KR20120107928A - Implementing a stand-up sequence using a lower-extremity prothesis or orthesis - Google Patents

Abstract

Knee supports or prostheses may be automatically used when appropriate to initiate a standing sequence based on the position of a person's knee relative to the person's ankle while the person is in a sitting position. When the knee is moved to a position in front of the ankle, at least one actuator of the support or prosthesis is operated to assist in raising the person from sitting position to standing position.

Description

IMPLEMENTING A STAND-UP SEQUENCE USING A LOWER-EXTREMITY PROTHESIS OR ORTHESIS

This application claims the priority of US Provisional Application No. 61 / 238,305, filed August 31, 2009.

The present invention relates generally to prosthetic prostheses, supports and exoskeleton devices and their components and methods for controlling them.

More than 100,000 people in the United States lose their legs by amputation each year. Hundreds of thousands of people are suffering from this breakdown worldwide. In addition, there are about 700,000 people in the United States each year who survive accidents that often cause various leg problems that restrict walking. Until recently, prosthetic and prosthetic systems had a positive work of non-conservation on each step needed to achieve economical walking behavior even on flat terrain—not to mention non-flat surfaces such as stairs. Employing nearly passive or low-power mechanisms that cannot perform conservative positive work.

Recognizing the requirements of the prosthetic prosthesis, the support, and the exoskeleton helps to understand the conventional biomechanics associated with the gait cycle of the subject. In particular, the function of the ankle of a person under sagittal plane rotation is described below for different movement conditions.

The mechanical properties of conventional passive ankle / foot prostheses (AFP), such as Ossur Flex-Foot®, remain essentially constant throughout the life of the device. US Patent Publication Application US 2006/0249315 (“'315 Application”) represents a significant improvement over these conventional AFPs. The '315 application, which is incorporated herein by reference in its entirety, can improve performance by dividing the walking cycle into five phases and optimizing the mechanical properties of the device independently for each of these five phases. It was recognized.

1A is a schematic diagram of different times of an object's walking cycle on the ground. The walking cycle is typically defined as starting with a heel strike of one foot and ending at the next heel landing of the same foot. The walking cycle is divided into two phases: the stance phase (approximately 60% of the walking cycle) and the subsequent swing phase (approximately 40% of the walking cycle). The flap represents a part of the walking cycle when your feet are off the ground. The standing starts at the heel landing when the heel contacts the ground and ends at the toe-off when the same foot falls off the ground. The stand is separated from the toe-lift when the same foot is raised from the ground. The standing phase is divided into three sub-phases: controlled plantar flexion (CP), controlled dorsiflexion (CD) and forceful plantar flexion (PP).

CP starts at heel landing shown at 102 and ends at foot-flat shown at 106. CP describes the process by which the heel and forefoot contact the ground. The results show that the CP ankle joint behavior is consistent with the linear spring response where the joint torque is proportional to the joint displacement relative to the equilibrium position of the joint position. However, the spring behavior is variable; Joint stiffness is continuously adjusted by the body at every step within the three sub-phases of stance and subsequent swing states.

After the CP period, the CD phase continues until the ankle reaches the state of maximum dorsal flexion, and begins a forceful plantar flexion (PP) as shown at 110. Ankle torque versus position during the CD period is described as a non-linear spring in which the stiffness increases with the rise of the ankle position. The ankle stores the elastic energy needed to propel the body upwards and forwards during the PP period during the CD.

The PP period begins after the CD and ends at the toe-lifting moment shown at 114. During PP, the ankle applies torque in response to the reflex response that propels the body upwards and forwards. The propulsion energy is then released in accordance with the spring energy stored during the CD phase to achieve high plantar flexion forces during subsequent standing. This propulsion behavior is necessary because the motion generated in the PP is more than the negative action absorbed during the CP and CD periods suitable for high walking speeds. The foot rises from the ground during the stray, from the toe-lifting shown at 114 to the next heel-landing shown at 118.

A separate description of ankle-foot biomechanics is presented in FIGS. 1B and 1C because the kinematic and motor patterns at the ankle during stair climbing / descending are different from the walking pattern of the flat. 1B shows ankle biomechanics of a person during climbing stairs. Controlled flexion 1 (CD) begins with a foot strike at the ventral flexion position shown at 130, continuing the flexion until the heel contacts the stair surface at 132. It is referred to as 1). At this time, the ankle can be modeled as a linear spring. The second phase is a forceful plantar flexion 1 (PP 1), which begins at the plantar-contacting moment (when the ankle reaches maximum ventral flexion at 132) and ends when the ventral flexion begins again at 134. The human ankle behaves as a torque actuator to provide extra energy to support the weight.

The third time period is controlled dorsal flexion 2 (CD 2), where the ankle flexes dorsally from 136 to the heel-off. At the CD 2 time, the ankle can be modeled as a linear spring. The fourth and final phase begins at the sole-lifting 136, continues while the foot leaves the stairs, and ends when the toes at the 138 leave the surface to begin the stir ending at 140. 2 (PP 2), acts as a torque actuator parallel to CD 2 to propel the body up and forward.

1C shows human ankle-foot biomechanics during the step down. The stance of stepping down is divided into three sub-phases: controlled dorsal flexion 1 (CD 1), controlled dorsal flexion 2 (CD2), and forceful plantar flexion (PP). CD 1 starts at the foot-landing shown at 150 and ends at the sole-contacting 152. At this time, the ankle of the person can be modeled as a variable damper. In CD 2, the ankle continues to dorsal flexion forward until the maximum dorsal flexion position shown at 154 is reached. Here, the ankle acts as a linear spring, storing energy for CD 2. During the PP, starting at 154, the ankle flexs plantar at 156 until the foot rises from the stairs. In this final PP period, the ankle releases stored CD 2 energy, pushing the body upwards and forwards. After the toe-lift at 156, the foot is controlled and positioned during the stir until the next foot lands at 158.

In the stair climb shown in FIG. 1B, the human ankle-foot can be effectively modeled using a combination of actuator and variable rigidity mechanism. However, in going down the stairs, shown in FIG. 1C, a variable damper needs to be included to model the ankle-foot compound, and the force absorbed by the person's ankle is less than the force released during the step climbing. Much bigger. Therefore, it makes sense to model the ankle as a combination of spring and variable dampers for descending stairs.

Conventional passive prostheses, supports and exoskeletons do not adequately reproduce the biomechanics of the walking cycle. They do not actively adjust impedance and apply reflected torque response; It is not a biomimetic technique because it does not climb or descend stairs or ramps from the ground, nor does it change terrain conditions. Therefore, there is a need for improved prosthetic prostheses, supports and exoskeleton devices, their components and methods for controlling them.

The present invention described herein relates generally to prosthetic prostheses, supports and exoskeleton devices. Representative use cases for various embodiments of the present invention include, for example, increased metabolism, permanent assistance to subjects with permanent limb pathology, or walking for wearers with temporary limb disorders. .

One aspect of the present invention relates to an active support or prosthetic device comprising a thigh member, a lower leg member and a knee joint for connecting the thigh member to the lower member. The apparatus also includes a rotating motor having a motor shaft output, a motor drive transmission assembly coupled to the motor shaft output, and a drive transmission assembly coupled to the output of the motor drive transmission, the output of the drive transmission assembly A portion is coupled to the lower member to apply torque to the knee joint to rotate the lower member relative to the thigh member. The device is also based on at least one sensor having at least one output that can be determined while the position of the knee joint relative to the ankle joint is in a seated position of the wearer of the device, and at least one output of the at least one sensor. To determine when the knee joint moves to a position before the ankle joint. In response to the determination, the controller controls the rotating motor to adjust the impedance, position or torque of the knee joint to assist in raising a person from a seated position to a stand-up position. .

Another aspect of the invention relates to a method of controlling a knee support or prosthesis having at least one actuator. The method comprises the steps of detecting a person's knee position relative to the person's ankle while the person is in a sitting position, determining when the knee is moved to a position in front of the ankle based on the results of the detecting step, And generating an output representative of the result. In response to the output, at least one actuator of the knee support or prosthesis is operated to assist in raising the person from sitting position to standing position.

1A is a schematic diagram of different times of a wearer's walking cycle on a flat surface.
1B is a schematic diagram of the different times of the walk cycle of the wearer climbing the stairs.
1C is a schematic diagram of the different periods of the wearer's walking cycle down the stairs.
2A is a schematic diagram of a method for determining ankle joint, heel, and toe trajectory of a prosthesis, support, or exoskeleton, in accordance with an exemplary embodiment of the present invention.
2B is a plot of experimental data showing ankle joint acceleration during walking.
3 is a schematic diagram illustrating a method for determining foot slope (heel height), in accordance with an exemplary embodiment of the present invention.
4 is a schematic diagram illustrating a method of determining the coordinates of the heel and toe with respect to the ankle joint within a foot frame of reference, in accordance with an exemplary embodiment of the present invention.
5 is a schematic diagram illustrating a method of evaluating a heel vector, in accordance with an exemplary embodiment of the present invention.
6A shows the inertial measurement unit-calculated ankle joint pivot trajectories in different walking situations.
6B illustrates a 2-D structure illustrating the flight trajectory of the ankle joint of the prosthetic device.
FIG. 6C illustrates how a step-slope separator may be installed using an ankle angle attack angle as a trajectory feature to distinguish between ramp and stair walking situations, in accordance with an exemplary embodiment of the present invention.
7A illustrates a method for positioning an ankle joint prior to foot-landing, in accordance with an exemplary embodiment of the present invention.
FIG. 7B illustrates how the method of FIG. 7A can be used to detect the presence of a staircase and the protrusion of a foot on a staircase's stairs, in accordance with an exemplary embodiment of the present invention.
7C illustrates a method for positioning an ankle joint in a ramp walk situation, in accordance with an exemplary embodiment of the present invention.
FIG. 7D illustrates how the method of FIG. 7B is adapted to use optimized impedance, in accordance with an exemplary embodiment of the present invention.
8 shows a method for determining an inertial-referenced spring balance based on the topographic angle at the sole-contact.
FIG. 9 shows the effect of walking speed on ankle torque versus ankle angle and illustrates how push-pull actuator control is applied to a suitably selected parallel elastic element.
10A illustrates a method for controlling a limb device, in accordance with an exemplary embodiment of the present invention.
10B is a schematic diagram of a model-based controller for implementing impedance and torque control in a prosthetic prosthetic device, in accordance with an exemplary embodiment of the present invention.
10C is a schematic diagram of a model-based controller for implementing torque control in a prosthetic prosthetic device, in accordance with an exemplary embodiment of the present invention.
FIG. 10D is a schematic diagram of a mechanical impedance relationship for operating the impedance control performed in FIG. 10A.
FIG. 10E is a schematic diagram illustrating an impedance and reflection relationship for operating the impedance and reflection control performed in FIG. 10B.
10F is a schematic diagram of how a zero moment pivot referenced ground reaction force is used to determine the restoring torque required to stabilize the inverted pendulum motion of a person wearing a prosthetic device.
FIG. 11A is a schematic diagram of the lower leg member, ankle joint and foot member of the ankle prosthesis showing ground reaction force and zero moment pivot. FIG.
11B-11D are schematic diagrams of the components of the ankle prosthesis showing the force and moment relationships among the components required to determine ground reaction force and zero moment pivot.
12A-12B show the biomechanical (F-θ) behavior of flat ankle prostheses as a function of walking speed during forceful plantar flexion.
12C-12D show the effect of foot transition on ground contact length.
12E shows how the velocity-dependent tables for the length of contact attenuation can use normalized ground contact lengths as a means of achieving biomimetic behavior during forceful plantar flexion.
12F shows how the estimated y-component of the zero moment pivot vector changes during a typical walking operation.
12G illustrates a method for applying attenuation factor to the performance of a device, in accordance with an exemplary embodiment of the present invention.
13A is a schematic diagram of a control system approach for a heel landing case, in accordance with an exemplary embodiment of the present invention.
13B is a schematic diagram illustrating a control system of a toe landing case, according to an exemplary embodiment of the present invention.
FIG. 13C illustrates a method for position control applied to an ankle prosthesis (eg, device 1700 of FIG. 17A), in accordance with an exemplary embodiment of the present invention.
14A illustrates a method for using phased terrain adaptation, in accordance with an exemplary embodiment of the present invention.
14B illustrates exemplary impedances for which ankle joint prostheses can be applied to three different gait situations.
15 is a schematic diagram of a limb biomechanical system, in accordance with an exemplary embodiment of the present invention.
16 illustrates a method of posture reconstruction for upper body posture, thigh posture and upper body / body weight-centric posture, in accordance with an exemplary embodiment of the present invention.
17A is a diagram of a prosthetic prosthetic device, in accordance with an exemplary embodiment of the present invention.
FIG. 17B is a view of a portion of the device of FIG. 17A showing a passive parallel elastic element. FIG.
FIG. 17C is a view of a passive parallel elastic element of the apparatus of FIG. 17B.
FIG. 17D is a diagram of a free-body diagram for the passive parallel elastic element of FIG. 17C, in accordance with an exemplary embodiment of the present invention. FIG.
FIG. 17E is a perspective view of the structural element (pyramid) of the apparatus of FIG. 17A, in accordance with an exemplary embodiment of the present invention. FIG.
FIG. 17F is a cross-sectional view of an alternative method for measuring axial force and moment applied to the lower member of FIG. 17A, in accordance with an exemplary embodiment of the present invention. FIG.
17G is a diagram of a method for calculating a planar moment vector and axial force using a circular arrangement of displacement sensors on a printed circuit assembly, in accordance with an exemplary embodiment of the present invention.
17H is a schematic diagram of a state and actuator controller for use with the devices of FIGS. 17A-17G, in accordance with an exemplary embodiment of the present invention.
17I is a schematic diagram of an electrical circuit equivalent of a prosthetic prosthetic device, in accordance with an exemplary embodiment of the present invention.
FIG. 17J is a schematic diagram of the electrical circuit of FIG. 17I including sensor measurements used to control the apparatus. FIG.
18A-18D are illustrations of passive series-elastic members, in accordance with an exemplary embodiment of the present invention.
19A is a view of a prosthetic prosthetic device incorporating a plate tandem spring, in accordance with an exemplary embodiment of the present invention.
19B-19C are diagrams of a prosthetic device using alternative tandem springs, in accordance with an exemplary embodiment of the present invention.
20A is a perspective view of a linear actuator that may be used in various prosthetic prostheses, supports and exoskeleton devices, in accordance with an exemplary embodiment of the present invention.
20B is a cross sectional view of the linear actuator of FIG. 20A.
21 is a perspective view of a linear actuator that may be used in the prosthetic prosthesis, the support and the exoskeleton device, in accordance with an exemplary embodiment of the present invention.
22A is a top view of a limb support or exoskeleton device (wearable robotic knee braces), in accordance with an exemplary embodiment of the present invention.
22B is a side view of the apparatus of FIG. 22A.
22C is a schematic illustration of the interior of the knee joint drive assembly of the devices of FIGS. 22A and 22B.
23A is a schematic diagram of a human balance problem on a sloped slope.
23B is a schematic diagram of acceptable solutions to a balancing problem based on various knee bends by the wearer.
23C is a schematic representation of the body showing how inherent sensing can be used to balance the wearer on a flat surface.
24A-24C are schematic diagrams of a method for balancing a wearer as the wearer wakes up in a chair, in accordance with an exemplary embodiment of the present invention.
25A shows the definition of a movement task.
25B shows the definition of hip impact force.
FIG. 26 shows the biomechanical characteristics of a normal human gait while walking.
27 illustrates a biomechanical mechanism in which quadriceps defects affect walking in the flat.
28 illustrates how the knee device is used to restore normal walking.
29A-29D show the standing sequence of a healthy person.
30A-30D illustrate problems in implementing the same standing sequence for patients with disabilities.
31A-31D illustrate how a knee device is used to assist in a standing sequence of a patient with a disability.

The inventors have recognized that during everyday days, a person's leg is used to perform and adapt to many different behaviors, with everyday walking such as climbing and descending stairs and walking on slopes. Ankle-foot components are most in direct contact with the underlying terrain and therefore require the most force and should exhibit the most terrain-adaptive behavior. The inventors have recognized that the performance of AFPs can be significantly improved by dynamically optimizing the mechanical properties of the device in different ways and dynamically controlling the device in different ways for each of these behaviors.

For example, if a person walks on flat ground, it is better to control the angle of the foot so that the heel is lower than the toe when the foot descends to the ground and touches it. However, when a person climbs up the stairs, it is better to control the angle of the foot so that the toes are lower than the heel when the foot descends and touches the next step.

This application describes various embodiments of AFPs that perform appropriately in each of these different situations by detecting the passing terrain and automatically adapting to the detected terrain. In some embodiments, the ability to control AFPs for each of these situations includes five basic capabilities: (1) the ability to determine the action performed; (2) the ability to dynamically control the properties of the AFP based on the actions performed; (3) the ability to dynamically drive AFP based on the actions performed; (4) determine the irregularities of the terrain texture (e.g., how rough the terrain is, how slippery the terrain is, whether the terrain is rough or smooth, whether there are any obstacles such as rocks in the terrain) and The ability to respond to them with traction control; And 5) mechanical design of the AFP capable of responding to dynamic control and dynamic drive.

The inventors have determined that an exemplary method of determining what action is being performed is to track the trajectory of a point on the lower leg (or shin) between the ankle and knee joints (typically the virtual center of rotation of the ankle joint). FIG. 6A shows shin trajectories corresponding to five different behaviors, with additional ramp trajectories to distinguish between steep and gentle ramps. The system can use this information to find out what actions are being performed by mapping the tracked trajectories to a set of actions.

By observing the trajectory of the lowering (shin), one can distinguish between flat terrain, climbing or descending stairs, or climbing or descending slopes. For example, if the system recognizes a trajectory, it will switch to the appropriate mode to dynamically control (drive) the AFP as previously built for that mode. If the trajectory does not exactly fit within the classification, the AFP controller optimizes the response to minimize the objective function in probabilistic control detection, or fuzzy logic or ad hoc based on the likelihood that the terrain corresponds to the classification. Will apply the control.

One suitable way of tracking the trajectory of the shin is performed by mounting an inertial measurement unit IMU on the front surface of the upper part of the lowering member (shank) and processing the signal output by the IMU. A suitable way to distinguish the various trajectories is to monitor the speed of the angle of attack of the ankle joint. This subject is explained in more detail below.

In addition to dynamically optimizing the mechanical properties and dynamically controlling the device in different ways for each of the different behaviors, the present inventors further fine-tune the control and properties of the AFP based on various parameters of the device's performance. it can be further improved by tuning).

