to the said attitude | position movement state.

  With the declining birthrate and aging population, nursing care for elderly people is becoming a social problem. Some care recipients may need to be accompanied when they go out because they may go missing or suddenly fall over. In addition, there are some people who are not confident in free behavior due to the body's weakness and tend to stay home even if they do not need care. The latter are said to be care-reserve reserves and will lead to an increase in care burden in the near future. It is a social responsibility to make such people have the joy of living by being able to go out alone and to make the latter people contribute to society again. To that end, it is necessary to construct a system that can monitor a person's operation state while protecting privacy, and can rush to rescue immediately when it becomes dangerous.

  Such a system has been proposed in the past, and a two-axis accelerometer is attached to the lumbar part of the human body, and the person is standing, sitting or hesitating from the acceleration change in the human body axis direction and the front-rear direction. A technique for determining whether a person is walking, walking, riding a train, a car, a bicycle, or the like (see, for example, Non-Patent Document 1). In addition, a technique is shown in which a triaxial accelerometer is attached to determine each state of walking, running, standing still, and falling from the three-axis acceleration change in the vertical direction, the front-rear direction, and the horizontal direction ( For example, see Patent Document 1). However, the technique of the above-mentioned document cannot accurately determine whether a person is in a specific operating state, particularly a stable state or a dangerous state, and a problem remains in reliability.

  In addition, prompt response is required for lifestyle-related diseases, which account for more than 60% of the causes of death in Japan. In particular, in the so-called working generations in their 40s to 50s, life-style related diseases such as malignant neoplasms, heart diseases, cerebrovascular diseases, and diabetes are important. Efforts have been focused on secondary prevention aimed at discovery and early treatment. Many life-style related diseases progress without being conscious, and when they are noticed, severe symptoms such as stroke and myocardial infarction occur, resulting in a decrease in the quality of life. Therefore, in recent years, it is expected to prevent the onset and progression of diseases by changing lifestyle habits, and the concept of primary prevention for the purpose of health promotion and onset prevention has been emphasized. The improvement of lifestyle habits means improvement of eating habits and exercise habits at home, and it is ideal for exercise habits to make daily exercise a comfortable exercise such as walking fast. And in order to realize this, it is indispensable to accurately grasp the amount of physical activity or energy consumption.

Conventionally, various methods and devices have been proposed for techniques for measuring the amount of physical activity in daily life, particularly techniques using acceleration sensors. For example, a method has been proposed in which an impulse MV of a subject's kinetic mass is obtained from detected acceleration, and calorie consumption is calculated from the obtained impulse (see, for example, Patent Document 2). However, since the magnitude of the acceleration generated when the foot contacts the ground cannot be ignored, the speed V obtained by integrating the acceleration detected by the acceleration sensor includes a large error. For this reason, a large error also occurs in the impulse MV of the kinetic mass, and it cannot be expected to obtain a highly accurate calorie calculation. In addition, a method has been proposed in which body motion is detected by an acceleration sensor to obtain an exercise component, and calorie consumption is calculated from the obtained exercise component (see, for example, Patent Document 3). Have difficulty. Therefore, there is a close relationship between the exercise form and the energy consumed, and therefore, it is not possible to calculate calories with high accuracy by a method that cannot accurately determine the exercise form.
Journal of the Japan Society of Mechanical Engineers 1996 Vol. 101No. 950p14-16 Japanese Patent Laid-Open No. 10-295649 JP 2001-258870 A JP 2004-141669 A

  The problem to be solved by the present invention is whether the person is standing, sitting, lying, running, walking, climbing stairs, descending stairs, or falling using a triaxial acceleration sensor attached to the human body. It is an object of the present invention to provide a posture / motion discrimination apparatus for a person who can accurately discriminate the movement. In addition, using a three-axis acceleration sensor attached to the human body, it is possible to accurately determine whether the person is standing, sitting, lying, running, walking, climbing stairs, or descending stairs. It is to provide an energy consumption calculation device that can accurately calculate the energy consumption according to the state.

  The present invention has been completed as a result of sincerity studies by the inventor in order to solve the above-described problems, and the gist thereof is as follows.

(1) In a human posture motion discriminating apparatus for discriminating a posture / motion state using acceleration detecting means mounted on the human body so that the X direction is front and back, the Y direction is right and left, and the Z direction is vertical, the X directions orthogonal to each other Acceleration detecting means for detecting acceleration G _X , Y-direction acceleration G _Y and Z-direction acceleration G _Z, and means for calculating a Z-direction attitude component G _SZ from the detected Z-direction acceleration G _Z , Means G _MXYZAVE for calculating the combined acceleration component of the three-direction acceleration from the three-direction acceleration, and the average value G of the combined acceleration component of the calculated Z-direction posture component G _SZ and the three-direction acceleration. provided standing from _MXYZAVE, sitting, running and posture state consisting of decubitus, walking, stair rise, stair descending, the first determining means for determining whether the one of the state of the operation state consisting overturning Door attitude motion determining apparatus of people characterized by.

(2) Means for calculating the X-direction posture component G _SX from the detected X-direction acceleration G _X , and standing, sitting, and lying from the calculated X-direction posture component G _SX and Z-direction posture component G _SZ (2) The human posture / motion discriminating device according to (1), further comprising: a second discriminating unit that discriminates which posture state of the human body.
(3) Means for calculating the change amount ΔG _SZ of the posture component in the Z direction from the detected acceleration G _{Z in the} Z direction, and the detected X direction acceleration G _X and the Y direction acceleration G _Y from the X direction and the Y direction It means for calculating a resultant acceleration G _XY acceleration, running from the variation .DELTA.G _SZ synthetic acceleration G _XY and Z directions of the orientation component of the acceleration of the calculated X and Y directions, walking, stair rise consists stairs descending normal The human posture / motion discriminating apparatus according to (1) or (2), further comprising third discriminating means for discriminating whether the motion state is a dangerous motion consisting of a motion or a fall.

(4) A fourth discriminating means for discriminating whether the vehicle is in the running state or the other normal operation state from the average value G _MXYZAVE of the motion component of the combined acceleration of the calculated accelerations in the three directions is provided. The human posture / movement discriminating apparatus according to any one of (1) to (3).
(5) means for calculating an average value G _MXAVE component of motion in the X direction from the acceleration G _X of the detected X-direction, a frequency analysis means for analyzing the frequency spectrum of the acceleration G _Y of the detected Y-direction, the frequency analysis Means for detecting the peak power value S _Y of the frequency spectrum of the acceleration in the Y direction analyzed by the means, means for calculating the average value G _MZAVE of the motion component in the Z direction from the detected acceleration G _{Z in the} Z direction, X-direction orientation component _{G SX,} the average value _{G MXAVE} the X-direction component of motion, the sum is multiplied by a predetermined coefficient determined in advance in each of the average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component (5) The method according to (4), further comprising: a fifth discriminating unit that discriminates whether the normal operation state of walking, stair climbing, or stair descent is based on the obtained calculation result. Attitude operation discrimination device.

(6) means for calculating an average value G _MXAVE component of motion in the X direction from the acceleration G _X of the detected X-direction, a frequency analysis means for analyzing the frequency spectrum of the acceleration of the detected three directions, by the frequency analysis means Means for calculating power spectrum values Y _MAX and Z _MAX of fundamental frequencies in the Y direction and the Z direction from the analyzed power spectra in the three directions, the calculated X direction orientation component G _SX , and the average value of the X direction motion components G _MXAVE and power spectrum values Y _MAX and Z _MAX are used as inputs to determine whether the normal operation state is walking, stair climbing, or stair descent by performing fuzzy inference using membership functions for each motion state determined in advance. A human posture / motion discriminating device according to (4), further comprising sixth discriminating means for discriminating.

(7) Furthermore, it has a means for calculating the walking rate R from the power spectrum in the three directions analyzed by the frequency analyzing means, and the sixth discriminating means has a predetermined operation state corresponding to the walking rate R. The human posture / movement discriminating apparatus according to (6), wherein it is discriminated whether the normal operation state is walking, stair climbing or stair descent by performing fuzzy inference using a membership function.
(8) Furthermore, an operation state input means for inputting whether the person wearing the acceleration detection means is in a normal operation state of walking, stairs climbing, or stairs descending, and the inputted true motion state. Membership function learning means for rewriting a membership function of each predetermined motion state based on the X-direction posture component G _SX , the X-direction motion component average value G _MXAVE , and the power spectrum values Y _MAX and Z _MAX The human posture / movement discriminating apparatus according to (6) or (7), wherein:

(9) The human communication apparatus according to any one of (1) to (8), further comprising: a wireless communication unit that transmits a determination result by the determination unit to an external device installed outside the apparatus main body. Posture motion discrimination device.

