JP2006110217A - Muscle activity estimation method, program, storage medium, muscle activity estimation system and muscle activity database - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To highly precisely estimate the muscle activity of a person in an art estimating the activity of the muscle of the person. <P>SOLUTION: Respective quantities of movement states of the joint and the muscle in an analysis site in a person are inputted as a quantity of joint movement state and a quantity of muscle movement state, and an increment of at least one physical characteristics out of the rigidity and the viscosity of an object in a state where a contact site of the person and the object in a state where a contact site in the person and the object physically come into contact with each other, from the physical characteristic of the object alone is inputted as an equivalent physical characteristic equivalent to the physical characteristic of the person alone (S1002). The muscle activity is so calculated as to substantially minimize the total value in the whole attending movement direction of the object of a difference between the input values and estimated values of the equivalent physical characteristic based on the input quantity of the joint movement state, quantity the muscle movement state, and equivalent physical characteristic (S1005). The estimated value is described by using a function using the muscle activity, the quantity of the joint movement state, and the quantity of the joint movement state as parameters, respectively. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a technique for estimating the activity of a human muscle, and particularly to a technique for improving the estimation accuracy.

  A technology for analyzing human behavior already exists. Human behavior is generally used as a term that includes static behavior or posture and dynamic behavior or motion. This kind of behavior analysis technology is implemented in fields such as medical analysis, anatomical analysis, ergonomic analysis, biomechanical analysis, and computer graphics for producing images in which virtual humans appear.

  This type of behavior analysis technology may be implemented in the field of numerically analyzing human behavior. In this case, this kind of behavior analysis technique is implemented, for example, to analyze a human behavior geometrically or to analyze it dynamically.

  Furthermore, this kind of behavior analysis technology is sometimes implemented for the purpose of measuring and analyzing human behavior using motion capture etc., but predicts the behavior that humans exhibit in a specified environment. It may be implemented for the purpose of doing this. However, in order to achieve the latter purpose, it is possible to refer to the result of actual measurement of human behavior using motion capture or the like.

  An example of a field that is implemented for the purpose of predicting human behavior by this kind of behavior analysis technology is a field in which a human predicts a behavior exhibited in a vehicle as a moving body.

  In this field, for example, an occupant that responds to an occupant hitting a vehicle structure because the vehicle hits or is about to hit an obstacle prior to or in parallel with the design of the vehicle. Behavior (displacement and load) is analyzed. The analysis result is fed back to the vehicle design, which is useful for designing a vehicle with improved collision characteristics.

  Further, in this field, a human who responds to an occupant operating an operating device (for example, a brake operating device, an accelerator operating device, a steering operating device) in the vehicle prior to or concurrently with the design of the vehicle. Behavior (displacement and load) is analyzed. The analysis result is fed back to the vehicle design, which is useful for designing an operating device with improved operational feeling.

  Japanese Patent Laid-Open No. 2004-228561 is based on an optimal control theory using a minimum muscle motion function as an objective function in order to reproduce a natural motion without a hassle when animating a human body motion by simulation. Disclosed is a technology for estimating human movement by simulation based on science.

Patent Document 2 describes the calculation of forward dynamics for a model with the aim of making it possible to create a detailed model of the human body with bone geometry data and muscle / tendon / ligament data using commercially available modeling software. A technique for executing (processing for calculating movement from muscle force) and inverse dynamics calculation (processing for calculating muscle force and ligament force necessary for movement) at high speed is disclosed. In this technology, dynamic calculation using a human body model is performed so that force balance is established between joint torque and muscle force.
Japanese Patent Laid-Open No. 11-85209 JP 2003-339673 A

  In order to analyze human behavior, it goes without saying that it is necessary to specify the external force that acts on the human being. However, humans generate a mechanism that generates force and outputs it to the outside, that is, muscle tissue. Therefore, it is also necessary to specify the output of the muscle tissue.

  Therefore, in order to improve the analysis accuracy of human behavior, it is important to perform human behavior analysis so that the characteristics of real human muscles are reflected. On the other hand, in the muscle characteristics, rigidity (also referred to as elasticity) as a physical quantity that correlates force and displacement and viscosity (also referred to as attenuation coefficient) as a physical quantity that correlates force and displacement speed are muscle physical characteristics. Exists as. Therefore, in order to improve the analysis accuracy of human behavior, it is effective to perform human behavior analysis in consideration of muscle physical characteristics.

  The muscle physics characteristics are not uniquely determined for each muscle type, and are not uniquely determined for each muscle length (posture) and muscle length change speed (posture change speed). The muscle activity exists as a physical quantity independent of the muscle movement state quantity such as the muscle length and the muscle length change speed, and the muscle physical characteristics are influenced by the muscle activity. Therefore, in order to perform human behavior analysis so that muscle physics characteristics are reflected sufficiently accurately, it is important to consider muscle activity on the assumption that muscle physics characteristics depend on muscle activity. is there.

  Muscle activity is not necessarily associated with commands from the human brain. Although muscle activity does not increase only with human active behavior, it is necessary to consider muscle activity, especially when it is necessary to analyze human behavior in consideration of active behavior. It is important to improve accuracy. The active behavior includes, for example, an operation in which a passenger in a vehicle colliding with an obstacle avoids danger by itself. Examples of the danger avoidance operation include posture, escape, reflex motion, and brake operation.

  However, in the conventional method for analyzing the human behavior, the muscle activity is not considered on the assumption that the muscle physical property depends on the muscle activity. Therefore, it is difficult to analyze human behavior sufficiently accurately with the conventional method.

  Hereinafter, this will be described more specifically.

  In the conventional method, the behavior of the human being to be analyzed is estimated on the assumption that the human will use the muscle strength efficiently without waste.

  However, the actual human behavior is not only for the purpose of efficiently performing the operation, but also for the purpose of enhancing the ease of controlling a specific part of the human, for example, the position of the hand. It may be performed for the purpose of maximizing the force at the hand in an emergency or the like. In the latter two cases, the activity of the muscle of the hand is higher than normal, and as a result, the muscle physics characteristics including the rigidity of the muscle are also higher than normal.

  However, since the conventional method analyzes human behavior without considering the above-described facts, it is difficult for this conventional method to accurately estimate the muscular strength for an arbitrary motion.

  In view of the circumstances described above, the present invention has been made with an object of estimating human muscle activity more accurately.

  The following aspects are obtained by the present invention. Each aspect is divided into sections, each section is given a number, and is described in a form that cites other section numbers as necessary. This is to facilitate understanding of some of the technical features that the present invention can employ and combinations thereof, and the technical features that can be employed by the present invention and combinations thereof are limited to the following embodiments. Should not be interpreted. That is, although not described in the following embodiments, it should be construed that it is not impeded to appropriately extract and employ the technical features described in the present specification as the technical features of the present invention.

  Further, describing each section in the form of quoting the number of the other section does not necessarily prevent the technical features described in each section from being separated from the technical features described in the other sections. It should not be construed as meaning, but it should be construed that the technical features described in each section can be appropriately made independent depending on their properties.

(1) A human being mainly composed of a plurality of muscles and a plurality of bones, and having a characteristic in which a physical characteristic that is at least one of rigidity and viscosity of each muscle changes according to the activity of the muscle, In a state where at least one part of a plurality of parts constituting a human being is a contact part and in contact with an object at the contact part, a plurality of muscles in the analysis part selected as an analysis target among the plurality of parts are respectively A muscle activity estimation method for estimating the activity shown by a computer as muscle activity,
The joint state and muscle motion state quantities at the analysis site are input as joint motion state quantities and muscle motion state quantities, and the rigidity of the object in a state where the contact site and the object are in physical contact with each other. And an input step of inputting, as an equivalent physical property equivalent to the physical property of the human alone, an increment of the physical property that is at least one of viscosity and viscosity from the physical property of the object alone;
Based on the input joint motion state quantity, muscle motion state quantity, and equivalent physical characteristics, the muscle activity is calculated based on the difference between the input value and the estimated value of the equivalent physical characteristics in all the motion directions of interest in the object. A muscle activity calculation step for calculating the total value so as to be substantially minimized, wherein the estimated value includes the muscle activity, the motion state quantity of the muscle, and the motion state quantity of the joint as variables; A method for estimating muscle activity, which includes:

  In a state where the contact site and the object of the human are in physical contact with each other, the physical characteristic that is at least one of rigidity and viscosity of the object increases from the physical characteristic of the object alone. The increment is equivalent to the physical property of a single person and functions as an equivalent physical property.

  On the other hand, a certain relationship is established among physical characteristics (at least one of rigidity and viscosity) of individual human, muscle activity, muscle motion state quantity, and joint motion state quantity. This will be described in detail later. Of these variables, the muscle motion state quantity and the joint motion state quantity can be easily obtained by various methods.

  Therefore, in the muscle activity estimation method according to this section, the muscle activity is calculated based on the input joint motion state quantity, muscle motion state quantity, and equivalent physical characteristics. Specifically, the muscle activity is calculated so that the sum of the difference between the input value of the equivalent physical characteristic and the estimated value in the movement direction of interest in the object is substantially minimized. Here, the estimated value of the equivalent physical property is described using a function having the muscle activity, the muscle motion state quantity, and the joint motion state quantity as variables, and the function is specified in advance.

  Therefore, according to this muscle activity level estimation method, it is possible to estimate the muscle activity level based on the joint motion state quantity, the muscle motion state quantity, and the equivalent physical characteristics.

