JP6110657B2 - Motion support system, voluntary movement recognition device, and voluntary movement recognition program - Google Patents

Description

  The present invention relates to a motion support system, a voluntary motion recognition device, and a voluntary motion recognition program. More specifically, the present invention relates not only to a case where a tremor patient intends “active motion” but also to “passive” such as holding a specific posture. The present invention also relates to a motion support system, a voluntary motion recognition device, and a voluntary motion recognition program that can reliably support only voluntary motion based on the intention of a tremor patient even when intending “motion”.

  There is a neurological disease called “tremor” in which the limbs vibrate against their will. As a kind of “tremor”, there is “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, for example, because hands holding various instruments necessary for daily life shake. In order to suppress such tremor, there are the use of drugs such as β-blockers and deep brain stimulation (DBS), but there are problems such as adverse effects on the body such as causing side effects.

  By the way, Patent Document 1 (Japanese Patent Laid-Open No. 11-253504) discloses an “upper limb movement assisting device” that assists the manual movement of the body. This “upper limb motion assist device” has a free end provided with a brace that is worn on the user's upper limb, and 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. In this “upper limb movement assist device”, force information changes 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 user, which is caused by shaking of the user's hand. In this case, the trajectory control of the free end is not performed.

  However, the “upper limb motion assisting device” of Patent Document 1 cannot assist the upper limb motion of the essential tremor patient for the following reason. 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 “upper limb motion assisting device” of Patent Document 1, the motion control of the free end is stopped at the time when a periodic change in force due to involuntary motion is detected, and the voluntary motion cannot be assisted. Because. In other words, the “upper limb movement assist device” of Patent Document 1 cannot perform appropriate exercise support for essential tremor patients in which involuntary exercise is mixed when voluntary exercise is performed.

  Patent Document 2 (Japanese Patent Application Laid-Open No. 2011-193941) discloses a “involuntary movement suppression system” that enables tremor patients to perform voluntary movement by suppressing the influence of involuntary movement. Has been. The “involuntary movement suppression system” includes a brace 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 movement of the brace. The involuntary movement suppression system which suppresses the involuntary movement which the said user does not intend by operation | movement of. The control device includes a database storing data for determining the presence or absence of voluntary movement, and a movement estimation means for estimating the presence or absence of 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. When the command means estimates that the voluntary motion is being performed by the motion estimation means, On the other hand, when it is estimated that the voluntary movement is not performed by the movement estimation means, the actuator is stopped.

  The “involuntary movement suppression system” of Patent Document 2 focuses on the fact that myoelectric potentials of tremor patients have periodicity, and based on the myoelectric potentials, only involuntary movements and voluntary movements in which involuntary movements coexist. And a signal processing algorithm that automatically identifies This signal processing algorithm uses the periodicity of myoelectric potentials of tremor patients and analyzes the spectrum for each frequency by processing the myoelectric potentials with a short time Fourier transform in order to automatically recognize voluntary movements. It is carried out. According to Patent Document 2, according to this algorithm, the discriminating rate of voluntary movement in a tremor patient is about 90%, which is said to bring high usability. Therefore, according to this “involuntary exercise suppression system”, it is possible to provide appropriate exercise support to essential tremor patients in which involuntary exercise is mixed when voluntary exercise is performed.

JP-A-11-253504 JP 2011-193941 A

  According to the “involuntary movement suppression system” of Patent Document 2 described above, it is possible to perform appropriate exercise support for essential tremor patients in which involuntary movement is mixed when voluntary movement is performed. The “involuntary movement suppression system” stops suppressing involuntary movement (tremor) by finding voluntary movement included in the myoelectric potential data obtained from the tremor patient.

  Further, the applicant of the present application removes the signal component of involuntary movement from the electromyographic data of a tremor patient etc., and extracts only the signal component of voluntary movement, A patent application has been filed for the “motion support device” used (Japanese Patent Application No. 2011-184380). This "myoelectric data processing device" estimates a signal component corresponding to an involuntary movement unintended by the subject from the electromyographic data related to the myoelectric potential of the subject, and the voluntary movement intended by the subject A myoelectric data processing device for identifying a signal component corresponding to the involuntary waveform obtained by storing a plurality of involuntary waveforms representing changes over time of the myoelectric potential during the involuntary movement, and the subject. For the acquired waveform representing the change over time of the myoelectric data, select the most approximate waveform among the involuntary waveforms, and calculate the similarity between the most approximate waveform and the acquired waveform A degree calculating means, an attenuation rate calculating means for obtaining an attenuation rate representing a ratio of the signal component of the involuntary movement in the myoelectric data based on the most approximate waveform and the similarity; and Before acquisition EMG data and a filtering means divided into the respective signal components.