(10) Standing, sitting, standing, running, walking, stair climbing, and stair descent using acceleration detecting means mounted on the human body so that the X direction is front and back, the Y direction is right and left, and the Z direction is vertical In an energy consumption calculation device that discriminates the posture movement state and calculates the energy consumption according to the posture movement state, the basal metabolic rate and running in the standing position, sitting position, and the standing position, walking, walking up the stairs, walking down the stairs Acceleration detection for detecting a regression equation storage means for storing a regression equation indicating a correlation between the rate and energy consumption, and an X-direction acceleration G _X , a Y-direction acceleration G _Y and a Z-direction acceleration G _Z which are orthogonal to each other Z means and, means for calculating an average value _{G MXAVE} component of motion of the orientation component _{G SX} and the X-direction in the X direction from the acceleration _{G X} of the detected X-direction, from the acceleration _{G Z} of the detected Z-direction Means for calculating an average value G _MZAVE the orientation component G _SZ and Z direction motion component countercurrent, and frequency analyzing means for each analyzing the frequency spectrum of the acceleration detected in three directions, were analyzed by the frequency analysis means Y It means for detecting a peak power value S _Y of the frequency spectrum of the direction of acceleration, to detect the peak frequency of the frequency spectrum of the acceleration of the peak frequency and the Z-direction of the frequency spectrum of the acceleration of the analyzed X-direction by the frequency analysis means Peak frequency detection means, means for detecting the walking rate R from the peak frequency of the detected frequency spectrum of the acceleration in the X direction and the peak frequency of the frequency spectrum of the acceleration in the Z direction, and acceleration in the three directions from the detected acceleration in the three directions It means for calculating an average value _{G MXYZAVE} component of motion of the combined acceleration of the calculated Z Any average value G _MXYZAVE movement component of the resultant acceleration of the orientation component G _SZ and 3 the direction of acceleration of the direction standing, sitting, running and posture state consisting of decubitus, walking, stair rise, the operating condition consisting of stairs descending A first discriminating means for discriminating whether the vehicle is in a standing state, and discriminating whether the posture state is standing, sitting or lying from the computed X-direction posture component G _SX and Z-direction posture component G _SZ Second discriminating means, and third discriminating means for discriminating whether the vehicle is in the normal operation state of running or other normal operation from the average value G _MXYZAVE of the motion component of the combined acceleration of the calculated three-direction acceleration If, computed X-direction orientation component _{G SX,} the average value _{G MXAVE} the X-direction component of motion, a predetermined coefficient determined in advance in each of the average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component multiply A fourth discriminating means for discriminating whether it is a normal operation state of walking, stair climbing or stair descent from the calculation result obtained by taking the sum, and a basal metabolic rate or regression to be applied according to the discrimination result by the discrimination means Energy consumption calculation comprising: a regression equation selection means for selecting an expression; and an energy consumption calculation means for calculating an energy consumption amount corresponding to the posture / motion state based on the selected basal metabolic rate or regression equation apparatus.

(11) Standing, sitting, standing, running, walking, stair climbing, and stair descent using acceleration detecting means mounted on the human body so that the X direction is front and back, the Y direction is right and left, and the Z direction is vertical. In an energy consumption calculation device that discriminates the posture movement state and calculates the energy consumption according to the posture movement state, the basal metabolic rate and running in the standing position, sitting position, and the standing position, walking, walking up the stairs, walking down the stairs Acceleration detection for detecting a regression equation storage means for storing a regression equation indicating a correlation between the rate and energy consumption, and an X-direction acceleration G _X , a Y-direction acceleration G _Y and a Z-direction acceleration G _Z which are orthogonal to each other Z means and, means for calculating an average value _{G MXAVE} component of motion of the orientation component _{G SX} and the X-direction in the X direction from the acceleration _{G X} of the detected X-direction, from the acceleration _{G Z} of the detected Z-direction Means for calculating the orientation component G _SZ countercurrent, and frequency analyzing means for each analyzing the frequency spectrum of the acceleration of the detected three directions, from the power spectrum of the three directions which are analyzed by the frequency analyzing means in the Y and Z directions Means for calculating the power spectrum values Y _MAX and Z _MAX of the fundamental frequency, means for calculating the walking rate R from the power spectrum in the three directions analyzed by the frequency analysis means, and three directions from the detected acceleration in the three directions. _Means for calculating an average value G _MXYZAVE of the motion component of the combined acceleration of the acceleration, and standing, sitting, and _臥 from the calculated posture component G _SZ of the Z direction and the average value G _MXYZAVE of the combined acceleration of the three-direction acceleration The first discrimination that discriminates between the posture state consisting of the position and the operation state consisting of running, walking, stair climbing, and stair descent Means, a second discriminating means for discriminating from the calculated X-direction posture component G _SX and the Z-direction posture component G _SZ whether the posture is in the standing position, the sitting position or the supine position, and the calculated three directions A third discriminating means for discriminating from the average value G _MXYZAVE of the movement component of the acceleration of the acceleration of the normal operation state of running and other normal operation, and the calculated X-direction posture component G _SX , Walking, staircase by performing fuzzy inference using membership function of each predetermined motion state corresponding to walking rate R, using average value G _MXAVE of power component in X direction and power spectrum values Y _MAX and Z _MAX as input A fifth discriminating means for discriminating whether it is in the normal operating state of rising or stair descent, and a basal metabolic rate or regression equation to be applied according to the discrimination result by the discriminating means And-option regression equation selection means, the energy consumption calculation unit, characterized in that it comprises an energy consumption amount calculation means for calculating the energy consumption in accordance with the posture operation state based on the basal metabolic rate or regression formula is selected.

(12) Furthermore, an operation state input means for inputting whether the person wearing the acceleration detection means is in a normal operation state of walking, stairs climbing or stairs descending, and the input true motion state. Membership function learning means for rewriting a membership function of each predetermined motion state based on the X-direction posture component G _SX , the X-direction motion component average value G _MXAVE , and the power spectrum values Y _MAX and Z _MAX The energy consumption calculation device according to (11), characterized in that:

(13) Any one of the above (10) to (12), characterized by comprising consumption amount storage means for sequentially storing the calculated energy consumption amount at regular intervals, and display means for displaying the calculated energy consumption amount. The energy consumption calculation device according to Item 1.
(14) The above (10), characterized by comprising means for integrating the calculated energy consumption, integrated consumption storage means for storing the integrated energy consumption, and display means for displaying the integrated energy consumption. The energy consumption calculation device according to any one of to (13).
(15) Provided with a regression equation input means for inputting a regression equation indicating the correlation between the basal metabolic rate or running, walking, stair climbing, and stair descent in the standing position, sitting position, and standing position and energy consumption. The energy consumption calculation device according to any one of (10) to (14), characterized in that it is characterized.

(A) According to the human posture / motion discriminating apparatus according to the present invention including the first discriminating means, it is possible to discriminate with high accuracy whether the person is in the posture state or the motion state.
(B) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means and the second determination means, whether the person is in a standing position, a sitting position, or a lying position, an operating state Can be determined with high accuracy.
(C) According to the human posture / motion discriminating apparatus according to the present invention including the first discriminating means and the third discriminating means, whether the person is in the posture state, in the fall state, or normal operation other than the fall It is possible to determine whether the state is in a state with high accuracy.
(D) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the second determination means, and the third determination means, the person is in a standing position, a sitting position, or a lying position Therefore, it is possible to determine with high accuracy whether the vehicle is in the fall state, the fall state, or the normal operation state other than the fall state.

(E) According to the human posture / motion discriminating apparatus according to the present invention including the first discriminating means and the fourth discriminating means, whether the person is in the posture state, the running state, or the normal operation other than the running It is possible to determine whether the state is in a state with high accuracy.
(F) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the second determination means, and the fourth determination means, the person can be in a standing position, a sitting position, or a lying position. It is possible to determine with high accuracy whether the vehicle is in a traveling state or in a normal operation state other than traveling.
(G) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the third determination means, and the fourth determination means, whether the person is in a posture state or in a fall state Therefore, it is possible to determine with high accuracy whether the vehicle is in a traveling state or a normal operation state other than traveling.
(H) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the second determination means, the third determination means, and the fourth determination means, the person is in a standing position, a sitting position, It is possible to determine with high accuracy whether the vehicle is in the prone position, whether it is in a fall state, whether it is in a traveling state, or in a normal operation state other than traveling.

(I) According to the human posture / motion discriminating apparatus according to the present invention including the first discriminating means, the fourth discriminating means, and the fifth discriminating means, whether the person is in the posture state, running, walking, stairs It is possible to determine with high accuracy whether the state is ascending or descending stairs.
(J) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the second determination means, the fourth determination means, and the fifth determination means, the person is standing, sitting, It can be discriminated with high accuracy whether it is in the prone position, or it is in the state of running, walking, stair climbing, or stair descent.
(K) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the third determination means, the fourth determination means, and the fifth determination means, is the person in a posture state? Therefore, it is possible to determine with high accuracy whether the vehicle is in a fall state or is in a state of running, walking, stair climbing, or stair descent.
(L) According to the posture / movement determination apparatus for a person according to the present invention including the first determination means, the second determination means, the third determination means, the fourth determination means, and the fifth determination means, It is possible to discriminate with high accuracy whether the vehicle is in a standing, sitting, or lying position, in a falling state, or in a running, walking, stair climbing, or stair descent state.
(M) According to the human posture / movement discriminating apparatus according to the present invention having the fifth discriminating means described above, it is possible to discriminate with high accuracy whether the person is in the state of walking, stair climbing or stair descent. However, according to the human posture / motion discriminating apparatus according to the present invention, which comprises sixth discriminating means for performing fuzzy inference using membership functions of predetermined motion states instead of the fifth discriminating means. It is possible to determine whether the person is walking, climbing the stairs, or descending the stairs with higher accuracy.
(N) Furthermore, according to the human posture / motion discriminating apparatus according to the present invention that self-detects the walking rate R, the fuzzy inference is performed using the membership function of each predetermined motion state corresponding to the walking rate R. Therefore, it is possible to determine with high accuracy whether the person is in a walking state, climbing stairs, or descending stairs.
(O) Furthermore, according to the human posture / motion discriminating apparatus according to the present invention including the membership function learning means, it is possible to easily create an optimal membership function for the person wearing the acceleration detecting means. It is possible to determine whether the state is walking, climbing stairs, or descending stairs with high accuracy.

(P) According to the energy consumption calculation device of the present invention, it is possible to accurately determine whether the person is standing, sitting, lying, running, walking, climbing stairs, or descending stairs. Therefore, the energy consumption amount corresponding to the posture movement state can be accurately calculated.
(Q) Furthermore, according to the energy consumption calculation apparatus according to the present invention that uses fuzzy reasoning to discriminate between walking, stair climbing, and stair descent, it is highly accurate whether it is in the operating state of walking, stair climbing or stair descent. Since the determination can be made, the energy consumption corresponding to each operation state can be calculated more accurately.

Hereinafter, the best mode for carrying out the present invention will be described with reference to FIGS.
FIG. 1 is a schematic diagram for explaining a mounting position and an axial direction of a triaxial acceleration sensor 11 that is an acceleration detecting means 10. In the present invention, as shown in FIG. 1, the three-axis acceleration sensor 11 is attached to the subject so that the X-axis direction is front and rear, the Y-axis direction is left and right, and the Z-axis direction is vertical. It is desirable that the triaxial acceleration sensor 11 be attached to the waist where the movement is the starting point and changes most with respect to the movement or posture change. More preferably, it is preferably attached to the lower part of the anterior iliac spine. It is also desirable to correct the angle so that the X-axis acceleration is 0G, the Y-axis is 0G, and the Z-axis is 1G when standing upright (standing).