For example, when a person walks slowly (eg, speeds less than 0.9 m per second), performance can be improved by increasing the impedance of the ankle joint to the impedance typically used in walking. Or, if a person walks fast (for example, a speed of 1.7 meters per second), performance can be improved by reducing the impedance of the ankle joint to the impedance typically used in walking.

In addition, if the controller determines that the ankle joint is not responding as expected when passing through normal terrain, the controller may determine the feature, texture, or irregularities of the terrain (eg, how sloppy the terrain is, how slippery the terrain is), It can be taken into account (and change the output of the controller) whether the terrain is rough or smooth, or if there are any obstacles such as rocks in the terrain.

The five abilities mentioned above (i.e. the ability to find out what actions are being performed; the ability to detect the presence of features, textures or irregularities of the terrain; the ability to dynamically control the features of the AFP; the ability to dynamically drive the AFP) And mechanical design of the AFP) are described in more detail below.

Determine the action performed

Inertial Posture and Trajectory Estimation

FIG. 2 illustrates a prosthetic, support or exoskeleton device (eg, FIG. 2) based on the inertial position of the lower member 220 coupled with the ankle joint 200 and the angle between the lower member 220 and the foot member 208. A schematic diagram of a method for determining ankle joint 200, heel 212, and toe 216 trajectories of device 1700 of 17A. Posture is the position and orientation of the coordinate system. The device 1700 includes an inertial measurement unit 204 that is coupled with the lowering member 220. The inertial measurement unit 204 includes a three-axis rate gyro for measuring the angular ratio and a three-axis accelerometer for measuring the acceleration. Positioning the inertial measurement unit above the lowering member 220 correlates measurements of the angular rate and acceleration for all three axes of the lowering member 220. The inertial measurement unit 204 provides a six-degree of freedom estimate of the lower member 220 posture, inertia (see world frame) orientation, and ankle-joint 200 (center of rotation of the ankle-foot).

In some embodiments, the lower member 220 posture is used to calculate the instantaneous position of the knee joint. By using knowledge of the ankle joint 200 angle [theta], the instantaneous posture of the bottom of the foot member 208, including the position of the heel 212 and the toe 216, can be calculated. In turn, this information may be used when the foot member 208 is flat to measure the topographic angle of the plane defined by the axis of rotation of the ankle joint / foot member. Mounting the inertial measurement unit 204 over the lower member 220 has an advantage over other potential positions. Unlike when mounted over foot member 208, mounting of retracting member 220 prevents physical abuse and prevents exposure to water. In addition, by removing the cable tether that would be needed if it was on the foot member 208, mechanical and electrical stability is ensured. As a result, the lowering member 220 is centered in the kinematic chain of the hybrid system (see FIG. 15) which facilitates the calculation of the thigh and upper body posture with minimal additional sensors.

), 위치(

) 및 속도(

)를 계산하기 위해 이용된다.

는 세계 프레임에 대한 발목 관절의 x, y 및 z 축들의 배향을 정의하는 4원법 또는 단위 벡터의 3X3 행렬에 의해 표현될 수 있다. 발목 관절(200) 좌표 프레임은 하퇴 부재(220)에 대한 배향으로 발목 관절 회전축의 중심에 위치되도록 정의된다. 이러한 중앙 지점으로부터, 위치, 속도 및 가속도가 계산될 수 있다. 예를 들어, 발(예를 들어, 뒤꿈치(212) 또는 발가락(216)), 발 부재-발목 관절 배향 변환에 대해서 살펴보면,

는 다음 관계식을 이용하여 위치를 추출하기 위해 이용된다:The inertial measurement unit 204 measures the orientation of the prosthetic prosthesis in the ground-referenced world frame.

), location(

) And speed (

Is used to calculate

Can be represented by a quadratic or 3 × 3 matrix of unit vectors that define the orientation of the x, y and z axes of the ankle joint with respect to the world frame. The ankle joint 200 coordinate frame is defined to be positioned at the center of the ankle joint rotation axis in an orientation with respect to the lower member 220. From this central point, position, velocity and acceleration can be calculated. For example, with respect to foot (eg, heel 212 or toe 216), foot member-ankle joint orientation transformation,

Is used to extract the location using the following relation:

식 1

Equation 1

here,

식 2

Equation 2

Where γ is the inertial lowering member angle,

식 3

Equation 3

Where θ is the ankle joint angle.

In the present embodiment, the inertial measurement unit 204, which includes a three-axis accelerometer and a three-axis rate gyro, has a front surface at the top of the lowering member 220 (eg, as shown in FIG. 17A). Is located on. Need to eliminate the effects of scale, drift and cross-coupling on world-frame orientation, velocity and position estimates introduced by numerical integration of accelerometer and rate gyro signals have.

Zero-speed update

Inertial navigation systems typically employ zero-speed update (ZVUP) periodically by averaging over an extended period of time, typically a few seconds to several minutes. This arrangement of the inertial measurement unit is hardly fixed in the prosthetic prosthesis. However, the sole is the only fixed position and only during the controlled dorsal flexion state of the walking cycle. An exemplary zero-rate update method, which is not affected by this limitation, used in various embodiments of the present invention, is further described below.

)를 디지털화한 후, 발목 관절 가속도(

)는 다음의 강체 동역학(rigid body dynamics) 방정식으로 추출된다:
To solve this problem, the orientation, velocity and position integration of the ankle joint is performed. Inertial Measurement Unit Acceleration (

) And then digitize the ankle joint

) Is extracted from the following rigid body dynamics equation:

식 4*

Equation 4

및

는 관성 측정 유닛 프레임 내에서 각각, 각도 비율 및 각 가속도의 벡터들이며, X는 벡터적을 나타낸다.here,

And

Are the vectors of the angular ratio and the angular acceleration, respectively, within the inertial measurement unit frame, and X represents the vector product.

은 표준 스트랩다운(strapdown) 관성 측정 유닛 적분 방법들을 이용하여 식 1 내지 식 3과 마찬가지로, 풀려진다. In accordance with the following relations known to those skilled in the art, the relations,

Is solved, as in equations 1 through 3, using standard strapdown inertial measurement unit integration methods.

식 5

Equation 5

식 6

Equation 6

식 7

Equation 7

식 8

Equation 8

식 9*

Equation 9

식 10

Equation 10

식 11

Equation 11

식 12

Equation 12

식 13

Equation 13

식 14

Equation 14

은 배향 행렬

과 상호 교환가능하게 이용될 것이다. In formulas 5 to 14, the matrix,

Silver orientation matrix

Will be used interchangeably.

The world frame-referenced ankle joint velocity and position is timely extracted after the timing of the previous zero-velocity update (i-th zero-velocity update) based on:

식 15

Equation 15

식 16

Equation 16

Here, ^ω p _ankle (t = ZVUP (i)) is reset to zero for all i.

Through experiments, using the logs of inertial measurement unit data obtained from an exemplary prosthetic prosthetic device (eg, the prosthetic prosthetic device 1700 of FIG. 17A), the z-acceleration is controlled in the dorsal flexion state. Approximately equal to 1 g (approximately 9.8 m / s ² ) and the change in z-acceleration is a predetermined value (<0.005 g ² ) —the time when the lower leg member 220 is rotated about the fixed ankle joint 200 In the following, it was determined that the inertial measurement unit-referenced acceleration was fast enough at approximately 50.75 seconds and 50.9 seconds (see FIG. 2). In other embodiments of the present technology, an appropriate quiet period may be detected on a portion of the foot. Knowledge of the acceleration, angular rate and angular acceleration of the ankle joint can be combined with knowledge of the detected ankle angle (angle between the foot and lower members), angle ratio and angular acceleration to calculate the acceleration of any point of the foot. have. Some points of the sole can often be used to perform zero speed updates in successive walking cycles. Once this speed is known, the speed of the ankle joint can be calculated later. This rate (not zero) can be used as a reference for which a zero rate update can be performed.

In the prosthetic prosthetic device, the settling period is almost always in a controlled dorsal flexion state, so that zero-speed updates can be performed for every step the wearer takes. During each zero-velocity update, the velocity error contribution from each of the three items is preferably evaluated as the world frame z around the x-axis (the vector aligned with the ankle joint rotation axis during the zero-velocity update in the previous step). Tip of the axis, δθ _x ; tilt of the world frame z-axis around the _y -axis (vector-defined vector of the world-frame vertical (as opposed to the gravity vector) and the world frame x-axis), δθ _y ; And an inertial measurement unit scale, δg, scaled along the vertical axis. The values of these items are used to correct the calculated pose, inertial orientation and previously calculated poses and inertial orientations of the different components of the device (eg, the lowering member 1712 of FIG. 17A).

에 위해 유도될 수 있다. M(t)는 다음의 관계식에 기초한다:While performing orientation, velocity and position integration, a sensitivity matrix, M (t), for velocity error is calculated, wherein the velocity error is a vector of errors,

Can be induced for. M (t) is based on the following relation:

식 17

Equation 17

Here, M (t) is numerically integrated to produce the overall terminal speed sensitivity, M ^* .

식 18

Equation 18

In some embodiments, the vector of errors is expanded to include accelerometer bias offsets if these errors are significant, thereby increasing the number of columns in M (t) and M ^* . In this case, M ^{* -1} takes the form of Penrose pseudo-inverse or optimal innovation gain, K ^* . K ^* can be calculated using standard optimal linear filtering methods. For those of ordinary skill in the art, other items may be included or used without loss of universality.

를 생성했을 α의 값은 다음에 기초하여 결정된다:In the zero-velocity update in step (i), the estimated non-zero ankle joint velocity,

The value of α that would generate is determined based on:

식 19

Equation 19

) 및 중력 크기(g)가 다음에 기초하여 결정된다:Where α is the innovation calibration vector. Non-zero velocity results in part from noise in accelerometers and angular rate measurements, but not all innovation corrections α apply. Instead, the calibration is scaled by the filtering constant (fractional), k, depending on the magnitude of the noise. In this case, the new orientation matrix (

) And gravity magnitude (g) are determined based on:

식 20

Equation 20

식 21

Equation 21

Where O _x (tip) and O _y (tilt) indicate incremental rotations of the tip and tilt around each of the x and y axes, and ZVUP _i ⁺ and ZVUP _i ⁻ indicate the periods after and before ZVUP, respectively. .

Linear estimators may be used to extend the zero velocity update to estimate accelerometer and rate gyro bias offsets. Constant angular alignment errors can be used to estimate the rate gyro bias about the axis (eg, about a given axis). In one embodiment, they are performed by generating linear probabilistic models of accelerometer and rate gyro bias and using zero-velocity update side residuals as inputs to the linear filter applied to these models.

The method described above is a method for continuously updating orientation and apparent gravity magnitudes. The initialization procedure is used in this embodiment when the prosthetic prosthesis is first loaded with force. In this way, the wearer will go one step forward and stop when requested by the device (eg, the vibration code sent by the device or alternative user interface), and then back one step back to its original position. In this process, the steps will be performed on the affected leg (for both amputated patients this calibration will be performed in a continuous manner as selected by the amputated patient). This scale adjustment will apply two ZVUPs, one to initialize the orientation and gravity magnitude, and the second to check the result. This will ensure the integrity of the inertial measurement unit signals, processing and controller communication.

The process described above achieves initialization of the inertial orientation. However, this is of general interest to perform full calibration of the IMU to account for a vector of error sources ε including bias offset, scaling and cross-sensitivity implemented in accelerometer and gyro signals. In manufacturing, a robot or other 6-degree of freedom machine may carry the IMU and apply the reference trajectory continuously as a means of measuring the effect of these error sources. The sensitivity matrix M ( ε ) of the sensed reference trajectories for each of the error sources can be easily calculated by one of ordinary skill in the art. By measuring the detected deviations from the rich set of reference trajectories—typically the deviation of the end-point of each trajectory segment—the vector ε is used to determine that the set of reference trajectories stimulates the influence of each error source. Can be estimated using regression or other linear estimation methods. The inventors have found that reference trajectories including closed paths such as polygons and circles in three orthogonal planes are sufficient to adjust the scale of the entire vector of the error source. These reference trajectories are also used to re-scale the major elements of the vector (accelerometer bias, scale and cross-sensitivity), for example, by walking in a sequence of closed patterns on the horizontal plane and sequentially rotating around the vertical axis. It may be performed by the wearer.

In some embodiments of the present invention, these principles of the method may include, for example, a thigh member and / or upper body of the wearer where the prosthesis, support, or exoskeleton treats or increases the performance of these parts of the wearer's body. The same applies to correct or minimize the accelerometer and rate gyro drift errors associated with the accelerometers and rate gyro located at. In one embodiment, the method relates to an accelerometer signal and a rate gyro signal output by an accelerometer and a rate gyro coupled with the thigh member of the prosthesis or support when the ankle joint is substantially fixed during the prosthetic or support walking cycle. Determining offset values. The method may also include measuring the angle of the lowering member relative to the thigh member. In another embodiment, the method also determines offset values for the accelerometer and rate gyro signals output by the accelerometer and rate gyro coupled with the upper body of the wearer when the ankle joint is nearly stationary during the walking cycle of the prosthesis or support. It includes a step. The method may also include measuring the angle of the thigh member relative to the upper body of the wearer. Therefore, the method can be extended to the thigh member and / or torso of the wearer by performing these measurements as shown in FIG. 16 and relying on linkage constraints and related methods. At the zero velocity update, the coupling device limitation enables propagation of articulation velocity references backwards from the ground-referenced zero velocity of the lowest link in the kinematic chain (eg, the coupling device defining the hybrid human-robot system). do. These velocity references can be used as input to posture realignment and gravity compensation as defined above.

Example ankle joint trajectories and terrain situation ( terrain  context)

가 다음에 기초하여 계산된다: After the inertial measurement unit offset is calculated and corrected (zeroed), the foot-slope β (alternatively referred to as heel height) is determined, for example, as shown in FIG. 3. From the figure it is easy to see β = − (θ + γ) when the wearer is standing in the sole-contacting state on the ground. By averaging over a period of approximately 10 seconds, an accurate estimate of β can be determined. The orientation component of the transformation that then defines the foot as the ankle coordinate system,

Is calculated based on:

식 22

Equation 22

As before, the transform component of this transform will remain zero.

및

는 새로운 발 좌표 시스템 내의 뒤꿈치 및 발가락의 벡터 좌표로서 정의된다. β의 회전 기여도가 이미 포함되었으므로, 이러한 벡터들의 z-컴포넌트는 동일하다. 이러한 벡터들의 x-컴포넌트가 모두 영이라는 점이 가정될 수 있다. 이러한 벡터들은 다음과 같은 형태를 취한다:Once the foot-slope is defined, it is necessary to determine the heel 212 and toe 216 coordinates in the foot coordinate system. In one exemplary method for determining this,

And

Is defined as the vector coordinate of the heel and toe in the new foot coordinate system. Since the rotational contribution of β has already been included, the z-components of these vectors are identical. It can be assumed that the x-components of these vectors are all zero. These vectors take the form:

식 23

Equation 23

식 24

Formula 24

Where z ₀ defines the z-coordinate of the bottom of the foot (shoes).

4 is a schematic diagram of a method for determining heel 212 and toe 216 coordinates for ankle joint 200 in a reference foot frame, in accordance with an exemplary embodiment of the present invention. In step (1) of the foot scale adjustment method defined in FIG. 4, the y-coordinate of the ankle joint 200 is referenced to ground reference (eg, a salient feature on a boundary line, rug or linoleum surface on a pavement). Aligned. We arbitrarily define these ground references to be the origin of the world coordinate system. In mathematical notation, this alignment takes the form:

식 25

Equation 25

는 단계들(2 및 3)에서 일어나는 움직임에 대한 개시 위치이다. 단계(2)에 있어서, 발가락(216)은 지면 참조 상에 위치된다. 수학적 표기법에 있어서, 이러한 정렬은 다음과 같은 형태를 취한다:here,

Is the starting position for the movement occurring in steps 2 and 3. In step 2, toes 216 are positioned on the ground reference. In mathematical notation, this alignment takes the form:

식 26

Equation 26

or

식 27

Equation 27

Similar relationships are determined during the alignment of step (3). If the above equations are solved independently, two different estimates of z ₀ are obtained. By combining the two constraints into one, the least squares estimate of y _heel , y _toe and z _o can be obtained.

The heel 212 and toe 216 calibration method described above comprises a series of steps, first of which a new pair of feet / shoes are worn. Such calibration can be performed, for example, at the prosthetic office.

)로서 설명된다. 다음과 같이, 두 개의 통계학적으로 구별되는 하퇴 부재(220)(γ_i,) 및 발목 관절(200) 각도(θ_i,) 위치에 대응하는, 두 개의 발목 위치 측정이 요구된다:In another exemplary method, the heel and toe vectors are then calculated. As shown in FIG. 5, the ankle joint 200 tracks the arc 500 at an initial stance between foot-landing and plantar-contacting. The radius and orientation (midpoint angle) of arc 500 completely determines the heel and toe vector. Mathematically, this is a series of ankle positions (recorded during initial standing)

It is explained as Two ankle position measurements are required, corresponding to two statistically distinct lowering member 220 (γ _i, ) and ankle joint 200 angle (θ _i, ) positions:

식 28

Equation 28

식 29

Equation 29

Then, by differentiating the equation, the vector solution is:

식 30

Equation 30

이 가역적일 것을 요구한다. 또한, 최적 선형 필터링 관점에서 볼 때, 이러한 "이득 행렬"은 통계학적으로 중요한 결과를 산출하는데 충분히 커야 한다.The solution is

Requires this to be reversible. In addition, in terms of optimal linear filtering, this "gain matrix" should be large enough to yield statistically significant results.

Given the fact that the prosthetic prosthesis performs significant vibrations during the initial stance, the above equation can be extended to N combinations of ankle joint position / angle measurements. The resulting N-1 equation can be solved to obtain the best estimate of the vector using least squares. The above equation applies likewise to solve for the toe vector when the toe contact initiates the initial stance.

6A shows various terrains: flat 620, uphill 5 ° ramp 624, downhill 5 ° ramp 628, uphill 10 ° ramp 632, downhill 10 ° ramp 636, uphill stairs 640, and Inertial measurement unit-calculated ankle joint pivot trajectory in different walking situations for wearer walking on downhill stairs 644. Context is the shape of the terrain and how the wearer interacts with the terrain.

