JP2011121152A - Motion assisting device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a motion assisting device for assisting motion by applying appropriate torque to a joint of a human body. <P>SOLUTION: The motion assisting device includes a memory module, a sensor for measuring an angle of the joint of the human body, a brace having a motor applying the torque to a specified joint, and a controller. The memory module stores a known trajectory group of n numbers of the joint angles from a first point in time to a second point in time, and known s-dimensional time-series data corresponding to respective known trajectories. The controller obtains a measured n-dimensional trajectory on time-series data of the n numbers of the measured joint angles, and calculates measured s-dimensional time-series data corresponding to the measured trajectories, where n>s. The controller compares the respective known time-series data till a middle point in time with the measured time-series data and identifies the known time-series data nearest to the measured time-series data to control the motor to allow the angle of the specified joint to follow the trajectory of the specified joint in the known trajectory corresponding to the identified known time-series data after a point in time set after the middle point in time. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a motion assisting device that assists human motion by applying torque to a human joint.

  Research has been made on motion assist devices that apply torque to human joints to assist human motion. Patent Document 1 discloses one such operation assisting device. The motion assisting device of Patent Document 1 includes a device having an actuator that applies torque to a human joint, and a database in which a plurality of minimum units (phases) constituting a human motion pattern are classified according to predetermined reference parameters. ing. For example, a joint angle is selected as the reference parameter. The motion assisting device measures a physical quantity corresponding to the reference parameter and compares the measurement result with the reference parameter to estimate the phase of the motion currently being performed by the person. That is, the motion assisting device determines the phase having the reference parameter closest to the measurement result as the estimated value of the motion pattern performed by the person. The motion assisting device assists human motion by applying torque to the human joint so as to realize the estimated phase.

JP 2005-95561 A

  A motion assisting device that assists human motion is required to apply an appropriate torque to a human joint for each motion pattern. The “phase” in Patent Document 1 is considered to be an operation pattern. In order to appropriately assist human movements, it is necessary to appropriately estimate the patterns of movements performed by humans. The present invention provides a motion assisting device that can appropriately estimate a motion pattern performed by a person and apply an appropriate torque to a human joint based on the estimated motion pattern.

  The human movement pattern is determined by the joint angle trajectory. Here, “trajectory” may be rephrased as “time-series data” or “change over time”. Therefore, the following procedure can be considered as a method of estimating a human motion pattern. First, n-joint angle trajectory data (known trajectory) realized in accordance with a predetermined motion of a person is stored for a plurality of different motions. Next, the joint angle during which a person is performing some motion is measured in real time. The known trajectory closest to the measured joint angle trajectory data (measured trajectory) is selected from the stored known trajectory group. The motion pattern represented by the selected known trajectory is determined as an estimated value of the motion pattern currently being performed by the person.

  However, the above method cannot always estimate the motion pattern accurately. There are two main factors. The first factor is due to the fact that the motion assisting device is intended for a person who cannot completely achieve the target motion pattern by himself. A joint that requires assistance may not move along the intended joint angular trajectory without assistance. That is, the trajectory data (measurement trajectory) of the joint angle group detected by the sensor does not necessarily coincide with the known trajectory that should be consistent. The second factor is due to the fact that each time an action performed by a person is repeated slightly changes. For example, in a walking motion, the trajectory drawn by the joint angle is slightly different for each walking cycle. Not only the walking motion but also the same motion often occurs that the trajectory of the joint angle slightly varies with each repetition. Therefore, for example, when an n-dimensional measurement trajectory composed of n joint angle trajectories is compared with each of a plurality of previously stored n-dimensional known trajectories, the measurement trajectory varies every repetition. Due to this, there is a possibility that a known known trajectory may be selected.