FIG. 2 is an example of a flowchart showing a discrimination method of the human posture / motion discrimination device 1 according to the present invention. In this example, the first discriminating means 40 discriminates between a posture state consisting of standing, sitting and lying and an operating state consisting of running, walking, stair climbing, stair descent, and falling. When the posture state is determined, the process proceeds to the second determination unit 41, and when the movement state is determined, the process proceeds to the third determination unit 42.
The second determining means 41 determines whether the posture is standing, sitting or lying, and the third determining means 42 is a normal operation consisting of running, walking, stair climbing, and stair descent and falling. It is determined which of the operation states of the dangerous operation consists of. When the third determining means 42 determines that the operation is normal, the routine proceeds to the fourth determining means 43.
In the fourth step 43, it is determined whether the vehicle is in a running state or a normal operation other than traveling, and when it is determined that the operation state is other than traveling, the process proceeds to the fifth determining unit 44, where walking, It is determined whether the stairs are moving up or down.

The first discrimination means 40 of the human posture / motion discrimination device 1 according to the present invention will be described. The In the first discriminating means 40, standing from the average value G _MXYZAVE movement component of the resultant acceleration in the Z direction orientation component G _SZ and 3 the direction of the acceleration as shown in FIG. 2, locus, consisting supine posture It is determined whether the state is an operating state consisting of running, walking, stair climbing, stair descent, or falling.
Orientation component G _SZ in the Z direction is a value calculated from the acceleration G _Z of the detected Z-direction by the following equation (1) by the acceleration detecting means 10, a calculation result with the acceleration G _Z of the detected Z-direction Z FIG. 3 shows the relationship with the orientation component _{GSZ in the} direction.
In the present invention, as shown in FIG. 4, the acceleration G _X (i) in the X direction, the acceleration G _Y (i) in the Y direction, and the acceleration G _Z (i) in the Z direction are detected at a sampling frequency of 100 Hz. At the same time, this is stored in the predetermined acceleration data storage means 20, and the acceleration calculation means 30 performs arithmetic processing for deriving data used for posture action discrimination for 512 pieces of stored acceleration data. The length N is not limited to this.

The average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions is a value calculated by the following equations (2) and (3) from the accelerations in the three directions detected by the acceleration detecting means 10.

In order to determine whether the posture state is composed of standing, sitting, and lying down and the movement state including running, walking, stair climbing, stair descent, and falling, the Z direction posture component G _SZ and the three directions The reason for using the average value G _MXYZAVE of the motion component of the combined acceleration of acceleration will be described.
FIG. 5 is a distribution diagram showing the relationship between each posture action and the posture component _{GSZ in} the Z direction. It can be seen that by comparing the calculated Z-direction posture component _GSZ with a predetermined threshold value, it is possible to discriminate the supine position from other posture operation states. FIG. 6 is a distribution diagram showing the relationship between the average value G _MXYZAVE of the motion component of the combined acceleration of each posture action and the acceleration in three directions. It can be seen that by comparing the average value G _MXYZAVE of the combined acceleration of the calculated three-direction acceleration with a predetermined threshold value, the posture state composed of standing and sitting and the other motion state can be discriminated.

The second determination means 41 of the human posture / movement determination apparatus 1 according to the present invention will be described. In the second discriminating means 41, as shown in FIG. 2, it is discriminated from the X-direction posture component G _SX and the Z-direction posture component G _SZ whether the posture state is standing, sitting or lying. .
The X-direction attitude component G _SX is a value calculated from the X-direction acceleration G _X detected by the acceleration detection means 10 according to the following equation (4). The Z-direction attitude component G _SZ is obtained by the acceleration detection means 10. is a value calculated by the equation (1) from the acceleration G _Z of the detected Z-direction.

The reason why the posture component G _SX in the X direction and the posture component G _SZ in the Z direction are used to determine whether the posture state is the standing position, the sitting position, or the supine position will be described.
FIG. 7 is a distribution diagram showing the relationship between the standing position, the sitting position, and the supine position and the posture component G _{SX in the} X direction and the posture component G _{SZ in} the Z direction. The calculated X-direction posture component G _SX and Z-direction posture component G _SZ are plotted, and it can be seen that the posture state of the standing position, the sitting position, or the supine position can be discriminated depending on which region is located.

The third discrimination means 42 of the human posture / motion discrimination device 1 according to the present invention will be described. In the third discriminating means 42, as shown in FIG. 2, from the combined acceleration G _XY of the acceleration in the X direction and the Y direction and the change amount ΔG _SZ of the posture component in the Z direction, from running, walking, stair climbing, and stair descent It is determined whether the normal operation or the dangerous operation consisting of a fall is present.
The combined acceleration G _XY of the acceleration in the X direction and the Y direction is a value calculated by the following equation (5) from the acceleration G _X in the X direction detected by the acceleration detecting means 10 and the acceleration G _{Y in the Y} direction. the variation .DELTA.G _SZ of the orientation component is the change amount at a predetermined time G _SZ computed by the equation (1).

The combined acceleration _GXY and the Z-direction attitude components of the acceleration in the X and Y directions to determine whether the operation state is normal operation consisting of running, walking, stair climbing, and stair descent, or dangerous motion consisting of falling. The reason why the change amount ΔG _SZ is used will be described.
FIG. 8 is a distribution diagram showing the relationship between each motion and the combined acceleration _GXY of the accelerations in the X and Y directions. It can be seen that the fall cannot be distinguished from other operating states only by the combined acceleration _GXY . However, since the change amount ΔG _SZ of the posture component in the Z direction at the time of the fall, more specifically at the moment of the fall, is obviously larger than the operation state such as running, the combined acceleration G _XY of the acceleration in the X direction and the Y direction. Further, by using the change amount ΔG _SZ of the posture component in the Z direction, the fall and the other operation state are determined.

The 4th discrimination | determination means 43 of the human posture movement discrimination | determination apparatus 1 which concerns on this invention is demonstrated. In the fourth determination means 43, as shown in FIG. 2, it is determined from the average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions _{whether the vehicle} is in a running state or other normal operation state. .
The average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions is a value calculated from the accelerations in the three directions detected by the acceleration detecting means 10 according to the expressions (2) and (3).

The reason why the average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions is used to determine whether the vehicle is in the running state or the normal operation other than that will be described.
As described above, FIG. 6 is a distribution diagram showing the relationship between the average value G _MXYZAVE of the motion components of the combined acceleration of each posture motion and the acceleration in the three directions. It can be seen that the running and other operating states can be discriminated by comparing the average value G _MXYZAVE of the motion component of the combined acceleration of the calculated accelerations in the three directions with a predetermined threshold value.

The fifth determination means 44 of the human posture / movement determination apparatus 1 according to the present invention will be described. In the fifth discriminating means 44, as shown in FIG. 2, the X-direction posture component G _SX , the X-direction motion component average value G _MXAVE , the Z-direction motion component average value G _MZAVE and the peak power value S _{It is} determined from the calculation results Z1 and Z2 obtained by multiplying each of _Y by a predetermined coefficient in advance whether the normal operation state is walking, stair climbing, or stair descent.
The X-direction posture component G _SX is a value calculated from the X-direction acceleration G _X detected by the acceleration detection means 10 according to the above equation (4).
The average value G _MXAVE of the motion component in the X direction is a value calculated from the acceleration G _X in the X direction detected by the acceleration detection means 10 according to the following equations (6) and (7).

The average value G _MZAVE of the motion component in the Z direction is a value calculated by the following equations (8) and (9) from the acceleration G _Z in the Z direction detected by the acceleration detecting means 10.

The peak power value S _Y is a peak power value of the frequency spectrum of the acceleration in the Y direction analyzed by the frequency analysis of the acceleration G _Y in the Y direction detected by the acceleration detecting means 10. For the frequency analysis, it is desirable to perform fast Fourier transform (FFT). In the present invention, FFT analysis is performed at 512 points.

Next, orientation component of the calculated X-direction _{G SX,} the average value _{G MXAVE} the X-direction component of motion, the two discriminant functions and the average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component elements Z1, Z2 is calculated by the following equations (10) and (11).

In order to determine which of the normal operation states of walking, stair climbing, and stair descent, the posture component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, and the average value G of the motion component in the Z direction _{The reason why the MZAVE} and the peak power value S _Y and the discriminant functions Z1 and Z2 having these as elements are used will be described.

FIG. 9 is a distribution diagram showing the relationship between the average value G _MXAVE of the motion component in the X direction during walking, stair climbing, and stair descent and the walking rate. It can be seen that the plot area during walking is higher than the plot area for stair climbing and descent, and the average value G _MXAVE of the motion component in the X direction can be used as a discriminating element for discriminating walking and other motion states. .

FIG. 10 is a distribution diagram showing the relationship between the average value _GMZAVE of the motion component in the Z direction and the walking rate when walking, climbing stairs, and descending stairs. The plot area at the time of stairs descent is higher than the plot area for walking and stairs rise, and the average value G _MZAVE of the motion component in the Z direction can be used as a discriminating element for discriminating between the stairs descent and other motion states. Recognize.

FIG. 11 is a distribution diagram showing the relationship between the posture component _GSX in the X direction and the walking rate at the time of walking, stair climbing, and stair descent, and FIG. 12 is a distribution diagram when performed on another subject. FIG. 13 is a distribution diagram showing the relationship between the peak power value _SY and the walking rate at the time of walking, stair climbing, and stair descent, and FIG. 14 is a distribution diagram when performing for another subject. There are individual differences in the posture component G _SX in the X direction during walking, stair climbing, and stair descent, and only the posture component G _{SX in the} X direction is used as a discriminating element for discriminating the stair climbing and other motion states. However, it can be seen that the determination accuracy can be improved by using the peak power value _SY together with the determination element.