6B illustrates a 2-D structure illustrating the flight trajectory of the ankle joint of the prosthetic device. If we treat the flat walk as the object of a ramp up / down ramp situation (the flat is a zero degree ramp), the situation distinction depends on the distinction from the ramp up / down of the stair up / down. This distinction is typically a plantar flexion (dorsiflexion) of the ankle joint 600 in staircase situations, as compared to a dorsal flexion (or neutral maintenance) to optimize foot-landing kinematics on a ramp walk. This is important because it is required to optimize foot-and-land kinematics. In the latter case, it only corresponds to a very steep downhill where the plantar flexed ankle is in the proper orientation.

FIG. 6C illustrates how an ankle angle attack angle Φ can be performed as a trajectory feature that distinguishes between steps and ramp walk situations in a combination of data recorded with a step-tilt separator. 6C is a plot of the estimated velocity vector attack angle of the ankle joint 600 of the device for each step taken by the wearer during the walking cycle. In this data, the amputation patient wearing the prosthetic device 1700 of FIG. 17A on his right foot walks thirty steps (meaning a walking cycle referenced to the right foot) in the following manner:

1. Steps 1 to 6: Uphill 5 ° ramp 6 steps

2. Step 7: Landing One Step

3. steps 8 to 9: downhill 10 ° ramp three steps

4. Gap recording

5. Steps 10-11: Two steps uphill

6. Step 12: One Step Landing

7. Step 14-17: 4 steps downhill 5 ° ramp

8. Steps (18 to 19): Two steps of flat

9. Steps (20-21): Two steps uphill 10 ° ramp

10. Step (22): Land after 10 ° ramp

11. steps (23 to 24): two steps downhill

12. Steps (25-31): Seven steps of flats.

The steps taken during this recording include climbing and descending ramps and stairs. 6C shows that by monitoring the ankle velocity attack angle ψ, the staircase can be distinguished from the ramp with the ankle floating in the air prior to foot-landing. If ψ falls below a small positive value in this record (and other similar records), foot 604 always lands on the stairs. In all other cases, the foot lands on the ramp regardless of ramp angles (0 °, -5 °, + 5 °, -10 °, + 10 °). Accordingly, ψ is an appropriate walking task situation discriminator that can be used by the processor to determine that the action is to be performed.

Alternative methods for distinguishing by step-tilt may be employed in other embodiments of the present invention. The posture (orientation in the inertial space) retraction member 608 (shank) and ankle velocity attack angle ψ may be used in one embodiment of the present invention to distinguish between stairs or ramps / flat. The trajectory of the ankle joint 600 in the yz plane (see FIG. 6A) can be used in an alternative embodiment of the present invention to distinguish by step-tilt.

Boom  Ankle Positioning

Stairway ramp discriminator provides real-time prediction of terrain slope angle, φ ^{^} (t). When the discriminator detects a step that includes flat land, φ ^{^} (t) = O. Otherwise, the slope angle is:

식 31

Equation 31

This slope angle corresponds to the minimum value at which the foot may not be in contact with the ground. φ ^{^} (t) is the minimum of two possible slope angles: the angle the heel makes to the toe position at the last step and the angle the toe makes to the toe position at the last step.

Once φ ^{^} (t) is known, various different methods can be applied to position the ankle in the manner employed for this predicted terrain slope. Two examples of these methods are described below. In one embodiment of the present invention, the discriminator methodology described above includes at least one of joint impedance, position, or torque of a prosthetic prosthesis, support, or exoskeleton device (eg, device 1700 of FIG. 17A) worn by the wearer. It is used to control. The method includes estimating the velocity vector attack angle (eg, the y-axis value of the data in FIG. 6C) of the ankle joint of the device during final approval. In one embodiment, the method also determines the foot member of the device relative to the position under the toe when the velocity vector attack angle has a predetermined sign (eg, a negative value in the case of the data of FIG. 6C). Adjusting the position. In another embodiment of the present invention, the method includes adjusting the position of the foot member of the device with respect to the position under the heel when the velocity vector attack angle has a sign opposite to a predetermined sign (positive sign). Include.

In some embodiments, the method is based on the expected force exerted on the lower member during the period between when the heel of the foot member is in contact with the lower terrain and when the foot member is positioned in the sole-contacting position with respect to the lower terrain. Adjusting the impedance of the device (eg, ankle joint impedance) to minimize the cost function.

7A illustrates a method for positioning ankle joint 700 prior to foot-landing. In this method, the ankle joint angle is optimized to minimize the cost function based on the expected force f (t) applied to the ankle joint 700 from the foot member 708 landing to the sole-contacting. The heel-first 716 and toe-first 712 strategies are evaluated and a strategy is selected that includes an optimal ankle joint 700 angle that minimizes cost function. 7A illustrates the method used.

In another embodiment, the method of FIG. 7A is performed as shown in FIG. 7B to detect the presence of a step and limit the attack angle optimization for toe landing only for steps with short landing zones. In order to climb or descend a steep and narrow combination of steps, the prosthetic device is programmed to maintain the trajectory of the volume swept by the foot during the volume climb without any contact between the foot and the step. In later stages, for example, if it is determined that there is no landing on the heel, the optimization is limited to be a toe-bottom solution. In this embodiment, the z-rotation is a rotation about the longitudinal axis of the descending member 704 (eg, the z-axis of FIG. 17A) of the device. If a person descends the stairs and rotates the foot member 708 in this manner, the landing area is limited and the foot member 708 is likely to have to be rotated to land directly on the stairs. In such a case, landing under toes 712 yields the minimum force solution available of the method of FIG. 7A. This z-rotation can be a signal system with a limited landing area and, when compared with the heel-bottom, makes the toe-bottom landing the safest alternative.

The complex impedance calculation employed in the method described above may be applied to any adaptive ankle position method as a means of minimizing the use of kicks or excessive braking force as the ankle joint 700 rotates to the underfoot state. FIG. 7D illustrates how the method of FIG. 7A is applied to use optimized impedance. Once the optimal attack angle ψ ^* is found, it is found that the linear and angular momentum of the ankle joint can be zeroed without a kick according to the optimal control (Γ ^* _c (t)). The corresponding ankle angle response θ ^* (t) is used as the equilibrium trajectory. The corresponding optimum impedance for this optimal trajectory can be extracted to accommodate uncertainties in momentum and local topographic angles.

A simpler method can also be used as shown in FIG. 7C. 7C illustrates a method of positioning the ankle joint in a ramp walk situation. In this method, the ankle joint 700 angle articulates to be in a slope-shaped foot-contacting position when the lowering member 704 is vertical (having a slope angle φ (t)). It is also useful to generalize this method to adjust the ankle angle to be linear with respect to the slope angle expected by the following relationship:

식 32

Equation 32

Using this relationship, the ankle angle can be adjusted to follow the wearer preferences.

In one of the two methods described above, the ankle joint angle 700 prior to foot-landing will be continuously controlled (steered) to match the desired ankle joint 700 angle until the foot is in contact with the ground.

Position impedance and torque control

의 임피던스를 변경)을 이동시킨다. 제어 시스템은 발목 균형 각도가 지형 각도(φ)와 동일하도록 보철의 발목 관절(804)의 임피던스를 설정하며; 제어 시스템은 국부 수직(850)과 정렬하도록 하퇴 부재(812, 정강이)의 배향을 회복시킨다. The next step includes reverting the orientation of the lower leg (shank) to local vertical alignment during the standing phase. FIG. 8 illustrates a method for determining an inertial-reference spring balance based on the topographic angle at the foot-contact of the prosthesis 800, eg, the prosthetic device 1700 of FIG. 17A. Prosthesis 800 has a foot member 808 having a toe 816 and a heel 820. The prosthesis also has an ankle joint 804 and a lower leg 812 (shank). The terrain angle φ is the input to the control system. The control system calculates the curve Γ-Θ (ankle joint) of FIG. 10A to maintain and improve the overall balance of the wearer (as shown in FIG. 10F) during controlled plantar flexion based on the change in terrain angle φ.

Change the impedance). The control system sets the impedance of the ankle joint 804 of the prosthesis such that the ankle balance angle is equal to the terrain angle φ; The control system restores the orientation of the lowering member 812 (shanks) to align with the local vertical 850.

의 임피던스를 변경)을 이동시켜, 평형점을 향해 하퇴 부재(812, 정강이)를 이동시키도록 발목 관절(804)에 명령함으로써 제어된 족저 굴곡 중에 착용자의 전체 균형을 유지 또는 향상시킨다.9 shows the effect of walking speed on ankle torque versus ankle angle during controlled dorsal flexion. The control system calculates the curve Γ-Θ (ankle joint 804) of FIG. 10A based on the change in terrain angle φ.

Change the impedance of the device) to instruct the ankle joint 804 to move the lowering member 812 (shank) toward the equilibrium point, thereby maintaining or improving the overall balance of the wearer during controlled plantar flexion.

10A illustrates a method for controlling a limb device, in accordance with an exemplary embodiment of the present invention. As shown in FIG. 10A, this is achieved in the control system by:

를 변경)에 대해 설명된 바와 같이, 발-착지 및 발바닥-닿기 사이의 시간 간격 사이의 충격을 완화시키도록 후기 유각기 임피던스를 조정(동적 강성 및 발목-각도 평형 각도) (단계(1000));1) FIG. 7A (controller moves curve Γ-Θ) based on minimization of negative-impact energy shock and hip shock during forceful plantar flexion (impedance of ankle joint)

As described above, adjust the late-stage drift impedance (dynamic stiffness and ankle-angle equilibrium angle) to mitigate the impact between the time interval between foot-landing and plantar-contacting (step 1000). ;

2) introducing a lifting force to the following foot—by applying a reflex response to the ankle (and knee) at or before the point of impact of the preceding foot (step 1004); And

3) Maintain an inertial-reference equilibrium angle at controlled dorsal flexion timing (as shown in FIG. 10F) to maintain balance (equilibrium) on the sloped terrain (step 1008).

10B is a schematic diagram of a controller for implementing impedance and torque control in a prosthetic prosthetic device (eg, device 1700 of FIGS. 17A-17E), in accordance with an exemplary embodiment of the present invention. FIG. 10E is a schematic diagram of the impedance and reflection relationship for performing the impedance and reflection control performed in FIG. 10B.

As shown, the spring, damping, and inertial components of impedance are defined in relation to the trajectory, θ _o (t). The impedance gain matrix and trajectory shown in FIG. 10B is adaptively loaded in real time from the state controller processor according to the timing in walking cycle, terrain situation, terrain texture and walking speed as described above.

)로서 채용함으로써 달성된다. 다른 구현예에 있어서, 이러한 반사 토크 입력은 발목-동역학의 모델-기반 계산을 이용하여 추정될 수 있다.Studies have shown that intact limbs show reflex responses resulting from non-linear positive torque (force) and non-linear positive joint velocity feedback. The reflection relationship as shown in FIG. 10E employs two ways of feedback. Other non-linear implementation examples of this amount of feedback relationship, including some linear and some non-linear, are obvious to those of ordinary skill in the art. In an exemplary embodiment, the positive torque feedback measures the torque of the shin of the ankle prosthesis and translates it into a non-linear feedback signal (

It is achieved by employing as). In another implementation, such reflected torque input can be estimated using model-based calculations of ankle-dynamics.

The inventors have found that the reflection of the biomimetic impedance and the standing angle combine when the effects of walking speed and terrain slope are considered as shown in FIG. 9. For this reason, in one preferred embodiment, the parallel elasticity of the prosthesis (eg, parallel, or K3 spring) is selected to show stiffness with respect to slow walking speed as shown. In biomimetic systems, the rigid component of the prosthesis is weakened at higher walking speeds, and the reflection response is steeper as shown in FIG. Despite this optimal biomimetic control and mechanical implementation, this response causes the actuator to push the parallel spring in controlled dorsal flexion and pull it in forceful plantar flexion. We call this bipolar, or push-pull, operation. In a non-optimal control and mechanical implementation example, the reflection is implemented with only unipolar, pull-force twice the size. The exemplary embodiment thus reduces the peak actuator force and motor current by two factors when the appropriate bilateral series spring response is selected, prolongs the actuator design-life by 8X and nearly two factors. By reducing the ball-nut speed. This has the enormous advantage of increasing actuator durability, reducing actuator weight-reducing the number of ball bearings and ball-nuts to achieve design life targets-and reducing acoustic noise.

10C is a schematic diagram illustrating a controller implementing impedance control in a prosthetic prosthetic device (eg, device 1700 of FIGS. 17A-17E), in accordance with an exemplary embodiment of the present invention. FIG. 10D is a perspective view illustrating a mechanical impedance relationship for controlling the impedance control performed in FIG. 10C. τ _M is the torque exerted on the ankle joint of the prosthetic prosthesis by a linear actuator. Despite the appropriate "high-gain" compensation, G _c (z), where z means individual-time signal conversion, the motor torque is 1) series-elastic actuator, 2) "K 3 " parallel elastic and 3 ) Will act as the sum of the torque exerted by the acceleration torque, Γ _c , the desired result, of the same ankle as the torque command. K ^{^} ₃ and K ^{^} _s are used to represent model estimates for these mechanical variables, and therefore reference to model-based control.

FIG. 10F is a schematic diagram illustrating the manner used to determine the restoring torque required to stabilize the inverted pendulum motion of a person wearing a prosthetic device. FIG. The torque Γ _CM is applied to the center of gravity of the system (eg, the combination of the person wearing the prosthesis and the prosthesis) to balance the wearer based on:

식 33

Equation 33

및

는 앞서가는 발 및 따라가는 발 각각의 무게 중심 및 제로 모멘트 피벗 사이의 세계 좌표 참조된 벡터를 나타낸다. 전술한 용어인 제로 모멘트 피벗은 지면 반력 분포의 모멘트가 영인 발의 관성-참조 지점을 의미한다. 우리는 또한 이러한 지점을 본 서류의 나머지 부분에서 상호 교환 가능한 압력 중심(CoP)로 참조할 것이다. Where f _l and f _t are the ground reaction forces acting on each of the leading foot and the following foot. v _CM is the velocity vector of the wearer's center of gravity. ZMP _l and ZMP _t Represents the zero moment pivot of the leading foot and the following foot.

And

Represents the world coordinate referenced vector between the center of gravity and the zero moment pivot of each of the leading foot and the following foot. The above term zero moment pivot refers to the inertia-reference point of the foot where the moment of the ground reaction force distribution is zero. We will also refer to these points as interchangeable pressure centers (CoPs) in the remainder of this document.

surface Reaction  And zero moment pivot

Ground reaction force (GRF) is a force exerted by the lower surface on the foot (or foot member of the prosthetic device). Ground reaction force is an important biomechanical input during standing. By recognizing the total ground reaction forces acting on the zero moment pivot (referred to herein as ZMP and CoP), the control system of the prosthetic prosthetic device (e.g., controller 1712 in FIG. It has a direct way to improve balance and optimize force transmission. US Pat. No. 7,313,463 to Herr et al. Further discloses ground reaction force and zero moment pivot positions (both incorporated herein by reference) for biomimetic movement and balance controllers used in prosthetics, supports and robots and methods. .

FIG. 11A shows a prosthesis (eg an example of how a GRF component (in particular, an vector from ankle joint 1104 to ZMP, ^ω r _ZMP , and a GRF vector, ^w F _GRF ) changes during a standing phase in a typical walking cycle. For example, a schematic diagram illustrating the lower leg member 1100, the ankle joint 1104, and the foot member 1108 of the device 1700 of FIG. 17A. GRF estimation in the study setting is often performed by applying a sensor to the sole of the shoe. However, such external sensing may not be practical for prosthetic and support devices, because the sensors used in the study setup cannot do this because reliable packaging means must withstand contact stress for millions of walking cycles. . In addition, such means often require customization of the shoes, which may not be suitable for the wearer.

In another embodiment of the present invention, the inherent sensing of the GRF is coupled to the inertial state and the retraction member force / torque input 1112 (eg, using the structural elements 1732 of FIGS. 17A and 17E). Performed in a new way.

11B, 11C, and 11D are schematic views of components of the apparatus 1700 of FIG. 17A. The foregoing figures also show the components necessary for determining ground reaction force and zero moment pivot (linear linear elastic actuator 1116 (eg, a combination of linear actuator 1716 and series elastic member 1724 of FIG. 17) and a parallel spring. 1120 (e.g., the passive elastic member 1724 of Figure 17A) shows the force and moment relationship ^ω r _ZMP and ^w F _GRF are calculated based on the following steps:

1. Update the inertia states of the lower member 1100 and the foot member 1108 using the inertial measurement unit and ankle joint 1104 angle input. Using rigid estimation, the world-referenced acceleration is further calculated at the lower member 1100 and the foot member 1108 of the lower member 1100 and the foot member 1108 and the center of gravity (CM) of angular velocity and acceleration. .

를 푼다.2. As a function of the force acting on the lower member 1100 as it is disassembled along the lower member 1100 axis.

Loosen

를 푼다.3. As a function of the moment exerted by each of the force and moment components acting on the lowering member 1100

Loosen

및

에 대한 값을 이용하고 발 부재(1100)에 가해지는 힘의 균형을 맞추어 ^wF_GRF를 푼다.4. Calculated in steps 2 and 3

And

Solve ^w F _GRF by using the value for and balancing the force applied to the foot member 1100.

5. Balance the moment around the ankle joint 1104 assuming that ^w F _GRF is applied at the foot-ground boundary (ie, ^ω r ^z _ZMP = 0).

6. Solve the ^ω r ^y _ZMP .

terrain As a texture  Ankle Joint Behavior

12A shows the biomimetic Γ-θ behavior of a flat prosthetic device (eg, device 1700 of FIG. 17A) as a function of walking speed. 12B shows the net work (zones under the Γ-θ curve) starting and performing from the ideal biomimetic response, especially when walking at high speeds, with the ankle joint torque weakening rapidly with angle during plantar flexion with force. To reduce the power consumption.

In a typical robotic system, trajectories or other reproducing means are used to convey repeatable and programmable responses. Such means are undesirable in prosthetic and support devices because the wearer's intention can vary in the middle of the regeneration segment. For example, the wearer may walk quickly and stop suddenly in front of an ice patch. Pre-programmed trajectories and the like are reproduced, and there is no easy way to stop them without causing rapid change in force and torque and causing a hazard. In fact, this is why inherent means are used.

