JP5467268B2 - Involuntary movement suppression system, movement recognition apparatus, and movement recognition apparatus program - Google Patents

Description

  The present invention relates to an involuntary movement suppression system, a motion recognition device, and a program for a motion recognition device, and more specifically, suppresses involuntary motion that is mixed when a tremor patient performs voluntary exercise, and voluntarily supports a tremor patient. The present invention relates to an involuntary movement suppression system, a movement recognition apparatus, and a program for the movement recognition apparatus that enable movement support that allows only movement.

  There is a neurological disorder called tremor in which the limbs vibrate against their will. As one type of tremor, there is an essential tremor in which a limb or the like periodically shakes when performing a certain motion or trying to maintain a certain posture. A patient with essential tremor causes problems such as difficulty in writing, difficulty in drinking a beverage in a container, and difficulty in eating, because hands holding various instruments necessary for daily life shake. In order to suppress such tremors, there are drugs such as β-blockers and deep brain stimulation (DBS). However, the effects on the body such as side effects are problematic.

  By the way, Patent Literature 1 discloses an upper limb motion assisting device that assists the manual motion of the body. This upper limb movement assist device has a free end provided with a brace to be worn on the user's upper limb, performs trajectory control of the free end based on the user's force information applied to the brace, It assists the movement of the upper limbs. Here, in the device, when the force information is changing at a low frequency cycle of about several Hz so as not to detect a force vector that is not related to the intention of the person generated by the shaking of the user's hand, The trajectory control of the free end is not performed.

JP-A-11-253504

  However, the device of Patent Document 1 cannot assist the movement of the upper limb of the essential tremor patient. That is, as described above, when an essential tremor patient tries to perform a motion intended by himself / herself, that is, voluntary movement, a tremor of the body unrelated to his / her own intention, that is, involuntary movement occurs. For this reason, in the apparatus of Patent Document 1, when the periodic change in force due to involuntary movement is detected, the motion control of the free end is stopped, and the voluntary movement cannot be supported. Appropriate exercise support cannot be performed for patients with essential tremor in which involuntary movements are mixed. For this reason, conventionally, there is no appliance or mechanical system for causing tremor patients to perform voluntary movements while suppressing the influence of involuntary movements.

  By the way, as a result of earnestly conducting experimental research, the present inventors pay attention to the fact that the myoelectric potential of tremor patients has periodicity, and based on the myoelectric potential, only involuntary movement and involuntary movement are mixed. We created a signal processing algorithm that automatically identifies the voluntary movement. Here, in order to automatically recognize voluntary movements using the periodicity of myoelectric potential of tremor patients, spectrum analysis was performed for each frequency by performing signal processing of myoelectric potential with short-time Fourier transform. . That is, the tremor patient performs voluntary exercise for a predetermined time, and acquires the myoelectric potential during the voluntary exercise and the non-operating time when the voluntary exercise before and after the voluntary exercise is performed every predetermined time. Each myoelectric potential was subjected to signal processing by short-time Fourier transform. At this time, a rectangular window was used as the window function, the window time was changed within each range of 5 to 200 (msec) and the frequency was within 1 to 300 (Hz), and the spectrum at each window time and each frequency was obtained. . As a result, when the window time is long (for example, 200 (msec)), there is a large change in the spectrum during operation and non-operation, and the change is particularly remarkable at a certain frequency (for example, 90 Hz). I found out. However, at this window time, the spectrum at the start of voluntary movement is very small at any frequency, and a time delay occurs until a large reaction is seen. Therefore, when the window time is short (for example, 20 (msec)), although the change in the spectrum during the operation and the non-operation is small, the spectrum changes with a slight time delay when attention is paid at the start of voluntary movement. I found out. Therefore, when the myoelectric potential obtained from the tremor patient is signal-processed by a short-time Fourier transform, the voluntary movement of the tremor patient can be achieved by using the above-mentioned two different window times that are specific to the patient. It was found that recognition was possible accurately and without delay. According to the experiments of the present inventors, the algorithm has a discriminating rate of voluntary movements in tremor patients of about 90%, resulting in high usability.