  According to the above “myoelectric data processing device”, the myoelectric data acquired from a subject such as a tremor patient who performs involuntary movements, a signal component corresponding to voluntary movement and a signal component corresponding to involuntary movement, Therefore, it is possible to remove any one of the signal components and extract only one of the other signal components from myoelectric data obtained from a subject who has both involuntary movement and voluntary movement. it can. As a result, the present invention can be applied to the control of a device that supports exercise only for voluntary movement by eliminating the influence of involuntary movement, a device that supports exercise for suppressing involuntary movement, and the like. The “motion support device” includes a brace worn near a patient's joint and a control device that controls the motion of the brace, and the “myoelectric data processing device” is used as the control device. . When a patient using this “motion support device” performs voluntary movement (for example, bending motion of a joint), if there is an influence of tremor, the influence of the involuntary movement by the tremor is The actuator of the appliance is driven on the basis of a command signal that is removed by the processing device and takes into account only voluntary movements. As a result, the bending motion intended by the patient is supported. The influence of tremor when the patient does not voluntarily exercise is almost eliminated by the “myoelectric data processing device”, and the actuator of the appliance is not driven. As a result, an operation unintended by the patient is prevented from being supported.

  However, considering the daily movements of tremor patients, voluntary movements are not only intended for specific “movements” but also for specific “holding postures”. It goes without saying that the daily movements of tremor patients include, for example, positive “movements” of flexion and extension of the elbow joint, but specific conditions of the elbow joint (flexion and extension at a certain angle) There is also a negative “action” of maintaining. However, the above “motion support device” can accurately support such an active “motion”, but cannot provide accurate support for a negative “motion”.

  The present invention has been made in consideration of the circumstances as described above, and the purpose of the present invention is not only when a tremor patient intends a specific “aggressive motion” but also, for example, a specific posture. Provides a motion support system, voluntary motion recognition device, and voluntary motion recognition program that can reliably support only voluntary motion based on the intention of a tremor patient even when a specific “passive motion” such as holding of the motion is intended It is to be.

  Other objects of the present invention which are not specified here will be apparent from the following description and the accompanying drawings.

(1) In a first aspect of the present invention, an operation support system is provided. This motion support system
A brace that is worn near the user's joint and operates to support movement of the body part by the joint;
A control device for controlling the operation of the brace,
The orthosis is attached to an arm operable only in the movement direction of the body part, an actuator that operates the arm, and a muscle part that operates the body part, and the myoelectric potential of the muscle part is determined in a predetermined time unit. A biological signal sensor to be acquired,
The arm moves the body part only in the direction of the voluntary movement intended by the user by driving the actuator, and is fixed so as to regulate the movement of the body part when the actuator stops.
The controller is
A first determination step of determining whether tremor has occurred in the user using a correlation coefficient calculated from an autocorrelation function based on a myoelectric signal obtained from the biological signal sensor;
If it is determined that tremor has occurred, whether the user is performing a specific action or maintaining a specific posture using the generation period of the grouped discharge calculated from the autocorrelation function And a second determination step for determining
If it is determined in the second determination step that the user is performing a specific movement, the actuator is driven to support the voluntary movement by eliminating the influence of the involuntary movement,
When it is determined in the second determination step that the user maintains a specific posture, the actuator is stopped to restrict the movement of the body part.

  Since the operation support system according to the first aspect of the present invention has the above-described configuration, after the occurrence of a tremor is recognized in the first determination step, the user is specified in the second determination step. The actuator can be driven or stopped after it is determined whether it is operating or maintaining a specific posture. Therefore, not only when the tremor patient intends a specific “aggressive motion”, but also when the tremor patient intends a specific “reactive motion” such as holding a specific posture, the voluntary option based on the tremor patient's intention Only the operation can be reliably supported.

  (2) In a preferred example of the motion support system according to the first aspect of the present invention, the acceleration data acquired by the acceleration sensor attached to the body part is used when making the determination in the first determination step. In this case, there is an advantage that it is easy to determine whether or not tremor has occurred.

  Further, in this example, the average value of the correlation coefficient when the tremor is not occurring in the user is equal to or greater than the average value of the correlation coefficient when the tremor is occurring in the user. Preferably, the threshold is determined by the range.

  In another preferable example of the motion support system according to the first aspect of the present invention, the joint angle data of the joint is used when making the determination in the second determination step. In this case, there is an advantage that it is easy to determine whether the user is performing a specific motion or maintaining a specific posture.

Further, in this example, an average value of the generation period of the grouped discharge is calculated, and the average when the user maintains a specific posture is equal to or more than the average value when the user is performing a specific action. It is preferable that the threshold value is determined within the range of the value or less.