가 발바닥-닿기에서 발가락-들기까지 어떻게 변화하는지를 도시한다.To extend the application of the ankle joint torque during forceful plantar flexion, the walking speed-dependent normalized ground contact length is used as a means to mitigate the vertex plantar flexion torque, Γ _o . Ground contact length is estimated by using an anomaly model of the foot inferred in the description associated with FIGS. 2A-5 and measuring the inertial attitude of the foot member during plantar flexion with controlled dorsal flexion and force. As shown in FIG. 12C, as the foot transitions from the sole-to-toe to the toe-lifting, the sections of the idealized foot fall below the terrain, allowing estimation of ground contact length. 12D is

Shows how the change from the sole to the toe-lift.

FIG. 12E shows how the velocity-dependent tables of the length of contact weakening enable the use of normalized ground contact length as a means of achieving biomimetic behavior during forceful plantar flexion. The tables can be calculated by dynamically measuring the ground reaction force and the foot absent posture of a non-cutting surgical patient in a controlled environment as a function of walking speed. The functional relationship between the relaxation function and the ground contact length can be calculated for each walking speed. These tables can be stored in the controller of the prosthetic device as a reference relationship. The function may be shaped to meet specific wearer needs if the prosthetic device fits the wearer.

As mentioned above, one of the motivations for using inherent feedback, which is the inverse of the apparent trajectory or playback means, is to accommodate a change in the wearer's intention (e.g., decide to stop quickly). Inherent sensing using ground contact length as a means of mitigating ankle joint torque is not general enough to accommodate a change in the wearer's intentions, including stopping and turning. 12G, in one embodiment of the invention implemented in a prosthetic device, the time-dependent damping factor (e- ^{t / τ} ) is used in series with the ground contact length attenuation. The time constant τ of this attenuation can be altered to eliminate the plantar flexion drive torque loaded with force to prevent the risk associated with the change of intention of the wearer. τ can typically be in the range of 50 to 100 msec.

Preferably, the prosthetic device allows the wearer to walk faster on all terrains with less effort. It is insufficient to accommodate only changes in terrain conditions (stairs, slope climbs / downs). Such changes in terrain texture may cause slippage (eg ice / snow), and the risk of sinking (mud, snow, sand, fine gravel) must be accepted. Inherent sensing of zero moment pivot trajectories can be used to optimize walking performance and / or eliminate risk while walking on changing terrain textures.

, 의 추정된 y-컴포넌트가 어떻게 변화하는지를 도시한다. 도시되는 바와 같이, 비-미끄러짐 조건

은 발바닥-닿기(3) 및 발가락-들기(4)의 조건 사이에서 단조롭게 증가해야 한다. 이는 이러한 시기에(보행 사이클이 진행함에 따라 증가하여) 지형 표면으로부터 위로 올라오는 것이 뒤꿈치이기 때문이다. 제로 모멘트 피벗의 속도가 음의 y-축을 따라 이동하는 경우, 발이 미끄러진다. 잠금-방지 브레이크(ABS)가 자동차에서 구현되는 것과 유사한 방식에 있어서, 보철 장치는 음의 제로 모멘트 피벗 속도의 적분으로부터 추출되는 감쇠 인자에 의해 토크를 감소시킬 수 있다. 일 실시예에 있어서, 소음 민감도를 감소시키기 위해, 소음 임계치 이하의 음의 속도만이 적분된다.12F is a zero moment pivot vector during normal walking movement,

We show how the estimated y-component of,,, changes. As shown, non-slip conditions

Should increase monotonically between the conditions of the sole-contacting 3 and the toe-lifting 4. This is because at this time (increasing as the walking cycle progresses), it is the heel to rise up from the terrain surface. If the velocity of the zero moment pivot moves along the negative y-axis, the foot slips. In a manner similar to that in which an anti-lock brake (ABS) is implemented in a motor vehicle, the prosthetic device can reduce the torque by a damping factor extracted from the integration of the negative zero moment pivot speed. In one embodiment, only the speed of sound below the noise threshold is integrated to reduce noise sensitivity.

는 지면-참조된, 세계 프레임에서의 발목 관절 속도의 z-컴포넌트를 나타낸다. 도 13b는 입각기가 지면과 접촉하는 발가락(1324)에 의해 개시되는 보행 이동을 도시한다.
13A and 13B provide a state control situation in an exemplary embodiment of the present invention applied to, for example, the device 1700 of FIGS. 17A-17E. Common walks involve cycling between two phases: the stray and the stance. 13A illustrates a control system that includes walking movement initiated by heel 1320 in contact with the standing ground.

Denotes the z-component of the ankle joint velocity in the ground-referenced world frame. FIG. 13B illustrates the walking movement initiated by the toe 1324 where the stance is in contact with the ground.

Example Control System Behavior for Driving a Prosthesis or Support During a Walk Cycle

13A and 13B show that the control system 1300 changes the ankle behavior as an ankle transition between states in the stator 1304 and the stance 1308. The control system 1300 applies the position control 1328 to the stator to position the ankle to avoid risk of crushing in the initial sting phase, and to select certain terrain conditions (slopes, steps, steps) in the staging phase. Optimize heel-toe landing attack angle (adaptive ankle positioning) for The control system 1300 controls the impedance and torque control 1332 in the stance-optimizing the inertia, spring and damping characteristics of the ankle-heel / toe landing, foot downward movement, peak energy storage (dorsiflexion with exponential hardening). ) As an ankle transition via forceful plantar flexion and toe-lift.

FIG. 13C illustrates a method for position control applied to a lower limb device (eg, device 1700 of FIG. 17A), in accordance with an exemplary embodiment of the present invention. Preferably, the wearer and / or controller of the device does not move the foot member 1348 forward until it is certain that the toe 1340 wishes to pass the terrain in front of the wearer. One exemplary way to do this is to wait until the toe 1340 of the foot member 1348 is a sufficient distance above the last position of the toe 1340 relative to the underlying terrain. In this embodiment, the control system 1300 has a gap distance (ε ₀ ) measured along the normal vector of the topographical surface between the toes 1340 of the foot member 1348 at time t and time t _{k -l} . Position control 1328 is applied by starting to rotate ankle joint 1340 only after it is determined to be larger. This minimizes the risk that the toes 1340 may be sluggish. In one embodiment, the position of the toes 1344 at two different times t and t _{k -l} are determined using inertial measurement unit measurements as described above. Those skilled in the art will know how to apply other ways of determining when it is appropriate to move the foot member 1348 forward. In some embodiments, the controller can determine an appropriate forward movement timing, for example, based on whether the volume of the foot at the dorsal flexion achieves the desired clearance with respect to the terrain surface.

In summary, in an embodiment of the present invention, the prosthetic device employs a stepwise terrain adaptation method adapted to the intention to achieve true biomimetic behavior during walking, including flat walking, stepping up / down and ramping up / down. Adopt. 14A schematically illustrates the process by which the stepwise adaptation method is performed. In the stir phase, the inertial measurement unit supplies a unique sensing input (opposite of the external nerve / myopia input), which causes the device to identify the terrain situation from the cue supplied by the stipple trajectory feature. . Adaptive angular ankle positioning refers to the joint combination of ankle angle, θ, to achieve a natural heel or toe landing that is optimized for the most similar topographical situation, as determined by topographical contextual distinction on the stator trajectory signal.

14B shows exemplary impedances that may apply ankle joint prosthesis to three different gait situations. 14B is a graph of the required ankle torque (in 1404, Nm / kg) versus ankle joint angle (1408 in degrees). The graph includes three curves 1412, 1416, and 1420. Curve 1412 shows ankle joint torque 1404 versus ankle joint angle 1408 for walking on a 5 degree uphill ramp. Curve 1416 shows ankle joint torque 1404 versus ankle joint angle 1408 for walking on a 5 degree downhill ramp. Curve 1420 shows ankle joint torque 1404 versus ankle joint angle 1408 for walking on a zero degree ramp (flat). The slope of the curve is equal to the stiffness (or generally impedance). The area enclosed by the closed Γ-θ curve corresponds to the amount of non-conservation work required for certain terrain conditions (eg slopes, stairs) and walking speed. As can be seen in the graph, the ankle joint prosthesis will need to provide more work to perform a walk of walk versus flat walk on an uphill slope because the area within curve 1412 is larger than the area within curve 1416.

hybrid  Generalization of Prosthetic Footprint System

15 is a schematic diagram of the prosthetic biomechanical apparatus 1500, in accordance with an exemplary embodiment of the present invention. In one embodiment, the device 1500 is a support device that increases the wearer's ability to walk. In another embodiment, the device 1500 is a support device attached to the wearer's body to support and / or fix the musculoskeletal malformations and / or abnormalities of the wearer's hips, thighs, shins and feet. In another embodiment, the device 1500 is an exoskeleton device attached to the wearer's body to assist or increase the wearer's limb biomechanical output (eg, increasing the wearer's limb strength and mobility).

The device 1500 is a connecting device constituted by a plurality of links (or members) and joints connecting the links. Device 1500 includes foot members 1508, L ₀ , which are connected to lower leg members 1516, L ₁ by ankle joint 1512. Device 1500 also includes thigh members 1524, L ₂ , which are connected to lowering member 1516 by knee joint 1520. The device also includes a hip 1528 that couples the thigh member 1524 to the upper body 1532, L ₃ of the wearer. Center of gravity 1504 is the center of gravity of the combination of device 1500 and wearer.

The foot member 1508 is in contact with the terrain 1536 underlying the foot member 1508 at the zero moment pivot 1540. Foot member 1508 includes a toe portion 1544 and a heel portion 1548. Each joint of device 1500 also includes an actuator having a generalized vector of torque (force) Γ _i , displacement ξ _i , and impedance K _i , where i = 0 corresponds to ankle joint 1512, i = l corresponds to the knee joint and i = 2 corresponds to the hip joint. Each articulation actuator may comprise a mechanical component (eg, ball-screw actuator or rotational coordination driver), human muscle, or both. Joint displacement usually takes the form of angular displacement (rotation), but may include, for example, a combination of linear and angular displacement found in conventional knee joints. The posture of the link, i, is represented by a 4x4 matrix that defines the position of the link origin of the coordinate frame and the unit vector in terms of the world coordinate frame, the unit vector of W.

The posture of each link, j, is a linking device constraint relationship--specifically, the posture of the link, i-1, is converted into a transformation defined by generalized displacement, ξ _j , and certain link variables (link length, distortion, and convergence angle). By multiplying by For example, if the postures of the shins are known, the postures of the feet, thighs and upper body can be calculated either directly by sensing them or using the inertial sensor, assuming that the generalized displacements for these connecting devices are known. The vector of sensor information unique for each link is compressed into what may be called a unique sensing unit (ISU). Examples of inherent sensors include direct or indirect measurement of generalized displacement; Measurement of angular rate and acceleration of the link (eg, using an inertial measurement unit); Measuring or estimating a component of a force or torque on the link; Multi-modal computer shaping (eg, range maps) or measurements of the output of specific neural pathways on or adjacent to a link.

The terrain is modeled as a surface function, an appearance function with α (x, y), z (x, y). In this case, the surface properties are sufficient to capture the surface's resilience / capacity to capture the surface energy and capture the ability of the foot to obtain a static frictional force on the surface as it relates to the work he is forced to contact with the surface and to be pushed away from it with the foot member. Plasticity, damping properties and coefficient of friction.

16 is a schematic diagram illustrating a method of determining a wearer's thigh member, hip member, and upper body posture, in accordance with an exemplary embodiment of the present invention. In a lower limb system employing a robotic knee prosthesis or support, the position of the hip can also be calculated by installing an inertial measurement unit over the thigh or by measuring the relative knee angle referenced to the lower leg member. If the inertial measurement unit is further employed on the upper body, the posture of the upper body can also be calculated instantaneously. Alternatively, the pose can be calculated by measuring two degrees of freedom hip joint displacement. Compensation for the upper body attitude prediction error resulting from the rate gyroscope and accelerometer drift on the upper body inertia measurement unit can be corrected during the down member zero-speed update through a chain process of speed limiting via the hybrid system connection device.

, 허벅지 자세

및 상체/바디 무게 중심 자세

의 예측을 교정하도록 이용되는 자세 복원 방법을 도시한다. 단계 1(1604)은, 도 2a 내지 도 5에 대하여 전술한 바와 같이, 하퇴 부재 자세를 결정하도록 하퇴 부재(1620, 링크 1) 상의 제로 속도 업데이트의 출력부를 포착한다. 허벅지 부재(1624, 링크 2) 및 상체 부재(1628, 링크 3) 각각에 대한 솔루션(단계 2 및 3)은 단계 1(1604)의 예를 따르지만, 이러한 경우에 있어서, 속도 제한은 영이 아니며, 이전 링크로부터 병진 및 회전 속도에 의해 예측된다.
Figure 16 shows the upper body posture limit

Thigh posture

And upper body / body center of gravity posture

A posture reconstruction method used to calibrate the prediction of P is shown. Step 1 1604 captures the output of the zero velocity update on the lower member 1620, link 1, as described above with respect to Figures 2A-5. The solution (steps 2 and 3) for each of the thigh members 1624, link 2 and upper body 1628, link 3 follows the example of step 1 1604, but in this case, the speed limit is not zero, Predicted by translation and rotational speed from the link.

Example Machine Design

17A shows a prosthetic prosthesis 1700, in accordance with an exemplary embodiment of the present invention. Device 1700 has a mounting interface 1704 that allows attachment to a wearer's complementary limb socket member. Device 1700 also includes structural element 1732 (herein referred to as pyramid) coupled with mounting interface 1704 and first end-point 1702 (referred to herein as shank) of the retraction member 1712. It includes. In some embodiments, the axial forces and moments applied to the retracting member of the device are determined based on measurements made using structural members (pyramids) engaged with the retracting member of the device. A pyramid is a meter structure that is a component of a prosthesis that engages a wearer's limb socket. In one embodiment, pyramid (structural element) measurements are used by the controller to determine the axial force and moment applied to the descending member. In this embodiment, the structural element 1732 engages with the first end-point 1722 of the retraction member 1712 having a combination of pins 1711. The pin 1711 passes through a combination of holes 1713 in the retracting member 1712 and a combination of holes 1715 in the structural element 1732 (shown in FIG. 17E).

The structural element 1732 has a top surface 1731 positioned towards the mounting interface 1704 and a bottom surface 1733 positioned towards the retraction member 1712. The lower member 1712 also engages with the foot member 1708 at the ankle joint 1740 at the second end-point 1744 of the lower member 1712. Ankle joint 1740 (eg, a rotating bearing) causes foot member 1708 to rotate about the x-axis with respect to lower leg 1717. The foot member includes a heel 1772 and a toe 1776.

The apparatus 1700 also includes a linear actuator 1716 having a first end-point 1736 and a second end-point 1748. Linear actuator 1716 produces linear motion 1703. The first end-point 1736 of the linear actuator 1716 couples (eg, with a rotary bearing) with the first end-point 1722 of the retracting member 1712. The apparatus 1700 also includes a first passive elastic member 1728 connected in series with the linear actuator 1716. The passive elastic member 1728 is coupled with the foot member 1708 and the second end-point 1748 of the linear actuator 1716. Passive elastic member 1728 engages foot member 1708 (eg, with a rotary bearing) at an adjacent end-point 1730 of passive elastic member 1728. The end-point 1726 of the passive elastic member 1728 couples (eg, with a rotary bearing) between the second end-points 1748 of the linear actuator 1716. Linear actuator 1716 applies torque about ankle joint 1740.

The apparatus 1700 also includes an alternative second passive elastic member 1724 having a first end-point 1756 and a second end-point 1760. The second passive elastic member 1724 provides a unidirectional spring force (providing parallel elasticity) parallel to the retraction member 1712. The first end-point 1756 of the second passive elastic member 1724 is coupled with the first end-point 1702 of the lower member 1712. The second end-point 1760 of the second passive elastic member 1724 engages with the foot member 1708. However, during plantar flexion, the springs do not engage, thus merely providing a unidirectional spring force to the device.

In some embodiments, the second passive resilient member 1724 is a non-compliant stop that stores little or no energy and limits further rotation of the ankle over a certain angle during forceful plantar flexion. .

17B and 17C show a portion of the limb device of FIG. 17A showing a second passive elastic element 1724. The second passive elastic element 1724 stores energy during dorsal flexion, but not plantar flexion. Elastic element 1724 has a double-cantilever coupling (clamped at position 1780 between first end-point 1756 and second end-point 1760). The elastic member 1724 has a tapered shape 1784 that allows the elastic member 1724 to efficiently store energy by maximizing the bending strain along the entire length (along the y-axis) of the elastic element 1724. In some embodiments, the normalized spring constant range is 0 to 12 Nm / rad / kg. At the top of the range, the energy storage is approximately 0.25 J / kg.

Increasing the cam / tilt of the elastic member 1724 allows the spring constant to be easily matched to the wearer's weight. The cam element 1788 is located at the second end-point 1760 of the elastic member 1724. The ramp element 1792 is located in the foot member 1708. Cam element 1788 engages ramp element 1792 during dorsal flexion; The cam element 1788 does not engage the ramp element 1792 during plantar flexion or engages with other portions of the apparatus 1700. Because cam element 1788 does not engage ramp element 1792 or engages with other portions of device 1700 during plantar flexion, elastic member 1724 stores energy only during ventral flexion. In one embodiment, the position of ramp element 1792 is screw-adjustable to allow the wearer or second party to match the ramp engagement of cam element 1788 to match energy storage characteristics to the wearer's walking habits. To help. The operator adjusts the position of the ramp element 1792 relative to the position of the cam element 1788 to change the energy storage characteristics of the passive elastic member 1724.

In an alternative embodiment, the actuator is engaged into the ramp to adjust the ankle joint angle at which the second passive elastic member 1724 (elastic member engagement angle) is engaged. This, for example, causes the ankle joint 1740 to bend back during the stator phase, without engaging the elastic member 1724 when the wearer climbs the ramp and stairs, and during running.

Passive elastic element 1724 also serves to increase the frequency in response to device 1700 when elastic element 1724 is engaged in dorsal flexion. The dynamics of the device 1700 in the dorsal flexion benefit from a fast response (bandwidth) series elastic actuator (ie, a combination of the linear actuator 1716 and the first passive elastic element 1728). The spring constant associated with the second passive elastic element 1724 increases the bandwidth of the device 1700 by the factor, β:

식 34

Equation 34

Where K ₃ is the spring constant of the second passive elastic member 1724 and K _s is the spring constant of the combination of the linear actuator 1716 and the first passive elastic element 1728. In one embodiment of the present invention, the second passive elastic element provides β of 1 to 3; Increase the bandwidth of device 1700 approximately 5 Hz to approximately 15 Hz.