  The present invention has been devised on the basis of the above problems and knowledge, and its purpose is to automatically and accurately determine whether or not a tremor patient is performing voluntary exercise, To provide an involuntary movement suppression system, a movement recognition apparatus, and a program for the movement recognition apparatus that can suppress involuntary movement mixed when the patient performs voluntary movement and contribute to causing the tremor patient to perform only voluntary movement. It is in.

(1) In order to achieve the above object, the present invention includes an appliance that is mounted near a user's joint and operates to support the movement of a body part by the joint, and a control device that controls the operation of the appliance. An involuntary movement suppression system that suppresses involuntary movement that is not intended by the user by the operation of the appliance,
The brace is mounted on the skin surface of a muscle part that operates the body part, an arm that can be operated only in the direction of movement of the body part, an actuator that operates the arm, and the muscle potential of the muscle part is predetermined. With a myoelectric potential sensor acquired in units of time,
When the actuator is driven, the arm moves the body part only in the direction of voluntary movement intended by the user, while when the actuator stops, the arm is fixed so as to be able to regulate the movement of the body part,
The control device includes a database in which data for determining the presence or absence of the voluntary movement is stored, and movement estimation means for estimating the presence or absence of the voluntary movement when detecting a myoelectric potential from the myoelectric potential sensor based on the data And command means for giving an operation command to the actuator according to the estimation by the motion estimation means,
In the command means, when it is estimated that the voluntary movement is performed by the movement estimation means, while driving the actuator, it is estimated that the voluntary movement is not performed by the movement estimation means, The actuator is stopped.

  (2) Further, in the motion estimation means, each time a myoelectric potential is acquired by the myoelectric potential sensor, a frequency of a time-series signal using a preset window function, window time, and frequency for the myoelectric potential. It is preferable to adopt a configuration in which a spectrum is calculated by performing signal processing by conversion into a region, and the presence or absence of the voluntary movement is estimated from the value of the spectrum.

  (3) Here, the motion estimation means recognizes, as the combination of the window time and the frequency, the first combination for discriminating between the voluntary motion and the time when it is not, and the start time of the voluntary motion. It is preferable to adopt a configuration in which a second combination is set, the spectrum is calculated for each of the first and second combinations, and the presence or absence of the voluntary movement is estimated based on the value of each spectrum.

(4) Further, the data is a correspondence relationship between a value of each spectrum by the signal processing in the first and second combinations and a torque at the time of operation of the arm,
In the motion estimation means, for each myoelectric potential obtained sequentially based on the data, the magnitude of torque during operation based on the user's voluntary motion is estimated,
The command means preferably employs a configuration in which a drive command is given to the actuator so that the arm operates with the torque estimated by the motion estimation means.

(5) Furthermore, the control device further includes data creation means for creating the data by the user operating the body part,
In the data creation means, a spectrum is calculated by the signal processing in which a window function is set in advance for a plurality of combinations of window time and frequency with respect to the myoelectric potential acquired every predetermined time by the myoelectric potential sensor. The first combination of the window time and the frequency at which the spectrum change during the operation of the body part appears for the longest time, and the window time and the frequency at which the spectrum change during the operation of the body part appears at the earliest time 2 is determined, and the correspondence between the spectrum value obtained as a result of performing the signal processing with the first and second combinations on the measured myoelectric potential and the presence / absence of a voluntary action is defined as the data. ,
The motion estimation means may employ a configuration in which the spectrum is calculated for each of the first and second combinations, and the presence / absence of the voluntary motion is estimated based on the value of each spectrum.

(6) Further, the present invention is a motion recognition device that estimates whether or not the motion is a voluntary motion intended by the user from a myoelectric potential obtained during the motion of the user's body part,
A database in which data for determining the presence or absence of the voluntary movement is stored; and a movement estimation means for estimating the presence or absence of the voluntary movement when detecting the myoelectric potential based on the data;
The motion estimation means calculates a spectrum by performing signal processing on the myoelectric potential by converting a time-series signal into a frequency domain using a preset window function, window time, and frequency, The configuration is such that the presence or absence of the voluntary movement is estimated based on the value.