(2) The input step further inputs a contact force acting at the contact point between the contact part and the object,
The muscle activity estimation method further includes:
A joint torque calculating step of calculating a joint torque acting around the joint at the analysis site by an inverse dynamic analysis method based on the input joint motion state quantity and the contact force;
The muscle activity calculation step calculates the muscle activity so that the total value is substantially minimized and a force balance is substantially established between the joint torque and the muscle strength of the muscle. The muscle activity estimation method according to item (1).

  The force generated from each human part is usually the resultant force of a plurality of forces generated by a plurality of muscles. On the other hand, it is ideal to estimate the muscle activity for each muscle, but the muscle activity for each muscle is estimated mainly from the resultant force of multiple muscles and the combined stiffness or viscosity of multiple muscles. In the analysis environment that must be performed, the muscle activity should be calculated on the premise of the balance of force between the muscle motion state quantity and the joint motion state quantity, that is, between the muscle force and the joint torque. Is desirable.

  Based on such knowledge, in the muscle activity estimation method according to this section, based on the joint motion state amount, the muscle motion state amount and the equivalent physical property, the difference between the input value of the equivalent physical property and the estimated value, The muscle activity is calculated so that the total value in all the motion directions of interest of the object is substantially minimized, and a force balance is substantially established between the joint torque and the muscle force.

  Therefore, according to this muscle activity level estimation method, it is easier to improve the estimation accuracy than when muscle activity level is estimated without considering force balance between joint torque and muscle strength.

(3) The joint torque calculation step includes:
(A) Based on the input joint motion state quantity and the contact force, an apparent joint torque that acts around the joint at the analysis site by the inverse dynamics analysis method, that is, an actual joint torque and a joint passive resistance torque, An apparent joint torque calculating step for calculating a composite of
(B) an actual joint torque calculation step of calculating the actual joint torque from the calculated apparent joint torque and the joint passive resistance torque,
The muscle activity calculation method according to (2), wherein the muscle activity calculation step calculates the muscle activity using the calculated actual joint torque.

  According to this muscle activity estimation method, it is easier to improve the calculation accuracy than when the joint torque is calculated without considering the joint passive resistance torque.

(4) The joint torque calculation step uses a skeletal model that physically represents the human on at least the plurality of bones on the computer, and gives a predetermined motion to the skeleton model, The muscle activity estimation method according to (2) or (3), including a step of calculating the joint torque.

  The “model” in this section and each of the following sections can be a whole body model that represents the whole human body or a local model that represents a part of the human body.

(5) The skeletal model is a multi-joint that expresses a plurality of parts constituting the human body, which is the human body, as a plurality of rigid segments, so that the plurality of rigid segments can rotate around the joint. The muscle activity estimation method according to item (4), including a rigid model.

  According to this muscle activity estimation method, the joint torque can be easily calculated by using the articulated rigid body model, compared to the case of using the finite element human body model.

(6) Further, a musculoskeletal model that physically represents the human on at least the plurality of muscles and the plurality of bones on the computer is used, and a predetermined action is given to the musculoskeletal model. Accordingly, the muscle activity state estimation method according to any one of (1) to (5), including a muscle motion state amount calculation step of calculating the muscle motion state amount.

(7) The musculoskeletal model expresses a plurality of portions constituting the human body, which is the human body, as a plurality of rigid segments, so that the plurality of rigid segments can rotate around the joint. The muscle activity estimation method according to item (6), including an articulated rigid body model that represents each of the plurality of muscles as a plurality of wire elements.

  According to this muscle activity estimation method, it is easier to calculate the muscle motion state quantity by using the articulated rigid body model than when using the finite element human body model.

(8) The joint motion state quantity includes at least one of an angle, an angular velocity, and an angular acceleration of the joint,
The muscle activity state estimation method according to any one of (1) to (7), wherein the muscle movement state quantity includes at least one of a length of the muscle and a change speed thereof.

(9) The human is mainly composed of a plurality of muscle element models and a plurality of bone element models, and a physical characteristic that is at least one of rigidity and viscosity assigned to each muscle element model is the muscle element model. The muscle activity estimation method according to any one of (1) to (8), wherein the muscle activity is expressed on the computer by a human body model having a characteristic that changes in accordance with the activity assigned to.

(10) The human being is an occupant riding a vehicle,
The object is a structure of the vehicle where the occupant is expected to collide when the vehicle collides with an obstacle,
The muscle activity estimated by the muscle activity estimation method is used to analyze the response behavior of the occupant to the collision with the structure. The muscle activity according to any one of (1) to (9) Degree estimation method.

(11) The human being is an occupant riding a vehicle,
The object is an operating device that is assumed to be operated by the occupant in the vehicle,
The muscle activity estimated by the muscle activity estimation method is used for analyzing the response behavior of the occupant responding to the operation of the operating device, according to any one of (1) to (9). Activity estimation method.

(12) The muscle activity estimation method according to (11), wherein the operation device includes at least one of a brake operation device, an accelerator operation device, a steering operation device, and a shift operation device.

  The “brake operating device”, “accelerator operating device”, “steering operating device”, and “shift operating device” in this section are each an operating member operated by an occupant (for example, a brake pedal, an accelerator pedal, a steering wheel, a shift lever). , Switch, dial, button, etc.).

(13) The occupant is a driver of the vehicle,
In the vehicle, at least one of its running state and behavior is automatically controlled regardless of the driver's intention,
The muscle activity estimation method is used to analyze the response behavior in consideration of an automatic control state of at least one of the running state and behavior of the vehicle. The method for estimating muscle activity as described in 1. above.

(14) A human being mainly composed of a plurality of muscles and a plurality of bones and having a characteristic in which a physical characteristic that is at least one of rigidity and viscosity of each muscle changes according to the activity of the muscle, In a state where at least one of a plurality of parts constituting a human being is a contact part and is in contact with an object at the contact part, a plurality of muscles in a selected part selected as an analysis target among the plurality of parts are respectively shown. A muscle strength estimation method for estimating muscle strength by a computer,
The muscle activity estimation method according to any one of (1) to (13);
A muscle strength estimation method including: a muscle strength calculation step of calculating muscle strength according to a predetermined relationship between muscle activity and strength based on the muscle activity estimated by the muscle activity estimation method.

  In this muscle strength estimation method, muscle activity is estimated according to the same principle as the muscle activity estimation method according to (1) above in a state where the muscle activity is treated as a variable physical quantity, and the estimated muscle activity is further estimated. The muscle strength is estimated using the degree.

  Therefore, according to this muscle strength estimation method, it is easier to improve the estimation accuracy than when muscle strength is estimated in a state where the muscle activity is treated as a fixed physical quantity.

(15) The predetermined relationship includes a predetermined relationship among muscle activity, maximum muscle strength, muscle length, and muscle length change rate for each muscle. Muscle strength estimation method.

(16) A human being mainly composed of a plurality of muscles and a plurality of bones, and having a characteristic in which a physical characteristic that is at least one of rigidity and viscosity of each muscle changes according to the activity of the muscle, At least one of a plurality of parts constituting a human being as a contact part, and selected as an analysis target among the plurality of parts in an analysis target period that is at least one of immediately before and immediately after colliding with an object at the contact part A human behavior estimation method for estimating, by a computer, the behavior shown in a selected part,
A muscle activity estimation method according to any one of claims 1 to 13,
An input step of inputting an external force and a displacement acting on the person in the analysis target period;
Based on the input external force and displacement and the muscle activity estimated by the muscle activity estimation method, the person is physically represented on the computer with respect to at least the plurality of muscles and the plurality of bones. A human behavior estimation method including: a behavior calculation step of calculating a behavior exhibited by the human in the period to be analyzed by using a human physical model.

  In this human behavior estimation method, muscle activity is estimated according to the same principle as the muscle activity estimation method according to the above (1) in a state where the muscle activity is treated as a variable physical quantity, and the estimated muscle Human behavior is estimated using activity.

  Therefore, according to this human behavior estimation method, it is easier to improve the estimation accuracy than when human behavior is estimated in a state where the muscle activity is treated as a fixed physical quantity.

  Furthermore, the fact that muscle activity is treated as a variable physical quantity means that it is possible to analyze human behavior in consideration of changes in muscle activity according to, for example, the presence or strength of human consciousness. Therefore, according to this human behavior estimation method, for example, it is easy to analyze human behavior including active behavior.

(17) A finite element human body model that expresses a human body, which is a human body, by a plurality of finite elements, in which attribution relationships between a plurality of parts constituting the human and the plurality of finite elements are set in advance. It is a human behavior analysis method that analyzes the behavior that humans show by using,
By classifying and combining the plurality of finite elements constituting the finite element human body model according to the belonging relationship, the plurality of portions are respectively defined as a plurality of rigid body segments, and the plurality of rigid body segments rotate around the joint. A skeletal model creation process for creating a skeletal model that expresses as possible,
In accordance with what expresses the plurality of muscles among the plurality of finite elements in the finite element model, on the created skeletal model, wire elements that express the plurality of muscles as wires are respectively represented by the rigid bodies of the muscles. A musculoskeletal model creation step for creating a musculoskeletal model by defining it with the position of the attachment point to the segment and the waypoint in the middle;
An information acquisition step for acquiring information necessary for analyzing the behavior exhibited by the human in a specified environment using the finite element human body model by using the created skeletal model and musculoskeletal model;
A finite element human body model restoring step of restoring the finite element human body model from the created musculoskeletal model so as to represent the acquired information;
A human behavior analysis method comprising: a behavior calculation step of calculating a behavior exhibited by the human in the designated environment by using the restored finite element human body model.