The second passive resilient member 1724 employs a dovetail feature 1796 to allow clamping at both end-points without using mounting holes. In one embodiment, the second passive elastic member 1724 is made of synthetic fiber material. The mounting holes can form stress strength and can cause fiber dropout in the passive elastic member 1724 that can compromise the strength of the spring. End-point clamp 1798 has a complementary shape that holds passive elastic element 1724 in place. In one embodiment of the invention, the epoxy is employed in the clamp to permanently secure the second passive elastic member 1724 in the end-point clamp. Epoxy joints will be likely to fail without dovetail features 1796.

Passive elastic element 1724 employs a tapered design that maximizes energy storage at element 1724 to ensure that the energy storage density is constant throughout its length for a given deflection. Referring to FIG. 17D, as a free-body diagram of the passive elastic element 1724, the roller force, F _roller , and the retraction member force, F _shank , are the same and opposite in direction by the central pivot. Show how to combine. In this embodiment, the roller force and the retraction member force are applied at the same distance from the central pivot. Forces at the end-point, F, combine to produce the central pivot force of FIG. 17F. Using a standard thin beam relationship, the moment acting at a distance of x from the center pivot changes linearly-starting at the value of FL at the center and falling to zero at x = L-where L is the force The length of the passive elastic element 1724 in between. The energy storage density along x is proportional to the product of the moment M (x) and strain at the surface ε ₀ (x).

식 35

Equation 35

식 36

Equation 36

For a given layup of the synthetic material, the surface strain is below the threshold, ε *. For a given moment, the energy density of the beam will be maximized if the surface strain is set for this threshold. In order to maintain the energy density constant and its maximum value, the optimal width of the beam, w * (x), is defined by the above relationship:

식 37

Equation 37

In one embodiment, the taper 1784 changes linearly from the center of the beam. By using this design method, we amplified the energy storage of the spring by a factor of two compared to the beam without taper 1784. Since the synthetic spring material is not uniform and the thin beam equation is not applicable, a calculation tool is used to estimate the energy storage density in the passive elastic member 1724. The shape that can store most of the energy is highly dependent on the fiber peel, the peel design, the thickness and the exact manner in which the passive elastic member 1724 is attached to the device 1700. However, we determined that linear taper delivers an optimal range of energy storage of approximately 10%. In an exemplary embodiment, the linear taper is used because it is relatively easy for the linear taper pattern to be cut from the plywood composite sheet using a water-jet process. In an alternative, less exemplary embodiment, a tapered spring can be used.

FIG. 17E shows a perspective view of an embodiment of the structural element 1732, referred to herein as a pyramid. The structural element 1732 couples between the mounting interface 1704 of the retracting member 1712 and the first end-point 1702. The structural element 1732 engages with the first end-point 1702 of the retraction member 1712 having a combination of pins 1711 (shown in FIG. 17A). The pin 1711 passes through a combination of holes 1713 in the retracting member 1712 and a combination of holes 1715 in the structural element 1732. The pin 1711 causes the degrees of freedom of rotation of the strain in the structural element 1732 to be incorrectly recorded as axial forces and moments in the structural element 1732. In the present embodiment, the structural element 1732 is " foot-down " by a controller 1762 state machine that measures the moment and axial load on the ankle joint 1740, eg, controls the device 1700 function. Positive (+) detection; Measurement of the moment applied for use by positive-feedback reflection control employed during forceful plantar flexion; And detection of the amount of slip for use of the safety system coupled to the controller 1762.

In this embodiment, the structural element 1732 is designed as a flexing element that amplifies the strain field induced by the lateral forces and lateral forces applied to the device 1700 during operation. The structural element 1732 is a large strain field (differential strain field) of opposite sign in the regions 1738 and 1742 around the center adapter mounting hole 1734 when an inward-side moment (a moment about the x-axis) is applied. ) This differential strain field does not appear only when axial forces are applied. The structural element 1732 includes one strain gauge 1782 and 1786 coupled to two moment-sensing regions 1738 and 1742, respectively, on the bottom surface 1733 of the structural element 1732. Gauges are applied to both sides of the Wheatstone bridge. Controller 1762 couples with the Wheatstone bridge to measure strain. Strain measurements are used to measure the moment of structural element 1732. In one embodiment, the sensitivity of the measurement is in the range of approximately 0.15 N-m, in which case the sensitivity defines the minimum resolvable change (signal-to-noise) when digitally sampled at 500 Hz.

In contrast to the moment-induced strain, the high strain is introduced by axial force along the medial-side axis in the regions 1746 and 1754 around the central adapter mounting hole 1734. This strain appears in the 0.76 mm thick region (regions 1746 and 1754) below the slots 1758 and 1770, respectively, machined along the medial-lateral axis. The section above the slot should be thick enough to deliver the moment load with the minimum strain in the thin lower section. The magnitude of the strain is significantly weaker in thin sections when only moment-load is applied. The structural element 1732 includes one strain gauge 1790 and 1794 attached to each of the two axial load-sensing regions 1746 and 1754, respectively, on the bottom surface 1735 of the structural element 1732. Gauges are applied to both sides of the Wheatstone bridge. Controller 1762 couples with the Wheatstone bridge to measure strain. Strain measurements are used to measure the axial force on the structural element 1732 and thus the axial force on the lower member 1712. Machined slots 1758 and 1770 amplify the axially-induced strain without compromising the structural stability of structural element 1732.

Since the structural element 1732 is in a critical chain of structural support between the wearer (not shown) and the remaining limb sockets of the device 1700, in one embodiment it is designed to withstand more than 60 N / kg of axial load It is preferable to be. In this embodiment, the sensitivity of the axial measurement is in the range of approximately 50N, which is below the threshold of approximately 100N typically used within the device 1700 to sense that the device 1700 is firmly positioned on the ground. During calibration of the device 1700, the 2 × 2 sensitivity matrix is determined so that true moment and axial forces can be extracted from the pair's strain measurements.

17F illustrates a cross-sectional view of an alternative method of measuring the axial force and moment applied to the retracting member, in accordance with an exemplary embodiment of the present invention. In the present embodiment, the structural element 1732 is curved to amplify the displacement of its bottom surface 1733 in such a way that axial forces and planar moments (two degrees of freedom) can be extracted in a redundant fashion. Adopt a design. In this embodiment, the device 1700 includes a displacement sensing device 1735 for measuring the deflection of the structural element 1732 to determine the moment (torque) and the axial force 1712 applied to the retracting member. .

In this embodiment, the displacement sensing device 1735 includes a printed circuit assembly (PCA) that employs one or more displacement sensors 1735 (eg, contact or non-contact displacement sensors). The sensor measures the distance between the sensor 1737 and the bottom surface 1733 of the structural element 1732 at each sense coordinate.

In one embodiment, the mutual inductance change of the coil printed on the PCA relative to the bottom surface 1733 of the structural element 1732 is used to measure the local surface strain (displacement). In the present embodiment, the reverse "eddy" current in the structural element 1732 acts to reduce the coil inductance as opposed to the distance between the coil and the bottom surface 1735 of the structural element 1732. Other displacement sensing techniques may be included, including non-contact capacitance and optical or contact-based sensors employing force-sensitive resistors, piezoelectric or strain-gauge integrated into the PCA. By sampling the displacement sensors of the array, the axial forces and moments can be estimated using the sensitivity matrix calculated during the off-line calibration process.

In this embodiment, the structural element 1732 is fastened to the retraction member 1712 by screws, eliminating the need for the pin 1711 employed in the embodiment shown in FIG. 17E. Screw fastening methods reduce weight and manufacturing complexity. This fastening method also amplifies the displacement measured at the center of the structural element 1732 where the displacement sensing device 1735 is located. 17G shows how the planar moment vector and the axial force can be calculated using a circular arrangement of displacement sensors on the printed circuit assembly. As shown, demodulation of the bias and sinusoidal displacement functions is used to estimate the moment and force. Other displacement sensor arrangements may be used as inherent means of moment and force.

Moment and force sensing is useful as a means to signal changes in walking state. In addition, the measurement of the retraction member 1712 moment acts as a feedback means by which the reflective behavior is achieved in the plantar flexion on which the force is loaded. In combination with inertia and actuator feedback, inherent moment and force measurements are used to calculate ground reaction forces and zero moment pivots useful for static friction force control and balance. For this reason, it is advantageous to package the inherent moments and forces that are sensed by the inertial measurement unit and the state control processing function. 17F shows how this function can be implemented on the PCA. This PCA is a FR- with a stable core material (e.g. Invar) that acts as a rigid intervening substrate between the top displacement sensing FR4-based layer and the bottom FR-4-based layer in combination with the signal processing layer. It can be implemented as a sandwich of four materials. Integrating these materials and functions within a single assembly eliminates the need for cables and other non-dependent means of interconnecting these functions. This integration also allows stand-alone tools to be used by prostheses to set up passive prosthetics and research, gait parameters, including energy recovery and gait statistics.

Referring to FIG. 17A, the apparatus 1700 includes a controller 1762 coupled with a linear actuator 1716 to control the linear actuator 1716. In this embodiment, the controller is installed in the housing 1764 of the device 1700 to protect it from the external environment. Battery 1768 in housing 1764 provides power to the device (eg, controller 1762 and various sensors associated with device 1700).

The device 1700 includes an inertial measurement unit 1720 that predicts the inertial posture trajectory of the ankle joint 1740, heel 1772, and toe 1776 for the previous toe-lifting position. Inertial measurement unit 1720 is electrically coupled to controller 1762 and provides an inertial measurement signal to controller 1762 to control linear actuator 1716 of device 1700. In one embodiment, the inertial measurement unit 1720 employs a three-axis accelerometer and a three-axis rate gyro. The 3-axis accelerometer measures local acceleration along three orthogonal axes. A three-axis rate gyro measures angular rotation along three orthogonal axes. Through the use of this well-established method for numerical integration, the position, velocity and posture of any point in the foot structure can be calculated.

In some embodiments, the inertial measurement unit 1720 is used to detect the presence of terrain slopes and steps and steps—the “attack angle of the foot relative to the underlying terrain prior to the spring equilibrium position of the contact and ankle joint at the standing position. Enable optimization of ". In some embodiments, the inertial measurement unit 1720 may include the walking speed of the wearer and the conditions of the terrain (such as terrain features, textures or irregularities (e.g., how dense the terrain is, how slippery the terrain is, rough or smooth). , Is there any obstacle, such as a rock, on the terrain). This allows the wearer to walk confidently in all terrain types. Inertia poses are captured or quaternized as a fixed, ground-referenced (world) coordinate frame-an orientation component of homogeneous transformation (three unit vectors defining the x, y and z axes in the world reference frame); Translational movement of ankle joint 1740 in a world frame; And three degrees of freedom orientation of the lower member 1712 at-captured as the velocity of the ankle joint 1740 in the world frame. In this embodiment, the inertial measurement unit 1720 is physically coupled with the lowering member 1712. In some embodiments, inertial measurement unit 1720 is coupled with foot member 1708 of device 1700.

FIG. 17H is a schematic diagram illustrating a state estimation and actuator controller (state and actuator control PCA-SAC) used in the apparatus of FIGS. 17A-17G, in accordance with an exemplary embodiment of the present invention. In this embodiment, the controller 1762 is a dual 40 MHz dsPIC to control and adjust the coordinate linear actuator 1716 (eg, rotation motor 504 of FIGS. 5A and 5B) and inertial measurement unit 1720. Employs a processor (manufactured by MicrocHiP ™). In this embodiment, space-vector modulation is employed to perform brushless motor control to produce an optimal pulse width modulation drive that maximizes motor RPM. Space vector modulation is a PWM control algorithm for multi-phase AC generation in which the reference signal is sampled regularly. The PWM of the signal or power supply includes modulation of three-phase motor winding voltage utilization (eg, rotary motor 504). After each sampling of the reference signal, one or more of a nonzero active switching vector and a zero switching vector adjacent to the reference vector are selected for the appropriate portion of the sampling period to synthesize the reference signal.

The controller 1762 receives various input signals, such as an inertial attitude signal 1781 from the inertial measurement unit 1720, a torque and axial force signal 1783 from the structural element 1732 strain measurement, ankle joint Ankle joint angle signal 1785 from the Hall-effect transducer located at 1740, right angle with motor position signal 1787, index and absolute motor position from encoder (eg, encoder 2040 of FIG. 20A) Phase encoder), strain signals 1789 from strain sensor 1704 of series elastic member 1728, and controller variables 1791 (e.g., device configuration data, wearer-specific tuning, firmware updates). ). The controller 1762 also outputs various signals, including device performance data 1793 (eg, real time data, error log data, real time performance data), and ankle status update 1795. In addition, the controller 1762 outputs a command to the linear actuator 1716, which is provided from the linear actuator 1716 (generally the signal 1797) to an actuator feedback signal, for example, a power electronic component for the linear actuator 1716. Receive a three-phase pulse width modulated signal, battery power for linear actuator 1716, and current feedback measurements and temperature measurements from linear actuator 1716.

This embodiment uses sensor feedback to identify the state change as the device 1700 transitions through the stance and stipple states. By using redundancy and various sensor measurements, the fault condition is also checked and the device 1700 is in an appropriate safe state. Using the on-board real time clock, time-tag the faults and store them in the on-board e ² PROM (error log). The contents of the error log are retrieved wirelessly by the prosthetic and / or manufacturer service personnel. In this embodiment, the motor drive PCA (MD) receives a pulse width modulation (PWM) command from the SAC PCA to convert current into the motor windings. The MD returns the sensed current and power information back to the SAC PCA to enable closed-loop control.

In this embodiment, the IMU PCA is nominally mounted in a sagittal plane (local plane parallel to the anterior portion of the tibia), a three-axis accelerometer, a dual-axis rate gyro (ω _z and ω _x ) and Employ a single-axis rate gyro (ω _y ). In this embodiment, the coordinate frame definition is used to define the front y-axis, the upper z-axis, and the x-axis defined as the vectors of the y and z axes (y X z). The IMU receives status information from the SAC at a system sampling rate of 500 Hz. This conveys the ankle state vector—in particular, the position and speed of the ankle pivot, the position of the heel and the position of the toe—from the previous step relative to the plantar-contacting position.

17I and 17J are schematic diagrams illustrating exemplary electrical equivalents of the apparatus 1700 of FIG. 17A. The electrical circuit symbol may comprise a mechanical element-a resistance representing a mechanical component with a linear damping torque with speed; Resistance indicative of a mechanical component having rotational inertia properties; And an inductor representing a mechanical component with linear spring quality. According to this circuit notation, current corresponds to torque and voltage corresponds to angular velocity.

은 압축 시 직렬 스프링(예를 들어, 도 17a의 수동 탄성 엘리먼트(1728))의 비틀림 스프링 상수이며;

은 인장 시 직렬 스프링의 비틀림 스프링 상수이며; K₃는 단일 방향 평행 스프링(예를 들어, 도 17a의 수동 탄성 부재(1724))의 비틀림 스프링 상수이며; 및 J_Ankle은 발목(예를 들어, 도 17a의 발 부재(1708)) 이하의 발 구조의 회전 관성이다. 모델 내의 전류(토크) 소스는 이하와 같이 정의된다:

은 착용자의 바디에 의해 하퇴 부재(예를 들어, 하퇴 부재(1712)) 상에 가해지는 알려지지 않은 토크이며; τ_motor는 작동기(예를 들어, 선형 작동기(1716))에 의해 가해지는 토크이며; 및

는 구조 엘리먼트(예를 들어, 도 17a 및 도 17e의 구조 엘리먼트(1732))를 이용하여 측정되는 토크이다.The circuit components are defined as follows: J _shank is the unknown equal inertia of the lower leg (shank) and the remaining limbs below the knee (eg, lower leg 1717 in FIG. 17A); J _Motor is the uniform motor and ball-screw transfer assembly inertia (eg, inertia of linear actuator 1716 of FIG. 17A);

Is the torsion spring constant of the tandem spring (eg, passive elastic element 1728 of FIG. 17A) upon compression;

Is the torsional spring constant of the serial spring at tension; K ₃ is the torsional spring constant of the unidirectional parallel spring (eg, passive elastic member 1724 of FIG. 17A); And J _Ankle is the rotational inertia of the foot structure below the ankle (eg, foot member 1708 in FIG. 17A). The current (torque) source in the model is defined as follows:

Is an unknown torque exerted on the lower member (eg, the lower member 1712) by the wearer's body; τ _motor is the torque exerted by the actuator (eg, linear actuator 1716); And

Is the torque measured using the structural element (eg, structural element 1732 of FIGS. 17A and 17E).

17i illustrates the importance of series and parallel springs as energy storage elements. The use of stored energy reduces the power consumption that may be required by the linear actuator. A further object of the K ₃ spring is also to function as a shunt across the ankle inertia which increases the ankle-spring resonance.

17J illustrates how the sensors employed in this embodiment to provide high-accuracy position and force control and achieve sensor redundancy and variability desirable for delivering a unique safety design. As shown, the ankle joint position, θ, is extracted from:

식 38

Equation 38

here

식 39

Equation 39

Redundant measurements of θ are achieved through the use of Hall-effect angle transducers to ensure that they are properly adjusted by the control system. In one embodiment, the Hall-effect converter includes a Hall-effect device located on the SAC PCA in the housing 1764 of the device 1700. The transducer also includes a magnet that is coupled with the foot member 1708. The field effect magnitude (signal output by the transducer) varies in a known manner in response to angular joint rotation (ie, the movement of the magnet relative to the Hall-effect device). The Hall-effect transducer is calibrated during manufacture of the apparatus 1700 by, for example, measuring the output of the transducer up to a known displacement of the Hall-effect apparatus with respect to the magnet. In another ankle angle measurement embodiment, the mutual inductance measured on the coil on the lowering member has a known relationship as a function of the ankle angle, where the inductance is not sensitive to the magnetic field generated by the motor in a linear actuator or other out-of-position field. In order to calculate the angular displacement. In addition, as shown in FIG. 17J, the ankle moment exerted by the wearer is also measured. This applies to the linear actuator to achieve reflection behavior (eg to increase stiffness).

18A, 18B, 18C and 18D show the passive elastic member 1728 of FIG. 17A, in accordance with an exemplary embodiment of the present invention. Passive elastic member 1728 provides bidirectional stiffness and is connected in series with linear actuator 1716 and foot member 1708. Passive elastic member 1728 is coupled with the second end-point 1748 of linear actuator 1716 at one end-point and with the foot member (not shown) at the other end-point. Passive elastic member 1728 includes a strain sensor 1704 coupled with the passive elastic member 1728 to measure strain in the passive elastic member 1728. In the present embodiment, the strain sensor 1704 is a strain gauge whose response is momentum about the force exerted by the linear actuator 1716 and the ankle joint 1740 exerted by the linear actuator 1716. It is scaled to measure. Strain gauge signals are measured using controller 1762 of FIG. 17A.

