JP7474466B2 - Motion discrimination device and motion discrimination program - Google Patents

Description

The present invention relates to a motion discrimination device and a motion discrimination program, and in particular to a motion discrimination device and a motion discrimination program suitable for controlling an assist device that assists human motion.

In recent years, an assist device with variable viscoelastic properties using pneumatic artificial muscles has been proposed, as shown in Patent Document 1. It is known that humans perform movements while changing the viscoelasticity of their muscles. The assist device shown in Patent Document 1 is configured to operate based on the variable viscoelastic properties that are the principle of human joint drive, and therefore has a high affinity with humans. In addition, by making the viscosity and elasticity variable in generating the assist force, the torque required for backdriving can be reduced, and a variety of movements can be assisted with high backdrivability.

JP 2019-126668 A

On the other hand, in order to operate the assist device, it is necessary to determine the movement of the wearer and output an assist force according to the movement. Known methods for determining the movement of the wearer include, for example, a method using the wearer's myoelectric potential, a method using an acceleration sensor, a method using ZMP (zero moment point), and the like.
Since myoelectric potential is an electrical signal, devices that use myoelectric potential can detect the wearer's intended movement in real time. However, there are problems with the large amount of noise and the time required to attach the electromyograph to the wearer. Devices that use acceleration sensors require only a small number of sensors, and the noise output from the sensors is relatively small. However, they do not have the real-time capability of the aforementioned electromyograph, and are therefore unsuitable for assistive devices.
In the case of ZMP-based systems, the ZMP of the wearer is measured using floor reaction force sensors attached to the soles of the feet. The ZMP (zero moment point) is the point where the combined force of gravity and inertia intersects with the road surface in the trajectory generation and control methods for bipedal robots. However, this requires the installation of additional sensors on the soles of the feet, which makes the system more complicated.
Therefore, in order to solve the above problems, the present invention aims to provide a motion discrimination device and a motion discrimination program that have less noise when discriminating the motion of an assist device, reduce the effort required for wearing, and have excellent real-time performance.

As a configuration of a motion discrimination device for solving the above problems, there is provided an assist device including a first orthosis attached to one part of a person connected via a joint, a second orthosis attached to the other part, a connector that rotatably connects the first orthosis and the second orthosis in response to the motion of the person's joint, an artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis, and a feature detection means that detects the effect of the person's motion on the first orthosis and the second orthosis as a feature, or an assist device including a first orthosis attached to one part of a person connected via a joint, a second orthosis attached to the other part, and an electric a connector in which the viscosity of a functional fluid changes based on an electrical signal, making the rotational resistance variable, and rotatably connecting a first orthosis and a second orthosis in response to the movement of the joints of the person; and a feature detection means for detecting the effect of the movement of the person on the first orthosis and the second orthosis as a feature. Alternatively, the assist device includes a first orthosis worn on one part of a person connected via a joint, and a second orthosis worn on the other part of the person, and the viscosity of a functional fluid changes based on an electrical signal, making the rotational resistance variable, and rotatably connecting the first orthosis and the second orthosis in response to the movement of the joints of the person. A motion discrimination device for discriminating motion of a person wearing an assist device or the like, the assist device or the like including an artificial muscle that generates a relative rotational force between a connecting body, a first orthosis, and a feature detection means that detects the effect of the motion of the person on the first orthosis and the second orthosis as a feature, the motion discrimination device including a comparison discrimination means that compares a time change in the feature detected by the feature detection means with a threshold value for discriminating motions that is set in advance, and discriminates the motion of the person, the comparison discrimination means including a balance state discrimination means that discriminates between motions that do not involve a movement of the body's center of gravity and motions that involve a movement of the body's center of gravity, and a balance state discrimination means that discriminates between motions that do not involve a movement of the body's center of gravity and motions that involve a movement of the body's center of gravity. The balance state determination means determines whether the legs are moving alternately based on the movement determined by the balance state determination means, the leg alignment state determination means determines whether the legs are aligned based on the movement determined by the balance state determination means, and the sitting/standing position determination means determines whether the position is sitting or standing based on the movement determined by the leg alignment state determination means.
According to this configuration, there is less noise in distinguishing movements, and human movements can be distinguished more easily without the hassle of wearing the device or impairing real-time performance.
In addition, the motion discrimination device may be configured so that the feature detection means detects a mutual reaction force generated between the first or second equipment and a person when a person wearing the first or second equipment attempts to move .
As an embodiment of the motion discrimination program for solving the above problem, there is provided an assist device including a first orthosis attached to one part of a person connected via a joint, a second orthosis attached to the other part, a connector that rotatably connects the first orthosis and the second orthosis in response to the motion of the person's joint, artificial muscles that generate a relative rotational force between the first orthosis and the second orthosis, and a feature detection means that detects the effect of the person's motion on the first orthosis and the second orthosis as a feature, or a motion discrimination program including the first orthosis attached to one part of a person connected via a joint, and a second orthosis attached to the other part. an assist device including a second orthosis to be worn, a connector in which the viscosity of a functional fluid changes based on an electrical signal to make the rotational resistance variable and which rotatably connects the first orthosis and the second orthosis in response to the motion of the joint of the person, and a feature amount detection means for detecting the effect of the motion of the person on the first orthosis and the second orthosis as a feature amount; or a first orthosis to be worn on one part of the person connected via a joint, and a second orthosis to be worn on the other part of the person, and a connector in which the viscosity of a functional fluid changes based on an electrical signal to make the rotational resistance variable and which rotatably connects the first orthosis and the second orthosis in response to the motion of the joint of the person. a connecting body that rotatably connects the first orthosis and the second orthosis, an artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis, and a feature detection means that detects the effect of the person's movement on the first orthosis and the second orthosis as a feature, the motion discrimination program discriminating between a motion of a person wearing an assist device or the like, the assist device or the like comprising: a connecting body that rotatably connects the first orthosis and the second orthosis; an artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis; and a feature detection means that detects the effect of the person's movement on the first orthosis and the second orthosis as a feature. The motion discrimination program causes a computer to execute a first step of comparing a time change in the feature detected by the feature detection means with a threshold value that is set in advance for discriminating between motions, and a first step of discriminating between motions that do not involve movement of the body's center of gravity and motions that involve movement of the body's center of gravity, a second step of discriminating whether the legs are moving alternately, a third step of discriminating whether the legs are aligned, and a fourth step of discriminating between a sitting position and a standing position, the second step of discriminating whether the legs are moving alternately based on the discrimination result of the first step, the third step of discriminating whether the legs are aligned based on the discrimination result of the first step, and the fourth step of discriminating between a sitting position and a standing position based on the discrimination result of the third step .
According to this aspect, there is less noise in distinguishing movements, and human movements can be distinguished more easily without the hassle of wearing the device or compromising real-time performance.
In addition, the feature may be a mutual reaction force generated between the first and second equipment and the person when a person wearing the first and second equipment attempts to move .