(3) In the 2nd viewpoint of this invention, the voluntary movement recognition apparatus is provided. This voluntary movement recognition device
A first determination step of determining whether tremor has occurred in the user using a correlation coefficient calculated from an autocorrelation function based on a myoelectric signal obtained from a biological signal sensor;
If it is determined that tremor has occurred, whether the user is performing a specific action or maintaining a specific posture using the generation period of the grouped discharge calculated from the autocorrelation function A second determination step for determining
If it is determined in the second determination step that the user is performing a specific motion, a command signal for driving the actuator is generated so as to support the voluntary movement by eliminating the influence of the involuntary movement. ,
When it is determined in the second determination step that the user maintains a specific posture, the actuator is stopped to generate a command signal that regulates the movement of the body part. Is.

  Since the voluntary movement recognition apparatus according to the second aspect of the present invention has the above-described configuration, after the occurrence of tremor is recognized in the first determination step, the user specifies the second determination step. It is possible to drive or stop the actuator after determining whether the operation is performed or whether a specific posture is maintained. Therefore, not only when the tremor patient intends a specific “aggressive motion”, but also when the tremor patient intends a specific “reactive motion” such as holding a specific posture, the voluntary option based on the tremor patient's intention Only exercise can be reliably supported.

  (4) In a preferred example of the voluntary movement recognition device according to the second aspect of the present invention, acceleration data acquired by an acceleration sensor attached to the body part is used when making the determination in the first determination step. . In this case, there is an advantage that it is easy to determine whether or not tremor has occurred.

  In this example, the average value of the correlation coefficient when the tremor is not occurring in the user is greater than or equal to the average value of the correlation coefficient when the tremor is occurring in the user. The threshold is preferably determined.

  In another preferred example of the voluntary movement recognition device according to the second aspect of the present invention, the joint angle data of the joint is used when making the determination in the second determination step. In this case, there is an advantage that it is easy to determine whether the user is performing a specific motion or maintaining a specific posture.

In this example, an average value of the generation period of the grouped discharge is calculated, and is equal to or higher than the average value when the user is performing a specific action, and is equal to or lower than the average value when the specific posture is maintained. It is preferable that the threshold value is determined in the range of.

(5) In the 3rd viewpoint of this invention, the voluntary movement recognition program is provided. This voluntary movement recognition program
A first determination step of determining whether tremor has occurred in the user using a correlation coefficient calculated from an autocorrelation function based on a myoelectric signal obtained from a biological signal sensor;
If it is determined that tremor has occurred, whether the user is performing a specific action or maintaining a specific posture using the generation period of the grouped discharge calculated from the autocorrelation function A second determination step for determining
If it is determined in the second determination step that the user is performing a specific motion, a command signal for driving the actuator is generated so as to support the voluntary movement by eliminating the influence of the involuntary movement. ,
If it is determined in the second determination step that the user is maintaining a specific posture, the actuator is stopped to generate a command signal that regulates the movement of the body part. It is characterized by functioning.

  Since the voluntary movement recognition program according to the third aspect of the present invention has the above-described configuration, after the occurrence of tremor is recognized in the first determination step, the user specifies in the second determination step. It is possible to drive or stop the actuator after determining whether the operation is performed or whether a specific posture is maintained. Therefore, not only when the tremor patient intends a specific “aggressive motion”, but also when the tremor patient intends a specific “reactive motion” such as holding a specific posture, the voluntary option based on the tremor patient's intention Only exercise can be reliably supported.

  According to the motion support system according to the first aspect of the present invention, the voluntary movement recognition apparatus according to the second aspect of the present invention, and the voluntary movement recognition program according to the third aspect of the present invention, It is possible to reliably support only the voluntary movement based on the intention of the tremor patient, not only when the movement movement is intended but also when the movement is intended to be a specific negative movement such as holding a specific posture. There is an effect that.

It is explanatory drawing which shows the whole structure of the operation | movement assistance system which concerns on 1st Embodiment of this invention. It is a flowchart which shows the outline of operation | movement of the control apparatus of the operation | movement assistance system which concerns on 1st Embodiment of this invention. It is a flowchart which shows the detail of operation | movement of the control apparatus of the operation | movement assistance system which concerns on 1st Embodiment of this invention. It is a wave form diagram which shows the myoelectric potential signal waveform produced | generated during the bending operation | movement of the elbow joint of a patient with essential tremor. It is a wave form diagram which shows the myoelectric potential in the state which the tremor has generate | occur | produced, and the state which has not generate | occur | produced. It is explanatory drawing which shows that the generation cycle of the signal which causes tremor changes according to the operation state of the elbow joint of a patient with essential tremor. It is explanatory drawing which shows the method of calculating the correlation coefficient of an autocorrelation function in the operation | movement assistance system which concerns on 1st Embodiment of this invention. It is explanatory drawing which shows the method of calculating the generation period of grouping discharge from an autocorrelation function in the operation | movement assistance system which concerns on 1st Embodiment of this invention. It is explanatory drawing which shows the simulation task of the drink operation | movement implemented in order to calculate the threshold value used with the operation | movement assistance system which concerns on 1st Embodiment of this invention.