In this embodiment, the passive elastic member 1724 is a formed carbon-fiber layup that delivers the desired bidirectional (bending function in both directions) normalized drive stiffness. In one embodiment, the passive elastic member 1724 preferably has a compression of 14-25 N-m / rad / kg and a tensile force of 4-8 N-m / rad / kg. Biomechanical forces and torques are aligned with the weight of the wearer. When scaling prosthetic and support devices, design variable details are typically normalized. For example, the serial and parallel resilience of such a device may be designed to provide a distinct value that is intended to be scaled to the weight or to cover some range of weight. The compression and tension range reflects the torque change resulting from the difference in the linear actuator moment arm for the ankle joint across the entire range of rotation—from maximum plantar flexion to maximum ventral flexion. The tandem spring constant is optimized to be relatively non-compliant during the stator dorsal flexion position control (while the spring is in compression) when the ankle is re-positioned to a toe-lift immediately following in the toe-lift walk. However, some compliance is maintained to isolate the linear actuator from the impact load.

18C and 18D, high rigidity is achieved by inserting the dorsal flexion rotational bottom limiting portion 1708 toward the end end point 1726 of the passive elastic member (spring) 1728. 1728). This restriction reduces the effective moment arm of the linear actuator 1716 upon bending of the serial spring 1728 during compression (toward the bend bend). In tension, the moment arm is effectively increased by placing the plantar flexion upper limit 1716 towards the adjacent end-point 1730 of the spring restraint. With longer moment arms, the spring beam will bend more freely, reducing the spring constant upon tensioning. In addition to the bidirectional stiffness characteristic, in some embodiments, the spring constant of the passive elastic member 1728 is optimized by design to minimize ball-screw rotational speed, and this embodiment of the elastic member 1728 has a symmetrical characteristic. Delivers higher compliance in tension than compression. Higher compliance in tension increases the energy storage of the tandem spring 1728 for the use of forceful plantar flexion. Energy is released at the first 100 ms involved in forceful plantar flexion, reducing the required energy contribution of the linear actuator 1716. In an embodiment of the invention that utilizes a ball-screw delivery assembly associated with a rotary motor for a linear actuator (e.g., ball-screw delivery assembly 2024 of FIGS. 20A and 20B), It has the further advantage of reducing the desired operating speed of the ball-nut assembly portion and the motor drive requirements of the rotating motor. The spring moves the foot member strongly without the need for high speed ball-nut positioning in this case. The optimized value of series elasticity is in the range of 3-4 Nm / rad / kg.

FIG. 19A illustrates a prosthetic prosthesis 1900 with a plate tandem spring 1928, in accordance with an exemplary embodiment of the present invention. The device 1900 has a mounting interface 1904 to allow it to be attached to the wearer's complementary limb limb socket member. The apparatus 1900 includes a retraction member 1912 that is coupled with the mounting interface 1904. The lower member 1912 also engages with the foot member 1908 at the ankle joint 1940 of the device 1900. Ankle joint 1940 causes the foot member 1908 to rotate about the x-axis relative to the lower member 1912. The foot member includes a heel 1972 and a toe 1976.

The apparatus 1900 also includes a linear actuator 1916 having a first end-point 1936 and a second end-point 1948. The first end-point 1936 of the linear actuator 1916 couples with the retracting member 1912. The apparatus 1900 also includes a passive elastic member 1928 that is connected in series with the linear actuator 1916. The passive elastic member 1928 is connected between the foot member 1908 and the second end-point 1948 of the linear actuator 1916. Passive elastic member 1928 is coupled with foot member 1908 at an adjacent end-point 1930 of passive elastic member 1928. The end-point 1926 of the passive resilient member 1928 is coupled with the second end-point 1948 of the linear actuator 1916. Linear actuator 1916 applies torque about ankle joint 1940.

The apparatus 1900 also includes a controller 1960 that is coupled with the linear actuator 1916 to control the linear actuator 1916. In this embodiment, the controller 1960 is located within the housing 1964 of the device 1900 and protected from the environment; Part of the housing is removed to expose the contents within the housing as shown. Battery 1968, coupled with device 1900, provides power to device 1900 (eg, controller 1960 and various sensors associated with device 1900).

The passive resilient member 1928 of FIG. 19A is a flat spring (eg, made of water-cut equipment). The flat spring reduces the cost of the passive elastic member 1900 and makes it easy to configure the spring constant to align with the weight of the wearer. In one embodiment, the spring is a ball-nut of linear actuator 1916 (see, eg, FIGS. 20A and 20B) due to the lack of parallelism between the axis of rotation of the ball-nut and the series passive elastic member 1928. It is divided in the longitudinal direction (along the y-axis) to reduce the moment out of the plane on the component of. In this embodiment, non-strain sensing is employed in the actuator torque feedback loop. Rather, the torque transmitted through the spring is measured spring deflection (measured ankle joint 1940), which is defined kinematically as an ankle joint 1940 angle that may result from a particular ball-nut position along the screw if the spring deflection is zero. ), Multiplied by the known spring constant of the flat spring.

19B and 19C illustrate alternative two-piece series-elastic springs of prosthetic device 1900, in accordance with an exemplary embodiment of the present invention. Device 1900 has a mounting interface 1904 that allows it to attach to the wearer's complementary limb socket member. The apparatus 1900 includes a retraction member 1912 that is coupled with the mounting interface 1904. The lower member 1912 also engages with the foot member 1908 at the ankle joint 1940 of the device 1900. Ankle joint 1940 causes foot member 1908 to rotate about the x-axis with respect to lower member 1912. The foot member includes a heel 1972 and a toe 1976. The apparatus 1900 also includes a linear actuator 1916 having a first end-point (not shown) and a second end-point 1948. The first end-point of the linear actuator 1916 couples with the retraction member 1912. The device 1900 is also a bearing that causes the foot member 1908 to rotate about the x-axis of the ankle joint 1940, which connects the foot member 1908 to the lower member 1912 at the ankle joint 1940. Coupling member 1988 (eg, bracket).

The apparatus 1900 also includes a passive elastic member 1928 that is connected in series with the linear actuator 1916. Referring to FIG. 19C, the passive elastic member 1928 has two member sections (eg, beam-shaped sections) 1994 and 1996. The elastic member 1928 also has a first end-point 1962 on the first member 1994 and a second end-point 1980 on the second member 1996. The elastic member 1928 also has an intermediate position 1996 where the two members 1994 and 1996 meet and the two members 1994 and 1996 pivot about each other about the x-axis. As the second member 1996 pivots towards the first member 1994, the elastic member stores energy in a compressed state during dorsal flexion (shown by arrow 1992).

The first end-point 1962 of the elastic element 1928 engages with the second end-point 1948 of the linear actuator 1916 having a bearing that allows rotation about the x-axis. The second end-point 1980 of the elastic element 1928 engages with a position on the coupling member 1988 having a bearing that allows rotation about the x-axis.

Example Linear Actuator

20A and 20B illustrate a linear actuator 2000 for use of various prosthetic prostheses, supports, and exoskeleton devices, in accordance with an exemplary embodiment of the present invention. 20A is a perspective view of a linear actuator 2000. 20B is a cross sectional view of the linear actuator 2000. The linear actuator 2000 can be used, for example, as the linear actuator 1716 of the device 1700 of FIG. 17A or the device 400 of FIG. 4. The actuator 2000 includes a motor 2004 and a screw delivery assembly 2024 (in this embodiment, ball-screw delivery assembly, also called ball-screw assembly) for delivering linear power along the A axis. The screw delivery assembly 2024 acts as a motor drive transmission to convert the rotational motion of the motor 2004 into linear motion. In one embodiment, the ball-screw delivery assembly 2024 is a custom ball-screw delivery assembly manufactured by Nook Industries (Cleveland and Ohio). The custom ball-screw transfer assembly has the following details: 14 mm x 3 mm pitch screw, thrust of 4000 N at 150 mm / s, and Ll rating life at 5 million cycles of the present application. In some embodiments, the screw delivery assembly is a lead-screw delivery assembly (also called a lead-screw assembly).

The actuator 2000 includes a rotary motor 2004 with a motor shaft output 2008. Motor shaft output 2008 has pulley 2032 that engages (eg, welds) with motor shaft output 2008. In one embodiment, the rotary motor 2004 is a high speed brushless motor (model EC30 motor manufactured by Maxon Motor AG, Maxon Precision Motors, Inc., of Fall River, Mass.). Motor 2004 includes an inductive incremental-absolute angle encoder 2040 integrated with motor 2004 to determine the angular alignment between the rotor and stator of rotary motor 2004. The encoder 2040 also provides the position feedback signal needed to provide the position of the screw 2060 of the linear actuator 2000 for control and “immediate” motor replacement and redundant position feedback monitoring.

The inductively-coupled encoding element of encoder 2040 allows the system to determine absolute rotor-stator alignment (eg, 10 bits of resolution per revolution) simultaneously with high-precision incremental rotor position feedback. By cross checking this redundant feedback element, it is possible to minimize the possibility that an encoder failure can cause ankle adjustment. The incremental encoder achieves a run out of 300 μrad or less to eliminate sensed speed variations when the ball-screw transfer assembly 2024 (see below) operates at a constant speed. As a result, the torque change is less applied by the actuator 200.

Rotary motor 2004 also includes an integral motor heat-sink 2048. In one embodiment, the heat-sink 2048 draws heat from the windings of the motor 2004 so that the wearer can walk at the peak level of the non-conservation operation without exceeding the motor coil temperature limit (typically 160 ° C.). Make sure Motor heating appears due to the resistance loss (i ² R loss) of the motor 2004 as the linear actuator 2000 transmits thrust. As the coil temperature rises, the coil resistance rises at a rate of 0.39% / 占 폚, further raising the coil temperature. Also, the motor K _t (measures torque as it is fitted to the motor current) typically drops to near 20% as the coil temperature increases to its limit. This requires additional current consumption to perform the same workload, further raising the coil temperature. The heat-sink in linear actuator 2000 reduces the coil temperature by 40% or more. Motor wear is markedly reduced when the lower motor coil operating temperature is maintained, as wear phenomena occur that leads to premature failure of motor winding insulation and the motor bearings are effectively reduced by a factor of 2X for every coil temperature reduction of 10 ° C. Increases. In addition, using a unique coil temperature sensing method, the motor simply reduces the forceful plantar flexion power (current) as the maximum rating approaches, and reduces battery power, for example, when reaching a predetermined limit of 150 ° C. Ultimately by breaking it can be protected from exceeding the absolute maximum rating of 160 ° C.

Robotic prostheses are usually driven to employ a small, lightweight motor to intermittently transmit power to the affected limb. In some scenarios, the power may be applied repeatedly at a high rate for an extended period of time. Motor copper and vortex current losses will cause excessive cumulative heating effects causing the motor winding temperatures to rise. As the copper winding resistance increases with the temperature ratio (0.39% / ° C), the copper losses will increase to amplify the heating effect. Critical winding temperature limits can sometimes be reached to the extent that additional temperature rise will cause permanent damage to the motor. Detecting the point of time when this temperature limit is reached is preferably performed by the control system.

Two conventional methods can be used to prevent or detect when the copper winding temperature limit is reached or reached. First, copper and vortex current losses are calculated while the control system is operating. These are used to drive a thermal model of the winding so that the winding temperature can be estimated. Occasionally, the ambient temperature is measured to make better winding temperature measurements. The advantage of this method is that it is implemented at low cost. The disadvantage is the difficulty in building and calibrating these coil temperature models. In addition, good measurement of the temperature around the motor is often difficult, resulting in many errors in winding temperature measurement.

In the second method, sometimes in combination with the first method, the temperature of the motor is measured with a thermistor applied outside the case or inside the motor. The advantage of this is that it can be measured directly. The disadvantage is that measurements can only be made at one point and the application of the sensor is expensive and often unreliable.

A more preferred approach is to detect temperature and mitigate potential overheating conditions. Here, when the measurement is performed by simply fixing the ankle at the fixed position (removing the back-emf effect from the resistance calculation), the motor winding resistance is measured at every step during the walking cycle. In one embodiment, the coil temperature is determined by applying a fixed current (alternatively a fixed voltage) to the motor windings and measuring the corresponding voltage (alternatively current) in the windings. To increase accuracy, voltage (or current) is applied in the forward and reverse directions and the difference in current (or voltage) is measured.

Motor drive electronics employ a PWM current control method, and all structures exist that perform these measurements. If the ankle is at rest (scale adjustment constant), the ratio of the difference between the winding resistances can be displayed, thus allowing the on-site winding resistance to be estimated at no cost. In a typical servo system, this measurement cannot be made because the actuator must remain in the closed-loop control state. However, in ankle prostheses, there are cases where ankle position does not require maintaining precise control at 5 milliseconds, which is typically required for measurement (slag). Once the winding temperature has been calculated in this manner, the control system can detect when the winding reaches a critical temperature. During this time, the drive power available for walking is reduced or eliminated together until the temperature is lowered to a safe level.

In some embodiments, the output of the temperature sensor 2052 is configured to control the torque output by the linear actuator 2000 based on the temperature of the motor 2004 (eg, the controller 1762 of FIG. 17A). Is provided).

The belt 2012 couples the pulley 2032 with the helix shaft 2060 of the ball-screw delivery assembly 2024 so that the rotational movement of the motor shaft output 2008 is caused by the ball-nut of the ball-screw delivery assembly 2024. To be converted to linear motion of the portion of assembly 2036. In some embodiments, two or more belts are applied in parallel to each drive the drive linear actuator 2000 ball-screw transfer assembly 2024 itself, such that the linear actuator 2000 breaks a single belt. To withstand failures. In this case, the belt break sensor 2056 senses the condition and approves belt integrity during operation (eg, during each walk cycle of the wearer using the prosthesis).

In one embodiment, an optical sensor (eg, a through-beam sensor) is used as the belt break sensor and the output signal of the optical sensor changes in a known manner upon belt break. In another embodiment of the invention, the capacitive sensor is used as a belt break sensor and the output of the capacitive sensor changes in a known manner upon belt break.

In one embodiment, pulley 2032 and belt (s) are not used as a device to convert rotational motion into linear motion. Rather, a combination of traction wheels is used as the delivery device. In this embodiment, the risk of belt failure is eliminated.

In one embodiment, in case of belt breakage, the controller of the device in which the linear actuator 2000 is used (eg, controller 1762 of the device 1700 of FIG. 17A) may cause the linear actuator 2000 to be repaired. Change the position of the foot member relative to the lower member to the safe position so that it can act as a manual ankle prosthesis until it is. In one embodiment, the controller shorts three electrical leads of the rotary motor 2004 in response to a belt break sensor that detects a failure of one or more of the plurality of belts. By shorting the three-phase electrical input lead, the motor 2004 introduces viscous drag on the motor shaft output 2008. During walking, viscous drag grabs the fixed rotor shaft output, allowing the device to act as a manual prosthesis. However, the device may be moved slowly in a manner that allows it to move to a non-fixed equilibrium position in which he stands or sits. Each input lead can be shorted to ground by its individual switch.

In one embodiment, the switch is operated by a rechargeable battery (a battery separate from the primary battery used to operate the device). By using a separate battery, the switch will be able to short the input leads (and place the device in safe mode) even if a breakdown occurs (or the primary battery is exhausted).

In one embodiment, spiral shaft 2060 includes a hollow exterior and actuator 2000 comprising a noise damping material to reduce the noise generated by actuator 2000. In one embodiment, spiral shaft 2060 is a 14 mm diameter stainless steel shaft, has a bore of 8.7 mm inner diameter, extends 64 mm of the length of the shaft, and is acoustically damped material manufactured by 3M (St. Paul, Minn.). Filled with ISODAMP®C-1002.

The actuator 2000 also includes a radial and thrust bearing 2028 that supports the belt 2024 tension due to the thrust of the rotating motor 2004 and the screw 2036. The load due to belt tension and thrust is presented statistically during the walking cycle.

Ball-nut assembly 2036 includes one or more recirculating ball-tracks 2042 that hold a plurality of ball bearings; Combinations that support the linear motion of the ball-nut assembly 2036. In one embodiment, five ball-tracks are used. Actuator 2000 couples actuator 20 (e.g., a coupling element 2020 (e.g., passive elastic member 1724 of Figure 17A) to the foot elastic member of the prosthetic device (e.g. For example, the second end point 1748 of the linear actuator 1716 of FIG. 17A is included.

21 illustrates a perspective view of a linear actuator 2100 used in various prosthetic prostheses, supports, and exoskeleton devices, in accordance with an exemplary embodiment of the present invention. Linear actuator 2100 may be used, for example, as linear actuator 1016 of device 1000 of FIG. 17A or device 400 of FIG. 4. Linear actuator 2100 is an example strain of actuator 2000 of FIGS. 20A and 20B.

The actuator 2100 includes a rotary motor 2004 with a motor shaft output 2008. The motor shaft output portion 2008 has a pulley 2032 welded to the motor shaft output portion 2008. Motor 2004 includes inductive incremental-absolute angle encoder 2040 integrated with motor 2004 to determine the angular alignment between rotor motor and stator. Rotary motor 2004 also includes an integral motor heat-sink 2048.

Instead of the single belt 2012 of FIGS. 20A and 20B, two belts 2104a and 2104b are used side by side. Each belt has the ability to drive the linear actuator delivery itself with a margin of 1.5 times the belt break, allowing the linear actuator 2100 to survive a single belt break failure. In one embodiment, in the case of a belt break, the controller of the device in which the linear actuator 500 is operated (eg, in a manner such that the device acts as a manual ankle prosthesis until the linear actuator 500 is repaired). Controller 1762 of device 1700 of FIG. 17A moves the ankle to a safe position. In one embodiment, the controller shorts three electrical leads of the rotary motor 504 in response to a belt break sensor that detects a failure of one or more of the plurality of belts. In this case, one or more belt break sensors detect the condition and move the ankle to a safe position in such a way that the system acts as a manual ankle prosthesis until the linear actuator is repaired.

The two belts 2104a and 2104b couple the pulley 532 to the helix axis of the ball-screw delivery assembly (eg helix axis 2060 of FIG. 20B) to allow rotational movement of the motor shaft output 2008. It is intended to translate into linear motion of the ball-nut assembly 2036 portion of the ball-screw delivery assembly. The actuator 2100 also includes radial and thrust bearings 2028 that support the tension of the belts 2104a and 2104b due to the thrust of the rotating motor 2004 and the spiral screw. Loads due to belt tension and thrust are statistically displayed during the walking cycle.