  The novel technique disclosed in this specification reduces the influence of the above two factors that hinder proper estimation of motion patterns. One technique for this is to represent a series of motion patterns in a known trajectory that includes two time spans, a preceding time span that does not require assistance and a subsequent time span that requires assistance, and is closest to the measurement trajectory between the preceding time spans. Identify known trajectories. In a preceding time span that does not require assistance, the measured trajectory should be close to the known trajectory corresponding to the motion pattern intended by the person. One of the techniques disclosed in this specification increases the possibility of identifying an appropriate known trajectory by comparing a measured trajectory with a known trajectory in a preceding time span that does not require assistance. That is, the influence of the first factor described above is reduced by such a technique.

  Another one of the new technologies disclosed in this specification adopts a variable reduced (reduced in dimension) from n dimensions by principal component analysis when identifying a known orbit closest to an n-dimensional measurement orbit. To do. Principal component analysis is a well-known technique in multivariate analysis, and provides a conversion function that converts an s-dimensional objective variable smaller than n while inheriting features represented by n-dimensional explanatory variables as much as possible. The reduction in dimensions by principal component analysis suppresses the influence of local fluctuations in n-dimensional explanatory variables, and gives an s-dimensional objective variable vector that well represents the global behavior. One of the techniques disclosed in this specification applies such a principal component analysis technique to an operation assisting device.

  One embodiment of the motion assist device disclosed in the present specification includes a storage module, a sensor, a brace, and a controller. The storage module stores the known trajectories from the first timing to the second timing of the angles of the n human joints realized in accordance with a predetermined motion of the human for each of a plurality of different motion groups. ing. Further, the storage module stores a known time-series data group of s-dimensional objective variable vectors obtained by inputting values at each time of n-dimensional known trajectories into a conversion function described later. Here, the conversion function is a function based on principal component analysis that outputs an s-dimensional objective variable vector smaller than n when n joint angles are input as explanatory variables. The conversion function is determined in advance by experiments or analysis.

  The sensor measures the angle of n human joints. The brace has an actuator that applies torque to a specific joint among n human joints. The controller controls the actuator. The controller executes measurement trajectory acquisition processing, measurement time series data calculation processing, recently known time series data identification processing, and tracking control processing. In the measurement trajectory acquisition process, the controller controls n actuators so as not to apply torque to the specific joint during the preceding time span, and includes n-series angle time series data measured by the sensor. Get the measurement trajectory of a dimension. Here, the preceding time span means a period between the first timing and an intermediate timing set after the first timing and at a timing earlier than the second timing.

  In the measurement time series data calculation process, the controller calculates the measurement time series data of the s-dimensional objective variable vector in the preceding time span obtained by inputting the value at each time of the n-dimensional measurement trajectory to the conversion function. In the recently known time series data specifying process, the controller compares the data in the preceding time span of each known time series data with the measured time series data, and specifies the known time series data closest to the measured time series data based on the comparison result. . In the follow-up control process, after the auxiliary start timing set after the intermediate timing, the controller determines that the measurement angle of the specific joint is in the specific joint trajectory included in the known trajectory corresponding to the specified known time series data. The actuator is controlled so as to follow. The follow-up control may employ a control law based on classical control theory or a control law based on modern control theory. Here, the auxiliary start timing is preferably substantially equal to the intermediate timing.

  The novel motion assisting device disclosed in the present specification sets a preceding time span in advance, and by applying time series data of n joint angles by a sensor without applying torque to a human joint during the preceding time span. Obtain a configured n-dimensional measurement trajectory. During the period when the measurement trajectory is acquired, no torque is applied to the specific joint of the person. In other words, the preceding time span is set in advance to a period in which a person can continue the intended operation without assistance from the device. Therefore, during the preceding time span, a person can operate smoothly without assistance. With such a configuration, the first factor is suppressed.

  Furthermore, the motion assisting device disclosed in this specification specifies a known trajectory closest to the measurement trajectory from among a plurality of n-dimensional known trajectory vectors. At the time of identification, the controller does not directly compare the n-dimensional known trajectory and the n-dimensional measurement trajectory, but instead of each of the known trajectory and the measurement trajectory, the s (n> s) dimension obtained by reducing the n-dimensionality Compare the time series data of the objective variable vectors.