  In general, in order to analyze human behavior, there are a method of performing a finite element analysis using a finite element human body model and a method of performing a simple dynamic calculation using an articulated rigid body model (multibody model).

  The former method has the advantage of being able to accurately analyze the deformation and displacement of each part of the analysis target. However, the amount of displacement of the analysis target is high, such as the problem of large computational load and active human movement. When analyzing a problem that each element representing the analysis target is greatly deformed, there is a problem that the analysis accuracy is remarkably lowered.

  On the other hand, the latter method is advantageous in that it has a light calculation load and can easily handle a large movement, and this advantage is a disadvantage of the former method.

  Therefore, in the human behavior analysis method according to this section, although the final analysis of human behavior is performed using a finite element human body model, it is the information necessary to improve the accuracy of the finite element analysis, and the same finite What is difficult to obtain using the element human body model is obtained using the skeletal model and the musculoskeletal model.

  Therefore, according to this human behavior analysis method, the accuracy of human behavior is improved by compensating for the shortcomings of finite element analysis using finite element human body models with the advantages of both skeletal model and musculoskeletal model, which are both articulated rigid body models And the reduction of calculation load can be easily achieved.

  Furthermore, according to this human behavior analysis method, a skeletal model is created from a finite element human body model, and a musculoskeletal model is created from the created skeletal model. While easily sharing characteristics with the human body model, it is easier to create than the case of creating from zero without relying on the finite element human body model.

(18) A program executed by a computer to implement the method according to any one of (1) to (17).

  If this program is executed by a computer, the same effect can be realized according to basically the same principle as in the method according to any one of the above items (1) to (17).

  The program according to this section can be interpreted so as to include not only a combination of instructions executed by a computer to fulfill its function, but also files and data processed in accordance with each instruction.

  In addition, this program may achieve its intended purpose by being executed by a computer alone, or may be intended to achieve its intended purpose by being executed by a computer together with other programs. it can. In the latter case, the program according to this section can be mainly composed of data.

(19) A recording medium on which the program according to item (18) is recorded so as to be readable by a computer.

  If the program recorded on this recording medium is executed by a computer, the same operation and effect as the method according to any one of (1) to (17) can be realized.

  This recording medium can adopt various formats, for example, a magnetic recording medium such as a flexible disk, an optical recording medium such as a CD and a CD-ROM, a magneto-optical recording medium such as an MO, and an unremovable storage such as a ROM. Any of these may be adopted.

(20) A human being mainly composed of a plurality of muscles and a plurality of bones and having a characteristic in which a physical characteristic which is at least one of rigidity and viscosity of each muscle changes according to the activity of the muscle, In a state where at least one part of a plurality of parts constituting a human being is a contact part and in contact with an object at the contact part, a plurality of muscles in the analysis part selected as an analysis target among the plurality of parts are respectively A muscle activity estimation system that estimates the activity shown as muscle activity,
The joint state and muscle motion state quantities at the analysis site are input as joint motion state quantities and muscle motion state quantities, and the rigidity of the object in a state where the contact site and the object are in physical contact with each other. And an input means for inputting an increment of the physical property that is at least one of viscosity and viscosity from the physical property of the object alone as an equivalent physical property that is equivalent to the physical property of the human alone,
Based on the input joint motion state amount, muscle motion state amount, and equivalent physical property, the muscle activity is calculated based on the difference between the input value of the equivalent physical property and the estimated value in all the motion directions of interest in the object. The muscle activity calculating means for calculating the total value so as to be substantially minimized, wherein the estimated value includes the muscle activity, the muscle motion state quantity, and the joint motion state quantity as variables, respectively. A muscle activity estimation system including:

  According to this muscle activity estimation system, muscle activity is estimated according to basically the same principle as the muscle activity estimation method according to the above item (1).

  The muscle activity degree calculation means can be implemented in a form constituted by a computer including a processor and a storage. In this aspect, for example, data representing the input joint motion state quantity, muscle motion state quantity, and equivalent physical characteristics are read from the storage, and the read data and the storage are stored in advance. Based on the relational expressions (for example, Expressions (4) and (5) in FIG. 16), an optimization calculation technique is used to obtain the input value of the equivalent physical characteristic and the estimated value of the equivalent physical characteristic. It is described using a function (for example, a function represented by Expression (7) to Expression (9) in FIG. 18) having the activity, the movement state quantity of the muscle, and the movement state quantity of the joint as variables. The muscle activity is calculated so that the total value of the difference from the object in all the movement directions of interest in the object is substantially minimized.

(21) The input means further inputs a contact force acting at the contact point between the contact part and the object,
The muscle activity estimation system further includes:
A joint torque calculating means for calculating a joint torque acting around the joint at the analysis site by an inverse dynamics analysis method based on the input joint motion state quantity and the contact force;
The muscle activity calculation means calculates the muscle activity so that the total value is substantially minimized and a force balance is substantially established between the joint torque and the muscle strength of the muscle. The muscle activity estimation system according to item (20).

  In this muscle activity estimation system, muscle activity is estimated according to basically the same principle as the muscle activity estimation method according to item (2). Therefore, according to this muscle activity estimation system, it is easy to estimate muscle activity with higher accuracy than when muscle activity is estimated without considering force balance between joint torque and muscle strength. .

(23) A human being mainly composed of a plurality of muscles and a plurality of bones and having a characteristic in which a physical characteristic that is at least one of rigidity and viscosity of each muscle changes according to the activity of the muscle, In a state where at least one part of a plurality of parts constituting a human being is a contact part and in contact with an object at the contact part, a plurality of muscles in the analysis part selected as an analysis target among the plurality of parts are respectively A muscle activity database in which the activity shown is read as a muscle activity by executing a search program by a computer,
(A) a joint motion state quantity which is a joint motion state quantity at the analysis site;
(B) Equivalent physical characteristics that are equivalent to the physical characteristics of the person alone, and are at least one of rigidity and viscosity of the object in a state where the contact portion and the object are in physical contact with each other. A physical property equal to the increment from the physical property of the object alone;
(C) a contact force acting at the contact point between the contact site and the object;
(D) A plurality of physical quantities including the muscle activity are recorded in advance in association with each other as variables, and a combination of the joint motion state quantity, the equivalent physical characteristic, and the contact force by executing the search program If a muscle activity is identified, a muscle activity database associated with the combination is identified and output by executing the search program.

  According to this muscle activity database, a combination of a plurality of physical quantities including joint motion state quantities, equivalent physical characteristics, and contact force is specified, and a normal search program is executed on the muscle activity database by a computer. Then, the muscle activity corresponding to the identified combination is read from the muscle activity database without any special analysis or calculation.

  The relationship between the joint motion state quantity, the equivalent physical property, the contact force, and the muscle activity recorded in the muscle activity database is calculated by, for example, the method according to any one of (1) to (16) described above. It is possible to obtain by performing using.

  The “joint motion state quantity” in this section is generally defined to include at least one of the joint angle, angular velocity, and angular acceleration.

(24) A human being mainly composed of a plurality of muscles and a plurality of bones and having a characteristic in which a physical characteristic that is at least one of rigidity and viscosity of each muscle changes according to the activity of the muscle. In a state where at least one part of a plurality of parts constituting a human being is a contact part and in contact with an object at the contact part, a plurality of muscles in the analysis part selected as an analysis target among the plurality of parts are respectively A muscle activity database in which the activity shown is read as a muscle activity by executing a search program by a computer,
(A) a joint motion state quantity which is a joint motion state quantity at the analysis site;
(B) a muscle motion state quantity that is a muscle motion state quantity at the analysis site;
(C) Equivalent physical characteristics that are equivalent to the physical characteristics of the person alone, and are at least one of rigidity and viscosity of the object in a state where the contact portion and the object are in physical contact with each other. A physical property equal to the increment from the physical property of the object alone;
(D) A plurality of physical quantities including the muscle activity are recorded in association with each other in advance as variables, and the joint motion state quantity, the muscle movement state quantity, and the equivalent physical characteristics are obtained by executing the search program. If a combination of these is specified, a muscle activity database associated with the combination is specified and output by executing the search program.

  According to this muscle activity database, a combination of a plurality of physical quantities including a joint motion state quantity, a muscle movement state quantity, and an equivalent physical characteristic is specified, and then a normal search program is executed for the muscle activity database. If the process is executed, the muscle activity corresponding to the identified combination is read from the muscle activity database without any special analysis or calculation.

  The relationship between the joint motion state quantity, the muscle motion state quantity, the equivalent physical property, and the muscle activity recorded in the muscle activity database is, for example, the method according to any one of (1) to (16) above. Can be obtained by implementing using a computer.

  The “joint motion state quantity” in this section is generally defined to include at least one of the joint angle, angular velocity, and angular acceleration. The “muscle movement state quantity” is generally defined to include at least one of muscle length and muscle length change rate.

  Hereinafter, one of more specific embodiments of the present invention will be described in detail with reference to the drawings.

  FIG. 1 conceptually illustrates in block diagram form a hardware configuration of a system 10 suitable for implementing a human behavior analysis method according to an embodiment of the present invention.

  This system 10 uses a human body model that can express the behavior of an occupant on the computer 20, so that when the vehicle on which the occupant is in contact with an obstacle (from the outside, the vehicle on which the occupant is on is impacted). The behavior, load (stress), and deformation that occur in each part of the occupant are analyzed by simulation.