1 is a schematic configuration diagram of an assist device according to an embodiment of the present invention; FIG. 13 is a diagram showing the configuration of a connector. FIG. 1 is a diagram showing the configuration of an artificial muscle. FIG. 13 is a diagram showing a connection state of frames. 13 is a graph showing changes in the angles of the left and right hip and knee joints actually obtained by motion capture. FIG. 2 is a flowchart showing a process performed by the movement discrimination device. This is a table summarizing the body types of the four subjects. This is the result of the discrimination rate calculated based on the discrimination signal results obtained by simulation. This is the result of the discrimination signal output by the algorithm. This is the algorithm used when a subject was fitted with the assist device and a motion discrimination experiment was conducted. 11 is a table summarizing parameters set when the motion discrimination experiment shown in FIG. 10 is performed. FIG. 11 is a diagram summarizing the results and success rates of discrimination signals obtained from walking motion experiments. FIG. 13 shows experimental results and success rates of rearing up. FIG. 13 is a diagram showing the definition of the direction of the mutual reaction force generated between the assist orthosis and the wearer. FIG. 13 is a diagram showing a motion discrimination algorithm for discriminating a motion of a wearer using a mutual reaction force generated between the prosthesis and the wearer.

The present invention will be described in detail below through the embodiments of the invention, but the following embodiments do not limit the invention as claimed, and not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention, but include configurations that are selectively adopted.

Figure 1 is a configuration diagram of one embodiment of an assist device suitable for application of the motion discrimination method according to the present embodiment. The assist device shown in Figure 1 is a so-called exoskeleton type assist robot that can assist human motions such as standing up from a sitting position, standing up from a standing position to a sitting position, and walking, and generally comprises a motion assist unit 4 that generates an assist force for human motions, and a control unit 8 that controls the motion of the motion assist unit 4.

An exoskeleton is one that is attached to the human body and does not use the human skeleton as a structural element in the assisting motion. When the connections between the bones in the human body are considered to be a single link mechanism, the device independently moves in a similar manner to the bones, performing bending and stretching movements at the joints.

The motion assistant units 4 are attached along the left and right outer sides of the body to form the exoskeleton described above. Each motion assistant unit 4 includes a frame 2, an artificial muscle 40, a wire 46, a pulley 44, and an angle sensor 48.

The frame 2 includes a waist frame 24 fixed to the waist, a thigh frame 20 fixed to the thigh, and a lower leg frame 22 fixed to the lower leg along the left and right outer sides of the body. The waist frame 24 and the thigh frame 20, and the thigh frame 20 and the lower leg frame 22 are connected to each other rotatably via connectors 42.
That is, the waist frame 24 and the thigh frame 20 correspond to a first orthosis attached to one part of a person and a second orthosis attached to the other part, and the thigh frame 20 and the lower leg frame 22 correspond to a first orthosis attached to one part of a person and a second orthosis attached to the other part. The waist frame 24, the thigh frame 20, and the lower leg frame 22 are each composed of a rigid body.

The length of the thigh frame 20 is set so that it extends from the base of the thigh along the outer side of the body to the knee joint, and connectors 42 are attached to the end on the waist side and the end on the foot side, connecting the thigh frame 24 and the lower leg frame 22.

Connector 42: 42 is attached to the thigh frame 20 so that it corresponds exactly to the joint on the hip side of the femur and the joint on the distal end of the leg. In other words, the hip frame 24 and thigh frame 20 are rotatable by the connector 42 when the hip is bent, and the thigh frame 20 and crus frame 22 are rotatable by the connector 42 when the knee is bent.

Using the thigh frame 20 as a reference, the waist frame 24 extends from the waist joint along the outer side of the body toward the rear in the front-rear direction of the human body. The lower leg frame 22 extends from the knee joint along the outer side of the lower leg. A sole plate 22A is provided on the distal end of the lower leg frame 22, extending toward the sole of the foot.
The waist frame 24, thigh frame 20, and lower leg frame 22 are fixed by wrapping the fixing belt 3 around the person with the sole of the foot stepping on the sole plate 22A provided on the lower leg frame 22. The fixing belt 3 is preferably, for example, a non-elastic band-like one, and more preferably, a hook-and-loop fastener that functions as a fixing means is provided at the end in the extension direction of the fixing belt 3, and a cushioning material (not shown) is provided on the surface that is wrapped around the person to provide a cushion against the human body.

FIG. 2 is an external perspective view and a cross-sectional view of the connector 42. As shown in FIG.
For example, an MR brake formed in a generally cylindrical shape that allows rotational movement is applied to the connecting body 42. As shown in Fig. 2, the MR brake includes a core 64 that is rotatably accommodated in an accommodating section 62 provided in the center of a case 60 formed in a cylindrical shape with a bottom. The accommodating section 62 is formed as a cylindrical recess recessed from one side surface. A plurality of flat, annular disks 63 having through holes through which the cores 64 pass are arranged in the accommodating section 62 so as to be stacked in the axial direction. Each disk 63 is attached to a cylindrical surface 62a so as not to rotate.
The core 64 includes a shaft portion 64A that is inserted into the housing portion 62, and a disk portion 64B that is provided so as to close the opening of the housing portion 62. The tip side of the shaft portion 64A of the core 64 is rotatably attached to the case 60. A coil 66 is provided on the outer periphery of the shaft portion 64A. Furthermore, a plurality of flat, annular disks 68 are non-rotatably attached to the outer periphery of the coil 66. The disks 68 are alternately arranged with the disks 63 provided on the housing portion 62 side.

A magnetic fluid 69 is sealed in the space of the housing portion 62 of the case 60 housing the core 64. In the MR brake, when no voltage is applied to the coil 66, the core 64 rotates freely relative to the case 60, and when a voltage is applied to the coil 66, the magnetic field generated by the coil 66 affects the magnetic fluid, causing resistance to the relative rotation between the disk 63 on the case 60 side and the disk 68 on the core 64 side, thereby braking the free rotation of the core 64, and when the maximum allowable voltage is applied to the coil 66, the magnetic fluid solidifies and the case 60 and the core 64 rotate integrally.
The MR brake is a device that uses a magnetorheological fluid and is configured to be able to change the coefficient of friction with a response speed of about 10 ms by applying a magnetic field to the magnetorheological fluid sealed inside. In other words, as described above, the MR brake has the characteristics of a high response speed of several tens of milliseconds, high output, and backdrivability.
Each connecting body 42 is connected to a motion control device 84 described below, and the rotational resistance of the core 64 relative to the case 60 is variably operated based on an adjusted voltage (signal) input from the motion control device 84 .

FIG. 4 is a diagram showing the connected state of the frames 24 , 20 , and 22 connected by the connector 42 .
The connecting body 42 that operates in this manner, for example, as shown in FIG. 4, constitutes a link mechanism in which the waist frame 24 and the thigh frame 20 are flexed and extended in accordance with the bending and extending of the waist, by fixing the case 60 and the core 64 to the waist frame 24 and the thigh frame 20, and in which the case 60 is fixed to the thigh frame 20 and the core 64 is fixed to the lower leg frame 22, by which the thigh frame 20 and the lower leg frame 22 are flexed and extended in accordance with the bending and extending of the knee.