  The stored known time-series data and measured time-series data reflect the overall movement of the n joints while having a lower dimension than the number n of joints representing the motion pattern. Therefore, the locus drawn by each known time-series data differs depending on the difference in operation pattern. Here, the “trajectory drawn by the known time series data” is a trajectory drawn by the known time series data of the s-dimensional objective variable vector over time in the s-dimensional space.

  According to the motion assisting device described above, since the comparison is performed using the reduced-order variable, it is possible to suppress the influence of the local variation of the time series data of the n joint angles on the comparison result. A known trajectory is selected that has a global behavior that most closely resembles the measurement trajectory. In such a configuration, it is not necessary to compare individual joint angles of n dimensions, and the influence of the second factor described above can be suppressed.

  The controller preferably cancels the follow-up control when the difference between the measured time series data and the identified known time series data is greater than a threshold value. The case where the difference between the measured time series data and the identified known time series data is larger than the threshold value means that the difference between the value on the known trajectory and the value on the measurement trajectory at the same timing is large. That is, such a configuration prevents the specific joint from being largely moved when the measurement angle of the specific joint and the target value of the specific joint included in the known trajectory are greatly different. The measurement angle means a joint angle measured by a sensor.

  ADVANTAGE OF THE INVENTION According to this invention, the operation | movement assistance apparatus which can estimate appropriately the pattern of the operation | movement which the person is performing, and can provide an appropriate torque to a human joint based on the estimated operation pattern is realizable.

The schematic diagram of the operation | movement assistance apparatus of an Example is shown. The block diagram of the controller of a movement assistance apparatus is shown. A schematic graph of the joint angle trajectory is shown. The flowchart figure of the process which a controller performs is shown.

  An operation assisting device 100 according to an embodiment will be described with reference to the drawings. In this embodiment, it is assumed that the user cannot move the left knee joint freely. The motion assisting device 100 according to the present embodiment is a device that assists the user's walking motion by applying an appropriate torque to the user's left knee joint. Hereinafter, the left knee joint that the motion assisting device 100 assists may be referred to as a specific joint. FIG. 1 shows a schematic diagram of the motion assisting device 100 in a state worn by the user. 1A shows a front view, and FIG. 1B shows a side view.

  The motion assisting device 100 includes a leg brace 12 that is worn along the left leg of the user, and a controller 40. The leg brace 12 is fixed to the user's waist by a support link 30.

  The leg brace 12 will be described in detail. The leg brace 12 is attached to the outside of the left leg from the user's thigh to the lower leg. The leg orthosis 12 has a multi-link mechanism having an upper link 14L, a lower link 16L, and a sole link 18L. The upper end of the upper link 14L is swingably connected to the support link 30 via the waist joint 20aL. The lower link 16L is swingably connected to the upper link 14L by a knee joint 20bL located outside the knee. The sole link 18L is swingably coupled to the lower link 16L by an ankle joint 20cL located outside the user's ankle. The upper link 14L is fixed to the user's thigh with a belt. The lower link 16L is fixed to the user's lower leg with a belt. The sole link 18L is fixed to the user's sole with a belt. The belt for fixing the sole link 18L is not shown. The support link 30 is fixed to the trunk (waist) of the user.

  When the user wears the leg brace 12, the waist joint 20aL, the knee joint 20bL, and the ankle joint 20cL are positioned coaxially with the pitch axis of the user's hip joint, the pitch axis of the knee joint, and the pitch axis of the ankle joint, respectively. . Each link of the leg brace 12L can swing according to the movement of the user's left leg. Each joint is attached with an encoder 21 for measuring a joint angle (an angle between links). The encoder 21 of the waist joint 20aL measures the angle of the user's left leg hip joint. The encoder 21 of the knee joint 20bL measures the angle of the user's left knee joint. The encoder 21 of the ankle joint 20cL measures the angle of the user's left ankle joint. Hereinafter, the encoder group 21 attached to each joint may be collectively referred to as a leg sensor 21. The leg sensor 21 included in the leg brace 12 corresponds to a sensor that measures the angle of the user's joint.