  As shown in FIG. 1, this system 10 is configured by connecting an input device 22 and an output device 24 to a computer 20. As is well known, the computer 20 includes a processor 30 and a storage 32 connected to each other by a bus 34. In the computer 20, a necessary program is read from the storage 32 and executed by the processor 30. At this time, data necessary for the execution is read from the storage 32, and data representing the execution result is obtained as necessary. It is stored in the storage 32 and saved.

  For example, although not shown, the input device 22 is configured to include a mouse and a keyboard as a pointing device. Although not shown, the output device 24 is configured to include a monitor that displays images such as characters and graphics on the screen.

  As shown in FIG. 1, the storage 32 is provided with a program memory 40 and a data memory 42. Various programs including a finite element analysis program are stored in the program memory 40 in advance. In the data memory 42, model data for defining a plurality of types of human body models that three-dimensionally represent the whole human body on the computer 20 is stored in advance.

  These multiple types of human models were restored as finite element models from finite element human models, skeletal models that are multi-joint rigid segment models, musculoskeletal models that are also multi-joint rigid segment models, and musculoskeletal models. Includes a restored human model.

  The finite element human body model is a model for approximately reproducing a human body by dividing the human body into a plurality of finite elements (hereinafter simply referred to as “elements”). On the other hand, the articulated rigid segment model is also called a mechanism model, and is a model that reproduces the motion rather than the shape of the human body by connecting a plurality of rigid segments so as to be rotatable around a plurality of joints. is there.

  In the finite element human body model, each tissue of the human body is represented by a plurality of elements, and attributes such as the shape, density, material property value, etc. of each element are defined. The physical quantities such as the positions of a plurality of nodes constituting each element (an example of geometric quantities) and material property values are all defined in a single global coordinate system.

  In this finite element human body model, the whole human body is modeled not only for skeletal tissues but also for connective tissues such as ligaments, tendons, muscles, etc., so that the whole human body can be analyzed with its anatomical shape, structure and characteristics. Is expressed faithfully. As a result, the physical response of each part to the external force is accurately expressed.

  Furthermore, the finite element human body model is assumed to have a variable parameter that defines at least one of the shape, position, mechanistic property, and mechanical property of at least one region thereof, and the variable parameter is Prior to human behavior analysis, it is specified by a designer or analyst who is a user.

  In the present embodiment, the finite element human body model is used to perform simulation analysis on the behavior of the occupant and the degree of deformation when an external impact is applied to the vehicle. This simulation analysis is performed by using a finite element method for the behavior of the occupant and the degree of deformation when an impact is applied to the vehicle from the outside.

  As is well known, the finite element method is one of computer-based numerical calculation methods. A continuum is divided into a large number of elements having a finite size, for example, meshes, and physical quantities of individual elements are divided. Is a method of numerically calculating the entire system by incorporating the above into the simultaneous material property equation. The material characteristic equation has a material characteristic coefficient as a coefficient for controlling the analysis characteristic of the finite element method.

  More specific configuration of the finite element human body model and human behavior based on the finite element human body model by the finite element method (for example, at the time of collision, acceleration / deceleration when riding on a roller coaster, and falling Since the method for analyzing the behavior is disclosed in Japanese Patent Application Laid-Open No. 2002-149719 of the present applicant, the duplicate description is omitted by quoting by reference.

  In FIG. 2, the finite element human body model is shown in a perspective view regarding only the lower limbs of the human body. As shown in FIG. 2, this finite element human body model reproduces not only the human skeleton but also soft tissues such as muscles and skin. The lower limb part in this finite element human body model is divided into a waist part 100, left and right thigh parts 102, 102, left and right shin parts 104, 104, and left and right foot parts 106, 106.

  In FIG. 3, the skeleton model is shown in a perspective view with respect to a portion corresponding to the portion shown in FIG. 2 in the finite element human body model. As shown in FIG. 3, this skeleton model reproduces only the skeleton of the human body. The lower limbs in this skeleton model are divided into a hip bone 110, left and right femurs 112, 112, left and right tibias 114, 114, left and right ribs 116, 116, and left and right foot bones 118, 118. . Each foot bone 118 is a combination of fine bones.

  In this skeleton model, each tissue of the human body is represented by a plurality of rigid segments, and these rigid segments are connected to each other so as to be rotatable at joints. The rigid segments include a hip bone 110, left and right femurs 112, 112, left and right tibias 114, 114, left and right ribs 116, 116, and left and right foot bones 118, 118.

  The surface of the shape of each rigid segment is composed of polygons. A plurality of vertices on a polygon are constituted by a plurality of nodes on a plurality of elements which are bases of the polygon. Therefore, the shape of each rigid body segment is represented by the corresponding polygon data. A local coordinate system is assigned to each rigid segment, and in the local coordinate system, the mass, moment of inertia, and polygon position of the corresponding rigid segment are defined.

  In FIG. 4, the musculoskeletal model is shown in a perspective view with respect to a portion corresponding to the portion shown in FIG. 3 in the skeletal model. As shown in FIG. 4, this musculoskeletal model is configured by adding muscles as wire elements (shown by thick solid lines in FIG. 4) to the skeletal model shown in FIG. Has only been reproduced. That is, this musculoskeletal model is a combination of the skeleton model described above and a muscle model (muscle element model) configured as a wire element.

  In this musculoskeletal model, each muscle is attached to a corresponding bone (a position where relative movement with respect to the bone is prevented, and is indicated by a black circle in FIG. 4) and a corresponding bone. A point where each muscle passes in the middle between attachment points (relative movement in the direction in which the muscle extends with respect to the bone is allowed, but relative movement in the direction intersecting with the bone is blocked. Defined). Anatomically, since muscles are attached to the surface of bone, each muscle model is also attached to the surface of each rigid body model (bone element model).

  FIG. 5 is a schematic diagram of the human behavior analysis method that is executed by the computer 20 in a process diagram. The computer 20 constitutes a human behavior analysis system in cooperation with the input device 22 and the output device 24, and a portion related to the estimation of muscle activity in the human behavior analysis system constitutes a muscle activity estimation system. ing. The outline of the human behavior analysis method will be described below in association with the above-described four types of human body models.

  As shown in FIG. 2, at the time of each execution of the human behavior analysis method, first, in step S1 (hereinafter, simply referred to as “S1”. The same applies to other steps), the entire finite element model is Each element is sequentially read from the data memory 42 to a working area (not shown), which is another storage area in the storage 32. Precisely, model data representing the entire finite element model is sequentially read out for each element data representing each element.

  Next, in S2, each element read out is classified and combined with other elements to create a skeleton model. This skeletal model is used to analyze the motion of the human body. The details of S2 are conceptually represented by a flowchart in FIG. 6 as a skeleton model creation program. This will be described later.

  Subsequently, in S3, a musculoskeletal model is created based on the skeletal model by adding a muscle to the created skeletal model as the wire element. The details of S3 are conceptually represented by a flowchart in FIG. 7 as a musculoskeletal model creation program. This will be described later.

  Thereafter, in S4, the muscle activity of each muscle in the created musculoskeletal model is calculated. If it is a real person, at the time of a collision, muscle strength is actively generated in the muscles of each part by holding, and the muscle activity increases as the human tension increases. As the muscle activity increases, the muscle stiffness and viscosity also increase. Therefore, the higher the human tension, the higher the muscle stiffness.

  Therefore, in order to accurately analyze human behavior at the time of collision using a human body model, it is necessary to estimate muscle activity for each of a plurality of human muscles and analyze the human behavior in consideration of the results. is important. On the other hand, instead of using human body model elements (finite elements) as a unit, human motions are analyzed based on units of rigid body segments that are rigidly connected to each other. Furthermore, it is possible to estimate the muscle activity by adopting a muscle model or an optimization method for the activity of each muscle from the analyzed motion.

  Against the background of the knowledge described above, in this S4, it is not a finite element human body model having a large number of elements, but either a skeletal model or a musculoskeletal model having a smaller number of rigid body segments, or in some cases. By using both, muscle activity is calculated for each muscle. The details of S4 are conceptually represented in the flowchart of FIG. 8 as a muscle activity calculation program. This will be described later.

  Subsequently, in S5, the finite element human body model is restored from the musculoskeletal model so that the calculated muscle activity is reflected. In the restoration, the muscle model (wire element) expressed in the musculoskeletal model is added as an element that does not exist in the original finite element human body model. The details of S5 are conceptually represented in the flowchart of FIG. 9 as a finite element human body model restoration program. This will be described later.

  Thereafter, in S6, by using the restored finite element human body model, the displacement, deformation and stress acting on the human at the time of collision are analyzed by the finite element method. Thereby, human behavior at the time of collision is analyzed with high accuracy including not only passive behavior but also active behavior.

  Here, the skeleton model creation program will be described in detail with reference to FIG.

  When executing this skeletal model creation program, first, in S101, among a plurality of elements (finite elements) constituting a finite element human body model, those representing bones of the human body are classified into a plurality of parts constituting the human body, respectively. The Among a plurality of elements constituting the finite element human body model, those expressing tissues other than bone, that is, skin, flesh and the like are excluded from the classification target.

  In the finite element human body model, a part to which each element belongs is determined in advance. That is, data defining the relationship between elements and parts is incorporated in advance in data representing a finite element human body model, and each element is classified into parts using the data. For each part, a plurality of elements classified into each part are combined with each other to form one rigid segment.