From these technical findings, the present inventor used the posture component G _{SX in} the X direction and the motion component in the X direction as a discriminating element for discriminating whether the normal operation state is walking, stair climbing, or stair descent. mean value _{G MXAVE,} it was adopted average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component. In order to make discrimination easier, discriminant functions Z1 and Z2 having these as elements are employed. FIG. 15 is a distribution diagram in which discriminant functions Z1 and Z2 are plotted when 10 test subjects walk, climb the stairs, and descend the stairs. In this example, the discriminant functions Z1 and Z2 are calculated by the following formulas (12) and (13). These coefficients are the average values of the calculated X-direction posture component G _SX and X-direction motion component. G _MXAVE, and is calculated by using the average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component multivariate analysis. The determination accuracy in this example is 99% when walking, 98% when climbing the stairs, and 99.8% when descending the stairs. Thus, previously subject walking, stair rise, X direction orientation component _{G SX} during stair descending, the average value _{G MXAVE} the X-direction component of motion, the mean value _{G MZAVE} and peak power value _{S Y} and Z direction motion component By calculating, performing multivariate analysis using these data, and calculating the coefficient of each discriminating element, the determination accuracy can be further improved. Therefore, it is desirable to include a coefficient input means 50 for inputting these coefficients and a coefficient storage means 51 for storing them.

  FIG. 16 is a block diagram showing an example of the configuration of the human posture / movement discriminating apparatus 1 according to the present invention. The acceleration data storage means 20 is not particularly limited as long as it can store the detected acceleration data, but it is desirable to use a RAM or a hard disk. The same applies to the discrimination result storage means 60 and the coefficient storage means 51 that store the discrimination results. Further, a CPU can be used as the acceleration calculation means 30, or can be realized by software.

As described above, (1) the posture state composed of standing, sitting, and lying down, running, and walking from the average value G _MXYZAVE of the combined acceleration of the calculated Z-direction posture component G _SZ and the acceleration in the three directions. A first discriminating means 40 for discriminating whether the operating state is a stair climbing, a stair descent, or a falling, and (2) the calculated X-direction posture component G _SX and Z-direction posture component G _SZ. Second discriminating means 41 for discriminating whether the posture state is standing, sitting or supine from (3) the combined accelerations G _XY and Z direction posture components of the calculated X and Y direction accelerations A third discriminating means 42 for discriminating from the change amount ΔG _{SZ of the} normal operation consisting of running, walking, stair climbing, and stair descent and the dangerous motion consisting of falling, and (4) calculating 3 Combined acceleration of direction acceleration And fourth determination means 43 for determining whether the average value G _MXYZAVE component of motion is in one operating state of the traveling and other normal operation, (5) of the calculated X-direction orientation component G _SX, the _X-direction mean value G _MXAVE component of _motion, the walking from the _Z direction component of motion average G _MZAVE and peak power value S _Y each predetermined coefficient multiplying operation result of adopting the sum determined in advance in of stairs rising, stair descent According to the human posture / motion discriminating apparatus 1 according to the present invention including the fifth discriminating means 44 for discriminating which of the normal operating states, the person (subject) wearing the acceleration detecting means is standing and sitting. It is possible to determine with high accuracy whether the vehicle is in a saddle position, running, walking, stair climbing, stair descent or falling.

  Next, another embodiment of the human posture / movement discriminating apparatus 1 according to the present invention will be described. FIG. 20 is an example of a flowchart showing another discrimination method of the human posture / motion discrimination device 1 according to the present invention. FIG. 28 is a block diagram showing another example of the configuration of the human posture / movement discrimination device 1 according to the present invention. In this example, instead of the fifth discriminating means 44 described above, by providing the sixth discriminating means 45 for performing fuzzy inference, it is possible to determine whether the person is in the state of walking, climbing stairs or descending stairs. Determine.

The sixth discrimination means 45 of the human posture / motion discrimination device 1 according to the present invention will be described. In the sixth discriminating means 45, as shown in FIG. 20, the X-direction attitude component G _SX , the X-direction motion component average value G _MXAVE , the power spectrum value Y _MAX and the power spectrum value Z _MAX are inputted in advance. By performing fuzzy inference using the membership function of each determined operation state, it is determined whether the normal operation state of walking, stair climbing or stair descent is present.
The X-direction posture component G _SX is a value calculated from the X-direction acceleration G _X detected by the acceleration detection means 10 according to the above equation (4).
The average value G _MXAVE of the motion component in the X direction is a value calculated from the acceleration G _X in the X direction detected by the acceleration detecting means 10 according to the equations (6) and (7).
As shown in FIG. 21, the power spectrum value Y _MAX is obtained by performing frequency analysis of the acceleration in the three directions detected by the acceleration detecting means 10 and calculating the basic value in the Y direction calculated from the power spectrum in the three directions analyzed by the frequency analysis. It is the power spectrum value of the frequency.
As shown in FIG. 21, the power spectrum value Z _MAX is obtained by performing frequency analysis of the acceleration in the three directions detected by the acceleration detecting means 10 and calculating the basic value in the Z direction calculated from the power spectrum in the three directions analyzed by the frequency analysis. It is the power spectrum value of the frequency.
For the frequency analysis, it is desirable to perform fast Fourier transform (FFT). In the present invention, FFT analysis is performed at 512 points. Further, although the fundamental frequency extraction method is not particularly limited, as shown in FIG. 21, at least in the normal operation state of walking, stair climbing, and stair descent, the fundamental frequency in the X direction and the fundamental frequency in the Z direction. Is almost the same (one fundamental frequency is not an integral multiple of one fundamental frequency), and ½ of these fundamental frequencies and the fundamental frequency in the Y direction are almost identical. Thus, the fundamental frequency can be extracted.

Next, the reason why the attitude component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, the power spectrum value Y _MAX, and the power spectrum value Z _MAX are used as inputs for fuzzy inference will be described.
The posture component G _{SX in the} X direction will be described. As described above, FIG. 11 is a distribution diagram showing the relationship between the posture component _GSX in the X direction and the walking rate at the time of walking, stair climbing, and stair descent, and FIG. 12 is a distribution diagram when performed on another subject. . Although there are individual differences from these distribution charts, if the posture component G _{SX in} the X direction is large, it can be used as a discriminating element for discriminating stair climbing and other motion states, and the posture component G _{SX in the} X direction If it is small, it can be seen that it can be used as a discriminating element for discriminating stairs descending from other operating states.

The average value G _MXAVE of the motion component in the X direction will be described. As described above, FIG. 9 is a distribution diagram showing the relationship between the average value G _MXAVE of the motion component in the X direction and the walking rate when walking, climbing stairs, and descending stairs. From this distribution map, the plot area at the time of walking is higher than the plot area of the stair climbing and stair descent, and the average value G _MXAVE of the motion component in the X direction is a discriminating element for discriminating walking and other motion states It can be used as.

The power spectrum value Y _MAX will be described. FIG. 22 is a distribution diagram showing the relationship between the power spectrum value Y _MAX and the walking rate when walking, climbing stairs, and descending stairs. From this distribution chart, the plot area at the time of stairs rising is higher than the plot area at the time of walking and stairs descending, and the power spectrum value Y _MAX is used as a discriminating element for discriminating the stairs ascent and other operating states I understand that I can do it.

The power spectrum value Z _MAX will be described. FIG. 23 is a distribution diagram showing the relationship between the power spectrum value Z _MAX and the walking rate when walking, climbing stairs, and descending stairs. From this distribution chart, the plot area at the time of stairs descending is higher than the plot area at the time of walking and ascending stairs, and the power spectrum value Z _MAX is used as a discriminating element for discriminating between stairs descending and other operating states. I understand that I can do it.

From these technical findings, (1) membership functions relating to the posture component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, the power spectrum value Y _MAX and the power spectrum value Z _MAX are shown in FIG. ~ (D) is created in advance, (2) and the fuzzy rules 1 to 3 shown below are created in advance, and (3) the acceleration detection means 10 attached to the human body is detected momentarily. X direction posture component G _SX , X direction motion component average value G _MXAVE , power spectrum value Y _MAX and power spectrum value Z _MAX Is used as input to perform fuzzy inference using the membership function of each motion state determined in advance. To determine whether the normal operating state of the shift, a discriminating means 45 of the sixth. Next, the fuzzy rule will be described.

Rule 1: if G _SX = Large and Y _MAX = Large then the stairs rise Rule 2: if G _MXAVE = Large then Walk Rule 3: if G _SX = Small and Z _MAX = Large the stairs rise

Rule 1 uses the membership function of FIG. 24 (a), calculates the fitness by applying the calculated orientation component G _SX in the X direction, and uses the membership function of FIG. 24 (c). The fitness is calculated by applying the calculated power spectrum value Y _MAX , and the fitness by AND (logical product) of these fitness is calculated.
Similarly, in the rule 2, the membership function shown in FIG. 24B is used, and the fitness is calculated by applying the calculated average value G _MXAVE of the motion components in the X direction.
Similarly, in the rule 3, the membership function of FIG. 24A is used, the fitness is calculated by applying the calculated posture component G _SX in the X direction, and the membership function of FIG. The fitness is calculated by applying the power spectrum value Z _{MAX used} and calculated, and the fitness by AND (logical product) of these fitness is calculated.

  Then, when the fitness for each rule is calculated, it is determined whether it is in the normal operation state of walking, climbing stairs, or descending stairs by taking the center of gravity of the logical sum (superposition) of the membership functions of each rule. Is done. The defuzzification method is not particularly limited, and for example, the Max-Min method can be used.

The method for creating the membership function is not particularly limited, and the frequency distribution of the posture component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, the power spectrum value Y _MAX and the power spectrum value Z _MAX It is desirable that the Mandani method can be used. The created membership function is stored in the membership function storage means 73. Note that the membership functions shown in FIGS. 24A to 24D are created from the frequency distribution. Specifically, each of the membership functions when walking, climbing stairs, and descending stairs at a walking rate of 100 step / min. Created from frequency distribution.