  A motor 32 (actuator) is attached to the knee joint 20bL. The motor 32 is located outside the user's knee joint. The motor 32 can relatively swing the upper link 14L and the lower link 16L. That is, the motor 32 can apply torque to the user's left knee joint. The motor 32 includes a clutch 33. When the clutch 33 is released, the torque transmission path from the motor 32 to the left knee joint is interrupted. In other words, when the clutch 33 is released, the knee joint 20bL can swing. When the clutch 33 is released, the user can freely swing the left knee joint.

  Reference numeral 40 denotes a controller that controls the motor 32. FIG. 2 shows a block diagram of the controller 40. The structure of the controller 40 will be outlined. The controller 40 incorporates a storage module 52 and stores various data therein. The controller 40 includes an estimation module 54 and a servo controller 56. The controller 40 includes a CPU as hardware, and the functions of the estimation module 54 and the servo controller 56 are realized by software and a CPU.

  The storage module 52 stores a plurality of different known trajectories. The known trajectory is a set of time-series data of the angles of the n human joints realized in accordance with a predetermined motion of the human. The storage module 52 further stores time series data (known time series data) of objective variable vectors obtained by converting the data of each known trajectory with a conversion function in association with each known trajectory. The known trajectory and the known time series data will be described in detail later.

  The estimation module 54 compares the time series data (measurement trajectory) of the joint angles measured by the encoder 21 (leg sensor) with each of the known trajectory groups stored in the storage module 52, and the known trajectory closest to the measurement trajectory. Select. A specific method of selection will be described later. The selected known trajectory is output to the servo controller 56. The servo controller 56 controls the motor 32 so that the measurement angle of the user's left knee joint follows the knee joint trajectory included in the known trajectory. The estimation module 54 estimates the action being performed by the user and selects a known trajectory for realizing the estimated action. The servo controller 56 controls the motor 32 so as to follow the operation (estimated known trajectory) in which the measurement angle of the user's knee joint is estimated. In this way, the motion assisting device 100 estimates the motion being performed by the user, and applies torque to the knee joint so as to assist the estimated motion.

  The known trajectory stored in the storage module 52 will be described. The known trajectory is a set of time-series data of angles of three human joints (hip joint, knee joint, and ankle joint) realized in accordance with a predetermined walking motion of the person. FIG. 3 shows an example of one known trajectory. The known trajectory of the present embodiment is composed of time series data of the hip joint, knee joint, and ankle joint of the left leg over a time span TS of one cycle of walking. The start point of the time span TS is represented by timing tA, and the end point of the time span TS is represented by timing tC.

  A symbol θhr in FIG. 3 indicates time series data of the hip joint included in the known trajectory. A symbol θkr indicates time series data of the knee joint included in the known trajectory. The symbol θar indicates time series data of the ankle joint included in the known trajectory. The graph of FIG. 3 schematically shows changes with time of the joint angles θhr, θkr, and θar, and is different from the actual graph of the joint angles. Timing tA corresponds to the start point of one cycle of walking, and timing tC corresponds to the end point of one cycle of walking. That is, each known trajectory includes data over a time span TS from the start timing tA to the end timing tC of one cycle of walking.

  The storage module 52 stores a known trajectory for each of a plurality of different operation groups. For example, one known trajectory represents a time-series data group of joint angles realized with a walking motion having a larger step than the known trajectory in FIG. 3, and the other known trajectory is more than the known trajectory in FIG. Represents a time-series data group of joint angles realized with a walking motion with a small stride. The known trajectory is prepared in advance. In the description of the present embodiment, three types of known trajectories are illustrated, but the number of known trajectories may be any number.

  The storage module 52 stores time series data (known time series data) of objective variable vectors obtained by converting the data of each known trajectory with a conversion function in association with the known trajectory. The conversion function will be described. The conversion function used in the present embodiment is a function that outputs an s-dimensional objective variable vector smaller than n when the angles of n joints are input as explanatory variables. In this embodiment, n = 3 and s = 2. Each known trajectory data is composed of time series data (θhr (t), θkr (t), and θar (t)) of the angles of the three joints of the left leg hip joint, knee joint, and ankle joint. ing. The conversion function is a function that outputs known time-series data (Ar (t), Br (t)) of a two-dimensional objective variable vector when values at the respective timings of these three time-series data are input as explanatory variables. is there. The conversion function used in this embodiment is shown in (Equation 1).