(7) Furthermore, the present invention is a program applied to a motion recognition device that estimates whether or not the motion is a voluntary motion intended by the user from a myoelectric potential obtained during the motion of the user's body part. There,
A spectrum is calculated by performing signal processing on the myoelectric potential by converting a time-series signal into a frequency domain using a preset window function, window time, and frequency, and the value of the spectrum is used to calculate the voluntary movement. The structure which makes the said motion recognition apparatus perform the procedure which estimates the presence or absence is taken.

  According to the present invention, it is possible to automatically and accurately determine whether or not a tremor patient or a cerebral palsy patient is performing voluntary movement, and involuntary movement mixed when a tremor patient or the like performs voluntary movement. This makes it possible to support exercise that causes tremor patients and the like to perform only voluntary exercise.

  In particular, according to the configurations of (2) to (5) above, a signal obtained by converting a time-series signal into a frequency domain with respect to myoelectric potentials of a tremor patient or the like using two types of window times that are patient-specific. By performing the process, voluntary movements such as tremor patients can be recognized accurately and without delay.

The schematic block diagram of the involuntary movement suppression system which concerns on this embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a schematic configuration diagram of a tremor suppression system as an involuntary movement suppression system according to the present embodiment. In this figure, the tremor suppression system 10 is a system that suppresses unintentional involuntary movements among movements performed by a user H who is a tremor patient. The tremor suppression system 10 includes a brace 11 that is mounted near the joint of the user H and operates to support the movement of the body part by the joint, and a control device 12 that controls the movement of the brace 11. .

  The brace 11 is adapted to be worn on the arm portion of the user H, and supports the bending motion of the elbow joint. The brace 11 is attached to the skin surface of the upper arm muscle for performing the bending operation, the arm 14 that can be operated only in the bending operation direction of the elbow joint of the user H, the motor 15 as an actuator that operates the arm 14. And a myoelectric potential sensor 17 for acquiring myoelectric potential of the upper arm muscle.

  The arm 14 includes an upper arm side arm 19 disposed along the outer side of the upper arm portion of the user H, a forearm side arm 20 disposed along the outer side of the forearm portion of the user H, and the upper arm side arm 19. And a forearm side arm 20, a connecting part 22 for connecting the upper arm side arm 20 in a relatively rotatable manner, an upper arm side mounting part 24 attached to the tip (upper end) side of the upper arm side arm 19 and fixed to the upper arm part of the user H, A forearm side mounting portion 25 that is attached to the distal end side of the arm 20 and can be fixed to the wrist of the user H is provided.

  The connecting portion 22 is arranged outside the vicinity of the elbow joint of the user H when the user H wears the arm 14. Further, the connecting portion 22 is provided with a stopper (not shown) that restricts the rotational operation range of the forearm arm 20 with respect to the upper arm side arm 19, and only by the stopper within the range of the bending operation of the user H. Thus, the forearm arm 20 can rotate.

  The upper arm side mounting portion 24 is formed in a substantially U shape (not shown) in plan view so that the periphery of the upper arm portion can be sandwiched inside, and the width of the inner portion where the upper arm portion is sandwiched is determined by the user H who wears it. It can be adjusted according to the thickness of the upper arm. The upper arm side mounting portion 24 is fixed to a predetermined position of the upper arm portion with a hook-and-loop fastener (not shown) so as not to fall off.

  The forearm side mounting portion 25 has substantially the same configuration as the upper arm side mounting portion 24 and is fixed to a predetermined position of the wrist with a hook-and-loop fastener (not shown).

  The motor 15 is provided in the connecting portion 22, and when the motor 15 is driven, the forearm arm 20 rotates with respect to the upper arm 19. On the other hand, when the motor 15 is stopped, the rotation operation of the forearm arm 20 with respect to the upper arm 19 is locked, and the user H with the arm 15 attached can freely move the forearm by locking the forearm arm 20. Can not work. In addition, at the time of data acquisition, which will be described later, before the operation support by the brace 11 is performed, the motor 15 is stopped, but the forearm arm 19 rotates freely according to the operation of the forearm portion of the user H. It is possible.