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

  FIG. 1 shows the overall configuration of an operation support system 1 according to an embodiment of the present invention. This motion support system 1 supports the life motion of the user U who is a tremor patient, eliminates the influence of the involuntary movement that the user U does not intend, and only the voluntary movement intended by the user U. Power assist about.

  As shown in FIG. 1, the movement support system 1 is worn near the elbow joint of a user U, and supports a movement of the arm by the elbow joint, that is, a bending operation of the elbow joint, and the brace 10. And a control device 20 for controlling the operation according to the above.

  This brace 10 is mounted on the right arm of the user U in the form as shown in FIG. 1, and can be operated only in the operating direction for bending the elbow joint of the user U, and an actuator (motor) that operates the arm 11. 12), a biosignal sensor (surface myoelectric potential sensor) 15 a that is attached to the skin surface of the biceps to move the elbow and acquires myoelectric data related to the surface myoelectric potential of the biceps, and the elbow And a biological signal sensor 15b (acceleration sensor) that is attached to the skin surface of the back of the hand and acquires acceleration data of the hand. Moreover, although not shown in figure, the goniometer for measuring the angle of the said elbow is affixed on the said elbow joint.

  The arm 11 includes an upper arm side arm part 11a arranged along the outer side of the upper arm part of the user U, a forearm side arm part 11b arranged along the outer side of the user U's forearm part, and an upper arm side arm part. 11a and the forearm side arm part 11b are comprised from the connection part 11c which connects so that relative rotation is possible. The upper arm side arm portion 11a and the forearm side arm portion 11b are relatively rotatable around the connecting portion 11c.

  The upper arm side mounting portion 13 is attached to the tip (upper end) of the upper arm side arm 11a, and the forearm side mounting portion 13 is attached to the tip of the forearm side arm 11b. The arm 11 is mounted on the outer side of the right arm by winding the upper arm side mounting portion 13 around the upper arm portion of the user U and fixing the upper arm side mounting portion 13 and winding the forearm side mounting portion 14 around the wrist. At this time, the connecting part 11c is arranged outside the vicinity of the elbow joint of the user U. By doing so, the arm 11 can be attached to the outside along the right arm and bent in the same manner as the right elbow is bent.

  The connecting portion 11c is provided with a stopper (not shown) that restricts the rotational operation range of the forearm arm portion 11b relative to the upper arm side arm portion 11a. With this stopper, the range of rotation of the forearm arm 11b is limited to the range of bending motion of the right elbow of the user U.

  The upper arm side mounting portion 13 is formed in a shape (for example, a U-shape, a cylindrical shape, a spiral shape, or the like) that can tighten the upper arm portion inward, and the width of the inner portion to be tightened is the upper arm portion of the user U It can be adjusted according to the thickness of the. The upper arm side mounting portion 13 is fixed so as not to drop off at a predetermined position of the upper arm portion with a surface fastener or the like (not shown).

  The forearm side mounting portion 14 has substantially the same configuration as the upper arm side mounting portion 13 and is fixed by a hook-and-loop fastener (not shown) so as not to drop off at a predetermined position on the wrist.

  The actuator 12 is a motor, for example, and is disposed in the connecting portion 11c. During the operation of the actuator 12, the forearm arm 11b is rotated with respect to the upper arm 11a. When the actuator 12 is stopped, the rotation of the forearm arm 11b with respect to the upper arm 11a is locked. At this time, the forearm of the user U is also locked and cannot move freely.

  In the present embodiment, a motor is used as the actuator 12, thereby rotating the forearm arm portion 11 b with respect to the upper arm side arm portion 11 a, but as long as the same effect is achieved, An actuator having any configuration other than the motor can be used.

  The biological signal sensor (myoelectric potential sensor) 15a includes a portion of the skin surface where the brachial muscle is used when the user U bends the elbow joint, that is, a biceps brachii muscle that serves as a main muscle during the bending motion. It is attached to the skin surface and adds additional muscle signals such as the triceps as needed. In these myoelectric potential sensors, myoelectric potential signals are acquired at a sampling period every predetermined time (for example, every 1 msec). The acquired myoelectric potential signal is sequentially transmitted to the control device 20 and used for operation control of the appliance 10.

  The biological signal sensor (acceleration sensor) 15b detects the acceleration of the movement of the right hand of the user U, and is attached to the back of the right hand of the user U.