  (Equation 1) is obtained from the time series data (θhr (t), θkr (t), and θar (t)) of the angles of the three joints using the principal component analysis method. Ar (t), Br (t)) is obtained. That is, the known time series data Ar (t) and Br (t) of the objective variable vector of (Equation 1) correspond to the main components. In (Equation 1), a1, a2, a3, b1, b2, and b3 are weight coefficients of each joint with respect to the principal component, and are coefficients given in advance. The known time-series data (Ar (t), Br (t)) of the two-dimensional objective variable vector obtained by this conversion function is two-dimensional, but three explanatory variables, that is, time-series data of known trajectories. The characteristics indicated by the whole of (θhr (t), θkr (t), and θar (t)) are well reflected. The storage module 52 stores the known time series data of the objective variable vector obtained by the above (Equation 1) for each known trajectory.

  FIG. 4 shows a flowchart of processing executed by the controller 40. The controller 40 starts the process of the flowchart of FIG. 4 every time one cycle of walking starts. That is, the controller 40 stores a start timing tA and an end timing tC of a known trajectory for one walking cycle, and has a timer for measuring elapsed time. The controller 40 also stores a preset intermediate timing tB. The intermediate timing tB is set between the start timing tA and the end timing tC. In other words, the intermediate timing tB is set to a timing earlier than the end timing tC. The usage of the intermediate timing tB will be described later. The start timing tA corresponds to the first timing, and the end timing tC corresponds to the second timing. Hereinafter, the time span TA from the start timing tA to the intermediate timing tB is referred to as a preceding time span TA, and the time span TB from the intermediate timing tB to the end timing tC is referred to as a subsequent time span TB.

  When the process of FIG. 4 is started at the start timing tA, the controller 40 first releases the clutch 33 and shuts off the power transmission path between the motor 32 and the knee joint 20bL (S2). By releasing the clutch 33, the knee joint 20bL becomes rotatable. The controller 40 maintains the release of the clutch 33 during the preceding time span TA from the start time tA to the intermediate timing tB. In other words, the process of releasing the clutch 33 corresponds to the controller 40 controlling the motor 32 so as not to apply torque to the left knee joint (specific joint) during the preceding time span TA.

  The controller 40 starts the timer as the process starts, and acquires the angle of each joint measured by the encoder 21 until the elapsed time from the start timing tA reaches the preceding time span TA (S4, S6: YES). Hereinafter, the measurement angle of the hip joint is represented by the symbol θhs (t). Similarly, the measurement angle of the knee joint is represented by the symbol θks (t), and the measurement angle of the ankle joint is represented by the symbol θas (t). The time series data of the measurement angles θhs (t), θks (t), and θas (t) of the hip joint, knee joint, and ankle joint correspond to the measurement trajectory.

  When the elapsed time from the start timing tA reaches the preceding time span TA (S6: NO), the controller 40 adds the measured angles θhs (t), θks (t), and θas ( t) is input to calculate time-series data of a two-dimensional objective variable vector (S8). A conversion function in which the measurement angles θhs (t), θks (t), and θas (t) are substituted is shown in (Expression 2).

  Similar to (Equation 1), (Equation 2) is two-dimensional by principal component analysis from time series data (θhs (t), θks (t), and θas (t)) of measurement angles of three joints. This is an equation for obtaining time series data (As (t), Bs (t)) of the main components. The time series data (As (t), Bs (t)) of the main component obtained by (Equation 2) corresponds to the measurement time series data. That is, the controller 40 calculates the measurement time series data from the data of the measurement trajectory using the conversion function. As described above, the controller 40 controls the actuator (clutch) so as not to apply torque to the user's knee joint (specific joint) during the preceding time span TA, while measuring the hip joint measured by the encoder 21. A three-dimensional measurement trajectory constituted by time-series data of knee joint and ankle joint angles is acquired.