  Next, in S102, a point to be a joint position is defined between two adjacent parts of the plurality of parts of the human body. The point is defined to coincide with any of a plurality of points in the finite element human body model. The point represents the joint position between two adjacent rigid body segments.

  Subsequently, in S103, the position of the center of gravity, the mass, and the moment of inertia are calculated for each part from the attribute information (shape, density, material property value, etc.) regarding the plurality of elements belonging to the part. When calculating the position of the center of gravity, mass, and moment of inertia, not only those that represent bones among a plurality of elements constituting a finite element human body model, but also tissue parts other than bone, such as skin, meat, etc. Things are also considered. As a result, each rigid segment in the skeletal model reflects the bone surface with respect to its shape, but with respect to the center of gravity, mass, and moment of inertia of each rigid segment, it reflects all the tissue parts that make up each site. .

  Thereafter, in S104, a local coordinate system is defined for each rigid body segment. Each local coordinate system coincides with the center of gravity of each rigid segment at its origin. Each local coordinate system has an x-coordinate axis, a y-coordinate axis, and a z-coordinate axis, and the directions of these coordinate axes coincide with the directions of a plurality of inertia main axes of each rigid segment. However, these coordinate axes are assigned to the principal axes of inertia in descending order of the value of the moment of inertia. The direction of the sign of each coordinate axis is adjusted so that each local coordinate system is a right-handed system.

  Subsequently, in S105, the coordinate system in which the geometric amount (position, direction, etc.) of each rigid segment is described is changed from the global coordinate system in which the finite element human body model is defined to the local coordinate system corresponding to each rigid segment. Is converted to That is, coordinate conversion is performed. The geometric amount of each rigid segment includes, for example, the shape, the position of the center of gravity, the moment of inertia information, the position of the joint, the movable direction of the joint, and the like.

  After that, in S106, a plurality of attribute information of each element that is not used when analyzing the human motion using the skeleton model (for example, material property values of each element) is prepared for the finite element analysis described later. And stored in a specific storage area of the storage 32.

  Subsequently, in S107, a plurality of elements classified into the respective parts are connected to each other to form a rigid body segment. During the movement of each rigid segment, the elements that comprise it move in unison with each other. Further, the plurality of rigid segments each configured in such a manner are connected to each other so as to be rotatable at corresponding joints.

  Thereafter, in S108, for each joint, a resistance torque (a torque generated structurally according to the joint angle) when the joint is passively moved is defined as a joint passive resistance torque. This joint passive resistance torque can be expressed as a function expression or a map table that defines a function of the joint angle.

  This completes one execution of the skeleton model creation program.

  Next, the musculoskeletal model creation program will be specifically described with reference to FIG.

  When executing the musculoskeletal model creation program, first, in S201, for each muscle to be attached to each rigid body segment of the skeletal model created earlier, a plurality of attachments in which each muscle is to be attached to the surface of each rigid body segment. The position of the point and the position of the point that is a via point between the attachment points and passes when each muscle extends along each rigid body segment are defined. The positions of these attachment points and via points are defined in the local coordinate system of each rigid segment.

  Next, in S202, the path through which each muscle in which the attachment point position and the via point position are defined extends along each rigid body segment, that is, the order in which each muscle passes through the plurality of attachment points and at least one via point is determined. Defined. Thereby, each muscle is arranged on the skeletal model.

  Subsequently, in S203, physical quantities of the arranged muscles are defined. The physical quantity includes the physiological cross-sectional area of each muscle, the penetration angle formed by muscle fibers and tendons in each muscle, and the like. The penning angle is shown in FIG. These physiological cross-sectional areas and pennation angles are used to calculate the maximum muscle strength of each muscle, as will be described later.

  The maximum muscle strength of each muscle can be calculated by the following equation using a constant k (about 5 to 10).

Maximum muscle strength = k × physiological cross-sectional area [cm ² ] × COS (penation angle [rad])

Thereafter, in S204 of FIG. 7, viscoelastic characteristics inside each muscle are defined by using, for example, the Hill model as the muscle model. According to the Hill model, the muscle strength of each muscle can be calculated by the equation (1) in FIG. This equation shows that the muscle strength F _{i of} each muscle is the muscle activity α _i , the maximum muscle strength f ^max _i , the function g of the muscle length l _i , and the function h of the muscle length change rate d (l _i ) / dt. It is estimated as the product of.

  Therefore, in order to define the viscoelastic characteristics of each muscle, specifically, a muscle length function g and a muscle length change rate h are defined.

  Thus, one execution of this musculoskeletal model creation program is completed.

FIG. 11 further shows that the rigidity K _{u, i} of each muscle is calculated by differentiating the muscle force f _{i of} each muscle with respect to the muscle length l _i . Therefore, if the relationship between the muscle strength f _{i of} each muscle and the muscle length l _i is found, the stiffness K _{u, i of} each muscle is induced.

  Next, the muscle activity calculation program will be specifically described with reference to FIG.

  When this muscle activity calculation program is executed, first, in S301, a human behavior to be analyzed is input to the created skeletal model. Specifically, the joint angle, angular velocity, and angular acceleration are input as behavior data for each joint with respect to the skeleton model. Such behavior data can be obtained by actually measuring a human behavior using an apparatus such as a motion capture for a real human being or by assuming a human behavior.

  Next, in S302, during the human behavior to be analyzed, the human is in contact with an object (for example, a steering wheel or a brake pedal) at a specific part (for example, a hand or a foot) thereof. The contact force acting between the object and the object is input to the skeleton model. The contact force is an external force against a human. The position where the contact force is input to the skeleton model is defined by the local coordinate system assigned to the corresponding rigid segment. This contact force can also be obtained by actual measurement or by assumption, similarly to the behavior data described above.

  Subsequently, in S303, an apparent joint torque acting on each joint is calculated by the inverse dynamic analysis method based on the input data. The apparent joint torque is equal to the sum of the actual joint torque (joint torque actually exhibited by humans) and the joint passive resistance torque.

  Thereafter, in S304, the joint passive resistance torque is calculated for each joint by substituting the joint angle input in S301 into the joint passive resistance torque function defined in S108 of FIG.

  Subsequently, in S305, the actual joint torque is calculated for each joint by subtracting the joint passive resistance torque calculated in S304 from the apparent joint torque calculated in S303.

  Thereafter, in S306, when the input joint angle, angular velocity, and angular acceleration are actually generated by the human based on the created musculoskeletal model, the muscle length and the muscle length change of each human muscle are generated. The speed is calculated.

  Subsequently, in S307, muscle strength and muscle activity are calculated by optimization calculation. The details of S307 are conceptually represented in the flowchart of FIG. 12 as an optimization program. This optimization program will be described later.

  S301 to S307 described above are repeated at predetermined time steps, and as a result, are repeated for all the times in the analysis target period. Thereby, during the analysis period, a substantially continuous human movement is analyzed. These S301 to S307 are used to analyze human behavior after the specific time if there is data representing the joint angle, angular velocity, angular acceleration, contact force, muscle length, and muscle length change speed at a specific time, respectively. It can be performed iteratively.

  Next, the optimization program will be specifically described with reference to FIG.

  First, the purpose of this optimization program will be outlined. Since the joint torque actively output by humans is the result of muscle strength, the corresponding muscle strength is calculated from the calculated joint torque. Is done. However, there are generally infinite combinations of muscle strengths that achieve a certain joint torque. This is because there are more muscles than joint degrees of freedom. Therefore, in this optimization program, in order to uniquely determine the muscular strength, a certain objective function is set and a solution is searched for using an optimization method.

  When this optimization program is executed, first, in S1001, the design variable is set to muscle activity. Although it is possible to set muscle strength instead of muscle activity as a design variable, in this case, formulation of constraint conditions in an analysis described later tends to be complicated.

  Next, in S1002, the equivalent stiffness at the point of interest (evaluation point) on the side of an object (for example, an operating device such as a brake operating device) with which a human contacts at a specific part (this is the “equivalent physical property” in the above (1)). ”Is measured.).

  The rigidity of a muscle in a specific part of a human depends on the degree of human tension, that is, the degree of muscle tension. The higher the degree of tension, the higher the muscle rigidity. On the other hand, it is difficult to directly measure the muscle stiffness of a human, but the muscle stiffness affects the stiffness of an object with which the human is in contact. Hereinafter, this will be specifically described with reference to FIG.

  FIG. 13 shows the interdependence and force (or torque) regarding the displacement between the elements that are directly connected to each other for the whole system of five elements: human muscle, joint, hand or foot surface, contact portion, and object. ) Are conceptually represented by two opposite arrows.

  It should be noted that, among these five elements, there are also an interdependency relationship regarding speed and an interdependence relationship regarding acceleration. However, in order to simplify the description, in the description of the present embodiment, the mutual relationship regarding displacement will be described. Only the interdependencies regarding dependencies and forces will be described representatively.

  In the example shown in FIG. 13, human muscles (for example, specific muscles physically connected to hands or feet) and joints (for example, specific joints physically connected to hands or feet) and Are examples of the “selected part” in the above item (1), and the hand or foot is an example of the “contact part” in the same term.

In FIG. 13, “J _u ” is a Jacobian matrix representing a coordinate system transformation between the muscle space and the joint space, and “J _h ” is the coordinate between the joint space and the hand or foot space. It is a Jacobian matrix representing system transformation. Further, “H” is a matrix representing a coordinate system conversion between the space of the hand or foot and the space of the contact portion, and “HS” is a coordinate system conversion between the space of the contact portion and the space of the object. It is a matrix that represents.