Ball-nut assembly 2036 includes a plurality of ball bearings; A recirculating ball-track that maintains a combination that supports the linear motion of the ball-nut assembly 2036. The actuator 2100 also couples the actuator 2100 (eg, a coupling element 2020 (eg, passive elastic member 1724 of FIG. 17A) to the passive elastic member of the foot member of the prosthetic device (eg, FIG. 17A). For example, the second end point 1748 of the linear actuator 1716 of FIG. 17A is included.

The actuator 2100 also includes a ball-screw assembly seal 2108. The ball-screw assembly seal 2108 protects the screw from, for example, contaminants (eg, sand, dust, corrosive materials, sticky materials). Such contamination ensures that the design life of the actuator cannot be determined.

Exemplary Prosthetic Supports (Wearable Robot Knee Braces)

22A, 22B and 22C are schematic diagrams illustrating a limb support or exoskeleton device 2200 (wearable robotic knee braces), in accordance with an exemplary embodiment of the present invention. Device 2200 is a knee brace that augments the wearer's limb function. 22A is a top view of the device 2200. 22B is a side view of the device 2200. 22C shows the interior of the knee joint drive assembly 2204 of the device 2200. Typical uses of the device 2200 include, for example, increased metabolism, permanent assistance of the wearer with permanent limb disorder, or walking of the wearer with temporary limb disorder.

An example of a metabolic augmentation use case is a wearer (eg, a soldier or other personnel) who, for example, must carry heavy loads and traverse rough terrain for extended periods of time. For this use, the knee correction device 2200 increases the wearer's own capabilities. Examples of permanent assisted use include wearers who suffer from permanent limb disorders (eg, knee tendon or meniscus degeneration) that are unlikely to walk. For this use, knee correction apparatus 2200 provides permanent assistance to the wearer. Examples of use cases involving walking of a wearer with a temporary limb disorder include a wearer recovering from an injury or other temporary condition. For this purpose, the knee correction apparatus 2200 is a programmable remote robotic tool employed by a physiotherapist to assist with recovery-through the movement of motor dynamics and progressively decreasing assistance during muscle memory and strength recovery. . In another embodiment, the method further comprises specifying a physiotherapy protocol that defines the level of assistance performed by the device to the wearer for a period of time and assisting the wearer by the device to assist walking of the limb disorder. Reducing the level. In some embodiments, the auxiliary level performed by the device decreases based on the wearer's impedance and torque contribution to the device.

Referring to FIGS. 22A and 22B, the apparatus 2200 includes a lower member 2216 (also referred to as a drive arm), a thigh member 2228, a lower cuff 2208, and an upper cuff 2212. Lower cuff 2208 engages lower member 2228. The lower cuff 2208 attaches the device 2200 to the wearer's shin. Reclining cuff 2212 engages thigh member 2228. Reclining cuff 2212 attaches device 2200 to the thigh of the wearer. The device 2200 includes a knee joint 2232 for connecting the thigh member 2228 to the lower member 2216. Knee joint 2232 (eg, a rotating bearing) causes lowering member 2216 to rotate about x-axis with respect to thigh member 2228.

Referring to FIG. 22C, knee joint drive assembly 2204 includes a linear actuator for driving knee joint drum 2232 through belt drive transfer 2236. The linear actuator is a rotary motor 2240 (eg, brushless motor) and ball-screw delivery assembly 2244 (eg, motor 2004 and ball-screw delivery assembly 2024 of FIGS. 20A and 20B). )to be. In the apparatus 2200, the rotational motion of the motor shaft output 2256 of the motor 2240 is converted into linear motion of the ball-nut assembly 2248 portion of the ball-screw delivery assembly 2244. Motor shaft output 2256 has a pulley 2260 that engages (eg, welds) with motor shaft output 2256. The motor 2240 includes an inductive increment-absolute angle encoder 2264 integrated with the motor 2240 to determine the angular alignment between the rotor and stator of the rotary motor 2240. The encoder also provides the position feedback signals necessary to control the screw 2252 position of the ball-screw delivery assembly 2244 and to monitor " instantaneous " motor replacement and redundant position feedback.

The belt 2268 couples the pulley 2260 with the helix shaft 2252 of the ball-screw delivery assembly 2244 so that the rotational movement of the motor shaft output 2256 causes the ball-nut of the ball-screw delivery assembly 2244 to be rotated. Converted to linear motion of part of assembly 2248.

In one embodiment, spiral axis 2252 includes a hollow portion that includes a noise damping material to reduce noise generated by knee joint drive assembly 2204. Knee joint drive assembly 2204 also includes radial and thrust bearings 2252 that support belt 2268 tension due to thrust of rotating motor 2240 and screw 2252. Loads due to belt tension and thrust are statistically displayed during the walking cycle.

Knee joint drive assembly 2204 also includes a spring elastic 2280, a spring cage 2284, a drive belt 2236, and a spring cage / belt connection 2288. In some embodiments, a drive band (eg, a thin piece of spring steel) is used instead of the drive belt 2236. In some embodiments, a drive cable (eg, a loop of strand material) is used in place of the drive belt 2236. The spring 2280 is a series passive elastic element, which acts in the same manner as the series elastic spring element 1728 of FIG. 17A. Spring cage 2284 provides a closed volume in which spring 2280 is located. Ball-nut delivery assembly 2248 engages with screws 2252. Ball-nut assembly 2248 also engages drive belt 2236. The linear movement of the screw 2252 causes the linear movement of the ball-nut assembly 2248. Linear motion in the ball-nut assembly 2248 causes linear motion in the drive belt 2236. The linear motion of the drive belt 2236 drives the knee joint 2232.

Apparatus 2200 controls controller 2292 (e.g., a printed circuit assembly with a linear actuator 2204, state and inertia measurement unit 2294 (e.g., inertia measurement unit 1720 of Figure 17A). And processing functions) to drive and control the operation of the device 2200. Referring to FIG. 22B, the device 2200 also includes a torque sensor 2220 coupled with the lower member 2216 to measure the torque applied to the lower member 2216 by the knee joint drive assembly 2204. Sensor 2220 acts as a feedback element in the control loop of controller 2292 to achieve high accuracy closed loop position, impedance, and torque (for reflection) control of knee joint 2232. In one embodiment, an array of force-sensitive transducers is mounted in the cuff structure to provide force measurements that are used to achieve rapid biomimetic response.

In some embodiments, a motor angle sensor (eg, encoder 2264) measures the motor position and the controller adjusts the rotating motor to adjust the position, impedance and torque of the knee joint 2232 based on the motor position. To control.

In some embodiments, the apparatus 2200 includes an angle sensor that determines the position of the drum 2232 of the belt drive transmission relative to the output of the motor drive transmission, and the controller is based on the position for impedance, position or torque. Control the rotating motor to adjust it. In some embodiments, the device 2200 includes a displacement sensor that measures the displacement of the series spring of the motor drive transmission to determine the force on the series spring, wherein the controller adjusts the impedance, position or torque based on the force on the spring. Control the rotating motor to adjust. In some embodiments, the inertial measurement unit 2294 engages the thigh member or the lower member to determine the inertial attitude of the lowered member, and the controller uses a rotating motor to adjust the impedance, position or torque based on the inertial position. To control. In some embodiments, the torque sensor 2220 measures the torque applied to the dropping member by the belt drive transmission, and the controller adjusts the impedance, position or torque based on the torque applied to the dropping member. To control. In some embodiments, the apparatus 2200 includes an angle sensor that determines an angle between the thigh member and the lower member, and the controller is adapted to adjust the impedance, position, or torque based on the angle between the thigh member and the lower member. Control the rotary motor.

In some embodiments, the apparatus 2200 includes a screw delivery assembly coupled with the motor shaft output to convert the rotational motion of the motor shaft output into a linear motion output by the screw delivery assembly instead of the motor drive transmission. do. In addition, the drive transmission assembly coupled with the output of the motor drive transmission is a redundant belt, band or cable drive transmission coupled with the screw delivery assembly, which rotates the lower member relative to the thigh member for the linear motion output by the screw delivery assembly. To a rotational motion that applies torque to the knee joint.

Unlike the prosthetic device 2000 of FIG. 20A, the knee correction device 2200 operates in parallel with the impulsive action of a person. For metabolic boosting and alternative applications, the knee correction control system will provide all the impedance and torque needed in the walking cycle. It is desirable to allow the wearer to walk all day without fatigue accumulation and much effort on the increased side (s) of the body. In gait applications, knee-calibration device 2200 supplies a programmed ratio of impedance and torque. In this application, knee-correction device 2200 acts as a remote robotic extension of a physiotherapist who oversees the wearer's gait.

In one embodiment of the knee correction control system, the physiotherapist creates a protocol to be performed remotely by the knee during the period between visits to the therapist. Using the air interface, patient condition is fed back to the physiotherapist to achieve telepresence. The protocol specifies the rate at which the assistance reduces the over time. As the knee straightening device reduces its assistance, the knee straightening device will be calculated through the biomechanical model of impedance and torque contribution by the wearer-reducing the assist according to the enhanced response to maintain the desired net biomimetic response. The biomechanical model will include solving the knee inversion kinetics—integrating the inertia rotation and acceleration of the lower leg, thigh member, and upper body. This six degrees of freedom information will be extracted from the inertial measurement unit in the thigh member and knee joint angular displacement. The zero-rate update for the inertial measurement unit will be achieved as described herein.

FIG. 26 illustrates the biomechanical characteristics of the gait of a typical person during walking, starting and ending at heel strike. It is divided into upright and stippled phases and involves all elements of movement in the hips, knees and ankles. Myositis diseases, such as IBM (inclusion body myositis), are associated with major quadriceps defects that negatively affect the patient's ability to walk safely and efficiently. Equally important, patients with IBM are unable to perform a safe transition between rest (standing or sitting) and gait.

27 illustrates a biomechanical mechanism by which quadriceps defects affect walking in the flat. The overall mechanical sequence is lacking in two respects with respect to FIG. First, if the stiffness (or mechanical impedance) of the ankle joint is inadequate in the initial stance to absorb the impact of the landing and apply a restorative torque to balance (as indicated by reference number 2710), the knee is initially It can be buckled in a standing position. The quadriceps's inability to actively strengthen the knees to absorb foot-landing limits the speed of walking, much like driving a brakeless car. Mechanical impedance is related to the stiffness of the joint and includes three components: a spring component defined by a spring constant and an equilibrium position that applies a linear or non-linear restoring torque in response to joint displacement; A damping component that applies a linear or non-linear viscous restorative torque in response to the joint velocity; And an inertial component that applies a linear restorative torque in response to joint acceleration and a second late and early elevation angle, wherein the quadriceps are the knees that interfere with rapid knee flexion (indicated by reference number 2720). Note that braking torque may not be applied.

The knee device shown in FIGS. 22A-22C can be used as a robot-assisted solution to address the deficiencies discussed above with respect to FIG. 27. FIG. 28 utilizes inherent sensing of the lower leg trajectory to reproduce the desired gait trajectory state, terrain, body posture and stability, and controls biomimetic impedance, extended torque and position on the knee joint of the patient according to the time of the walking cycle. By applying, it shows the use of the knee device to restore normal walking.

In an early stance (indicated by reference numeral 2810), the knee device provides increased stiffness to absorb the landing energy and provide a stable platform for the following leg to leave the ground. In late standing (as indicated by reference 2820), such as ankle plantar flexion, the knee device uses the inherent measure of thigh and shin orientation and wearer's walking speed to allow the wearer to metabolize-efficient walking. Apply biomimetic, recursive torque, the same as can be provided by a fully functioning quadriceps. Later, if the knee device detects that the lower leg is in the stator (indicated by reference 2830), the knee is bent and high impedance is applied to absorb the lower leg when the foot lands on the ground (braking). something to do. The result is safe, metabolic-efficient walking.

surface Reaction  And balance using zero moment pivot

Figure 23a illustrates a general problem of balancing over the slope of a variable (positive or negative) slope. This problem entails a multi-link, "inverted pendulum" problem, which can be handled with a non-linear feedback control implementation. In this solution, knowledge of the link angle and mass characteristics of the links (in this case, leg segments, upper body, head and arms) is used to clearly stabilize the multi-link system. However, such clear input is not included in most embodiments of the prosthetic prosthesis, support or exoskeleton device, and, if not impossible, would be difficult to reliably implement and package for the wearer. Also, in some cases, the wearer will have one intact leg, so that part of the stabilization will be achieved outside of the prosthetic prosthesis, support or exoskeleton, and the prosthetic prosthesis, support or exoskeleton will function as an intact leg. Increase

23B also shows a continuum for an acceptable solution to the balance problem. Specifically, there is an infinite number of bent-knee solutions that can be fully accepted and even at will, such as to lift a heavy cargo or box or to balance while playing a game. Thus, we confirm that the desired solution employs a unique sense of compensation (for prosthetic prosthetics, supports or exoskeleton) of the intact balance-manufactured body component to achieve equilibrium in line with human intentions.

The solution employed in some embodiments of the prosthetic prosthesis, support or exoskeleton device uses a simplified representation of the problem as modeled in FIG. 23C. In this indication, inherent sensing of the lower member inertia state, ankle joint angle and inertia-referenced ground reaction force drives ankle torque (eg, the torque provided to the ankle joint by a linear actuator of the prosthetic device). Is used as stabilization feedback. The body is modeled as a series mass (only one shown in the figure) on a mass, thin, buckling beam with no time-varying stiffness and moment of inertia.

Balance is achieved based on the following details. Desired equilibrium is achieved when the following conditions are met:

1. ^W F _GRF aligns with world z;

2. The line connecting the zero moment pivot and ankle joint is aligned with the world z unit vector; And

3. All time derivatives of inertia lowering member angle, γ, and ankle joint angle, θ, are zero.

The feedback control law that drives each of these conditions in parallel is extracted based on:

식 40

Expression 40

here

식 41

Equation 41

Optimize the secondary cost index, J, where

식 42

Expression 42

And

식 43

Equation 43

Here, the component of k is chosen to emphasize the link angle dynamic contribution to the cost index. In this embodiment, the control law solution is provided by the linear-secondary regulator (LQR) methodology. In non-professional terms, this means that the setting of a controller to control a machine or process is found by using the mathematical algorithm described above and by minimizing the cost function with the weight factor supplied by a person. "Cost" (function) is often defined as the sum of the deviations of the major measurements from their desired values. Indeed, such an algorithm finds such a controller setting that minimizes unwanted deviations, for example deviations from the desired work performed by the wearer's prosthesis. The magnitude of the control action itself is often included in the sum to maintain the energy extended by the control action itself, which is often limited. In fact, the LQR algorithm optimizes the controller based on the engineer's details of the weighting factors. The LQR algorithm is, at its core, an automated way of finding an appropriate state-feedback controller.

Use of a secondary cost index is not necessary; In one embodiment, the use of a secondary cost index as an optimization criterion is used to analyze and presently tailor target frame work for wearers of prosthetic prostheses to achieve an acceptable feel as a system work to balance the wearer in different terrain. Create It is not uncommon to find that a control engineer prefers alternative conventional methods such as full state feedback (also known as stick placement) to find a controller in the use of the LQR algorithm. In accordance with these, the engineer has a much clearer connection between the adjusted parameters and the resulting change in controller behavior.

Wearer assistant standing up in a chair

24A, 24B, and 24C illustrate a method of applying a balance control law to assist a wearer of a prosthetic prosthetic device to take place in a chair, in accordance with an exemplary embodiment of the present invention. This timed Get-Up and Go (TUG) is often used as an experimental means of validating dynamic and functional balance. The wearer stands up from the chair, walks forward three meters, crosses the floor line, comes back and is given verbal instructions to sit back. To achieve good "TUG" performance, leg prostheses often have "stand" and "sit" buttons to create the behavior of the prosthesis' control system. In one embodiment of the prosthetic prosthetic device employing the principles of the present invention, there is no clear requirement to set the behavior situation, for example by pressing a button. Sit, stand and sit behaviors are identified by sensors inherent in the prosthetic device. During control behavior, standing and sitting are simply part of keeping the wearer in balance.

24A, 24B and 24C illustrate how the unique balance control algorithm operation multiplies the wearer as the wearer wakes up in the chair. Referring to FIG. 24A, the start from the sitting position to the standing position includes three states. First, your feet fall off the ground or touch lightly. The prosthetic device (eg, device 1700 of FIGS. 17A-17E) may include a weight of the wearer; Inertial orientation of the lower and foot members; And ground reaction forces (eg, as determined for FIG. 11A). As such, the device “knows” or senses that the wearer is sitting. As the wearer starts to stand up, an increase in ground reaction force appears, and the state of the foot (foot-contact) is recognized through inertial measurement unit measurements and ankle joint angle sensor measurements. Inherent balance control law enforcement is disclosed. During this second state, the imbalance detected by the imbalance of ground reaction forces drives the lowering member forward as a means of pulling the upper body (center of gravity) over the ankle joint (eg, increases the torque applied to the ankle joint 1740). Drive forward by a controller 1762 that instructs linear actuator 1716.

Referring to FIG. 24B, inherent balance control keeps the wearer in equilibrium in front of the chair. 24C shows the wearer in mid-standing equilibrium, ready to walk, if necessary. As shown, the intention of the wearer, more specifically, the sit / stand behavior, is inferred by the sensing inherent in the prosthetic device. The cost of implementation and the complexity of explicit context switching (button press) are avoided. Prosthetic devices compensate and augment body function in a natural way.

Ankle torque induced by ground reaction force (GRF) is a preferred way to achieve exponential hardening during mid-standing. Unlike the use of torque on the lower leg (eg, torque measured using the structural element 1732 of FIG. 17A), the GRF-calculated ankle torque measures the torque applied to the ankle joint by the ground. GRF is often measured by force plates in pedestrian research settings and is used as a measure of how intact ankles interact with the ground during walking. GRF establishes what biomimetic ankle behavior is in different terrain situations. The advantage of using GRF as a means of achieving exponential cure is that it is easily performed by the ability to be measured for biomimetic references. In addition, since it is extracted from inherent inertial sensing (eg, using the inertial measurement unit 1720 of FIG. 17A), the use of such measurements ensures invariance to terrain orientation.