The relationship in which the case 60 and the core 64 in the connector 42 are attached to the waist frame 24, the thigh frame 20, and the lower leg frame 22 is not limited.
Furthermore, the connecting body 42 is not limited to an MR brake. As long as the connecting body 42 operates as described above, for example, an ER brake or the like in which a functional fluid whose viscosity changes based on an electrical signal is sealed may be used.

The waist frame 24 is provided with a drive mechanism 5 for rotating the thigh frame 20 relative to the waist frame 24, and the thigh frame 20 is provided with a drive mechanism for rotating the thigh frame 20 relative to the waist frame 24.

A pulley 44 is fixed to the lower leg frame 22, which constitutes the drive mechanism 5 and applies an assist force to the thigh frame 20 and the lower leg frame 22. The pulley 44 is attached to the thigh frame 20 and the lower leg frame 22 so that its center is positioned coaxially with the rotation axis of the connecting body 42.

As shown in FIG. 1, the drive mechanism 5 includes an artificial muscle 40, a spring 41, and a wire 46. The drive mechanism 5 is attached to a drive mechanism attachment part 25 provided on the upper body side of the waist frame 24, and a drive mechanism attachment part 21 provided on the upper body side of the thigh frame 20.

Figure 3 is a diagram showing the configuration of artificial muscle 40. As shown in Figure 3(a), artificial muscle 40 is formed in a cylindrical shape when viewed from the outside, and is an actuator that can expand and contract in the axial direction by supplying and discharging a fluid.

For example, an axial fiber-reinforced artificial muscle configured as shown in FIG. 3 can be used as such an artificial muscle 40. An axial fiber-reinforced artificial muscle is configured by inserting a carbon fiber sheet 52 into a cylinder 50 made of natural latex rubber, for example. The carbon fiber sheet 52 has multiple carbon fibers 51 oriented along the axial direction of the cylinder 50, and restrains the axial extension of the cylinder 50. Both ends of the cylinder 50 are closed by terminal members 53A and 53B, and an air chamber 56 is formed on the inner periphery of the cylinder 50. One of the terminal members 53A has an air circulation port 55 that communicates with the air chamber 56, and is configured to be connectable to a tube extending from an air supply and exhaust device 80 described later. In addition, multiple rings 54 are provided at equal intervals in the axial direction on the outer periphery of the cylinder 50.

3(b) by supplying air to the air flow port 55, the artificial muscle 40 expands in the radial direction to have multiple humps defined by the ring 54, and contracts in the axial direction due to the binding force of the carbon fiber sheet 52, generating a traction force. Also, by exhausting the air from the air chamber 56, the artificial muscle 40 operates to return (extend) to its natural length, as shown in FIG. 3(a).
The artificial muscle 40 is not limited to the axial fiber reinforced artificial muscle described above, and may be a McKibben type.

The spring 41 is a so-called coil spring, and one end is attached to the drive mechanism attachment portion 21; 25 so as to move in parallel with the artificial muscle 40. The other end of the spring 41 is connected via a pulley 44 to an end of a wire 46 that is attached to the other end of the artificial muscle 40.
The wire 46 is hung without slack around the pulley 44, and connects the artificial muscle 40 and the spring 41, which are arranged in an antagonistic relationship. In this way, the spring 41 applies resistance to the traction force caused by the contraction of the artificial muscle 40. Note that instead of the spring 41, an artificial muscle similar to the artificial muscle 40 may be arranged in an antagonistic relationship.

The wire 46 may be made of metal or synthetic fibers made of twisted organic fibers. It is preferable that the wire 46 is made of a material that has little stretch when tension is applied. Also, when durability is taken into consideration, metal or non-metallic inorganic fibers are preferable, and from the standpoint of weight, synthetic fibers made of twisted organic fibers are preferable.

Therefore, when the artificial muscle 40 is contracted, the wire 46 wound around the pulley 44 pulls the spring 41, causing the pulley 44 to rotate passively. The drive mechanism 5 attached to the waist frame 24 is attached to the waist frame 24 so that the spring 41 is on the front side and the artificial muscle 40 is on the rear side, as shown in Figure 1, and the drive mechanism 5 attached to the thigh frame 20 is attached to the thigh frame 20 so that the spring 41 is on the front side and the artificial muscle 40 is on the rear side.
As the pulley 44 rotates, the thigh frame 20 rotates together with the pulley 44 relative to the waist frame 24, assisting the bending and stretching of the waist, and the lower leg frame 22 rotates together with the pulley 44 relative to the thigh frame 20, assisting the bending and stretching of the waist. In other words, the contraction of the artificial muscle 40 imparts a rotational force of the thigh frame 20 relative to the waist frame 24, and a rotational force of the lower leg frame 22 relative to the thigh frame 20.
Furthermore, the wire 46 and the pulley 46 may be replaced with a timing belt and a timing pulley, respectively. By using a timing belt, the rotational force of the thigh frame 20 relative to the waist frame 24 and the rotational force of the lower leg frame 22 relative to the thigh frame 20 when the artificial muscle 40 contracts can be transmitted efficiently.

By using an artificial muscle 40 with such a configuration, the actuator can be made lighter and the assist force can be changed according to the standing up motion. The artificial muscle 40 has a characteristic that the output characteristic (traction force) for the contraction stroke is large at the beginning of the contraction and gradually decreases toward the end of the contraction. This characteristic corresponds exactly to the human characteristic that a large force is required at the beginning of the standing up motion and the required force gradually decreases as the standing up motion progresses, and is preferable as a power source (driving force generating means) for obtaining an assist force. In addition, the traction force can be changed by adjusting the pressure of the fluid supplied inside the artificial muscle 40, and since it is made of an elastic material and has flexibility, it can obtain a driving force close to that of human muscle, and its light weight can reduce the burden on the person when wearing it. In the following explanation, the operating state of the artificial muscle 40 is shown by axial extension and contraction.

As shown in FIG. 1, the control unit 8 includes angle sensors 48A; 48B that detect the rotation angle, an air supply and exhaust device 80 for supplying compressed air to the artificial muscle 40 and exhausting compressed air from the artificial muscle 40, a motion control device 84 that controls the operation of the air supply and exhaust device 80 (supply of compressed air to the artificial muscle 40 and exhaust of compressed air from the artificial muscle 40) and the operation of the connecting body 42, and a motion discrimination device 86 that discriminates the motion of a person.