  Here, either an assist mode for performing exercise support by the brace 11 or a data acquisition mode for acquiring data before the support is selected by input from an input unit (not shown), and based on the selection. Thus, the mode of rotation of the forearm arm 20 is determined. That is, when the support mode is selected, the drive of the motor 15 is controlled by the control device 12, and when the motor 15 is driven, the forearm arm 20 rotates relative to the upper arm 19 while the motor 15 is stopped. Sometimes, the rotation of the forearm arm 20 is locked. On the other hand, when the data acquisition mode is selected, the motor 15 stops, but the forearm arm 20 can be freely rotated. In the present embodiment, the motor 15 rotates the forearm arm 20 with respect to the upper arm 19 by the motor 15. However, as long as the same action is achieved, instead of the motor 15, a cylinder, etc. Other actuators can also be applied.

  The myoelectric potential sensor 17 has two upper and lower arm relations between the front and back of the skin surface where the brachial muscle is used, which is used when the user H flexes the elbow joint, that is, the main muscle and the antagonistic muscle during the flexing motion. One muscle and one triceps are attached to the skin surface. In these myoelectric potential sensors 17, myoelectric signals, that is, myoelectric potentials are acquired every predetermined time (in this embodiment, every 1 msec), and the myoelectric signals are sequentially transmitted to the control device 12.

  The control device 12 is configured by a computer including an arithmetic processing device such as a CPU and a storage device such as a memory and a hard disk, and a program for causing the computer to function as the following units is installed.

  The control device 12 includes, for each myoelectric state, data creating means 27 for creating data for determining the presence or absence of voluntary movement intended by the user H, a database 28 for storing the data, and a database 28. Based on the stored data, a motion estimation means 29 for estimating whether or not a voluntary movement is performed when detecting a myoelectric potential from the myoelectric potential sensor 17, and an operation command to the motor 15 according to the estimation by the motion estimation means 29. Command means 30 to perform.

  In the data creation means 27, when the data acquisition mode is selected, user-specific data for determining the presence or absence of voluntary exercise is created in the following procedure.

  First, the user H wears the brace 11 and selects the data acquisition mode, so that the forearm arm 20 can be freely rotated so that the user H can freely operate the forearm. Then, the user H performs the elbow joint bending motion several times. At this time, before and after the bending operation, the user H or an attendant presses a switch (not shown) to specify the start time and the end time of the bending operation that is an arbitrary movement. Then, the myoelectric potential of the user H is measured by the myoelectric potential sensor 17 at predetermined time intervals.

  Therefore, the data creation means 27 performs signal processing by means of short-time Fourier transform (hereinafter referred to as “STFT”), which is one of frequency domain conversion methods for time series signals, with respect to the measured value of the myoelectric potential at each time. The data is obtained as follows. The STFT here is to obtain a spectrum that is a complex number by multiplying the myoelectric potential while shifting the window function in time and Fourier transform, and when using the STFT, an appropriate window function, window time, It is necessary to determine the frequency. As the window function, a rectangular window with a small amount of periodic reduction of the myoelectric signal during operation is used. On the other hand, for the window time and frequency, a plurality of combinations of window time and frequency having a window time in the range of, for example, 5 to 200 (msec) and a frequency in the range of, for example, 1 to 300 (Hz) are prepared. For the myoelectric potential measured every predetermined time, the calculation process by the STFT is performed, and the spectrum is calculated. Then, from the obtained spectrum, a combination of a window time and a frequency (hereinafter referred to as “first combination”) in which the spectrum change during the bending operation appears for the longest time is determined. Further, a combination of window time and frequency (hereinafter referred to as “second combination”) in which the spectrum change during the bending operation appears at the earliest time is determined from the obtained spectrum. Here, the first combination is useful for distinguishing between voluntary movements and other times, and the second combination is useful for recognizing the start of voluntary movements.

  And it corresponds that the spectrum of each time obtained as a result of performing STFT with the 1st combination with respect to the measured myoelectric potential is in operation arbitrarily. Similarly, the second combination is also associated with a voluntary operation. From this information, by using a learning machine such as a neural network, the correspondence between each spectrum value by the STFT in the first and second combinations and the presence / absence of an optional action is obtained as data, and stored in the database 28. Remembered.

  When the use mode is selected, the motion estimation unit 29 determines the presence or absence of voluntary exercise by the following procedure every time the myoelectric potential of the user H is measured.