The standing support may also be implemented using the knee support shown in FIGS. 22A-22C or more generally any active knee support or prosthesis. The seemingly simple task of raising people from sitting position to standing position is actually a complex task that incorporates the involvement of the knee and hip extensors to apply lifting torque and restorative balance according to a stepwise sequence of movements. . 29A-29D are (a) sitting shown in FIG. 29A; (b) repositioning the knee over the foot in preparation for the standing shown in FIG. 29B; (c) transition from sitting to standing shown in FIG. 29C; And (d) a normal standing sequence of a healthy person, involving four periods of time, including balance control after the body is shown up in FIG. 29D.

30A-30D illustrate problems in implementing the same standing sequence for a person wearing an active leg support or prosthesis since some of the same four periods are not performed. More specifically, starting from a sitting position, as shown in FIG. 30A, most support or prosthetic wearers maintain sufficient force on their arms to place themselves in a chair, which is shown in FIG. 30B. As above, the knee and upper body are pre-positioned over the ankle and embodied timing (b). However, the lack of force in the quadriceps and hip extensors prevents rotation at a time c to raise the upper leg and upper body out of the chair, as indicated by X 3010 in FIG. 30C. In addition, if there is a lack of sufficient extensor torque in the knees and hips, a patient with improper knee and / or hip muscle strength will slow down and balance the upper body once the standing position is achieved, as indicated by X 3020 of FIG. 30D. It is virtually impossible to keep them.

31A-31D illustrate how an active knee device, such as an active support (shown in FIGS. 22A-22C) or an active knee prosthesis, can be used to assist in the standing sequence of a patient suffering from limb disorder. This requires the ability to recognize that the user is currently sitting and the ability to distinguish between the user's intention to maintain the sitting position and the user's intention to start standing.

One indicator that the user is currently sitting is when the thighs are substantially horizontal, as shown in FIG. 31A. This thigh pose uses an inertial sensor mounted on the knee device that measures the gravity vector against the thigh coordinate system, since the gravity vector in the sitting position lies mainly in the xy plane (orthogonal to the z-axis) of the thigh. Can be determined. Another indicator of being in a sitting position is based on the angle of rotation of the shin relative to the thigh (which can be obtained using an angle encoder) because the ankle joint is typically in front of the knee in the relaxed sitting position. For example, by maintaining a low joint impedance consistent with a relaxed sitting position, these conditions may be measured by the system and dependent on the system to remain in sitting mode.

These same sensors can then be used to detect the user's intention to transition to a standing position by detecting when the ankle joint is located closer to the bottom of the knee joint as shown in FIG. 31B. If such a situation is detected, the system preferably attempts to verify the user's intention to stand by experimentally starting the standing routine. Once the system receives positive feedback from the user, the standing routine continues. However, if the system does not receive positive feedback from the user, the standing routine is stopped.

This experimental initiation of the standing routine can be implemented by causing the knee device to apply torque 3110 that gradually increases with the estimated upward and forward speeds of the hip joint. Positive feedback may be provided by the user trying to stand up with the arm to move the hip joint vertically to assist the standing knee device. In this situation, the hip will move more vertically, and the gravity vector will begin to execute along the z-axis. The system interprets this state as if the user actually wants to stand. In response, the system further applies torque to keep the user up.

Alternatively, the positive feedback from the user may be in the form of EMG signals measured from the thigh and / or hip muscle tissue of the wearer using surface or implanted electrodes. More specifically, if the device detects that the knee joint moves forward toward the ankle joint, the standing routine can be initiated by the wearer bending the quadriceps and / or the hip extensor. The device may use conventional techniques to measure these EMG signals, amplify and filter each muscle output, and extract signal features such as magnitude, change, and / or frequency. These extracted features can then be used to identify the user's intention to stay seated or initiate a standing routine.

Optionally, a foot pressure sensor can be used to detect the ZMP, an upper body mounted IMU can be used to detect the CMP, and a ground reaction force vector is used as the feedback for the wearer to maintain balance on the lower leg. Can be. These pressure sensors communicate with a controller in the knee device using a suitable wireless interface (eg, Bluetooth). As the patient reaches the standing state, the knee device applies a restoring torque 3120 to assist the wearer to balance while in the standing state, as shown in FIG. 31D. The application of knee torque will invalidate the ZMP-CMP over time by steering the force vector before and after the wearer's center of gravity. Examples of suitable approaches to implementing this can be found in US Pat. No. 7,313,463, which is incorporated herein by reference.

It should be noted that while sitting in a chair, the user may sometimes trigger a trial initiation of the standing routine by changing posture in various ways. However, the torque of the applied knee device is relatively small at first, and will quickly disappear with time (eg, 1 second to 2 seconds) if no positive feedback is received from the user. An example of no positive feedback in a non-myoelectric system is when the user does not attempt to raise the body using the arm and the hip joint does not begin to move. An example of no positive feedback in a myoelectric system is when the user does not substantially activate the knee and / or hip extensors so as to be sure that the user's intention is a standing intent. If no positive feedback is received, the system will abort the standing routine and return to a relaxed sitting state.

Optimization methods

25A and 25B show 1) a transition operation to move weight from the following foot to the leading foot during the simultaneous standing phase of the walking cycle, W _t , 2) minimizing hip impact force and force ratio, or 3) cost (purpose) function Is a schematic diagram illustrating controlling a prosthetic device based on a stochastic optimization of minimizing a combination of the two. 25A shows a simplified model used to calculate the transition task. 25B shows a simplified model used to calculate hip impact force and force ratio.

The term stochastic refers to a person's intention; Biomechanical feedback (including walking speed); It assumes that the terrain situation, and the probability function for the terrain feature, minimizes the expected value of the objective function subject to hip impact force and force ratio constraints. Optimization is achieved by changing the impedance, torque and position control variables in the control algorithm. In fact, the transmission energy is minimized by minimizing the negative impact of landing force on the hybrid system energy and maximizing the positive impact of ground force by reflection, and hip impact force limitations are satisfied.

The foregoing optimization can be implemented in real time by inducing "evolutionary" perturbations in key components contributing to biomimetic behavior and measuring the transfer energy from such evolutionary perturbations. The delivered energy may be estimated using a biomechanical model to augment the inertial measurement unit feedback, or in special cases, a temporary inertial measurement unit sub-system (IMU mounted on a belt-shaped body around the upper body and / or the upper leg) It may be used to easily estimate the upper body posture and body center of gravity velocity. Using the Fletcher-Powell method (or other suitable optimization method known to those of ordinary skill in the art), the intelligent evolution of the variable can be derived and the optimization can be calculated. This optimization continues to change due to the walking effect of augmentation. By continually applying this evolutionary perturbation, optimization can be achieved based on continuity. In addition, in the case of initial adjustment or medical examination of the prosthesis or support, this evolutionary optimization may occur at fairly short intervals, i.e. 5 to 10 minutes.

The following describes different phases of the gait cycle of a subject, and in one embodiment, the steps performed by the ankle-joint prosthesis in accordance with the principles of the present invention provide for detecting the operation and control of the ankle-joint prosthesis. will be.

Controlled Plantar  curve

On impact, check the ground reaction force and the zero moment pivot that correspond to the part of the foot that is expected to reach the ground first (from the terrain discrimination model). Confirm the presence of a corresponding change in ankle angle (or ankle torque) and fixation of the appropriate end-point of the foot. After the impact, find a condition where the local terrain slope corresponding to the inertial sole-contact angle is significantly smaller than expected. Saturate ankle spring restoring force and increase damping when it is detected. For topographic differentiation, based on biomechanical model feedback, the topography hypothesis (slope versus staircase) is checked and the wearer is sloppy. For example, a step in the staircase may be measured as a large negative force in the y-direction instead of a large z-force applied to the center of the front portion of the foot. For terrain textures, the heel or front part of the foot will be impacted first. The non-elastic component of the depression associated with this impact will be calculated. On hard ground this depression should be ignored-only elastic strains (foot modules, linear actuators) will be observed. For mud or soft ground, topographic plasticity will be observed by looking at the trajectory of the impacted foot segment. Terrain plasticity will be used as a mitigator for the net work performed on this walking cycle. Slip can also be detected by indicating the forward velocity of the foot segment impacted after the impact. An escalator or vehicle can be detected by indicating that the shin angle does not rotate according to the forward speed of the foot and signaling that the wearer is well balanced and walking over the moving surface. In the case of impedance, the control of the ankle joint applies the optimum impedance using the estimated terrain-referenced velocity attack angle (y) lower limb momentum, estimated terrain slope and terrain characteristics. In reflex control, if a slip is detected, a balance-restored reflection will be generated to move the knee over the ankle. In balance control, optimal balance will typically be achieved by inertia referencing the spring equilibrium after the local terrain slope estimate has been updated at the sole-contact. If the terrain is slippery, the balancing algorithm induces a positive torque "reflection" to "pull" the shin forward, assisting the wearer as the operator positions the knee above the ankle,-the body center of gravity is estimated Align with ground reactions.

Controlled Domination  curve

Once the plantar-contact is detected, the controller inertically references the spring equilibrium angle for this local topographical slope so that if the wearer is standing in alignment with gravity on this slope, the restoring torque is applied by the ankle under static conditions at all. Do not At this point, the local terrain situation is now correctly known. The foot reference coordinates at this “foot-contacting” position are also defined to be used to evaluate the impact of the terrain texture. For terrain textures, the algorithm updates the terrain feature model, in particular to measure plasticity of the surface and its slipping by measuring how the impacted foot segment moves between foot-and-foot contact. Use integrated measurements of slip and strain for reference. These measurements can be used to mitigate ankle impedance and net work (reflection torque of late plantar flexion). In addition, when a "slip" is detected between foot-and-foot contact, the algorithm performed on the controller also determines the slippage between the slippery surface and the escalator / transportation means by observing the shin angular velocity (the way the knee moves relative to the ankle joint). Distinguish. In some cases, since a reliable "ankle joint at zero speed" would not be available at all at this pace, a zero-speed update could not be scheduled. If the terrain is slippery, certain measurements will need to be applied by the balance function. If the foot lands on a moving escalator or conveying means, the rated impedance can be used on a new inertia frame. In impedance control, the control system can apply an optimum impedance that maintains an inertial-reference equilibrium angle; Generate walking speed-dependent stiffness (lower stiffness at faster walking speeds) to enable higher levels of net work; Reduces stiffness at high-plastic surfaces. In reflex control, if a slip is detected, a balance-restored reflection will be created to move the knee over the ankle. In balance control, optimal balance will typically be achieved by inertia referencing the spring equilibrium after the local terrain slope estimate has been updated at the sole-contact. If the terrain is slippery, the balancing algorithm may induce a “pull” or positive torque “reflection” forward of the shin to assist the wearer as the operator positions the knee over the ankle—the ground reaction force from which the body center of gravity is estimated Will be aligned with.

Empowered Plantar  curve

Model monitor slipping and sinking into the surface identify ankle torque limits that can be used for efficient walking in these conditions. For terrain textures, terrain feature estimates are improved in this state and used as inputs for impedance, reflection, and balance functions. For impedance control, the rated impedance variable will be changed to accommodate changes in walking speed, terrain surface properties, and strain and foot slip. A particular "force field"-typically a non-linear actuator force that increases exponentially as the ball-nut reaches a predefined end-stop limit-is applied by the motor controller so that the K3 spring energy (inside the parallel elastic member) ) Is not to exceed the lower limit of the fracture limit. In the case of reflection control, the reflection amplitude will be adjusted to account for the net work “setpoint” from the biomechanical model in combination with the extent to which the terrain can support the manufacture of this net work. In balance control, optimal balance will typically be achieved by inertia referencing the spring equilibrium after the local terrain slope estimate has been updated at the sole-contact. If the terrain is slippery, the balancing algorithm introduces a positive torque "reflection" to "pull" the shin forward to assist the wearer as the wearer places his knee above the ankle-the body center of gravity is estimated. Will be aligned with ground reactions.

Early Embossed

In the case of early bees, immediately after the toes leave the ground, the model monitors the inertia trajectories of the ankles, heel, and toes, and determines when the ankles can bend back to a neutral position without obstruction by terrain. The model calculates the optimal trajectory with the appropriate impedance gain and forward-supply torque to move the ankle to the neutral position (avoid tingling risk) in the fastest, most efficient and stable manner. In the case of topographic distinction, the model begins to keep track of the volume of the sweeping ("non-contacting" foot) and the volume of the foot being moved, so that the toe-down solution is the only viable solution (eg, shallow staircase). Or landing on the ledge), announcing the adaptive ankle positioning function at later stages. In the case of impedance control at initial slope, the neutral value of impedance is applied by the controller. The force-field function is applied to ensure that the linear actuator will not impinge on a rigid stop (end of movement)-a condition under which the actuator will adhere to it (end of movement). In the case of impedance control of the initial angle noticed by the hybrid biomechanical model, the controller controls the impedance to create a trajectory that drives the equilibrium position (ankle angle set point) exponentially to the desired neutral position. The forward feed torque function is applied to reduce the interaction between impedance characteristics and ankle angles following errors that can lead to overshoot and alarms, for example.

review Embossed

In the case of topographic distinction, the model maintains a track of the "clear" volume through which the foot is moved, so that when the toe-down solution is the only feasible solution, ie, shallow staircase or ledge, the late ankle adaptive ankle positioning function. Notify. More generally, the ankle trajectory is monitored and pattern recognition functions are used to determine the likelihood that the foot will land on a stair / resist opposite to the inclined surface. One simple way found to distinguish between the two conditions is to measure the angle at which the ankle velocity is perpendicular to the vertical; In various experiments, it was determined that this angle was less than or equal to 10 degrees and the foot would land on a horizontal step. In the case of impedance control notified by the topographical model, the ankle trajectory (balance) will be modified by the controller as needed to avoid risk of wandering. For example, if the topographical feature provides the maximum likelihood of climbing stairs, additional dorsal flexion is ordered to ensure that the toes do not hold the stairs or ledges. As before, hybrid biomechanical models project continuous and updateable equilibrium trajectories that can be tracked in a secure and stable manner with tight tolerances. In the late-standing state, the biomechanical model calculates the optimal equilibrium angle and ankle impedance that will minimize the objective function including some combination of transfer energy and knee-hip impact force. These optimizations can be implemented through state machine ROM index tables. Or, in an exemplary embodiment, the state controller function will perform optimization in real time using an approximation of rigid body dynamics to calculate and optimize the objective function.

Strains, modifications, and other embodiments disclosed herein will be apparent to those of ordinary skill in the art without departing from the spirit and scope of the claimed invention. Accordingly, the invention will be defined not by the foregoing description, but by the spirit and scope of the following claims.

200: ankle joint
204: inertial measurement unit
208: foot member
212: heel
216: toes
220: retreat member

Claims (20)

As an active support or prosthetic device,
Thigh member;
Lower member;
A knee joint for connecting the thigh member to the lower member;
A rotary motor including a motor shaft output;
The motor drive transmission assembly coupled to the motor shaft output;
A drive transmission assembly coupled with an output of the motor drive transmission assembly, the output of the drive transmission assembly engages with the lower member to apply torque to the knee joint to rotate the lower member relative to the thigh member;
At least one sensor having at least one output from which the position of the knee joint relative to the ankle joint may be determined when the wearer of the device is in a seated position; And
Determine when the knee joint moves to a position before the ankle joint based on the at least one output of the at least one sensor, and in response to the determination a person from the sitting position to a standing positioin A controller to control the rotating motor to adjust the impedance, position or torque of the knee joint to assist in raising
Comprising, an active support or prosthetic device. The method of claim 1,
Wherein the at least one sensor comprises an inertial sensor measuring a gravity vector for the thigh coordinate system,
Active support or prosthetic device. The method of claim 1,
The at least one sensor detecting the rotational angle of the lower leg relative to the thigh,
Active support or prosthetic device. The method of claim 1,
The controller is programmed to maintain a low joint impedance in response to determining that the wearer is in a sitting position,
Active support or prosthetic device. The method of claim 1,
And the controller is programmed to verify the standing intent of the wearer by experimentally starting the standing routine. The method of claim 5, wherein
The controller is programmed to continue to run the standing routine when positive feedback is received and to stop the standing routine if no positive feedback is received. The method of claim 1,
The controller controls the rotating motor to assist in raising a person from the sitting position to a standing position by gradually increasing torque in accordance with the estimated upward and forward speeds of the hip joint,
Active support or prosthetic device. The method of claim 7, wherein
The controller controls the rotating motor to apply a restoring torque to assist the wearer to balance while in the standing state when the patient reaches a standing state,
Active support or prosthetic device. The method of claim 1,
Further comprising at least one pressure sensor configured to measure a force applied to the foot of the wearer,
Active support or prosthetic device. The method of claim 1,
The controller controls the rotating motor to assist in raising a person from the sitting position to an upright position by increasing knee torque in accordance with EMG signals measured from at least one of the knee and hip muscles,
Active support or prosthetic device. A method of controlling a knee support or prosthesis having at least one actuator,
The knee support or prosthesis is worn by a person,
The method comprising:
Detecting the position of the person's knee relative to the person's ankle while the person is in a sitting position;
Determining when the knee is moved to a position before the ankle based on a result of the detecting, and generating an output indicative of the determination; And
In response to said output, operating at least one actuator of said knee support or prosthesis to assist in raising said person from said sitting position to a standing position,
How to control knee supports or prostheses. The method of claim 11,
Further comprising measuring a gravity vector for the thigh coordinate system,
How to control knee supports or prostheses. The method of claim 11,
Detecting a rotational angle of the man's lower leg relative to the thigh of the man,
How to control knee supports or prostheses. The method of claim 11,
Maintaining a low joint impedance in response to determining that the person is in a sitting position,
How to control knee supports or prostheses. The method of claim 11,
Verifying the standing intent of the person by experimentally starting the standing routine,
How to control knee supports or prostheses. The method of claim 15,
Determining whether positive feedback is received; And
If it is determined in the determining that positive feedback is received, continuing to perform the standing routine; And
If it is determined in the determining step that positive feedback is not accepted, further comprising the step of stopping the standing routine,
How to control knee supports or prostheses. The method of claim 11,
Further comprising increasing torque in accordance with the estimated upward and forward velocity of the hip joint,
How to control knee supports or prostheses. The method of claim 17,
Applying a restoring torque to assist the person to balance in the standing state,
How to control knee supports or prostheses. The method of claim 11,
Further comprising measuring a force applied to the foot of the person,
How to control knee supports or prostheses. The method of claim 11,
Further comprising increasing knee torque in accordance with EMG signals measured from at least one of the knee muscle and the hip muscle,
How to control knee supports or prostheses.

Images