JP2006075456A - Wearing type support system based on human body model - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a wearing type support system used for sports, rehabilitation, or the like. <P>SOLUTION: Information obtained from a sensor for measuring the contact force (floor reaction) of a human being and a floor, and a sensor system for measuring the joint angles of the knees, feet, knees, hips, arms, and the like of the human being, is adapted to a human model constructed within a computer of the wearing type support system, to estimate biological information in the moving state of the human being. The wearing type support system controlling the driving moment of a driving device mounted to each joint can be constructed based on the estimate. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a wearable support system used for sports, rehabilitation, and the like.

  Currently, the declining birthrate and aging population are advancing rapidly among developed countries. In particular, in Japan, about one-quarter of the total population is expected to be 65 years old or older in 2015. As a result, the decline in the working population cannot be avoided, and above all, the shortage of people engaged in work such as nursing care cannot be avoided. Against this background of the times, in recent years, systems for assisting caregivers' power, assisting bathing, meal support devices, walking support machines, etc. as systems for encouraging care for the elderly and independence of the elderly Various welfare and nursing care devices have been developed. In particular, the present invention proposes a system that uses a plurality of sensors to acquire and estimate human motion and biological information based on sensor information and supports the motion.

Conventionally, wearable walking support machines, exoskeleton-type power assist suits, and the like have been developed as systems for estimating human movement and supporting the movement with some kind of brace.
H. Kawamoto et al., Power AsiistMethod for HAL-3 Estimating Operator's IntentionBased on Motion Information, Proceedings of the 12th International IEEE Workshopon Robot and Human Interactive Communication, RO-MAN 2003, p.2A2, (2003) K. Kiguchi et al., A 3DOF Exoskeltonfor Upper-Limb Motion Assist-Consideration of the Effect of Bi-Articular Muscles, Proceedings of IEEE International Conference on Robotics and Automation, pp.2424-2429, (2004) Takeshi Koyama et al., Human Assist System for Care (12th Report, Control System Design for Improving Assist Effect), Proceedings of the Japan Society of Opportunities Robotics and Mechatronics '03, 2A1-3F-D6, (2003)

In these systems, it is necessary to estimate human muscle strength, and biological signals such as changes in surface myoelectric potential and muscle surface hardness are used for the estimation. However, joint motion is generated in a complex relationship between the resultant force of all the muscle forces surrounding it, gravity, inertial force, external force, etc., so the assist device must be controlled based on information from several types of muscles. Is not easy. This is because a problem arises in the appropriateness and accuracy of determining the support moment etc. from a small amount of information.
In addition, when using a myoelectric potential generated after a human generates motion, a support moment must be derived, and there will be a time delay before it is generated by some kind of brace. It becomes difficult. In addition, a system for measuring myoelectricity becomes large-scale, and there are many problems to actually put such a system into practical use.

  In the present invention, information obtained from a sensor that measures contact force between the human and the floor (floor reaction force), a sensor system that measures joint angles of human knees, feet, knees, hips, arms, etc. is stored in the computer. By adapting to the constructed human model, joints such as feet, knees, hips, and arms in a human motion state without directly using biological signals such as surface myoelectric potential and muscle surface hardness Build a system to estimate biological information including moments.

    If the above estimation system is used, human movements can be supported in several forms. For example, consider a wearable walking support machine to support human walking. As shown in FIG. 1, the wearable walking support machine is a system constructed from a link, a gear, and a motor and servo brake for supporting a human leg. In such a system, it is possible to calculate a human joint angle or the like by using a rotation angle sensor or the like called an encoder or a potentiometer for a normal link movement or the like. In addition, force sensors and pressure sensors can be attached to the soles of the feet to measure the force (floor reaction force) that a person applies to the floor through the legs.

  By actually mounting such a system on a human leg, the joint angles, angular velocities, and floor reaction forces of each human joint are obtained from the sensors mounted on them, and the information is built in the computer. By applying to the human model, the joint moment acting on each joint can be obtained. Based on such information, the joint moment to be supported by the user (for example, the support moment for reducing the influence of gravity) is calculated, and the support moment is calculated by a motor or servo mounted on the wearable walking support system. It is possible to output the desired moment using the brake and actually support human movement.

  If such a system can be realized, unlike the method using myoelectric potential, the support moment is determined based on the human model, so that it is possible to provide high-speed motor support without any delay in support, etc. By improving the accuracy of the model, the performance of the support system can be dramatically improved. As an application range of this system, it is possible to enable long-term exercise by using it for walking support for elderly people and persons with disabilities and sports that require endurance (for example, skiing).

  Also, without using a wearable support system such as the one described above, use a motion capture system that measures the human motion and movement by measuring the floor reaction force by embedding force sensors and pressure sensors in shoes. If possible, it is possible to estimate biological information such as a human joint moment from the information obtained from the human body model built in the computer.

  Here, in various movements realized by sports, rehabilitation, etc., it is judged whether the current human movement is appropriate or abnormal from the biological information such as the joint moment estimated by the estimation system. It becomes possible to do. At this time, if the judgment information is conveyed to humans using sound or light, for example, in sports, etc., it is possible to compare detailed biological information such as exercise of a professional player and the accompanying movement of the center of gravity with own exercise etc. It becomes like this. Depending on the obstacle, this transmission means is not limited to sound and light, but may be vibration or a method of applying some kind of stimulation to the skin. In rehabilitation and the like, it is possible to compare the ideal motion with the current motion, and to construct a system that suppresses excessive motion that requires excessive joint moments.