FIG. 25 is a graph showing the determination accuracy of the sixth determination means 45 using fuzzy inference, and the test conditions are as follows.
(1) Subjects: 24 (2) Subject motion conditions: Walking, stair climbing, and stair descent motions at a specified walking rate of 60-130 step / min (10 step / min increments) (3) Fuzzy rule used: rule 1-3
(4) Membership function used: Membership function shown in FIGS.

  The membership function used as described above is a membership function created from the respective frequency distributions when walking, climbing stairs, and descending stairs at a walking rate of 100 step / min. The subject has a walking rate of 60 to 130 step / min. It was confirmed that each operation state could be discriminated with an accuracy of about 80% or more even when walking, climbing stairs, or descending stairs at 10 step / min increments.

  Therefore, the present inventor aims to further improve the discrimination accuracy, and automatically detects the walking rate R of the person using the acceleration detected every moment from the acceleration detecting means 10 attached to the human body, and automatically detects the walking. The present inventors have come up with a discrimination method that improves the discrimination accuracy of walking, stair climbing, and stair descent by performing fuzzy inference using a predetermined membership function corresponding to the rate R.

  Regarding the detection of the walking rate R, the frequency spectrum of the detected acceleration in the X direction and the acceleration in the Z direction is subjected to frequency analysis (FFT analysis) using the frequency analysis means 33, and the frequency spectrum of the analyzed acceleration in the X direction is analyzed. The walking rate R can be detected by detecting the peak frequency and the peak frequency of the frequency spectrum of the acceleration in the Z direction, comparing both frequencies, and selecting the lower peak frequency. Table 1 to be described later shows the determination accuracy when the method is adopted. It can be seen that the walking rate at the time of walking, stair climbing, and stair descent can be detected with higher accuracy than Table 1.

  For the frequency analysis, it is desirable to perform fast Fourier transform (FFT) using frequency analysis means 33. In the present invention, FFT analysis is performed at 512 points. The peak frequency extraction method is not particularly limited. As shown in FIG. 21, at least in the normal operation state of walking, stair climbing, and stair descent, the fundamental frequency in the X direction and the fundamental frequency in the Z direction. Are almost the same (one fundamental frequency is not an integral multiple of one fundamental frequency), and ½ of these fundamental frequencies and the fundamental frequency in the Y direction are almost identical. Thus, the peak frequency can be extracted. This is effective when there is a lot of noise in the acceleration data detected from moment to moment by the acceleration detection means 10 worn on the human body, and no peak appears at the original frequency, that is, the frequency indicating the walking rate even after frequency analysis. .

As described above, (1) walking, stair climbing by performing fuzzy inference using the posture component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, the power spectrum value Y _MAX and the power spectrum value Z _MAX as inputs. Using the basic principle of the sixth discriminating means 45 that can discriminate which of the normal operating states of stairs descending, and (2) acceleration detected momentarily from the acceleration detecting means attached to the human body It has been explained that the determination accuracy can be further improved by automatically detecting the walking rate R of the person and performing fuzzy inference using the membership function corresponding to the automatically detected walking rate R. However, the determination accuracy is improved. The method is not limited to this.

For example, in the fifth determination means 44 for determining whether the normal operation state of walking, stair climbing or stair descent is the same as the sixth determination means 45, the average value G _MZAVE of the motion component in the Z direction is used. It was. As shown in FIG. 10 showing the relationship between the average value G _MZAVE of the movement component in the Z direction and the walking rate R during walking, stair climbing, and stair descent, the plot area at the time of stair descent is that of walking and stair climbing. This is based on the technical knowledge that the average value G _MZAVE of the movement component in the Z direction can be used as a discriminating element for discriminating the descent from the staircase and other operation states. Therefore, also in the 6th discrimination | determination means 45 which performs a fuzzy reasoning, the determination precision can be improved by adding the rule shown below to the above-mentioned fuzzy rules 1-3. FIG. 24 (e) shows an example of the membership function of _GMZAVE .
Rule 4: if G _MZAVE = large then down the stairs

Further, the determination accuracy can be improved by normalizing the power spectrum value Y _MAX . FIG. 26A is a schematic diagram illustrating the power spectrum value Y _MAX in walking, stair climbing, and stair descent before normalization, and FIG. 26B is a schematic diagram illustrating the power spectrum value Y _MAX after normalization. As shown in FIG. 26 (a), the power spectrum value Y _MAX of a specific (minor) person is lower than the average person's power spectrum value Y _MAX , and the distributions in walking, stair climbing, and stair descent overlap. Therefore, even if fuzzy inference is performed as it is, the determination accuracy is limited. However, if normalization is performed according to the following equation (14), the power spectrum value Y _{MAX for} walking, stair climbing, and stair descent is as shown in FIG. 26B, and improvement in the determination accuracy of these operations can be expected. In the equation (14), _BY is an average value of the power spectrum values Y _MAX for each operation. Therefore, the accuracy of determination can be improved by adding the following rules to the fuzzy rules 1 to 3 described above. Incidentally, _{Y MAX} normalized to FIG 24 (f), i.e., an example of a membership function _{C Y.}
Rule 5: if _{C Y} = large then stairs rise

The above is a method of performing fuzzy inference using a predetermined membership function based on a frequency distribution, etc., but further determination is made by providing the posture motion discriminating apparatus 1 with the learning function (function) of the membership function. Accuracy can be improved. As shown in FIG. 16, acceleration data that changes from moment to moment is input to the posture movement determination device 1 according to the present invention, and a calculation value G _SX or the like is obtained by a G _SX calculation unit 31 or the like as the acceleration calculation unit 30. It is calculated and each posture action is discriminated using the calculated value G _SX or the like. Accordingly, by inputting whether the person wearing the acceleration detecting means 10 is in a normal operation state of walking, stair climbing, or stair descent, the inputted true motion state and calculation of G _{SX and the} like at that time are performed. The relationship with the value can be associated. That is, it is possible to grasp the calculation values such as G _SX in the person (subject) wearing the acceleration detecting means 10 in the walking, stairs ascending, and stairs descending. For example, if a calculated value such as G _SX is within a certain range, it can be determined by learning that the person is in a walking state.

The learning function (function) of the membership function utilizes this, and an operation state input means for inputting whether the person wearing the acceleration detection means 10 is in a normal operation state of walking, climbing stairs or descending stairs. 71, membership storage means based on the input true motion state, the input X-direction posture component G _SX , the average value G _MXAVE of the X-direction motion component, and the power spectrum values Y _MAX and Z _MAX The learning function can be realized by including membership function learning means 72 for rewriting the membership function of each predetermined operation state stored in 73. Thereby, according to the human posture movement discriminating apparatus 1 according to the present invention having the membership function learning ability, it is possible to easily create an optimal membership function for the person wearing the acceleration detecting means 10, It is possible to determine whether the state is walking, climbing stairs, or descending stairs with high accuracy. Incidentally, as described above, the membership functions is not limited thereto, it can be used in addition to the membership functions of the G _MZAVE and C _Y, etc. described above.

  The operation state input means 71 is not particularly limited, and any form can be used as long as it can input an operation state of walking, climbing stairs, or descending stairs electrically or mechanically. Further, the membership storage unit 73 and the membership function learning unit 72 are not particularly limited, and for example, a RAM or a hard disk can be used. For rewriting the membership function in the membership function learning means 72, a CPU can be used, or it can be realized by software.

In addition, the structural example of the discrimination | determination means in the human posture movement discrimination | determination apparatus 1 which concerns on this invention is not limited to an above-described thing.
For example, according to the human posture / motion discriminating apparatus 1 according to the present invention including the first discriminating means, it is possible to discriminate with high accuracy whether the person is in the posture state or the motion state.
Further, according to the human posture / motion discriminating apparatus 1 according to the present invention including the first discriminating means and the second discriminating means, whether the person is in the standing position, the sitting position or the lying position, the operating state Can be determined with high accuracy.
Furthermore, according to the human posture / motion discriminating apparatus 1 according to the present invention including the first discriminating means and the third discriminating means, whether the person is in the posture state, the fall state, or the normal motion other than the fall It is possible to determine whether the state is in a state with high accuracy.
Alternatively, according to the posture / movement determination apparatus 1 of the person according to the present invention including the first determination means, the second determination means, and the third determination means, the person can be in a standing position, a sitting position, or a lying position. Therefore, it is possible to determine with high accuracy whether the vehicle is in the fall state, the fall state, or the normal operation state other than the fall state.

Also, according to the human posture / motion discriminating apparatus 1 according to the present invention including the first discriminating means and the fourth discriminating means, whether the person is in the posture state, the running state, or the normal operation other than the running It is possible to determine whether the state is in a state with high accuracy.
Furthermore, according to the human posture / movement discriminating apparatus 1 according to the present invention comprising the first discriminating means, the second discriminating means, and the fourth discriminating means, the person can be in a standing position, a sitting position, or a lying position. It is possible to determine with high accuracy whether the vehicle is in a state, a traveling state, or a normal operation state other than traveling.
Alternatively, according to the human posture / motion discriminating apparatus 1 according to the present invention comprising the first discriminating means, the third discriminating means, and the fourth discriminating means, whether the person is in the posture state or in the fall state Therefore, it is possible to determine with high accuracy whether the vehicle is in a traveling state or a normal operation state other than traveling.
And according to the human posture / movement discriminating apparatus 1 according to the present invention comprising the first discriminating means, the second discriminating means, the third discriminating means, and the fourth discriminating means, the person is standing, sitting, It is possible to determine with high accuracy whether the vehicle is in the prone position, whether it is in a fall state, whether it is in a traveling state, or in a normal operation state other than traveling.

Also, according to the human posture / motion discriminating apparatus 1 according to the present invention comprising the first discriminating means, the fourth discriminating means, and the fifth discriminating means, whether the person is in the posture state, running, walking, stairs It is possible to determine with high accuracy whether the state is ascending or descending stairs.
Furthermore, according to the human posture / movement discriminating apparatus 1 according to the present invention comprising the first discriminating means, the second discriminating means, the fourth discriminating means, and the fifth discriminating means, the person is standing and sitting. It is possible to discriminate with high accuracy whether the vehicle is in the recumbent position, or is in the state of running, walking, climbing the stairs, or descending the stairs.
Alternatively, according to the human posture / motion determination device 1 according to the present invention including the first determination unit, the third determination unit, the fourth determination unit, and the fifth determination unit, whether the person is in the posture state or not. Therefore, it is possible to determine with high accuracy whether the vehicle is in a fall state or is in a state of running, walking, stair climbing, or stair descent.