  Next, the controller 40 calculates the difference D between the known time series data and the measured time series data over the period of the preceding time span TA (S10). A formula for calculating the difference D is shown in (Expression 3).

  The difference D represented by (Expression 3) corresponds to the mean square error between the known time series data and the measured time series data. The controller 40 calculates the difference D for each of a plurality of stored known time series data. Next, the controller 40 specifies the minimum difference D from the plurality of differences D. The controller 40 compares the magnitude of the minimum difference Dmin with a predetermined threshold value (S12). The threshold value is determined in advance by experiments or the like. When the minimum difference Dmin is smaller than the threshold (step S12: YES), the controller 40 specifies known time series data (recent known time series data) corresponding to the minimum difference Dmin, and is known corresponding to the latest time series data. A trajectory (recently known trajectory) is specified (S14). The estimation module 54 executes the processes in steps S4 to S14. The estimation module 54 outputs the selected recently known trajectory to the servo controller 56. Recently known time-series data corresponds to known time-series data closest to the measured time-series data among a plurality of stored known time-series data. The recently known trajectory corresponds to the known trajectory closest to the measurement trajectory among the stored known trajectories. Each joint angle trajectory after the subsequent time span TB described in the recently known trajectory corresponds to an estimated value of the user's walking pattern after the intermediate timing tB.

  Here, the known time-series data (Ar (t), Br (t)) is data over the time span TS for one walking cycle. To calculate the difference D, from the start timing tA to the intermediate timing tB. Note that only known time series data during the previous time span TA is used. That is, the controller 40 compares the data in the preceding time span of each stored known time series data with the measured time series data, and specifies the latest known time series data closest to the measured time series data.

  The estimation module 54 outputs a recently known trajectory at an intermediate timing tB. The recently known trajectory output by the estimation module 54 includes time-series data up to the end time tB of the time span TA of one walking cycle. The output time series data of each joint angle in the subsequent time span TB from the intermediate timing tB to the end timing tC in the recently known trajectory corresponds to the estimated value of the operation to be performed by the user.

  The servo controller 56 that recently received the known trajectory starts the follow-up control. Note that the timing for starting the follow-up control is after the intermediate timing tB. Most preferably, the follow-up control is started at an intermediate timing.

  The servo controller 56 outputs a message notifying the user of the start of follow-up control. The message may be a sound (buzzer sound), or may be a lighting of a light provided in a range visible to the user (that is, light). By outputting a message, the user can be notified of the start of control.

  At the start of the follow-up control, the servo controller 56 first engages the clutch 33 and prepares the motor torque to be transmitted to the user's knee joint (S16). Next, the servo controller 56 determines that the knee joint measurement angle θks (t) is the knee joint (specific joint) angle trajectory (time-series data) θkr (t) included in the recently known trajectory output by the estimation module 54. The motor 32 is controlled so as to follow (S20). More specifically, the servo controller 56 follows the motor 32 so that the measurement angle θks (t) of the knee joint follows the time-series data θkr (t) of the knee joint angle in the succeeding time span TB of the recently known trajectory. Control. A formula for calculating the motor command torque T (t) output from the servo controller 56 is shown in (Equation 4).

  In (Equation 4), the coefficient P represents a proportional gain. When the motor 32 is driven by the motor command torque T (t), torque is applied to the user's knee joint so that the knee joint measurement angle θks (t) follows the target value θkr (t). (Equation 4) represents a very simple control law. Instead of (Equation 4), PD control, PID control, or any other feedback follow-up control law may be employed.

  The servo controller 56 continues the follow-up control during the subsequent time span TB (S22: YES, S20). When the end timing tC is reached, that is, when the elapsed time from the start of control reaches the subsequent time span TB (S22: NO), the controller 40 prepares for assistance for the next walking cycle and performs the processing of the flowchart of FIG. Start again.