The angle sensors 48A, 48B are feature detection means according to this embodiment, and are provided in the movement assisting unit 4 so as to be able to detect the rotation angle of the thigh frame 20 relative to the waist frame 24, and the rotation angle of the lower leg frame 22 relative to the thigh frame 20. The angle sensors 48A, 48B are constituted by, for example, an encoder, and output the detected rotation angles to the movement control device 84 and the movement discrimination device 86.
The arrangement of each drive mechanism 5 is not limited to the above-mentioned form, and it is sufficient if it can rotate the pulley 44 provided corresponding to the joint. Instead of the waist frame 24 and thigh frame 20, it may be attached to, for example, the waist frame 24 and the lower leg frame 22, or the thigh frame 20 and the lower leg frame 22, and can be modified as appropriate.

The air supply and exhaust device 80 controls the pressure of the fluid in the artificial muscle 40, for example, based on a signal output from the operation control device 84. Specifically, it controls the supply and exhaust of air in the air chamber 56 of each artificial muscle 40 so that the pressure in the air chamber 56 of each artificial muscle 40 becomes the pressure set in the operation control device 84. For example, a pressure sensor for measuring the pressure of the artificial muscle 40, which constitutes part of the function of the air supply and exhaust device 80, and an air pressure valve for controlling the supply and exhaust of air to the artificial muscle 40 are provided in each drive mechanism 5.

The operation control device 84 is a so-called computer, and is equipped with a processing means such as a CPU provided as a hardware resource, storage means such as a ROM and a RAM, and an input interface to which input means such as an operation panel or switches that are manually operated by the user are connected.

The motion control device 84 includes an artificial muscle control unit and an MR brake control unit, and the storage means stores programs and target value setting data for controlling the drive of the connecting body 42 consisting of the artificial muscle 40 and the MR brake. The calculation processing means executes each process according to the program stored in the storage means, causing the motion control device 84 to function as an artificial muscle control unit, an MR brake control unit, etc.

The command value setting data is data for setting command values for operating the artificial muscles 40 and the connecting body 42 in the assist operation, and is set for each operation to be assisted. The command value setting data described below was created for the purpose of assisting the bending and extending operation of the knee joint. The target value setting data includes target stiffness data (also simply referred to as stiffness data) and target viscosity data (also simply referred to as viscosity data).

It is known from knowledge about human viscoelasticity control in human joint movements that viscoelasticity is proportional to the load torque of the movement. In particular, it is known that when the movement begins and ends (starting stage), the activity level of the antagonist muscles is lowered to reduce viscoelasticity, and when the movement stops (stopping stage), the viscoelasticity is increased by simultaneous activity of the antagonist and activating muscles. Therefore, the target stiffness value and target viscosity value in the assist operation are set to increase and decrease according to the increase or decrease in the torque of the joint in the human movement. For example, they are set to be proportional to the torque τ and to increase only when the movement stops. The stop of the movement can be determined when the angular velocity and angular acceleration of the joint have opposite signs in the movement of the human joint.

The operation control device 84 outputs a signal for controlling the pressure in the artificial muscle 40 and a voltage value to be output to the connecting body 42 based on the angle measured by the angle sensor 48 and the above command value data.

According to the drive mechanism 5 configured as described above, by supplying air to the artificial muscle 40, the wire 46 moves while pulling the spring 41, rotating the pulley 44. The rotation of the pulley 44 causes the thigh frame 20 to rotate relative to the waist frame 24, and the lower leg frame 22 to rotate relative to the thigh frame 20. By adjusting the amount of air or air pressure supplied to the artificial muscle 40, it is possible to control the rotation angle θ of the thigh frame 20 relative to the waist frame 24 and the rotation angle θ of the lower leg frame 22 relative to the thigh frame 20.
Furthermore, by applying a voltage to the connector 42 when the pulley 44 rotates, it is possible to generate resistance to the rotation of the pulley 44. In other words, it is an actuator that generates a viscous force to drive the artificial muscle 40, and functions as a viscous force applying means.

The motion discrimination device 86 is a so-called computer, and includes a processing unit such as a CPU provided as a hardware resource, a storage unit such as a ROM or RAM, and an input interface to which input units such as an operation panel or a switch operated manually by a user are connected, and functions as motion discrimination unit for discriminating human motion. The storage unit stores programs for causing the motion discrimination device 86 to function as each unit described below, and also stores thresholds for comparison with feature quantities for discriminating human motion by executing the programs.

The movement discrimination device 86 discriminates the movement of the wearer based on the angle sensor 48 provided in the movement assistance unit 4. In this embodiment, the angle sensors 48 will be described by distinguishing between the angle sensor 48A that detects the angle of the hip joint and the angle sensor 48B that detects the angle of the knee joint. In the following description, the angle detected by the angle sensor 48A provided corresponding to the hip joint may be referred to as the waist joint angle.
The hip joint angle detected by angle sensor 48A is indicated as _θH , the knee joint angle detected by angle sensor 48B is indicated as _θK , and their angular velocities are indicated as _θ'H and _θ'K . Furthermore, to distinguish between the left and right angle sensors 48A and 48B, the angle sensor provided on the right side of the body is marked with an R, and the angle sensor provided on the left side is marked with an L. In other words, the hip joint angles are indicated as _θHR and _θHL , the knee joint angles are indicated as _θKR and _θKL , and their angular velocities are marked with "'".
For ease of explanation, the movements that are discriminated by the movement discrimination device 86 are five: walking, standing up/standing down, standing with legs apart (hereinafter referred to as standing with legs apart), sitting, and standing.

The balance state discrimination means 86A discriminates between statically balanced motion (hereinafter, SBM), which is a characteristic of human movement, and dynamically balanced motion (hereinafter, DBM). SBM refers to motion that does not involve movement of the body's center of gravity, such as sitting, standing, and standing with arms outstretched. DBM refers to motion that involves movement of the body's center of gravity, such as walking, standing, and running.

The balance state determination means 86A distinguishes between SBM and DBM by comparing the angular velocity _θ'H , which is the time change in the hip joint angle _θH detected by the angle sensors _48A1L and _48A1R provided corresponding to the base of the leg of the action-assistance unit 4, with a threshold value _θ'Hthre .
The specific processing in the balance state discrimination means 86A involves calculating an angular velocity _θ'H indicating the movement state of the hip joint from changes in the angle _θH successively input to the movement discrimination device 86 from the angle sensors _48AL and _48AR , reading out from the storage means a threshold value _θ'Hthre set for discriminating between SBM and DBM and comparing it with the threshold value, and, for example, determining that _θ'H >_θ'Hthre is Yes as DBM, and determining that _θ'H ≦ _θ'Hthre is No as SBM.