Similarly, according to the posture / movement determination apparatus 1 for a person according to the present invention including the first determination means, the fourth determination means, and the sixth determination means, whether the person is in a posture state, running, walking, It is possible to determine with high accuracy whether the stairs are rising or falling.
Furthermore, according to the human posture / movement discriminating apparatus 1 according to the present invention comprising the first discriminating means, the second discriminating means, the fourth discriminating means, and the sixth discriminating means, the person is standing and sitting. It is possible to discriminate with high accuracy whether the vehicle is in the recumbent position, or is in the state of running, walking, climbing the stairs, or descending the stairs.
Alternatively, according to the posture / motion determination apparatus 1 of a person according to the present invention including the first determination means, the third determination means, the fourth determination means, and the sixth determination means, whether the person is in the posture state. Therefore, it is possible to determine with high accuracy whether the vehicle is in a fall state or is in a state of running, walking, stair climbing, or stair descent.

  In addition, it is desirable to include wireless communication means 70 for transmitting the result of determination by the determination means to an external device installed outside the apparatus main body. Thereby, a weak person such as an elderly person can act freely with peace of mind such as going out alone. In case of emergency, it is also possible to carry out quick relief activities. Therefore, the care recipient is also given the joy of living, and the care-reserving reserve army who tends to stay behind can be provided with opportunities for social contribution, and the burden on the caregiver can be significantly reduced.

Next, the energy consumption calculation apparatus 101 according to the present invention will be described. The energy consumption calculation device according to the present invention determines the posture operation state consisting of standing, sitting, lying, running, walking, stair climbing, and stair descent as the first step, and the posture operation state determined as the second step. The energy consumption corresponding to is calculated.
FIG. 17 is an example of a flowchart showing a posture motion discrimination method that is the first stage, and FIG. 19 is a block diagram showing an example of the configuration of the energy consumption calculating apparatus 101 according to the present invention. In this example, the first discriminating means 140 discriminates between a posture state consisting of standing, sitting and lying and an operating state consisting of running, walking, climbing stairs, and descending stairs. When the posture state is determined, the process proceeds to the second determination unit 141, and when the posture state is determined, the process proceeds to the third determination unit 142.
The second discrimination means 141 discriminates whether the posture is standing, sitting or lying, and the third discrimination means 142 is in any state of running or other normal operation. When it is determined that the operating state is other than traveling, the process proceeds to the fourth determining unit 143 to determine which of the normal operating state is walking, stair climbing, or stair descent.

The first determination unit 140 of the energy consumption calculation apparatus 101 according to the present invention will be described. In the first discriminating means, as shown in FIG. 17, a posture state composed of a standing position, a sitting position, and a supine position based on an average value G _MXYZAVE of a combined acceleration component of Z-direction posture component G _SZ and acceleration in three directions. And whether it is in an operation state consisting of running, walking, stair climbing, or stair descent.
The posture component G _SZ in the Z direction is a value calculated by the above equation (1) from the acceleration G _Z in the Z direction detected by the acceleration detecting means. Further, the average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions is a value calculated by the equations (2) and (3) from the accelerations in the three directions detected by the acceleration detecting means.
In addition, in order to discriminate between the posture state consisting of standing, sitting, and lying down and the operation state consisting of running, walking, stair climbing, and stair descent, the Z direction posture component G _SZ and the three directions The reason for using the average value G _MXYZAVE of the motion component of the combined acceleration of acceleration and the principle thereof are the same as those of the first discriminating means 40 of the human posture / motion discriminating apparatus 1 according to the present invention.

The second determination unit 141 of the energy consumption calculation apparatus 101 according to the present invention will be described. In the second determination means, as shown in FIG. 17, it is determined from the X-direction posture component G _SX and the Z-direction posture component G _SZ whether the posture state is standing, sitting or lying.
The X-direction posture component G _SX is a value calculated from the X-direction acceleration G _X detected by the acceleration detection means according to the equation (4), and the Z-direction posture component G _SZ is detected by the acceleration detection means. wherein the direction of the acceleration G _Z (1) is a value calculated by formula.
The reason why the posture component G _SX in the X direction and the posture component G _SZ in the Z direction are used to determine whether the posture is in the standing position, the sitting position, or the supine position and the principle thereof are described in detail in the following. This is the same as the second discriminating means 41 of the posture / motion discriminating apparatus 1.

The 3rd discrimination means 142 of the energy consumption calculation apparatus 101 which concerns on this invention is demonstrated. In the third determining means, as shown in FIG. 17, it is determined from the average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions whether the vehicle is in the running state or the other normal operation state.
The average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions is a value calculated by the equations (2) and (3) from the accelerations in the three directions detected by the acceleration detecting means.
The reason why the average value G _MXYZAVE of the motion component of the combined acceleration of the three directions of acceleration is used to determine whether the vehicle is in the running state or the normal operation other than that, and the principle thereof are described in detail. This is the same as the fourth discriminating means 43 of the posture / motion discriminating apparatus 1.

The 4th discrimination | determination means 143 of the energy consumption calculation apparatus 101 which concerns on this invention is demonstrated. In the fourth discriminating means, as shown in FIG. 17, the X-direction posture component G _SX , the X-direction motion component average value G _MXAVE , the Z-direction motion component average value G _MZAVE and the peak power value S _Y From each of the calculation results Z1 and Z2 obtained by multiplying each by a predetermined coefficient and determining the sum, it is determined which of the normal operation states of walking, stair climbing and stair descent.
The X-direction posture component G _SX is a value calculated by the equation (4) from the X-direction acceleration G _X detected by the acceleration detecting means. In addition, the average value G _MXAVE of the motion component in the X direction is a value calculated by the equations (6) and (7) from the acceleration G _X in the X direction detected by the acceleration detecting means. The average value G _MZAVE of the motion component in the Z direction is a value calculated by the equations (8) and (9) from the acceleration G _Z in the Z direction detected by the acceleration detecting means. The peak power value S _Y is a peak power value of the frequency spectrum of the acceleration in the Y direction analyzed by the frequency analysis of the acceleration G _Y in the Y direction detected by the acceleration detecting means.
In addition, in order to determine which of the normal operation states of walking, stair climbing, and stair descent, the X direction posture component G _SX , the X direction motion component average value G _MXAVE , and the Z direction motion component average The reason for using the value G _MZAVE and the peak power value S _Y , and the discriminant functions Z1 and Z2 having these as elements and the principle thereof are the same as the fifth discriminating means 44 of the human posture / motion discriminating device 1 according to the present invention. The same. Therefore, it is desirable that the energy consumption calculation apparatus according to the present invention also includes the coefficient input means 150 for inputting the coefficients of the respective discriminating elements and the coefficient storage means 151 for storing them.

  Next, another embodiment of the energy consumption calculation apparatus 101 according to the present invention will be described. FIG. 27 is an example of a flowchart showing another determination method of the energy consumption calculation apparatus 101 according to the present invention, and FIG. 29 is a configuration diagram showing another example of the configuration of the energy consumption calculation apparatus 101 according to the present invention. is there. In this example, in place of the above-described fourth discriminating means 143, by providing fifth discriminating means 144 for performing fuzzy inference, it is possible to determine whether the person is in a walking state, stairs up, or stairs down. Determine.

The 5th discrimination means 144 of the energy consumption calculation apparatus 101 which concerns on this invention is demonstrated. In the fifth discriminating means, as shown in FIG. 27, the X-direction posture component G _SX , the average value G _MXAVE of the X-direction motion component, the power spectrum value Y _MAX and the power spectrum value Z _MAX are determined in advance as inputs. In addition, by performing fuzzy inference using the membership function of each operation state, it is determined whether the normal operation state of walking, stair climbing, or stair descent is present.
The X-direction posture component G _SX is a value calculated from the X-direction acceleration G _X detected by the acceleration detection means 10 according to the above equation (4).
The average value G _MXAVE of the motion component in the X direction is a value calculated from the acceleration G _X in the X direction detected by the acceleration detection means 110 according to the equations (6) and (7).
As shown in FIG. 21, the power spectrum value Y _MAX is obtained by performing frequency analysis of the acceleration in the three directions detected by the acceleration detecting means 110 and calculating the basic value in the Y direction calculated from the power spectrum in the three directions analyzed by the frequency analysis. It is the power spectrum value of the frequency.
As shown in FIG. 21, the power spectrum value Z _MAX is obtained by performing frequency analysis of the acceleration in the three directions detected by the acceleration detecting means 110 and calculating the basic value in the Z direction calculated from the power spectrum in the three directions analyzed by the frequency analysis. It is the power spectrum value of the frequency.

Note that in order to determine which of the normal operation state is walking, stair climbing, or stair descent, the posture component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, the power spectrum value Y _MAX and the power The reason why the spectrum value Z _MAX is used and the principle thereof are the same as those described in the sixth determination unit 45 of the human posture / movement determination apparatus 1 according to the present invention.
Further, regarding fuzzy reasoning in general, for example, the fuzzy rule to be used, the membership function to be used and the creation method thereof, the membership learning function, etc., the sixth discrimination means 45 of the human posture movement discrimination device 1 according to the present invention. This is the same as described in.

  The energy consumption calculation apparatus 101 according to the present invention stores a basal metabolic rate in standing, sitting and lying positions and a regression equation indicating the correlation between the walking rate and energy consumption in running, walking, stair climbing, and stair descent. A regression equation storage unit 180, a regression equation selection unit 181 for selecting a basal metabolic rate or a regression equation to be applied according to a discrimination result by the discrimination unit, and a posture motion state based on the selected basal metabolic rate or the regression equation The energy consumption calculating means 182 for calculating the energy consumption corresponding to the above will be provided, which will be described.