  On the other hand, when the minimum difference Dmin is larger than the threshold value in step S12 (step S12: NO), the servo controller 56 cancels the follow-up control (S24). At this time, the controller 40 releases the clutch 33 of the motor 32 and frees the user's knee joint.

  The advantages and points to be noted of the motion assisting device 100 will be described. The motion assisting device 100 compares the measured time series data in the preceding time span TA of one walking cycle TS with each of the stored known time series data groups, and recently known time series data closest to the current user's action. Is identified. The known trajectory corresponding to the identified recent time series data corresponds to the estimated value of the action performed by the user. In the case of the present embodiment, the preceding time span TA is set to the free leg period in which the leg to be assisted is floating. Since the load is small during the swing leg period, it is not necessary to apply auxiliary torque to the knee joint. The motion assisting device 100 estimates the user's motion pattern in the preceding time span TA where no assist torque is applied.

  When selecting the known trajectory closest to the measurement trajectory, the motion assisting device of the embodiment uses s (2) smaller than n based on principal component analysis instead of the angle time-series data group of n (three) joints. Use time-series data of objective variable vectors with reduced dimensions. The objective variable vector has a lower dimension than that of the original trajectory, but well represents the global behavior of the original trajectory. By using such reduced time series data, the influence of local measurement angle data fluctuations when using an n-dimensional raw orbit (joint angle orbit) is suppressed, and the measurement time series is reduced. Known time series data close to the data can be appropriately identified from a plurality of known time series data.

  The servo controller 56 controls the motor 32 so that the measurement angle of the human knee joint (specific joint) follows the change over time of the angle of the specific joint included in the known trajectory corresponding to the selected known time series data. To do. As described above, since the positional relationship between the timing of the measurement angle and the timing on the stored angle time-series data is specified, the target value of the specific joint at each control timing is also specified.

  The motion assisting device of the embodiment is a device that assists the user's walking motion. The novel technique disclosed in the present specification can be applied not only to a device that assists walking, but also to a motion assist device that assists the movement of an arm. For example, the technology disclosed in the present specification is also preferably applied to an operation assisting device that assists repetitive work performed by factory workers.

  Specific examples of the present invention have been described in detail above, but these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and changes of the specific examples illustrated above. The technical elements described in this specification or the drawings exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. In addition, the technology exemplified in this specification or the drawings can achieve a plurality of objects at the same time, and has technical usefulness by achieving one of the objects.

12, 12L, 12R: Leg brace 21: Encoder (leg sensor)
32: Motor (actuator)
33: Clutch 40: Controller 52: Storage module 54: Estimation module 56: Servo controller 100: Operation assist device

Claims (2)

It is a movement assist device that assists human movement by applying torque to human joints,
The known trajectories from the first timing to the second timing of the n joint angles realized in accordance with a predetermined motion of a person are stored for each of a plurality of mutually different motion groups, and n When the angle of the joint is input as an explanatory variable, a transformation function based on principal component analysis that outputs an s-dimensional objective variable vector smaller than n is input to a value at each time of an n-dimensional known trajectory. A storage module storing a known time series data group of objective variable vectors;
a sensor for measuring the angle of n human joints;
a brace having an actuator for applying torque to a specific joint of n human joints;
A controller for controlling the actuator,
The controller
N measured by the sensor while controlling the actuator so that torque is not applied to the specific joint in the preceding time span between the first timing and the intermediate timing set earlier than the second timing To obtain an n-dimensional measurement trajectory composed of time-series data of the joint angles of
Calculating measurement time series data of an s-dimensional objective variable vector obtained by inputting a value at each time of an n-dimensional measurement trajectory into the conversion function;
Compare the data in the previous time span of each known time series data with the measured time series data, identify the known time series data closest to the measured time series data,
After the auxiliary start timing set after the intermediate timing, the actuator follows the specific joint trajectory included in the known trajectory corresponding to the specified known time-series data so that the measurement angle of the specific joint follows. A motion assisting device characterized by controlling.   The motion assisting device according to claim 1, wherein the controller cancels the tracking control when a difference between the measured time series data and the identified known time series data is larger than a threshold value.

Images