The Jacobian matrix J _u is a function of the muscle length and the muscle length change speed. However, since the muscle length and the muscle length change speed are a function of the joint angle, the Jacobian matrix J _u eventually becomes a function of the joint angle. Yes. Similarly, the Jacobian matrix _Jh is also a function of the joint angle.

FIG. 13 further conceptually represents the dependency between displacement and force for each of these five elements with two arrows pointing in the opposite direction. Specifically, for the element of muscle, the stiffness that is a physical quantity that relates the force to the displacement is shown as K ₁ , and conversely, the compliance that is a physical quantity that associates the displacement to the force is shown as K ₁ ⁻¹ . The remaining joints, hands or feet and the object elements are also illustrated accordingly. For the contact portion, it is assumed that the concept of rigidity does not exist.

  In addition, in each of these five elements, there is an inertia that relates a velocity and a force, and an inertia that associates an acceleration with a force. However, in order to simplify the explanation, in the explanation of the present embodiment, Only the stiffness relating displacement to force will be representatively described.

Of the five elements, between the two elements that are physically directly linked to each other, the influence of the stiffness of one element affects the other element. Specifically, the muscle adds a converted value of rigidity of the muscle alone (reflecting coordinate conversion between the local coordinate systems) to the joint. As a result, the apparent stiffness K ₂ ′ of the joint is equal to the sum of the stiffness K _{2 of the} joint alone and the increment ΔK ₂ resulting from the stiffness K _{1 of the} muscle alone.

In addition, the joint is added to the hand or foot with a converted value of the apparent stiffness of the joint (reflecting the coordinate conversion between the local coordinate systems). As a result, the hand or stiffness K ₃ apparent foot _'is a rigid K ₃ of hands or feet alone a increment [Delta] K ₃ joints apparent stiffness K _2' becomes equal to the sum of those due to.

In addition, the hand or foot adds an apparent stiffness conversion value (reflecting coordinate conversion between local coordinate systems) to the contact portion. Those result, the rigidity K ₄ over only the contact portion due to _', since the contact portions alone of the rigid K ₄ is assumed to be 0, the stiffness K ₃ apparent hand or foot a increment [Delta] K _4' Is equal to the sum of

Further, the contact portion adds an apparent stiffness conversion value (reflecting coordinate conversion between the local coordinate systems) to the object. As a result, the apparent rigidity K ₅ ′ of the object is equal to the sum of the rigidity K _{5 of the} object alone and the increment ΔK ₅ resulting from the apparent rigidity K ₄ ′ of the contact portion.

Its increment [Delta] K ₅ has a stiffness K ₁ of the muscle alone, the stiffness K ₂ of the joint alone, and a combination of the stiffness K ₃ of hands or feet alone reflects the rigidity of the human sole. On the other hand, if subtracting the stiffness K ₅ of the object alone from a rigid K _{5 'over} only object to it is possible to obtain the increment [Delta] K _5, the rigidity K ₅ of the object alone to get the actual measurement example Is possible. Therefore, it is possible to obtain the increment ΔK ₅ and thus to induce the rigidity of the person alone.

  In this way, the increase in the stiffness of the object in contact with a person from the rigidity of the object in non-contact with the person, that is, the rigidity of the object alone, is affected by the rigidity of the human muscle. Has occurred. Thus, the increment is referred to as equivalent stiffness in the sense that it is equivalent to human muscle stiffness alone.

  Referring to FIG. 14, leg muscles when a human driver steps on a brake pedal 202, which is an example of an operating member of a brake operating device 200, which is an example of an operating device mounted on a vehicle. The method for measuring the equivalent stiffness of the object will be specifically described by taking the method for measuring the stiffness as an example.

  As shown in FIG. 14, one end portion of the brake pedal 202 is a rotating portion 206 that is rotatably connected to the fixing member 204. On the other hand, the other end portion of the brake pedal 202 is a pad portion 208 on which a foot should be put and stepped on. The brake pedal 202 is urged by the linear spring 210 around the rotating portion 206 in a direction from the depressed position to the initial position.

FIG. 14A shows the brake operating device 200 in an initial state. In this state, when a force F is applied to the brake pedal 202 by an actuator (not shown) and the rotation angle θ of the brake pedal 202 is measured by a sensor (not shown) each time the force F is increased, a solid line graph rising to the right is obtained. To be acquired. The gradient of this graph represents the rigidity K ₀ of the brake operating device 200 alone as the object.

FIG. 14B shows the brake operation device 200 in the depressed state. In this state, when a force F is applied to the brake pedal 202 by the actuator and the rotation angle θ of the brake pedal 202 is measured by the sensor each time the force F is increased, a solid line graph rising to the right is acquired. The The slope of this graph represents the apparent rigidity K ₁ of the brake operating device 200. Those obtained by subtracting the rigidity K ₀ from the apparent stiffness K ₁ are equivalent stiffness of the foregoing.

  FIG. 15 shows a combination of the rigidity of the person and the rigidity of the brake operating device 200 as an object, that is, the apparent rigidity measured from the object, the rigidity of the object alone, and the rigidity of the person alone, that is, the above-described equivalent The relationship with the measurement value of rigidity is expressed by Expression (3) expressed by a vector including a plurality of independent stiffness directions and the same number of elements at the point of interest r (for example, the contact point between a person and an object). ing.

  If the equivalent rigidity is measured in this way, it is possible to measure the rigidity of a person alone by performing appropriate coordinate transformation.

  Subsequently, in S1003 of FIG. 12, the objective function represented by Expression (4) in FIG. 16 is set. The variable vector F (α) in this objective function is expressed by the equation (5) in the figure.

The variable vector is the element represented by the product of the weighting coefficient w _i and the muscle activity α _i for each muscle, and the measured value of equivalent stiffness ^r K _{h, measured, j} and the human alone for each stiffness direction. It includes an element represented by the product of the difference between the estimated value ^r K (α) _{h, estimated, j of} rigidity (hereinafter referred to as “estimation error”) and the weight coefficient w. The objective function is such that the sum of the products of the weighting factors w _i and the muscle activity α _i is minimized for all the muscles to be noticed, and the weighting factor w and the estimation error for all the stiffness directions to be noticed. Is set so that the muscle activity α _i that minimizes the sum of the products is calculated by the optimization calculation.

However, if the total value is calculated as a simple sum of a plurality of components, the calculated value depends on the sign of the plurality of components. Therefore, in this embodiment, in order to avoid such a situation, the sum of squares of the products of the weighting factor w _i and the muscle activity α _i is minimized for all the muscles to be noticed, and at the same time, it should be noted. The objective function is set so that the muscle activity α _i that minimizes the sum of squares of the products of the weight coefficient w and the estimation error is calculated by optimization calculation for all the stiffness directions.

The relationship between the muscular strength f _i and the muscular activity α _i for each rigidity direction is expressed by the equation (1) in FIG. In this equation (1), there are components other than the muscle strength f _i and the muscle activity α _i , all of which are determined prior to the subsequent optimization calculation. In the same figure, the relationship between the rigidity of a person alone, that is, the muscle rigidity K _{u, i,} and the muscle activity α _i is expressed by Expression (2). In the present embodiment, by using these equations (1) and (2), an estimated value ^r K (α) _{h, estimated, j} of the rigidity of a person alone is calculated.

However, in order to calculate the estimated value ^r K (α) _{h, estimated, j} of the rigidity of a person alone, a Jacobian matrix ^c defining a coordinate transformation between the contact point between the person and the object and the generalized coordinates J _h is interposed. This is represented by Expression (7) in FIG. Further, since the Jacobian matrix ^c J _h depends on the posture of the musculoskeletal model, that is, the joint angle, in order to calculate the estimated value ^r K (α) _{h, estimated, j} , the muscle length and the muscle length change rate are calculated. It is necessary to consider not only the motion state quantity but also the joint motion state quantity called the joint angle.

Thereafter, in S1004 of FIG. 12, a constraint condition to be considered in the optimization calculation is set. As the constraint conditions, there are a relational expression between the joint torque τ and the muscle force f and a range of the muscle activity α _i (0 or more and 1 or less). If the optimization calculation described later is performed so that this constraint condition is satisfied, the muscle activity α _i may be estimated so that force balance is substantially satisfied between the joint torque τ and the muscle force f. Guaranteed.

In FIG. 17, a relational expression between the joint torque τ and the muscle force f is shown as an expression (6). This equation (6) represents that the joint torque τ is induced as the product of the Jacobian matrix _Ju and the muscle strength f.

The Jacobian matrix J _u is a matrix for coordinate conversion from muscle space to joint space, and is shown in FIG. 13 together with its transposed matrix J _u ^T. Jacobian matrix J _u depends on the attitude That joint angle musculoskeletal model. The muscle strength f is a function of the muscle activity α.

Subsequently, in S1005 of FIG. 12, optimization calculation is performed based on the set objective function and constraint conditions, whereby the muscle activity α _i is calculated for each muscle. Parameters referred to for the calculation are weighting factor w, measured value of equivalent stiffness ^r K _{h, measured, j} , maximum muscle strength f ^max , muscle length and muscle length change rate, and muscle length function g. And the muscle length changing speed function h, the joint angle on which the Jacobian matrix ^c J _h depends, and the joint torque τ.

  The optimization calculation is generally called “constrained optimization”, and “sequential quadratic programming” and “correction executable direction method” are already known as general techniques. The specific method of the optimization calculation can be appropriately changed according to various uses.