The leg movement determining means 86B determines whether the legs are moving alternately.
When a person is walking, the left and right legs move alternately in the front-rear direction of the body. Therefore, the leg movement determination means 86B determines whether the legs are moving alternately based on the movement of the left and right thighs, which is a characteristic of walking.
Specifically, the leg action discrimination means 86B calculates angular velocities _θ'HL and _θ'HR , which are the time changes of angles θHL and _θHR of the left and right hip joints detected by the angle sensors _48AL and _48AR at the same time, and calculates the product of the calculated angular velocities _θ'HL and θ'HR. It then _judges whether the calculated numerical value is positive or negative, and determines whether the value is negative (angular _{velocities θ'HL} > 0 and _θ'HR < 0, or _θ'HL < 0 and _θ'HR > 0) as walking, and whether the value is positive (angular velocities _θ'HL > 0 and _θ'HR > 0, or _θ'HL < 0 and _θ'HR < 0) as both legs are moving together, and determines whether the action is standing up or falling down.
That is, it is determined whether the detected hip joint angular velocity _θ′HL >0 and _θ′HR <0, or _θ′HL <0 and _θ′HR >0 is satisfied, and if so, the answer is Yes and the motion is determined as walking. If not, the answer is No and the motion is determined as standing up/falling down.

The balance state determination means 86A and leg movement determination means 86B described above can distinguish between walking and standing up/falling down. When the balance state determination means 86A determines SBM, the leg alignment determination means 86C determines whether the legs are aligned, in other words, whether the legs are not aligned, and thus determines whether the person is standing with the legs apart, sitting, or standing. The leg alignment determination means 86C determines whether the legs are aligned based on the movement of the lower legs of the left and right legs, which is a characteristic of the SBM state.
As a specific process in the leg alignment state discrimination means 86C, the angle difference δθ _diff =|θ _KR -θ _KL | between the left and right knee joints detected by the angle sensors 48B _L and 48B _R is calculated, a threshold value δθ _diffthre set for discriminating the leg spread state is read from the storage means and compared, and when δθ _diff is greater than the threshold value δθ _diffthre , it is determined that the feet are not aligned as Yes, and when δθ _diff is equal to or less than the threshold value δθ _diffthre , it is determined that the feet are aligned as No. That is, the detected angle difference between the left and right knee joints is δθ _diff =|θ _KR -θ _KL |, and when δθ _diff is greater than the threshold value, it is determined that the legs are in a standing position with the legs widely spread apart, and when δθ _diff is equal to or less than the threshold value, it is determined that the legs are in a standing position or a sitting position.

The sitting/standing discrimination means 86D discriminates between sitting and standing. Figures 5(a) and 5(b) are graphs showing changes in the angles of the left and right hip joints and knee joints actually obtained by motion capture. The values of θ _HR , θ _HL , θ _KR , and θ _KL when sitting are characteristically larger than those when standing and walking. Therefore, threshold values θ _Hthre and θ _Kthre for discriminating between sitting and standing are set for the hip joint angles θ _HR and θ _HL and the knee joint angles θ _KR and θ _KL , and the sitting and standing positions can be discriminated by comparing the values.
The specific processing in the sitting/standing position discrimination means 86D involves comparing both the left and right hip joint angles _θHL , _θHR detected by angle sensors 48A _-L and 48A- _R with a threshold value _θHthre , and comparing both the left and right knee joint angles _θKL , _θKR detected by angle sensors 48B _-L and 48B- _R with a threshold value _θKthre , and discriminating the position as sitting if the left and right hip joint angles _θHL , _θHR are _θHL > _θHthre , _θHR > _θHthre and the left and right knee joint angles _θKL , _θKR are _θKL > _θKthre , _θKR > _θKthre , and discriminating the position as standing otherwise.
In the following description, the signal representing the operation determined by each means will be referred to as a determination signal.

6 is a diagram showing a process (flowchart) of the action determination device 86 according to this embodiment. The action determination device 86 executes the following process based on the angles input from the angle sensors 48A _L ; 48B _L and 48A _R ; 48B _R and the change in the angles over time.
S101: Determine whether the legs are moving, i.e., whether the patient is in SBM or DBM.
Specifically, the hip joint angular velocity _θ'H is compared with a threshold _θ'Hthre for the hip joint angular velocity _θ'H , and if the hip joint angular velocity _θ'H is greater than the threshold _θ'Hthre ( _θH > _θHthre ) (Yes), it is determined that the leg is moving, i.e., the user is in DBM, and the process proceeds to S102. If the hip joint angular velocity _θ'H is equal to or less than the threshold _θ'Hthre ( _θH ≦ _θHthre ) (No), it is determined that the leg is not moving, i.e., the user is in SBM, and the process proceeds to S103.

S102: It is determined whether the legs are moving alternately.
Specifically, the direction of the angular velocity _θ'HL of the left hip joint is compared with the direction of the angular velocity _θ'HR of the right hip joint to check whether the angular velocity _θ'HL of the left hip joint and the angular velocity _θ'HR of the right hip joint are in opposite directions ( _θ'HL >0 and _θ'HR <0, or _θ'HL <0 and _θ'HR >0).If this condition is met (Yes), it is determined that the person is walking.If this condition is not met (No), it is determined that the person is standing up/falling down because both legs are moving together.

S103: Determine whether the legs are not together, that is, the operating state in the SBM.
Specifically, the difference between the angles of the left and right hip joints δθ _diff =|θ _KR -θ _KL | is calculated, and if δθ _diff is greater than a threshold value δθ _diffthre for determining whether the legs are not together (Yes), the legs are determined to be widely apart and the person is determined to be standing with the legs apart. On the other hand, if δθ _diff is equal to or less than the threshold value δθ _diffthre (No), the legs are determined to be together and the process proceeds to S104.

S104: The sitting position and the standing position are determined based on the angles of the hip joint and the knee joint.
Specifically, the hip joint angles θ _HR and θ _HL are compared with a threshold value θ _Hthre set as a threshold value for determining angles in a sitting position, and the knee joint angles θ _KR and θ _KL are compared with a threshold value θ _Kthre set as a threshold value for determining angles in a sitting position, and if the hip joint angles θ _HR and θ _HL and the knee joint angles θ _KR and θ _KL are greater than the respective threshold values θ _Hthre and angle threshold value θ _Kthre (Yes), the position is determined as sitting. If the hip joint angles θ _HR and θ _HL and the knee joint angles θ _KR and θ _KL are equal to or smaller than the respective threshold values θ _Hthre and threshold value θ _Kthre (No), the position is determined as standing.
The result of the discrimination is then output to the operation control device 84 as a discrimination signal.

Based on the above-mentioned motion discrimination algorithm, human movements obtained through motion capture were applied, and simulations were conducted to verify whether the movements could be discriminated.
To verify the simulation, a preliminary experiment was conducted to measure the data necessary as input for the motion discrimination simulation using motion capture.
The necessary data are the hip joint angles θ _HR and θ _HL and the knee joint angles θ _KR and θ _KL during human motion.

In addition, in the motion capture measurements, the following three movements were performed to investigate the effects of transitions in human movements:
The first movement is walking (movement 1).
The second movement involves maintaining a sitting position for 2 seconds, standing up for 2 seconds, maintaining a standing position for 2 seconds, and finally walking (movement 2).
The third movement involves maintaining a standing position for two seconds, standing up over one second, and then maintaining a sitting position for two seconds (movement 3).
These movements were performed seven times by each of the four subjects, and data with relatively little noise was used for the simulation. Figure 7 shows a table summarizing the body types of the four subjects.