FIG. 18 is a distribution diagram showing the relationship between the minute oxygen intake V _O2 and the walking rate during walking, stair climbing, and stair descent. It can be seen that a close correlation is established between the minute oxygen intake _VO2 and the walking rate, and this relationship can be expressed by a regression equation. In addition, if the walking rate can be detected, it can be seen that the minute oxygen intake V _O2 can be calculated by the regression equation. Although the vertical axis of FIG. 18 is the partial oxygen intake amount at the time of V _O2, the oxygen concentration of the intake gas Ibis, determine the carbon dioxide concentration and gas volume determines the oxygen content of the oxygen content and in the breath during inspiration The minute oxygen uptake V _O2 and the carbon dioxide output are obtained, the respiratory exchange ratio is obtained, the calorie K generated per liter of oxygen is obtained, the minute oxygen uptake V _O2 (l / min) and the calorie K (kcal) ) To calculate the energy consumption Q (kcal / min).

  For detecting the walking rate R, frequency analysis (FFT analysis) of the detected X-direction acceleration and Z-direction acceleration is performed, and the peak frequency of the analyzed X-direction acceleration frequency spectrum and the Z-direction acceleration are analyzed. The walking rate can be detected by detecting the peak frequency of the frequency spectrum, comparing the two frequencies, and selecting the lower peak frequency. Table 1 shows the determination accuracy when the method is adopted. It can be seen that the walking rate at the time of walking, stair climbing, and stair descent can be detected with higher accuracy than Table 1.

  FIG. 19 is a block diagram showing an example of the configuration of the energy consumption calculation apparatus 101 according to the present invention. The acceleration data storage means 120 is not particularly limited as long as it can store the detected acceleration data, but it is preferable to use a RAM or a hard disk. Discrimination result storage means 160 for storing discrimination results and coefficient storage means 151, and further, basal metabolic rate in standing position, sitting position and prone position, and correlation between walking rate and energy consumption in running, walking, stair climbing, and stair descent The same applies to the regression equation storage means 180 that stores the regression equation indicating Further, a CPU can be used as the acceleration calculation means 130, and can also be realized by software.

  Therefore, according to the energy consumption calculation apparatus 101 according to the present invention, it is possible to determine the posture / operation state including standing, sitting, lying, running, walking, stair climbing, and stair descent using the determination unit. Can also detect the walking rate during each movement, so that the basal metabolic rate and running, walking, climbing stairs, walking rate in descending stairs stored in the regression equation storage means 180 Based on the regression equation indicating the correlation with the energy consumption amount, the energy consumption amount corresponding to each posture movement state can be calculated with high accuracy.

  The energy consumption calculation apparatus 101 according to the present invention includes a consumption storage unit 183 that sequentially stores the calculated energy consumption every predetermined time, and a consumption display unit 184 that displays the calculated energy consumption. Is desirable. Thereby, energy consumption can be confirmed in real time. The consumption display means 184 is not particularly limited, but it is desirable to use a display device using liquid crystal for portability.

  The energy consumption calculation apparatus 101 according to the present invention includes a consumption integration unit 185 that integrates the calculated energy consumption, an integrated consumption storage unit 186 that stores the integrated energy consumption, and an integrated energy consumption. It is desirable to provide an integrated consumption amount display means 187 for displaying. Thereby, the energy consumption integrated amount can be confirmed in real time. The consumption display means 187 is not particularly limited, but it is desirable to use a display device using liquid crystal for portability.

  In addition, the energy consumption calculation apparatus 101 according to the present invention is a regression equation that indicates the correlation between the basal metabolism or the walking rate in running, walking, rising stairs, and descending stairs and energy consumption in standing, sitting, and standing positions. It is desirable to include regression equation input means 179 for inputting. Since there is an individual difference in the basal metabolic rate or the energy consumption accompanying exercise, the energy consumption with higher accuracy can be calculated by providing the regression equation input means.

It is a schematic diagram for demonstrating the mounting position and axial direction of a 3-axis acceleration sensor. It is an example of the flowchart which shows the discrimination | determination method of the human posture movement discrimination | determination apparatus based on this invention. It is a schematic diagram for demonstrating the attitude | position component and motion component of an acceleration waveform. It is a schematic diagram for demonstrating the calculation processing method of acceleration data. FIG. 7 is a distribution diagram showing the relationship between each posture action and a posture component _{GSZ in} the Z direction. Is a distribution diagram showing the relationship between the average value G _MXYZAVE component of motion of the combined acceleration of accelerations in the orientation operation and three directions. FIG. 6 is a distribution diagram showing a relationship between a standing position, a sitting position, and a supine position and a posture component G _{SX in the} X direction and a posture component G _{SZ in} the Z direction. It is a distribution map which shows the relationship of each operation | movement and the synthetic acceleration _GXY of the acceleration of a X direction and a Y direction. It is a distribution map which shows the relationship between the average value _GMXAVE of the movement component of the X direction at the time of a walk, stairs rise, and stairs descent, and a walk rate. It is a distribution map which shows the relationship between the average value _GMZAVE of the movement component of the Z direction at the time of a walk, stairs raise, and stairs descent, and a walk rate. It is an example of the distribution map which shows the relationship between the attitude _| position component _GSX of the X direction at the time of a walk, stairs raise, and stairs descent, and a walk rate. It is another example of the distribution map which shows the relationship between the attitude _| position component _GSX of the X direction at the time of a walk, stairs rise, and stairs descent, and a walk rate. Walking, stair rise, which is an example of a distribution diagram showing the relationship between the peak power value S _Y and walking rate during stair descent. Walking, stair rise, which is another example of a distribution diagram showing the relationship between the peak power value S _Y and walking rate during stair descent. It is the distribution figure which plotted discriminant function Z1 and Z2. It is a block diagram which shows an example of a structure of the person's attitude | position discrimination | determination apparatus based on this invention. It is an example of the flowchart which shows the attitude | position determination method of the energy consumption calculation apparatus which concerns on this invention. It is a distribution map which shows the relationship between minute oxygen intake _VO2 and a walking rate. It is a block diagram which shows an example of a structure of the energy consumption calculating device which concerns on this invention. It is an example of the flowchart which shows another discrimination | determination method of the human posture movement discrimination | determination apparatus which concerns on this invention. It is a schematic diagram for explaining a power spectrum value _{Y MAX} and _{Z MAX.} Walking, stair rise is a distribution diagram showing the relationship between the power spectral value Y _MAX and walking rate during stair descent. Walking, stair rise is a distribution diagram showing the relationship between the power spectral value Z _MAX and the walking rate during stair descent. It is a schematic diagram for showing an example of a membership function. It is a graph which shows the determination precision of the 6th discrimination | determination means 45 using a fuzzy reasoning. It is a schematic diagram for _{demonstrating} normalization of the power spectrum value YMAX. It is an example of the flowchart which shows another discrimination method of the energy consumption calculation apparatus which concerns on this invention. It is a block diagram which shows another example of a structure of the person's attitude | position determination apparatus based on this invention. It is a block diagram which shows another example of a structure of the energy consumption calculation apparatus which concerns on this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Posture movement discrimination device 10 Acceleration detection means 11 3-axis acceleration sensor
12 A / D converter 20 the acceleration data storage unit 30 acceleration calculating means 31 G _SX calculating means 32 G _MXAVE calculating unit 33 frequency analyzer 34 S _Y detection means 35 G _SZ calculating means 36 .DELTA.G _SZ calculating means 37 G _MZAVE calculating means 38 G _XY calculation means 39 G _MXYZAVE calculation means 40 First determination means 41 Second determination means 42 Third determination means 43 Fourth determination means 44 Fifth determination means 45 Sixth determination means 50 Coefficient input means 51 Coefficient Storage means 60 Discrimination result storage means 70 Wireless communication means 71 Operation state input means 72 Membership function learning means 73 Membership function storage means 74 Y _MAX calculation means 75 Z _MAX calculation means 76 Walking rate calculation means

DESCRIPTION OF SYMBOLS 101 Energy consumption calculation apparatus 110 Acceleration detection means 111 3-axis acceleration sensor
112 A / D converter 120 the acceleration data storage unit 130 acceleration calculating means 131 G _SX calculating means 132 G _MXAVE calculating unit 133 frequency analysis means 134 S _Y detector 135 G _SZ calculating means 137 G _MZAVE calculating means 139 G _MXYZAVE calculation means 140 first 1 discriminating means 141 second discriminating means 142 third discriminating means 143 fourth discriminating means 144 fifth discriminating means 150 coefficient input means 151 coefficient storage means 160 discrimination result storage means 170 wireless communication means 171 operation state input means 172 membership function learning unit 173 membership function storage means 174 Y _MAX operating means 175 Z _MAX operating means 176 walking rate calculating means 179 regression equation input unit 180 regression equation storage unit 181 regression equation selection means 182 et Energy consumption calculating means 183 consumption storage unit 184 consumption display means 185 consumption amount integrating means 186 integrated consumption storage means 187 integrated consumption display means

Claims (15)