Thereafter, in S1006, the muscle strength f _i is calculated for each muscle by substituting the calculated muscle activity α _i into the equation (1) shown in FIG.

  Thus, one execution of the optimization calculation program is completed.

  FIG. 19 is a block diagram showing a calculation flow performed in order to estimate muscle activity in this embodiment. Hereinafter, this calculation flow will be described with reference to FIG.

  If the musculoskeletal model is used in the meaning including the skeletal model, this calculation is divided into an upstream process performed using the musculoskeletal model and a downstream process called optimization calculation.

  In the upstream process, by using the musculoskeletal model, human motion is analyzed in order to acquire data necessary for optimization calculation. In the downstream process, optimization is performed based on the objective function and constraints by using the data acquired in the upstream process and the input value of the equivalent physical property (an example of this is the equivalent stiffness). Calculations are performed, thereby estimating muscle activity.

  FIG. 20 shows a distribution of calculated values of muscle activity α of each muscle, taking the lower limb shown in FIG. 4 as a musculoskeletal model as an example. As shown in FIG. 20, in this example, a human depresses the brake pedal with the right foot. In the present embodiment, it is possible to calculate the muscle activity α as a continuous value within a region of 0 or more and 1 or less. However, in FIG. 20, for convenience of illustration, a muscle having a calculated value of 0, a muscle having a value of 1, and a muscle having an intermediate value between 0 and 1 are “α = 0” and “α = 1”, respectively. And “0 <α <1”.

  Next, the finite element human body model restoration program will be described in detail with reference to FIG.

  When executing this finite element human body model program, first, in S401, geometric quantities (node positions, etc.) of a plurality of elements (finite elements) belonging to each rigid segment constituting the musculoskeletal model, and geometric quantities of each muscle ( (Attached point position, via point position, etc.) are expressed in the global coordinate system. The coordinate system for describing these geometric quantities is converted from the local coordinate system assigned to each rigid segment to the musculoskeletal model to the global coordinate system assigned to the original finite element human body model. is there.

  Next, in S402, the material characteristic value, etc., among the attribute information of each element, which was not used in the preceding operation analysis (S303 and S306 in FIG. 8), is stored in advance in the specific storage area of the storage 32. Related information is read from the storage area.

  Subsequently, in S403, the musculoskeletal model is restored to the finite element human body model by using the geometric quantity subjected to the coordinate transformation and the read related information.

  Thereafter, in S404, by using the wire element as the muscle model added to the musculoskeletal model, the muscle model is added as a wire element to the restored finite element human body model. In the restored finite element human body model, each muscle model is defined with respect to the attachment point position and the via point position of the muscle and the order in which the positions are arranged. With this definition, the geometric feature amount of each added muscle model is specified in the restored finite element human body model.

  Subsequently, in S405, the calculated muscle activity or strength is given to each restored muscle for the restored finite element human body model. It is possible to analyze not only passive behavior but also active behavior.

  This completes one execution of the finite element human body model restoration program.

  Thereafter, in S6 of FIG. 5, by using the restored finite element human body model, the passive behavior and the active behavior exhibited by the human in the designated environment are analyzed by the finite element method. This finite element analysis can be carried out, for example, by executing a commercially available solver (analysis computer program) or an equivalent program by the computer 20. Commercially available solvers include, for example, “PAM-CRASH (a program manufactured by ESI of France and sold by ESI of Japan)” and “LS-DYNA (LSTC (Livermore)) which are explicit dynamic finite element analysis programs. Software Technology Corporation) is a program manufactured and sold by Japan Research Institute).

  When analyzing a human's active behavior by the finite element method, when using a solver originally having a muscle element such as the above-described LS-DYNA, the muscle element can be used directly.

  On the other hand, even when using a solver that does not have muscle elements, the muscle strength direction is determined by the muscle placement, and the muscle strength magnitude is determined by the muscle activity. By setting as, it is possible to express muscle strength on the human body model. Therefore, it is possible to use a wire element or the like, or it is possible to set only the force as a direct boundary condition.

  As is clear from the above description, in this embodiment, the human muscle activity is accurately estimated, and then the human behavior is analyzed in consideration of the estimated value. Therefore, according to this embodiment, it is possible to analyze human behavior in consideration of human active behavior. Thus, for example, when it is necessary to analyze the behavior of an occupant in a vehicle that collides with an obstacle, the occupant must exhibit active behavior such as, for example, posture, escape, reflective motion, and brake operation. It is possible to analyze human behavior with high accuracy.

  By the way, in the conventional finite element human body model, there is a lack of a muscle model which is an actuator for actively operating a human, and the human's active motion is analyzed using this conventional finite element human body model. Was difficult. To analyze finite element analysis in the first place, it is necessary to analyze a problem that the calculation load is large, or a problem that the displacement of the analysis target is large and each element representing the analysis target is greatly deformed, such as active operation. There is a problem that the analysis accuracy is remarkably lowered. Therefore, until now, it has been considered that the finite element human body model is not suitable for active behavior analysis.

  Theoretically, if a muscle model is additionally defined in a finite element human body model and the muscle activity assigned to the muscle model is specified, behavior analysis including active motion is used for the finite element human body model. Can be done. However, in order to specify the muscle activity level, it is necessary to analyze the active movement to be analyzed and obtain the activity state of each muscle. As mentioned above, the finite element human body model is suitable for motion analysis. Not.

  Therefore, in this embodiment, a skeletal model or a musculoskeletal model, that is, an articulated rigid body model (multibody model) is used instead of the finite element human body model for motion analysis. The multi-body model is characterized by a light computational load and easy handling of large exercises. In the present embodiment, specifically, joint torque as a joint motion state quantity is calculated by using a skeletal model, and muscle length and muscle as muscle motion state quantities are calculated by using a musculoskeletal model. The long rate of change is calculated.

  In the present embodiment, the muscle activity is estimated by the optimization calculation by using the calculated values of the joint motion state amount and the muscle motion state amount. The estimated muscle activity is input to the finite element human body model, so that the original finite element human body model is converted into an activated finite element human body model with each element given muscle activity. Converted.

  In this embodiment, the conversion to the activated finite element human body model is realized by a method of restoring from the musculoskeletal model to the finite element human body model. However, the finite element human body model is a muscle for all muscles in the first place. When the activity level is set as a variable parameter, activation of the finite element human body model is realized by simply specifying the muscle activity level for each muscle of the original finite element human body model.

  As shown in FIG. 19, in the present embodiment, the contact force (or restraining force) acting between a person and an object (for example, an operation member such as a brake pedal) and the actual or assumed of the person. The amount of motion state describing the motion (the amount of motion state given to the skeleton model in order to calculate a predetermined physical quantity using the skeleton model, hereinafter simply referred to as “human motion state amount”), and muscle activity If the initial value for the degree estimation and the equivalent physical characteristic are specified, the muscle activity is estimated by the calculation using the musculoskeletal model and the optimization calculation.

  Therefore, if the initial value for muscle activity estimation can be arbitrarily set, muscle activity is uniquely determined according to the combination of contact force, human motion state quantity, and equivalent physical characteristics. become.

  Therefore, data that systematically expresses the relationship between the contact force, the human motion state quantity and the equivalent physical property, and the muscle activity estimated corresponding to the combination thereof is recorded on, for example, a solid recording medium. When the muscle activity database is constructed, when it becomes necessary to refer to the muscle activity after that, the computer 20 only executes a normal search program for the muscle activity database. Is acquired.

  That is, if this muscle activity database is used in combination with a simple search program, the muscle activity can be obtained with high accuracy without executing the aforementioned muscle activity calculation program by the computer 20 one by one. . This muscle activity database is an example of the “muscle activity database” according to the item (22).

  By the way, as shown in FIG. 19, the human motion state quantity is used for the calculation of the joint torque, the calculation of the muscle motion state quantity, and the calculation of the joint motion state quantity. Therefore, when the joint torque, the muscle motion state quantity, and the joint motion state quantity are specified directly or by another method, it is not necessary to input the human motion state quantity to estimate the muscle activity. .

  Of these joint torques, muscle movement state quantities, and joint movement state quantities, joint torque can be estimated from muscle movement state quantities, and muscle movement state quantities can be estimated from joint movement state quantities. is there. Therefore, if the joint motion state amount is specified, it is possible to omit specifying the joint torque and the muscle motion state amount directly.

  Therefore, in order to further improve the usability of the above-described muscle activity database, the relationship recorded in the muscle activity database includes contact force, joint motion state quantity, equivalent physical property, and muscle activity estimation. It is desirable to change the relationship with the value.

  By the way, in the above-described musculoskeletal model, a plurality of real muscles in human beings are expressed using a plurality of muscle models (muscle element models), respectively, and a plurality of muscle activities to be assigned to the muscle models respectively. May depend on each other.

  For example, in the present embodiment, the muscle element model is physically associated with the bone element model at the attachment point and via point in order to express the balance between the joint torque around the joint and the moment due to the muscle force in the musculoskeletal model. . These attachment points and via points serve as the basis for calculating the muscular moment arm.

  However, it is not appropriate to think about the physical points of attachment points and via points for all real streaks. For example, for the greater gluteus muscle in the human waist, it is not appropriate to calculate the moment arm of the muscular strength by thinking about the attachment points and via points because the muscles are attached to the hip bone on the surface.

  Therefore, in the present embodiment, one continuous muscle is actually represented by a plurality of muscle element models connected to each other in the musculoskeletal model. However, since these muscle element models originally represent one real muscle, they should be handled so as to be mutually dependent on the muscle activity.