For movement 1, the movement was recorded using a motion capture camera (frame rate 100 fps). Force plates were placed on the floor to measure the reaction force and moment acting on the soles of the subjects' feet. The measurement data was recorded on a measurement PC via an A/D converter.

For movements 2 and 3, a chair was prepared so that the subject's knee angle was 90 degrees, and a force plate was placed on top of it. The motion capture camera was positioned in the same way as for movement 1. 26 markers were attached to skeletal landmarks on each body segment of the subject, and measurements were taken with the motion capture camera.

Based on the data measured in the above environment, a musculoskeletal simulation was performed using nMotion (Nac Image Technology Co., Ltd.), and the joint angle data obtained from this were used in the simulation. The simulation was performed using Matlab/Simulink (Mathworks.Inc.), and the data of the subject's θ _HR , θ _HL , θ _KR , and θ _KL in three movements measured in a previous experiment were used as input. The sampling time was 10 ms.

The discrimination rate by the simulation is defined as in the following formula (4.1).
t _judge / t _motion [%] Equation (4.1)
Here, t _judge [s] is the time during which a discrimination signal corresponding to a specific motion is output, and t _motion [s] is the time during which the motion to be measured is actually being performed.
In addition, the threshold values of the joint angles and angular velocities used in this simulation were the minimum values of the joint angles of each subject during each movement.

Figure 8 shows the discrimination rate calculated based on the discrimination signal results obtained by simulation. Figure 9 shows the discrimination signal results output by the algorithm. Note that the results shown in Figure 9 show only the results of subject A as a representative example, since the trends in the simulation results were similar for all subjects.

As shown in Figure 8, in motion 1, the average discrimination rate for all subjects was 81.4%. Also, according to Figure 9(a), the discrimination signal became "standing position with legs apart" at regular intervals. This is because the angular velocity fell below the threshold. Before and after the stance phase changes to the swing phase, the legs need to push off the ground, and at that time the angular velocity of the joints decreased, falling below the threshold.

In motion 2, the SBMs for sitting and standing had a higher discrimination rate than the DBMs for standing and walking. Also, according to FIG. 9(b), the conditions were met for sitting and standing, and appropriate discrimination signals were output. On the other hand, the "sitting" position was discriminated in the middle of standing up. This is thought to be due to the following reasons. When standing up, the hip joint is bent in order to shift the center of gravity forward. After that, the knee joint is extended to stand up, and at this time, the angular velocity of the hip joint becomes smaller than that at the initial center of gravity shift, and is thought to be below the threshold value of the angular velocity. Note that the discrimination rate was not 100% in the SBM, which is thought to be due to the misalignment of the markers during measurement. After the transition to walking, the same tendency as motion 1 was shown, and similarly, discrimination other than walking was made at the changeover between the stance phase and the swing phase.

9C, the results for the fall of motion 3 were similar to those for the rise of motion 2. Furthermore, the discrimination rate for SBM was higher than that for DBM.
From the above results, it was confirmed that the above-mentioned motion discrimination method (algorithm) can be used to discriminate between SBM and DBM motions, and its usefulness was also confirmed.

Next, the algorithm was verified based on the hip and knee joint angles input from an actual assist device.
The assist device used was shown in Figure 1. It was made based on variable viscoelastic properties. In order to eliminate complex elements, it does not have any viscous elements and operates based on variable elastic properties.

Fig. 10 and Fig. 11 summarize the experimental conditions and experimental method using the assisting orthosis of the algorithm according to this embodiment. Specifically, Fig. 10 shows an algorithm when a motion discrimination experiment was conducted with the assisting orthosis shown in Fig. 1 worn by a subject. Fig. 11 is a table summarizing the parameters set when conducting the motion discrimination experiment shown in Fig. 10.
[Experimental conditions]
The subject performed each of the following movements five times according to the operator's instructions.
The first movement is walking movement starting from a sitting position. The subject stands up from the sitting position at the instruction of the operator, and starts walking at the next instruction of the operator. When starting to walk, the subject is instructed to swing up the right foot. The walking range was set to 5 m.
The second action is to stand up from the sitting position at the command of the next operator and to stand up again at the command of the next operator and to sit down again.
At the operator's instruction, the subject stood up from the seated position and assumed a standing position.
In the two motion experiments, the subjects practiced the motion for 15 minutes before starting the experiment. The operator gave instructions at his/her discretion.
The timing of movements was recorded by having the subject press a switch when the movement began.
Each drive mechanism 5 is provided with a pressure sensor for measuring the pressure of the artificial muscle 40 and an air pressure valve for controlling the supply and exhaust of air to the artificial muscle 40. Signals from the pressure sensor and angle sensors 48A, 48B are recorded via an A/D converter, and the operation of the action assistant unit 4 is controlled based on these signals. The air pressure valve is operated via a D/A converter.

The success rate is then assessed.
The subject begins the movement at the specified timing. The algorithm determines whether the movement is the same as the subject's, and if the prosthesis performs that movement, it is judged as a "success." If the movement is different or if the prosthesis does not perform the movement, it is judged as a "failure." For each movement, the success rate was evaluated as (number of successful attempts)/(total number of attempts).
The thresholds for the joint angles and angular velocities were set in advance through preliminary experiments.

The algorithm used in this experiment was almost the same as that shown in Fig. 6. The term "almost the same" means that, while in the previous explanation, a distinction was made between rising and falling edges, in this experiment, the rising and falling edges were not distinguished.
In the algorithm used in this experiment, instead of detecting rising and falling edges, a distinction was made between rising and falling edges based on the results of discrimination between previous and following movements.
Specifically, when the algorithm determines that an action is "standing up/falling," if the action immediately prior to that is determined as "sitting," the action is determined as "standing up," and if the action immediately prior to that is determined as "standing," the action is determined as "falling."

Furthermore, in the motion verification experiment, chattering occurred in the motion discrimination results when the motion transitioned. Therefore, the algorithm was modified so that the motion would continue for a while immediately after the motion transition. The modified algorithm is shown in Figure 10.
The duration of each action is denoted as t _sit [s], t _standing [s], t _sitting [s], t _stance [s], t _{stance_wio} [s], and t _gait [s].
These times were set in an experiment to confirm the operation. Figure 11 shows the parameter values used in this algorithm (program).

Experimental Results The results of two experiments, walking and standing up, are explained below.
12(a), (b), and (c) show the results of the discrimination signals in this experiment. The table in Fig. 12(d) shows the success rate.
The signal results in this experiment can be broadly divided into three patterns.
Pattern 1: walking, Pattern 2: falling signal while walking, Pattern 3: standing signal while walking.

In pattern 1 (Fig. 12(a)), when the subject began to walk, he periodically repeated "walking" and "standing with legs apart" as shown in Fig. 9(a). This gave the same results as the simulation.