  The control device 20 uses the myoelectric potential signal acquired from the biological signal sensor (myoelectric potential sensor) 15a attached to the upper arm and the acceleration signal acquired from the biological signal sensor (acceleration sensor) 15b attached to the back of the hand, The actuator 12 of the appliance 10 is controlled.

  The myoelectric signal is used for assisting the operation of the tremor patient. The acceleration signal is used to know whether or not tremor has occurred. The signal from the goniometer is used to monitor the elbow condition (angle).

  The control device 20 is composed of an operation processing device such as a CPU and a combination device including a storage device such as a memory and a hard disk, and a program (an arbitrary motion recognition program) for causing the computer to function as follows is installed. ing.

  As shown in FIG. 1, the control device 20 includes a data buffering means 21, an angle change / posture holding determination means 22, a signal separation means 23, a motion estimation means 24, and a command means 25. ing.

  The data buffering means 21 is acquired and sent at the same sampling cycle as the myoelectric potential signal acquired and sent by the biological signal sensor (myoelectric potential sensor) 15a at a sampling period of every predetermined time (for example, every 1 msec). The acceleration signal and the elbow angle signal from the goniometer are temporarily held. Thereby, the motion of the tremor patient can be detected. These signals are sent to the angle change / posture holding determination means 22 and the signal separation means 23.

  The angle change / posture holding determination means 22 determines whether or not the bending angle of the right elbow of the user U has changed from the myoelectric potential signal, the acceleration signal, and the elbow angle signal. It is also determined whether or not the posture of the right elbow has been maintained (for example, holding a certain bent state). The determination result is sent to the signal separation unit 23 and the command unit 25.

  The signal separation means 23 separates a signal corresponding to voluntary movement and a signal corresponding to involuntary movement from the myoelectric potential signal acquired and transmitted by the biological signal sensor (myoelectric potential sensor) 15a as described later. . A signal corresponding to the voluntary movement is sent to the motion estimation means 24.

  The motion estimation unit 24 estimates the motion intended by the user U based on the signal corresponding to the voluntary movement separated by the signal separation unit 23. The estimation result is sent to the command means 25.

  The command means 25 changes the angle of the right elbow intended by the user U obtained by the angle change / posture holding judgment means 22, or the right elbow intended to be held by the user U. A command signal is sent to the actuator 12 so as to maintain the posture (angle). In addition, a command signal is sent to the actuator 12 so that the motion estimated by the motion estimation means 24 and intended by the user U is performed.

  Next, the operation of the control device 20 will be described.

  The outline of the operation of the control device 20 is as shown in FIG. That is, first, in step S1, the motion of the user U, that is, the tremor patient is detected. This step is executed by the data buffering means 21.

  In step S2, voluntary movement and involuntary movement are separated by biological signal processing. This step is executed by the angle change / posture holding determination means 22 and the signal separation means 23.

  In step S <b> 3, only the voluntary movement is reflected on the actuator 12 of the appliance 10. This step is executed by the motion estimation means 24.

  In step S4, the involuntary movement by the appliance 10 is suppressed by eliminating the signal due to the involuntary movement, and at the same time, the voluntary movement by the actuator 12 (including not only the specific movement but also the holding of the specific state) is performed. Thus, both are compatible.

  Next, the control operation of the actuator 12 by the control device 20 will be described in detail with reference to an experimental example.

  FIG. 4 shows an elbow joint angle and a surface myoelectric potential in the biceps (hereinafter referred to as “biceps”) when a patient with essential tremor holds a plastic bottle (550 g) with his right hand and then flexes the right elbow. FIG.

  It is known that the myoelectric potential of a patient with essential tremor shows a characteristic waveform shape called “grouped discharge” when symptoms of tremor are observed. The waveform shown in the region surrounded by the rectangular broken line in FIG. 4 is the waveform shape of “grouped discharge”. It can be seen that a state in which the amplitude near 0V is periodically displayed and a state in which a large amplitude is seen are repeated.

  In recent years, a technique for controlling a wearable robot such as the brace 10 described above using myoelectric potential as an input signal has been put into practical use. Here, in many robots that are being put into practical use, the expected users are caregivers, elderly people whose muscle strength has declined, and the like, and thus no grouped discharge is generated in the myoelectric potential. In other words, since an unintended large amplitude signal is not mixed as in an essential tremor patient, the intention of motion can be estimated at least by removing high frequency noise. On the other hand, if a signal with a large amplitude is measured unintentionally as in a patient with essential tremor, it is difficult to estimate the patient's movement intention (how it really wants to move regardless of tremor). Robot control is also difficult. However, according to the motion support system 1 of the present embodiment, it is possible to estimate the motion state that the patient wants to perform from the myoelectric potential of the essential tremor patient including “grouped discharge”.

  In the operation support system 1 of this embodiment, the patient's operation state is classified into the following three.