  In other words, similarly to the data creation means 27, the acquired myoelectric potential is subjected to arithmetic processing by an STFT having a window function as a rectangular window. At this time, STFT is performed for the first and second combinations obtained by the data creating means 27, and the spectrum at the time of the first and second combinations is obtained. Based on the data stored in the database 28, it is estimated at this time whether or not the user H is performing voluntary exercise.

  In the command unit 30, when the motion estimation unit 29 estimates that the user H is performing voluntary exercise, the command unit 30 issues a drive command to the motor 15, while the user H does not perform voluntary exercise. If estimated, a stop command is issued to the motor 15.

  With the above configuration, when the user H tries to perform voluntary exercise, the motor 15 is driven, the forearm arm 20 rotates relative to the upper arm 19, and the forearm arm 20 rotates, The elbow joint of the user H is bent in a form that matches the intention of the user H to voluntary movement. At this time, the involuntary movement mixed in the voluntary movement is not reflected in the movement of the arm 14 but is regulated by the arm 14. As a result, the user H can support only the voluntary movement by the movement support by the arm 14. Will be able to. On the other hand, when the user H does not perform voluntary exercise, the motor 15 stops and the rotation operation of the forearm arm 20 is locked, and the user H cannot freely flex the elbow joint. Involuntary movements that are not accompanied are regulated.

  Therefore, according to such an embodiment, it is possible to reduce the influence of involuntary movement mixed when a patient with essential tremor performs voluntary movement, and to perform exercise support only for voluntary movement accurately and without delay. The effect that you can do it.

  In addition, the tremor suppression system 10 according to the present invention is not limited to exercise support for essential tremor patients, and can also support exercise for other tremor patients. Furthermore, involuntary exercise such as cerebral palsy is possible. It can also be used as an involuntary movement suppression system that supports mixed voluntary movements.

  In the above-described embodiment, the orthosis 11 is attached to the arm and the exercise of the arm is supported. However, the present invention is not limited to this, and the body 11 is attached to another body part such as a leg. The tremor suppression system 10 can also be applied as exercise support for a part. In short, the tremor suppression system 10 can support exercise of various body parts that are operated by joints.

  In the data creation means 27 and the motion estimation means 29, signal processing by short-time Fourier transform is performed on the acquired myoelectric potential. However, in the present invention, frequency domain conversion for other time series signals such as wavelet transform is performed. A technique may be used, and in this case, the same effect as that of the above embodiment can be obtained.

  Furthermore, in the above-described embodiment, the data of the user H is created by the data creation means 27 before using the operation support by the appliance 11, but the first and second combinations are used without using the data creation means 27. Data of the correspondence between the spectrum value and the presence / absence of an optional action may be stored in the database 28 in advance. In this case, the first and second combinations are specified, or a plurality of combinations are prepared and selected according to the state of the user H.

  In addition, once the data is created by the data creation means 27, when the same user H uses the tremor suppression system 10 next time, creation of data in the data acquisition mode can be omitted.

  Further, in the data creation means 27, a torque sensor is provided in the connecting portion 22, or the torque is taken into account by the user H simulating various different operations in which the torque is set in advance in the data acquisition mode. The data can be obtained. That is, from the myoelectric potential obtained from the user H, by learning the correspondence between each spectrum value by the STFT in the first and second combinations and the torque at the time of the operation of the arm 14 by the movement of the user H It can also be converted into data. In this case, the motion estimation means 29 estimates the magnitude of torque during operation based on the voluntary movement of the user H for each of the sequentially obtained myoelectric potentials, and the command means estimates the motion estimation means 29. A drive command is issued to the motor 15 so that the arm 14 operates with a high torque.

  Whether the movement is an arbitrary movement intended by the user H from the myoelectric potential during the movement of the body part of the user H by the database 28 and the movement estimation means 29 or in addition to the data creation means 27. It can also be configured as a motion recognition device that estimates whether or not, and in addition to the motion support described above, it can be applied to various applications that require automatic determination of the presence or absence of voluntary motion.

  In addition, the configuration of each part of the apparatus in the present invention is not limited to the illustrated configuration example, and various modifications are possible as long as substantially the same operation is achieved.