In pattern 2 (Figure 12(b)), when the subject began to walk, the discrimination signal became "falling." This is thought to be because when the subject started to move his legs, he pulled his hips back, causing the brace to lean forward relatively. This resulted in the angular velocities of both legs of the brace having the same sign. In other words, it was determined that the legs were not moving alternately, resulting in a "falling" signal.

In pattern 3 (Figure 12(c)), the discrimination signal remained in "standing position" even when the subject began to walk. This is because the subject raised his right leg, but the orthosis was not tightly attached to the subject. As a result, the orthosis was unable to follow the subject's movement even though he was moving. As a result, the joint angular velocity fell below the threshold, and SBM was discriminated.

13(a) and (b) show the experimental results of standing up, and Fig. 13(c) shows the success rate. As with the walking experiment, the results are divided into patterns. Pattern 1: successful pattern, Pattern 2: falling in the middle of standing up.
For pattern 1 (FIG. 13(a)), an algorithm is in operation that distinguishes between a "rising edge" and a "falling edge" based on the immediately preceding operation in FIG.
In pattern 2 (Fig. 13(b)), the signal changes from "standing up" to "falling" in the middle of "standing up". This is thought to be due to the following reasons. When the subject tries to stand up, the discrimination signal becomes "standing up". At this time, the "standing up" program is executed, but the subject has not yet reached a standing position even after this program is finished. Therefore, the signal becomes "falling" instead of "standing up".

As explained above, a simulation was performed using human joint angles obtained by motion capture based on the above-mentioned movement discrimination algorithm, and verification was further performed on an actual device based on the simulation results, and the results showed that the method is effective in discriminating human movements based on angles.

In the above embodiment, the angle output from the angle sensor is used as a feature, and the movement of the person wearing the assist device is determined based on the angle and the change in the angle, but this is not limited to this.
For example, a movement can be determined based on the mutual reaction force between the assist device and the wearer.

FIG. 14 shows the definition of the direction of the mutual reaction force generated between the assist orthosis and the wearer.
FIG. 15 shows a motion discrimination algorithm for discriminating the motion of the wearer using the mutual reaction force generated between the prosthesis and the wearer.
The motion discrimination method according to this embodiment utilizes the reaction force generated in the prosthesis against the force exerted by the wearer trying to move. For example, if the prosthesis is assisting a "sitting" posture, the wearer will pull the prosthesis when the wearer tries to stand up. This feature can be utilized to develop the motion discrimination method.

In this movement discrimination method, as in the above-mentioned algorithm based on changes in joint angles, five movements are assumed: "standing", "sitting", "standing up", "standing down", and "walking".
As shown in FIG. 14, the force output by the force sensor provided on the waist is Fw, and the forces output by the force sensor provided on the crotch are Fh_R and Fh_L with the subscripts Right and Left.

This motion discrimination method starts with discrimination from waist movement. When Fw exceeds a certain threshold, it can be determined that the wearer is attempting to bend his/her waist, and therefore motions can be distinguished between standing and walking and other motions. The threshold value at this time can be obtained, for example, by investigating in advance the reaction force that occurs when the wearer starts to move and the prosthesis starts to follow that movement.

First, we will explain how to distinguish between standing and walking. If Fw does not exceed a certain threshold and the reaction forces of the right and left feet alternate between positive and negative, it is determined that the person is walking. If this is not the case, there is no reaction force in any part, so it can be determined that the person is standing.

Next, the sitting, standing, and falling positions are discriminated.
If the force output by the crotch sensor caused by the weight of part of the upper body and the crotch when sitting is set as Fh_sit_thre, then when standing up, the force acts in the direction that pulls the sensor, so Fh<Fh_sit_thre. When standing down, the force acts in the direction that compresses the sensor, so Fh>Fh_sit_thre. It can be considered that when sitting, Fh=Fh_sit_thre.

Figure 15 is a flowchart showing the motion discrimination algorithm based on the mutual reaction forces generated between the prosthesis and the wearer described above.

As described above, the information detected by the sensor provided in the action-assistance unit 4, which moves dynamically in response to the person's movements, can be compared with other pieces of information based on the characteristics of the person's movements, or with a threshold value that has been set in advance for the information, thereby making it possible to distinguish the person's movements.
Then, by outputting a discrimination value (signal) corresponding to the motion discriminated by the motion discrimination device 86 to a motion control device 84 that controls the orthosis, the artificial muscles 40 and the connecting bodies 42 can be controlled so that the orthosis performs an assist motion or a backdrive motion. In other words, the motion of the wearer can be discriminated by comparing information from all the sensors attached to the motion assistance unit 4 with each other or with a threshold value.

The motion discrimination device 86 may be configured as part of the motion control device 84, or vice versa.

In the above embodiment, the motion discrimination device 86 according to this embodiment is described using an assist device that assists the motion of a person's lower limbs, but the part is not limited to the lower limbs, and can also be applied to discriminating the motion of parts of the upper limbs such as the arms and hands.

In addition, in the above embodiment, the assist device 1 has been described as exerting an assist force on a person through the operation of the artificial muscles 40 and the function of the connecting body 42, but the assist device to which the above-mentioned motion discrimination device 86 is applied is not limited to one that exerts an assist force on a person through the operation of the above-mentioned artificial muscles 40 and the function of the connecting body 42, and may be one that exerts an assist force on a person only through the artificial muscles 40, or one that exerts an assist force through the function of the connecting body 42.

1 Assist device, 2 Frame, 3 Fixation belt,
4 Operation assistant unit, 5 Drive mechanism, 8 Control unit,
20 thigh frame, 22 lower leg frame, 24 waist frame, 40 artificial muscle,
41 spring, 42 connector, 44 pulley, 46 wire, 48 angle sensor,
50 Cylinder, 51 Carbon fiber, 53A Terminal member, 56 Air chamber,
60 case, 64 core, 69 magnetic fluid,
80 Air supply/exhaust device, 84 Operation control device, 86 Operation determination device,
86A balance state determination means, 86B leg action determination means,
86C leg alignment state determination means, 86D sitting/standing position determination means.