(1) A state in which no tremor has occurred (2) A state in which a tremor has occurred and is in a specific motion (3) A tremor has occurred and a specific posture is maintained In other words, whether a tremor has occurred or not (Determination A), and whether it is operating in a situation in which tremor is occurring or whether it is maintaining a posture (Determination B) is performed. The step of performing determination A corresponds to the first determination step of the present invention, and the step of performing determination B corresponds to the second determination step of the present invention.

(Determination A) (First Determination Step)
First, whether or not tremor has occurred is determined from information on the strength of periodicity included in the myoelectric potential waveform. The upper part of FIG. 5 shows a signal that has been processed by a band-pass filter with a cutoff frequency set to 30 to 300 Hz after rectifying the myoelectric potential signal. The lower part of FIG. 5 is an enlarged view of the waveform when tremor is occurring and the waveform when tremor is not occurring in the myoelectric signal shown in the upper part. Comparing the two, the waveform when tremor does not occur is not very periodic, whereas the waveform when tremor occurs indicates very strong periodicity.

  Therefore, based on this feature, by measuring the “periodic strength” of the waveform, it is possible to separate myoelectric potential when tremor occurs and when it does not occur. The strength of periodicity can be quantified from the difference in the correlation coefficient of the autocorrelation function. The state in which tremor is occurring and the state in which no tremor has occurred can be classified by the magnitude of the correlation coefficient.

(Determination B) (Second Determination Step)
On the other hand, when it is determined that a tremor has occurred, the determination as to whether the user U is performing a specific action or maintaining a specific posture is based on the “cycle of occurrence of grouping discharge”. To do. FIG. 6 summarizes the results of analyzing the “group discharge occurrence period” using the autocorrelation function. From this analysis result, it can be seen that the generation period of the grouped discharge is different between when the specific operation is performed and when the specific posture is maintained. Therefore, using this feature, it is possible to distinguish between a state in which a specific action is being performed and a state in which a specific posture is maintained when tremor is occurring.

  In the motion support system 1 of the present embodiment, after the states are separated as described above, the following different controls are performed in each state to support the performance of the essential tremor patient.

(1) No tremor → normal motion estimation (only removal of high frequency noise)
(2) A state in which a tremor has occurred and a specific action is performed → involuntary movement is eliminated, and only voluntary movement is supported (using the above-described movement support apparatus of Japanese Patent Application No. 2011-184380)
(3) A tremor has occurred and a specific posture is maintained → A brake is applied to stop the actuator operation (fixed to a specific posture)
Subsequently, the period of the waveform and the strength of the periodicity are derived.

  In the operation support system 1 of the present embodiment, the period of the myoelectric signal waveform and the strength of periodicity are quantified using a method called “autocorrelation function”. First, a procedure for deriving the waveform period and the strength of periodicity using the autocorrelation function will be described.

  First, a myoelectric data vector is created.

と、時刻ｔからある時間幅τずらした時刻（ｔ−τ）から窓時間κだけ切り出した離散データ

の相関係数を、時間幅τを変更しながら算出する。そのために、まず、筋電のデータベクトルＥ_ｗとＥ_τｓを準備する。 In the autocorrelation function, discrete data cut out from a certain time t by a window time κ (for example, 250 msec)

And discrete data cut out by a window time κ from a time (t−τ) shifted by a certain time width τ from the time t

The correlation coefficient is calculated while changing the time width τ. For this purpose, first, EMG data vectors E _w and E _τs are prepared.

次に、筋電位信号の波形の周期と周期性の強さを以下のようにして導出する。 Next, the correlation between the two data vectors is calculated. The shift time τ is changed in the range of 0 to 250 msec, and a vector C composed of τ correlation coefficients is generated.

Next, the period of the myoelectric signal waveform and the strength of the periodicity are derived as follows.

  The correlation coefficient vector C is graphed as shown in FIG. The horizontal axis is the shift time τ, and the vertical axis is the correlation coefficient. Assuming that the waveform has periodicity at a certain time T, the graph created from the calculation of the autocorrelation function is the first peak when the shift time is T as shown in FIG. 7 (the peak when τ = 0 is Excluding the first peak) appears. This indicates that the waveform when the waveform is shifted by T is more similar to the waveform before the shift than the other shift times. Due to the characteristics of such a graph, generally, the shift time T when the first peak is taken is treated as the period of the waveform. In addition, the magnitude of the correlation coefficient CT at the time of the first peak is an index indicating the strength of periodicity because it indicates the high degree of similarity.

  As described above, in the operation support system 1 of the present embodiment, in order to perform the discrimination A (first determination step), the CT value calculated from the autocorrelation function is used as a discrimination index, and the discrimination B (second determination). In order to perform (step), T calculated from the autocorrelation function is used as a discrimination index.

  Next, a determination method using these determination indexes will be described.