In a human posture / motion discriminating apparatus that discriminates a posture / motion state using acceleration detecting means mounted on the human body so that the X direction is front and back, the Y direction is right and left, and the Z direction is vertical,
Acceleration detecting means for detecting an acceleration G _X in the X direction orthogonal to each other, an acceleration G _{Y in the} Y direction, and an acceleration G _{Z in the} Z direction;
Means for calculating a posture component G _SZ in the Z direction from the detected acceleration G _{Z in the} Z direction;
_Means for calculating an average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions from the detected accelerations in the three directions;
From the calculated Z-direction posture component G _SZ and the average value G _MXYZAVE of the combined acceleration motion component from the posture state consisting of standing, sitting, and lying down and running, walking, stair climbing, stair descent, falling A human posture / motion discriminating device comprising first discriminating means for discriminating which of the following motion states is present.
It means for calculating the X direction orientation component G _SX from the acceleration G _X of the detected X-direction,
Claim, characterized in that it comprises calculating the X direction orientation component G _SX and Z direction standing from orientation component G _SZ, locus, a second determination means for determining whether the one of the orientation states decubitus The human posture / movement discriminating apparatus according to 1.
Means for calculating a change amount ΔG _SZ of the posture component in the Z direction from the detected acceleration G _{Z in the} Z direction;
Means for calculating a combined acceleration G _XY of the acceleration in the X direction and the Y direction from the detected acceleration G _X in the X direction and the acceleration G _{Y in the} Y direction;
Based on the combined acceleration G _{XY of the} calculated acceleration in the X and Y directions and the amount of change ΔG _{SZ in} the posture component in the Z direction, either normal operation consisting of running, walking, stair climbing, stair descent, or dangerous motion consisting of falling The apparatus according to claim 1, further comprising a third discriminating unit that discriminates whether or not the person is present.
4. A fourth discriminating means for discriminating whether the vehicle is in a running state or other normal operation state from an average value G _MXYZAVE of the motion component of the combined acceleration of the calculated accelerations in the three directions. The human posture motion discrimination device according to any one of 1 to 3.
It means for calculating an average value _{G MXAVE} the X direction motion component from the acceleration _{G X} of the detected X-direction,
A frequency analysis means for analyzing the frequency spectrum of the acceleration G _Y of the detected Y-direction,
Means for detecting a peak power value S _Y of the frequency spectrum of the acceleration of the analyzed Y direction by the frequency analysis means,
_Means for calculating an average value G _MZAVE of motion components in the Z direction from the detected acceleration G _{Z in the} Z direction;
Computed X-direction orientation component _{G SX,} the average value _{G MXAVE} the X direction motion component is multiplied by a predetermined coefficient determined in advance in each of the average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component sum 5. The human posture / movement discriminating apparatus according to claim 4, further comprising: a fifth discriminating unit that discriminates whether a normal operation state of walking, stair climbing, or stair descent is obtained based on a calculation result obtained from the above.
It means for calculating an average value _{G MXAVE} the X direction motion component from the acceleration _{G X} of the detected X-direction,
A frequency analysis means for analyzing a frequency spectrum of the detected acceleration in three directions;
Means for calculating power spectrum values Y _MAX and Z _MAX of fundamental frequencies in the Y direction and the Z direction from power spectra in three directions analyzed by the frequency analyzing means;
_Performing fuzzy inference using membership functions of predetermined motion states with the calculated X-direction attitude component G _SX , X-direction motion component average value G _MXAVE , power spectrum values Y _MAX and Z _MAX as inputs 5. The human posture / motion discriminating apparatus according to claim 4, further comprising sixth discriminating means for discriminating whether the normal operation state is walking, stair climbing, or stair descent.
Furthermore, a means for calculating the walking rate R from the power spectrum in the three directions analyzed by the frequency analyzing means,
Whether the sixth determination means is in a normal operation state of walking, stair climbing, or stair descent by performing fuzzy inference using a membership function of each motion state corresponding to the walking rate R. 7. The human posture / motion discriminating device according to claim 6, wherein the human posture / motion discriminating device is discriminated.
Furthermore, an operation state input means for inputting whether the person wearing the acceleration detection means is in a normal operation state of walking, stair climbing, or stair descent;
Based on the input true motion state, the input X-direction posture component G _SX , the average value G _MXAVE of the X-direction motion component, the power spectrum values Y _MAX and Z _MAX , 8. The human posture / movement discriminating apparatus according to claim 6 or 7, further comprising membership function learning means for rewriting the membership function.
9. The human posture / movement discriminating apparatus according to claim 1, further comprising a wireless communication unit that transmits a discrimination result of the discriminating unit to an external device installed outside the apparatus main body.
Posture action state consisting of standing, sitting, standing, running, walking, stair climbing, and stair descent using acceleration detection means mounted on the human body so that the X direction is front and back, the Y direction is left and right, and the Z direction is vertical In the energy consumption calculation device that calculates the energy consumption according to the posture movement state,
A regression equation storage means for storing a regression equation indicating a correlation between a basal metabolism amount in a standing position, a sitting position and a standing position and a walking rate in running, walking, stair climbing, and stair descent and energy consumption;
Acceleration detecting means for detecting an acceleration G _X in the X direction orthogonal to each other, an acceleration G _{Y in the} Y direction, and an acceleration G _{Z in the} Z direction;
_Means for calculating an average value G _MXAVE of the X-direction posture component G _SX and the X-direction motion component from the detected X-direction acceleration G _X ;
_Means for calculating an _attitude value G _{SZ in the} Z direction and an average value G _MZAVE of the motion component in the Z direction from the detected acceleration G _{Z in the} Z direction;
Frequency analysis means for analyzing the detected frequency spectrum of the acceleration in the three directions,
Means for detecting a peak power value S _Y of the frequency spectrum of the acceleration of the analyzed Y direction by the frequency analysis means,
Peak frequency detection means for detecting the peak frequency of the frequency spectrum of the acceleration in the X direction analyzed by the frequency analysis means and the peak frequency of the frequency spectrum of the acceleration in the Z direction;
Means for detecting the walking rate R from the peak frequency of the detected frequency spectrum of the acceleration in the X direction and the peak frequency of the frequency spectrum of the acceleration in the Z direction;
_Means for calculating an average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions from the detected accelerations in the three directions;
Computed Z direction orientation component G _SZ and three directions standing from the average value G _MXYZAVE component of motion of the combined acceleration of accelerations, sitting, running and posture state consisting of decubitus, walking, stair rise consists stairs descending movement First determining means for determining which state of the state is present;
A second discriminating unit for discriminating from the calculated X-direction posture component G _SX and Z-direction posture component G _SZ whether the posture state is a standing position, a sitting position, or a supine position;
Third discriminating means for discriminating from the average value G _MXYZAVE of the motion component of the combined acceleration of the calculated three-direction acceleration whether the vehicle is in a normal operation state of running or other normal operation;
Computed X-direction orientation component _{G SX,} the average value _{G MXAVE} the X direction motion component is multiplied by a predetermined coefficient determined in advance in each of the average value _{G MZAVE} and peak power value _{S Y} and Z direction motion component sum A fourth discrimination means for discriminating whether it is in a normal operation state of walking, stair climbing or stair descent from the calculation result obtained by taking
A regression formula selection means for selecting a basal metabolic rate or regression formula to be applied according to the discrimination result by the discrimination means;
An energy consumption calculation device comprising: energy consumption calculation means for calculating an energy consumption according to the posture movement state based on the selected basal metabolic rate or regression equation.
Posture action state consisting of standing, sitting, standing, running, walking, stair climbing, and stair descent using acceleration detection means mounted on the human body so that the X direction is front and back, the Y direction is left and right, and the Z direction is vertical In the energy consumption calculation device that calculates the energy consumption according to the posture movement state,
A regression equation storage means for storing a regression equation indicating a correlation between a basal metabolism amount in a standing position, a sitting position and a standing position and a walking rate in running, walking, stair climbing, and stair descent and energy consumption;
Acceleration detecting means for detecting an acceleration G _X in the X direction orthogonal to each other, an acceleration G _{Y in the} Y direction, and an acceleration G _{Z in the} Z direction;
_Means for calculating an average value G _MXAVE of the X-direction posture component G _SX and the X-direction motion component from the detected X-direction acceleration G _X ;
Means for calculating a posture component G _SZ in the Z direction from the detected acceleration G _{Z in the} Z direction;
Frequency analysis means for analyzing the frequency spectrum of the detected acceleration in three directions,
Means for calculating power spectrum values Y _MAX and Z _MAX of fundamental frequencies in the Y direction and the Z direction from power spectra in three directions analyzed by the frequency analyzing means;
Means for calculating the walking rate R from the power spectrum in three directions analyzed by the frequency analysis means;
_Means for calculating an average value G _MXYZAVE of the motion component of the combined acceleration of the accelerations in the three directions from the detected accelerations in the three directions;
Computed Z direction orientation component G _SZ and three directions standing from the average value G _MXYZAVE component of motion of the combined acceleration of accelerations, sitting, running and posture state consisting of decubitus, walking, stair rise consists stairs descending movement A first determination means for determining which of the states,
A second discriminating unit for discriminating from the calculated X-direction posture component G _SX and Z-direction posture component G _SZ whether the posture state is a standing position, a sitting position, or a supine position;
Third discriminating means for discriminating from the average value G _MXYZAVE of the motion component of the combined acceleration of the calculated accelerations in the three directions whether the vehicle is in a running state or other normal operation state;
Using the calculated posture component G _{SX in} the X direction, the average value G _MXAVE of the motion component in the X direction, and the power spectrum values Y _MAX and Z _MAX as input, membership functions of predetermined motion states corresponding to the walking rate R are used. A fifth discriminating means for discriminating whether it is in a normal operation state of walking, stair climbing or stair descent by performing fuzzy reasoning;
A regression formula selection means for selecting a basal metabolic rate or regression formula to be applied according to the discrimination result by the discrimination means;
An energy consumption calculation device comprising: energy consumption calculation means for calculating an energy consumption according to the posture movement state based on the selected basal metabolic rate or regression equation.
Furthermore, an operation state input means for inputting whether the person wearing the acceleration detection means is in a normal operation state of walking, stair climbing, or stair descent;
Based on the input true motion state, the input X-direction posture component G _SX , the average value G _MXAVE of the X-direction motion component, the power spectrum values Y _MAX and Z _MAX , 12. The energy consumption calculation device according to claim 11, further comprising membership function learning means for rewriting the membership function.
Consumption storage means for sequentially storing the calculated energy consumption at regular intervals;
The energy consumption calculation apparatus according to any one of claims 10 to 12, further comprising display means for displaying the calculated energy consumption.
Means for integrating the calculated energy consumption;
An integrated consumption storage means for storing the integrated energy consumption;
The energy consumption calculation device according to claim 10, further comprising display means for displaying the accumulated energy consumption.
It is provided with a regression equation input means for inputting a regression equation indicating a correlation between a walking rate and energy consumption in running, walking, stair climbing, and stair descent in standing, sitting, and standing positions. The energy consumption calculation device according to any one of claims 10 to 14.