10 Tremor suppression system (involuntary movement suppression system)
DESCRIPTION OF SYMBOLS 11 Orthosis 12 Control apparatus 14 Arm 15 Motor 17 Myoelectric potential sensor 27 Data preparation means 28 Database 29 Motion estimation means 30 Command means H User

Claims (7)

A brace that is mounted near the user's joint and operates to support the movement of the body part by the joint, and a control device that controls the movement of the brace, and is not intended by the user due to the movement of the brace. An involuntary movement suppression system that suppresses involuntary movement,
The brace is mounted on the skin surface of a muscle part that operates the body part, an arm that can be operated only in the direction of movement of the body part, an actuator that operates the arm, and the muscle potential of the muscle part is predetermined. With a myoelectric potential sensor acquired in units of time,
When the actuator is driven, the arm moves the body part only in the direction of voluntary movement intended by the user, while when the actuator stops, the arm is fixed so as to be able to regulate the movement of the body part,
The control device includes a database in which data for determining the presence or absence of the voluntary movement is stored, and movement estimation means for estimating the presence or absence of the voluntary movement when detecting a myoelectric potential from the myoelectric potential sensor based on the data And command means for giving an operation command to the actuator according to the estimation by the motion estimation means,
In the command means, when it is estimated that the voluntary movement is performed by the movement estimation means, while driving the actuator, it is estimated that the voluntary movement is not performed by the movement estimation means, An involuntary movement suppression system, wherein the actuator is stopped.   In the motion estimation means, every time a myoelectric potential is acquired by the myoelectric potential sensor, the myoelectric potential is converted into a frequency domain of a time series signal using a preset window function, window time, and frequency. The involuntary movement suppression system according to claim 1, wherein a spectrum is calculated by performing signal processing, and the presence or absence of the voluntary movement is estimated from a value of the spectrum.   In the motion estimation means, as the combination of the window time and the frequency, a first combination for distinguishing between voluntary motion and a time when it is not, and a second combination for recognizing the start time of the voluntary motion. The involuntary movement suppression system according to claim 2, wherein the spectrum is set for each of the first and second combinations, and the presence or absence of the voluntary movement is estimated based on the value of each spectrum. The data is a correspondence relationship between the value of each spectrum by the signal processing in the first and second combinations and the torque at the time of operation of the arm,
In the motion estimation means, for each myoelectric potential obtained sequentially based on the data, the magnitude of torque during operation based on the user's voluntary motion is estimated,
4. The involuntary movement suppression system according to claim 3, wherein the command means commands the actuator to drive the arm with the torque estimated by the movement estimation means. The control device further comprises data creation means for creating the data by operating the body part by the user,
In the data creation means, a spectrum is calculated by the signal processing in which a window function is set in advance for a plurality of combinations of window time and frequency with respect to the myoelectric potential acquired every predetermined time by the myoelectric potential sensor. The first combination of the window time and the frequency at which the spectrum change during the operation of the body part appears for the longest time, and the window time and the frequency at which the spectrum change during the operation of the body part appears at the earliest time 2 is determined, and the correspondence between the spectrum value obtained as a result of performing the signal processing with the first and second combinations on the measured myoelectric potential and the presence / absence of a voluntary action is defined as the data. ,
3. The involuntary movement suppression according to claim 2, wherein the movement estimation unit calculates the spectrum for each of the first and second combinations and estimates the presence or absence of the voluntary movement based on a value of each spectrum. system. A motion recognition device that estimates whether or not the motion is a voluntary motion intended by the user from a myoelectric potential obtained during the motion of the user's body part,
A database in which data for determining the presence or absence of the voluntary movement is stored; and a movement estimation means for estimating the presence or absence of the voluntary movement when detecting the myoelectric potential based on the data;
The motion estimation means calculates a spectrum by performing signal processing on the myoelectric potential by converting a time-series signal into a frequency domain using a preset window function, window time, and frequency, A motion recognition device, wherein the presence or absence of the voluntary motion is estimated from a value. A program applied to a motion recognition device that estimates whether the motion is a voluntary motion intended by the user from a myoelectric potential obtained during the motion of a user's body part,
A spectrum is calculated by performing signal processing on the myoelectric potential by converting a time-series signal into a frequency domain using a preset window function, window time, and frequency, and the value of the spectrum is used to calculate the voluntary movement. A motion recognition device program for causing the motion recognition device to execute a procedure for estimating presence or absence.

Images