  When performing discrimination, it is necessary to set a threshold value for each discrimination index (T, CT). These threshold values may be determined, for example, by measuring in a task that simulates a meal in advance and analyzing the measured data using an autocorrelation function. This is because, when interviewed by the subject, the opinion that trembling was a problem due to the movement of liquid (drinking water with a cup, drinking tea with a cup, drinking soup with a spoon, etc.) This is because of consideration.

  Here, a task simulating the movement of drinking water was set up. Specifically, as shown in FIG. 9, the task was to hold a plastic bottle placed on a desk, transport it to the mouth, and then return it to its original position. The reason for using the PET bottle is that the contents can be easily prevented from spilling by closing the cap.

  The equipment used was an electromyograph, goniometer, and accelerometer. The electromyograph was attached to the subject's biceps using dry electrodes. A goniometer was affixed to the elbow joint to measure the angle of flexion and extension of the elbow. The accelerometer was attached to the back of the hand to measure the tremor expressed as movement.

  Next, the threshold value is determined after processing the data acquired in this way with an autocorrelation function. Hereinafter, the determination A and the determination B will be described separately.

(Determination A) (First Determination Step)
First, the state where tremor has occurred and the state where it has not occurred are separated. In order to perform the separation, the value of the hand acceleration read from the accelerometer is used. And in each of the state where tremor has occurred and the state where it has not occurred, the average value of the correlation coefficient is calculated, and the threshold value is equal to or greater than the average value of the correlation coefficient in the state where tremor has not occurred, A threshold value that is equal to or lower than the average value of the correlation coefficient in the state in which it occurs is determined.

(Determination B) (Second Determination Step)
First, the state of operating and the state of maintaining the posture are separated. The elbow joint angle read from the goniometer is used to perform the separation. Similar to the determination A, the average value of the period T is calculated in each of the divided states, and a threshold value equal to or higher than the average value of the period during movement and a threshold value equal to or lower than the average value during posture maintenance are set.

  When this method is put to practical use, it is only necessary to make a determination by comparing the value of the period T calculated from the set threshold value and the measured myoelectric potential and the value of the correlation coefficient CT of the peak as needed.

  When the state separation (discrimination A and B) is completed as described above, the appliance 10 is controlled as follows. This control is executed according to the flowchart shown in FIG.

(No tremor has occurred)
If it is determined in step S11 that no tremor has occurred, the myoelectric potential of the essential tremor patient is the same as that of the healthy person, so that the process proceeds to step S14 and processing for the healthy person is applied. be able to. For this reason, when tremor has not occurred, the measured myoelectric potential is processed by a band pass filter (pass band: 30 Hz to 300 Hz) and then rectified, and a low pass filter (cutoff frequency: 10 Hz) is applied to the rectified waveform. To remove the high frequency noise. Thereafter, the process proceeds to step S16, and the operation intended by the patient is supported by turning the actuator 12 on and off. For example, a known learning machine such as a “time delay neural network” is used.

(Treasure is occurring and in motion)
If it is determined in step S11 that a tremor has occurred, the process proceeds to step S12, and it is determined whether a specific action is being performed or a specific posture is being maintained. If it is determined that a specific operation is being performed, the process proceeds to step S15 to perform signal processing for removing the grouping discharge (that is, tremor). This is because a grouping discharge different from the patient's intention is generated by tremor. This signal processing is the same as the signal processing of the operation support apparatus described in Japanese Patent Application No. 2011-184380. Thereafter, the process proceeds to step S16, and the operation intended by the patient is supported by turning the actuator 12 on and off.

(The tremor is occurring and the posture is maintained)
If it is determined in step S12 that the specific posture is being maintained, the process proceeds to step S13. In this case, what the patient really wants to do is to maintain a certain posture. Therefore, a brake (not shown) provided on the actuator 12 of the appliance 10 is operated to fix the actuator and maintain the posture, that is, control to support the trembling arm.

  As described above, in the operation support system 1 according to an embodiment of the present invention, after the occurrence of tremor is recognized in the first determination step, the user U performs a specific operation in the second determination step. It is possible to drive or stop the actuator 12 after determining whether or not a specific posture is maintained. Therefore, not only when the tremor patient intends a specific “aggressive motion”, but also when the tremor patient intends a specific “reactive motion” such as holding a specific posture, the voluntary option based on the tremor patient's intention Only the operation can be reliably supported.

  Moreover, since the voluntary exercise | movement recognition apparatus which concerns on one Embodiment of this invention respond | corresponds to the control apparatus 20 of the said movement assistance system 1, it is clear that the effect similar to the said movement assistance system 1 is acquired.

  Since the voluntary movement recognition program according to the embodiment of the present invention is stored in the control device 20 of the motion support system 1 and operates, it is apparent that the same effect as the motion support system 1 can be obtained.