  A human body model during walking according to an embodiment of the present invention is shown in FIG. 2 and is approximated by a link model here. This model consists of a cuboid foot link, a lower leg link and thigh link of a truncated cone, and an upper body link that combines a torso part combined with a truncated cone and an ellipsoidal head. The joints connecting these links are hinge joints that can only rotate around the forehead horizontal axis. The weight, length, and shape of each link are determined based on the measured values of the subject, the weight distribution of each part of the human body, and the physique survey document.

Next, the control algorithm according to the present invention for deriving the support force / moment according to the leg state during walking will be described. It is known that walking can be approximated as movement on a sagittal plane. Also, as shown in FIG. 2, the sagittal plane passing through the toes is O _i −Z _i −X _i , the origin O _i is the toes, the Z _i axis is perpendicular to the ground, and the X _i axis is the ground A coordinate system that is parallel to is set, and equations of motion for translational components and rotational components are derived for each link.

Calculate the right knee joint moment τ _kr by _combining the motion equations of the right foot link and right lower leg link, and calculate the left knee joint moment τ _kl by _combining the motion equations of the left foot link and left lower leg link. To do. The calculated knee joint moment is expressed by Equation 1. Here, the subscript i represents the right leg when i = r, and the left leg when i = l.

I _fi and I _si are the moments of inertia of the foot link and crus link, X _si and X _fi are the translational movements of the crus link and foot link centroids, respectively, and I _si is the crus link. , I _ai represents the length from the toe to the ankle joint, and G _fi and G _si represent the center of gravity positions of the foot link and the crus link, respectively. G represents the acceleration due to gravity and τ _GRFi represents the moment due to the floor reaction force.

Here, the support knee joint moment τ _si is designed as follows based on the gravity and floor reaction force terms of the knee joint moment calculated by Equation (1).

Here, α (0 = <α <= 1) is a support rate. The support rate α is a parameter that determines how much the gravity term and floor reaction force term of the calculated knee joint moment are supported, and the larger this is, the larger the support joint moment τ _si is. By adjusting this parameter, support is provided according to the walking ability of each wearer. If this algorithm is applied to a wearable walking support machine, a support knee joint moment can be calculated that is commensurate with the leg movement during walking.

  If the control system in the embodiment of the present invention is applied to the wearable walking assist device shown in FIG. 1, the support knee joint moment calculated based on the floor reaction force information is used in the human body model in the embodiment of the present invention. This enables natural walking support that does not interfere with leg movement during walking.

Next, a wearable walking support device according to another embodiment of the present invention will be described. It is attached to the human leg and supports walking, and is composed of a knee brace with a knee joint called a gear-type dual hinge, a back-drivable linear actuator consisting of a ball screw and a DC motor, and sensors. . The structure is such that the thrust generated by the linear actuator acts on the knee brace frame and transmits the knee joint support moment to the human knee joint.
As sensors, pressure-sensitive sensors are used for measuring the reaction force, and pressure sensors are provided at the parts of the shoe sole that touch the thumb ball and the heel, and potentiometers are provided at the ankle, knee and hip joints for joint angle measurement.

A model diagram of this embodiment of the present invention is shown in FIG. The wearer's body is approximated to a link model from the wearer's body measurement values, and then the walking motion is performed on the treadmill. Here, for the convenience of the wearable walking support machine, θ _{ti is set} to 90 [deg] for the inclination of both foot links. In addition, the floor reaction force is measured by the pressure sensor on the thumb ball and the heel of both foot soles.

  In this example, the support rate was set to 0.1, and the walking speed on the treadmill was set to 2.0 km / h corresponding to about half of the average walking speed of healthy persons. FIG. 4 shows the joint angle, floor reaction force, and support joint moment of the right leg at this time. It can be seen that the support joint moment corresponding to the joint angle and floor reaction force is generated. In addition, it can be seen from FIG. 4 that the support joint moment generated smoothly supports the wearer's walking.

  As described above, according to the present invention, it is obtained from a sensor that measures the contact force (floor reaction force) between a human and the floor, a sensor system that measures joint angles of human knees, feet, knees, waist, arms, and the like. By adapting the obtained information to a human model built in the computer, it is possible to construct a wearable support system that estimates biological information in a certain human movement state.

A wearable walking support machine according to one implementation of the wearable support system of the present invention. Human body link model Anthropometry Temporal changes in joint angle (angle), floor reaction force (GRF) and support joint moment (support torque) during walking.

Claims (5)

    Joints such as feet, knees, hips, and arms in a certain human motion state based on floor reaction force information that the human is in contact with the floor and the joint angles of the human knees, feet, knees, hips, and arms Wearable support system characterized by estimating and controlling the joint moment applied to the body.     Human body model calculation system that builds a human model, floor reaction force measurement sensor that obtains information that the human is in contact with the floor, and the angle of each joint such as the human knee, foot, knee, waist, and arm The human body model calculation system is used to estimate and control the joint moment applied to each joint such as feet, knees, hips, and arms in a certain human motion state using the human body model calculation system. The wearable support system according to claim 1.     2. The apparatus according to claim 1, further comprising: a driving device that applies a rotational driving force to each joint such as the knee, foot, knee, waist, and arm; and a braking device that controls the driving force of the driving device. Wearable support system.     The joint state is transmitted to a wearer of the wearable support system based on information on joint moments applied to each joint such as a foot, a knee, a waist and an arm in a certain motion state of the human. Wearable support system.     5. The wearable support system according to claim 4, wherein the means for transmitting is sound, light, vibration, or stimulation to the skin.