Claims (10)

A first orthosis attached to one part of a person connected via a joint;
A second device attached to the other part; and
a connector that rotatably connects the first orthosis and the second orthosis in response to a movement of a human joint;
An artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis;
a feature amount detection means for detecting an effect of a person's movement on the first orthosis and the second orthosis as a feature amount;
A motion discrimination device for discriminating a motion of a person wearing an assist device comprising:
a comparison and discrimination means for comparing a time change of the feature amount detected by the feature amount detection means with a preset threshold value for discriminating a motion, and discriminating a motion of a person ;
The comparison and determination means
a balance state determination means for determining whether a movement does not involve a movement of the center of gravity of the body or whether a movement involves a movement of the center of gravity of the body;
a leg movement determination means for determining whether the legs are moving alternately;
a leg alignment determining means for determining whether the legs are aligned;
A sitting/standing position discrimination means for discriminating between a sitting position and a standing position,
the leg movement determining means determines whether the legs are moving alternately based on the movement determined by the balance state determining means,
the leg alignment state determination means determines whether the legs are aligned based on the movement determined by the balance state determination means;
The motion discrimination device is characterized in that the sitting/standing discrimination means discriminates between a sitting position and a standing position based on the motion discriminated by the leg alignment state discrimination means, thereby discriminating the motion of the person . A first orthosis attached to one part of a person connected via a joint;
A second device attached to the other part; and
A connector in which the viscosity of a functional fluid is changed based on an electrical signal, and the rotational resistance is variable, and the connector rotatably connects the first orthosis and the second orthosis in response to the movement of a human joint ;
a feature amount detection means for detecting an effect of a person 's movement on the first orthosis and the second orthosis as a feature amount;
A motion discrimination device for discriminating a motion of a person wearing an assist device comprising:
a comparison and discrimination means for comparing a time change of the feature amount detected by the feature amount detection means with a preset threshold value for discriminating a motion, and discriminating a motion of a person ;
The comparison and determination means
a balance state determination means for determining whether a movement does not involve a movement of the center of gravity of the body or whether a movement involves a movement of the center of gravity of the body;
a leg movement determination means for determining whether the legs are moving alternately;
a leg alignment determining means for determining whether the legs are aligned;
A sitting/standing position discrimination means for discriminating between a sitting position and a standing position,
the leg movement determining means determines whether the legs are moving alternately based on the movement determined by the balance state determining means,
the leg alignment state determination means determines whether the legs are aligned based on the movement determined by the balance state determination means;
The motion discrimination device is characterized in that the sitting/standing discrimination means discriminates between a sitting position and a standing position based on the motion discriminated by the leg alignment state discrimination means, thereby discriminating the motion of the person . A first orthosis attached to one part of a person connected via a joint;
A second device attached to the other part; and
A connector in which the viscosity of a functional fluid is changed based on an electrical signal, and the rotational resistance is variable, and the connector rotatably connects the first orthosis and the second orthosis in response to the movement of a human joint;
An artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis;
a feature amount detection means for detecting an effect of a person's movement on the first orthosis and the second orthosis as a feature amount;
A motion discrimination device for discriminating a motion of a person wearing an assist device comprising:
a comparison and discrimination means for comparing a time change of the feature amount detected by the feature amount detection means with a preset threshold value for discriminating a motion, and discriminating a motion of a person ;
The comparison and determination means
a balance state determination means for determining whether a movement does not involve a movement of the center of gravity of the body or whether a movement involves a movement of the center of gravity of the body;
a leg movement determination means for determining whether the legs are moving alternately;
a leg alignment determining means for determining whether the legs are aligned;
A sitting/standing position discrimination means for discriminating between a sitting position and a standing position,
the leg movement determining means determines whether the legs are moving alternately based on the movement determined by the balance state determining means,
the leg alignment state determination means determines whether the legs are aligned based on the movement determined by the balance state determination means,
The motion discrimination device is characterized in that the sitting/standing discrimination means discriminates between a sitting position and a standing position based on the motion discriminated by the leg alignment state discrimination means, thereby discriminating the motion of the person . The motion discrimination device according to any one of claims 1 to 3, characterized in that the feature detection means detects the rotation angle of the connecting body. The motion discrimination device described in any one of claims 1 to 3, characterized in that the feature detection means detects a mutual reaction force generated between the first or second equipment and a person when a person wearing the first or second equipment attempts to move . A first orthosis attached to one part of a person connected via a joint;
A second device attached to the other part; and
A connector in which the viscosity of a functional fluid is changed based on an electrical signal, and the rotational resistance is variable, and the connector rotatably connects the first orthosis and the second orthosis in response to the movement of a human joint;
a feature amount detection means for detecting an effect of a person's movement on the first orthosis and the second orthosis as a feature amount;
A motion discrimination program for discriminating a motion of a person wearing an assist device comprising:
comparing the time change of the feature amount detected by the feature amount detection means with a preset threshold value for determining an action;
A first step of distinguishing between a motion that does not involve a movement of the center of gravity of the body and a motion that involves a movement of the center of gravity of the body;
a second step of determining whether the legs are alternating;
a third step of determining whether the legs are aligned;
and a fourth step of distinguishing between a sitting position and a standing position .
the second step determining whether the legs are moving alternately based on the determination result of the first step;
the third step is to determine whether the legs are aligned based on the result of the first step;
The fourth step is to distinguish between a sitting position and a standing position based on the discrimination result of the third step . A first orthosis attached to one part of a person connected via a joint;
A second device attached to the other part; and
a connector that rotatably connects the first orthosis and the second orthosis in response to a movement of a human joint;
An artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis;
a feature amount detection means for detecting an effect of a person's movement on the first orthosis and the second orthosis as a feature amount;
A motion discrimination program for discriminating a motion of a person wearing an assist device comprising:
comparing the time change of the feature amount detected by the feature amount detection means with a preset threshold value for determining an action;
A first step of distinguishing between a motion that does not involve a movement of the center of gravity of the body and a motion that involves a movement of the center of gravity of the body;
a second step of determining whether the legs are alternating;
a third step of determining whether the legs are aligned;
and a fourth step of distinguishing between a sitting position and a standing position .
the second step determining whether the user is moving the legs alternately based on the determination result of the first step;
the third step is to determine whether the legs are aligned based on the determination result of the first step;
The fourth step is to distinguish between a sitting position and a standing position based on the discrimination result of the third step . A first orthosis attached to one part of a person connected via a joint;
A second device attached to the other part; and
A connector in which the viscosity of a functional fluid is changed based on an electrical signal, and the rotational resistance is variable, and the connector rotatably connects the first orthosis and the second orthosis in response to the movement of the joint of the person;
An artificial muscle that generates a relative rotational force between the first orthosis and the second orthosis;
a feature amount detection means for detecting an effect of a person's movement on the first orthosis and the second orthosis as a feature amount;
A motion discrimination program for discriminating a motion of a person wearing an assist device comprising:
comparing the time change of the feature amount detected by the feature amount detection means with a preset threshold value for determining an action;
A first step of distinguishing between a motion that does not involve a movement of the center of gravity of the body and a motion that involves a movement of the center of gravity of the body;
a second step of determining whether the legs are alternating;
a third step of determining whether the legs are aligned;
and a fourth step of distinguishing between a sitting position and a standing position .
the second step determining whether the user is moving the legs alternately based on the determination result of the first step;
the third step is to determine whether the legs are aligned based on the determination result of the first step;
The fourth step is to distinguish between a sitting position and a standing position based on the discrimination result of the third step . The motion discrimination program according to any one of claims 6 to 8, characterized in that the feature is the rotation angle of the connecting body. An action discrimination program as described in any one of claims 6 to 8, characterized in that the feature is a mutual reaction force generated between the first and second equipment and a person when a person wearing the first and second equipment attempts to move .

Images