(Other variations)
The above-described embodiment shows an example in which the present invention is embodied. Therefore, the present invention is not limited to this embodiment, and it goes without saying that various modifications can be made without departing from the spirit of the present invention.

  For example, in the above-described embodiment, an example in which the present invention is applied to the motion of the elbow joint of a tremor patient is used, but the present invention is not limited to this. It can be applied to any joint other than the elbow joint.

DESCRIPTION OF SYMBOLS 1 Operation support system 10 Equipment 11 Arm 11a Upper arm side arm part 11b Forearm side arm part 11c Connection part 12 Actuator 13 Upper arm side mounting part 14 Forearm side mounting parts 15a, 15b Biosignal sensor 20 Control device 21 Data buffering means 22 Angle change / Attitude hold discrimination means 23 Signal separation means 24 Motion estimation means 25 Command means U User

Claims (11)

A brace that is worn near the user's joint and operates to support movement of the body part by the joint;
A control device for controlling the operation of the brace,
The orthosis is attached to an arm operable only in the movement direction of the body part, an actuator that operates the arm, and a muscle part that operates the body part, and the myoelectric potential of the muscle part is determined in a predetermined time unit. A biological signal sensor to be acquired,
The arm moves the body part only in the direction of the voluntary movement intended by the user by driving the actuator, and is fixed so as to regulate the movement of the body part when the actuator stops.
The controller is
A first determination step of determining whether tremor has occurred in the user using a correlation coefficient calculated from an autocorrelation function based on a myoelectric signal obtained from the biological signal sensor;
If it is determined that tremor has occurred, whether the user is performing a specific action or maintaining a specific posture using the generation period of the grouped discharge calculated from the autocorrelation function And a second determination step for determining
If it is determined in the second determination step that the user is performing a specific movement, the actuator is driven to support the voluntary movement by eliminating the influence of the involuntary movement,
When it is determined in the second determination step that the user maintains a specific posture, the actuator is stopped to restrict the movement of the body part.   The motion support system according to claim 1, wherein acceleration data acquired by an acceleration sensor attached to the body part is used when making the determination in the first determination step.   The motion support system according to claim 1, wherein joint angle data of the joint is used when making the determination in the second determination step. The threshold value is determined in a range not less than the average value of the correlation coefficient when no tremor occurs in the user and not more than the average value of the correlation coefficient when tremor occurs in the user. The operation support system according to claim 2. An average value of the occurrence period of the grouping discharge is calculated, in the range not less than the average value when the user is performing a specific motion, and not more than the average value when maintaining a specific posture, The operation support system according to claim 3, wherein the threshold value is determined. A first determination step of determining whether tremor has occurred in the user using a correlation coefficient calculated from an autocorrelation function based on a myoelectric signal obtained from a biological signal sensor;
If it is determined that tremor has occurred, whether the user is performing a specific action or maintaining a specific posture using the generation period of the grouped discharge calculated from the autocorrelation function A second determination step for determining
If it is determined in the second determination step that the user is performing a specific motion, a command signal for driving the actuator is generated so as to support the voluntary movement by eliminating the influence of the involuntary movement. ,
When it is determined in the second determination step that the user maintains a specific posture, the actuator is stopped to generate a command signal that regulates the movement of the body part. Voluntary movement recognition device.   The voluntary movement recognition apparatus according to claim 6, wherein acceleration data acquired by an acceleration sensor attached to the body part is used when making the determination in the first determination step.   The voluntary movement recognition apparatus according to claim 6, wherein joint angle data of the joint is used when performing the determination in the second determination step. The threshold value is determined in a range not less than the average value of the correlation coefficient when no tremor occurs in the user and not more than the average value of the correlation coefficient when tremor occurs in the user. The voluntary movement recognition apparatus according to claim 7. An average value of the occurrence period of the grouping discharge is calculated, in the range not less than the average value when the user is performing a specific motion, and not more than the average value when maintaining a specific posture, The voluntary movement recognition apparatus according to claim 8, wherein a threshold value is determined. A first determination step of determining whether tremor has occurred in the user using a correlation coefficient calculated from an autocorrelation function based on a myoelectric signal obtained from a biological signal sensor;
If it is determined that tremor has occurred, whether the user is performing a specific action or maintaining a specific posture using the generation period of the grouped discharge calculated from the autocorrelation function A second determination step for determining
If it is determined in the second determination step that the user is performing a specific motion, a command signal for driving the actuator is generated so as to support the voluntary movement by eliminating the influence of the involuntary movement. ,
If it is determined in the second determination step that the user is maintaining a specific posture, the actuator is stopped to generate a command signal that regulates the movement of the body part. Voluntary movement recognition program characterized by functioning.

Images