  In reality, a plurality of muscles that are independently separated from each other may depend on each other in terms of muscle activity due to the sharing of motor nerves. For example, when a driver quickly depresses a brake pedal with a foot to perform a sudden braking operation in a vehicle, the muscle activity of the leg muscles increases, but the muscle activity of the hand muscles also increases accordingly. There is a case. A plurality of muscles that depend on each other with respect to muscle activity, that is, a plurality of muscles that are activated in conjunction with each other, such as a combination of the muscles of the legs and hands, are referred to as muscle groups. A plurality of muscles activated in conjunction with each other are a plurality of muscles having co-contraction properties.

  In any case, the plurality of muscle element models in the musculoskeletal model includes a plurality of muscle element models that are mutually dependent on the muscle activity as a muscle element model group. Specifically, in the example of the greater ankle muscle, a plurality of muscle element models that represent the greater ankle muscle jointly constitute one muscle element model group, and the sudden braking operation described above is performed. In the example, the muscle element model that expresses the muscle of the foot and the muscle element model that expresses the muscle of the hand constitute one muscle element model group.

  For each muscle element model group, in order to calculate the muscle activity in the above-described optimization calculation, it is desirable that the constraint condition in S1004 in FIG.

In FIG. 21, the constraint condition is expressed by Expression (10). The matrix J _uc that describes the constraint conditions specifies a plurality of muscle element models that are mutually dependent on the muscle activity, and each of the muscle activity models indicated by the plurality of muscle element models that are mutually dependent on the muscle activity. Specify the ratio of degrees. For example, if the first to third muscle element models among N muscle element models depend on each other in terms of muscle activity and the dependency ratio is 1: 1: 1, the matrix J _uc is It is configured as shown in FIG.

  As is clear from the above description, in the present embodiment, S4 (excluding S1006) in FIG. 5 constitutes an example of the “muscle activity estimation method” according to the item (1), and S4 in FIG. (Including S1006) constitutes an example of the “muscle strength estimation method” according to the above-mentioned item (14), and the portion of the computer 20 that executes S4 (excluding S1006) in FIG. This is an example of the “muscle activity estimation system”.

  Further, in the present embodiment, S301 and S306 in FIG. 8 and S1002 in FIG. 12 together form an example of the “input step” in the above item (1), and S1001 and S1003 to S1005 in FIG. Together, they constitute an example of the “muscle activity calculation step” in the same section.

  In addition, in this embodiment, in S306 in FIG. 8, the muscle length and the muscle length change rate as the muscle motion state amount are calculated using the joint motion state amount based on the musculoskeletal model. However, this calculation can be easily replaced by an input of a muscle movement state quantity that is actually measured or acquired by another method. Focusing on this ease of replacement, in the present embodiment, it is reasonable to consider that the calculation of the muscle movement state quantity corresponds to an example of the input (broad sense) of the muscle movement state quantity.

  Further, in the present embodiment, the part of the computer 20 that executes S301 and S306 in FIG. 8 and S1002 in FIG. 12 constitutes an example of the “input means” in the above item (20), and S1001 and S100 in FIG. The part that executes S1003 to S1005 constitutes an example of “muscle activity calculation means” in the same section.

  Further, in the present embodiment, S301, S302, and S306 in FIG. 8 and S1002 in FIG. 12 jointly constitute an example of the “input step” in the item (2), and S303 in FIG. In FIG. 12, and S1001 and S1003 to S1005 in FIG. 12 together constitute an example of the “muscle activity calculation step” in the same term.

  Furthermore, in the present embodiment, the portion of the computer 20 that executes S301, S302, and S306 in FIG. 8 and S1002 in FIG. 12 constitutes an example of the “input means” in the above item (21), and in FIG. The part that executes S303 constitutes an example of “joint torque calculating means” in the same term, and the part that executes S1001 and S1003 to S1005 in FIG. 12 constitutes an example of “muscle activity degree calculating means” in the same term. It is.

  Further, in the present embodiment, the function expressed by the equation (7) in FIG. 18 constitutes an example of “function” in the item (1) and an example of “function” in the item (20). -ing

  Although one embodiment of the present invention has been described in detail with reference to the drawings, this is an exemplification and is based on the knowledge of those skilled in the art including the aspects described in the section of [Disclosure of the Invention]. The present invention can be implemented in other forms with various modifications and improvements.

1 is a block diagram conceptually showing a hardware configuration of a human behavior analysis system suitable for implementing a human behavior analysis method including a muscle activity estimation method according to an embodiment of the present invention. FIG. It is a perspective view which shows the leg part among the finite element human body models in FIG. It is a perspective view which shows the leg part among the skeleton models in FIG. It is a perspective view which shows the leg part among the musculoskeletal models in FIG. 2 is a flowchart conceptually showing a human behavior analysis program in FIG. 1. 6 is a flowchart conceptually showing details of S2 in FIG. 5 as a skeleton model creation program. 6 is a flowchart conceptually showing details of S3 in FIG. 5 as a musculoskeletal model creation program. 6 is a flowchart conceptually showing details of S4 in FIG. 5 as a muscle activity calculation program. 6 is a flowchart conceptually showing details of S5 in FIG. 5 as a finite element human body model restoration program. FIG. 8 is a plan view showing human muscles and tendons, focusing on the direction of force application, in order to explain the execution contents of S203 in FIG. 7. FIG. 8 is a diagram illustrating a muscle model as a numerical analysis model expressed by equations (1) and (2) in order to explain the execution contents of S204 in FIG. 10 is a flowchart conceptually showing details of S307 in FIG. 8 as an optimization calculation program. In order to explain the significance of the equivalent stiffness in S1002 in FIG. 12, a coupled physical system in which a human hand or foot and an object are in physical contact with each other is briefly described with a focus on displacement and force transmission direction. FIG. FIG. 13 is a side view showing a brake operating device and a graph showing a relationship between a rotation angle θ and a force F in order to explain the equivalent stiffness measurement method in S1002 in FIG. It is a figure explaining the analysis principle which S1002 in FIG. 12 employ | adopts by Formula (3). It is a figure which shows the objective function used in the optimization calculation in S1005 in FIG. 12 by Formula (4) and (5). It is a figure which shows the constraint condition used in the optimization calculation in S1005 in FIG. 12 by Formula (6). It is a figure which shows the estimated value ^rK ((alpha)) _{h, estimated, j} of the rigidity of only a person in Formula (5) of FIG. 16 by Formula (7) _{thru | or} (9). It is a block diagram showing the flow of the muscle activity calculation performed by execution of the muscle activity calculation program shown in FIG. It is a perspective view which shows the human lower limb model which is an example of the execution object of the muscle activity calculation program shown in FIG. 8 with the analysis result of the muscle activity about the lower limb model. It is a figure which shows a more desirable example of the optimization calculation in S1005 in FIG. 12 by Formula (10).

Explanation of symbols

10 system 20 computer 100 lumbar region 102 thigh 104 shin 106 foot 110 lumbar 112 femur 114 tibia 116 rib 118 shin

Claims (22)

A person composed mainly of a plurality of muscles and a plurality of bones and having physical characteristics that change at least one of rigidity and viscosity of each muscle in accordance with the activity of the muscles constitutes the person. The degree of activity exhibited by each of a plurality of muscles in an analysis site selected as an analysis target among the plurality of sites in a state where at least one of the plurality of sites is a contact site and is in contact with an object at the contact site Is a muscle activity estimation method for estimating a muscle activity by a computer,
The joint state and muscle motion state quantities at the analysis site are input as joint motion state quantities and muscle motion state quantities, and the rigidity of the object in a state where the contact site and the object are in physical contact with each other. And an input step of inputting, as an equivalent physical property equivalent to the physical property of the human alone, an increment of the physical property that is at least one of viscosity and viscosity from the physical property of the object alone;
Based on the input joint motion state quantity, muscle motion state quantity, and equivalent physical characteristics, the muscle activity is calculated based on the difference between the input value and the estimated value of the equivalent physical characteristics in all the motion directions of interest in the object. A muscle activity calculation step for calculating the total value so as to be substantially minimized, wherein the estimated value includes the muscle activity, the motion state quantity of the muscle, and the motion state quantity of the joint as variables; A method for estimating muscle activity, which includes: The input step further inputs a contact force acting at the contact point between the contact site and the object,
The muscle activity estimation method further includes:
A joint torque calculating step of calculating a joint torque acting around the joint at the analysis site by an inverse dynamic analysis method based on the input joint motion state quantity and the contact force;
The muscle activity calculation step calculates the muscle activity so that the total value is substantially minimized and a force balance is substantially established between the joint torque and the muscle strength of the muscle. The method for estimating muscle activity according to claim 1. The joint torque calculation step includes:
(A) Based on the input joint motion state quantity and the contact force, an apparent joint torque that acts around the joint at the analysis site by the inverse dynamics analysis method, that is, an actual joint torque and a joint passive resistance torque, An apparent joint torque calculating step for calculating a composite of
(B) an actual joint torque calculation step of calculating the actual joint torque from the calculated apparent joint torque and the joint passive resistance torque,
The muscle activity estimation method according to claim 2, wherein the muscle activity calculation step calculates the muscle activity using the calculated actual joint torque.   The joint torque calculating step uses a skeleton model that physically represents the human on at least the plurality of bones on the computer, and gives a predetermined motion to the skeleton model, whereby the joint torque is calculated. The method for estimating muscle activity according to claim 2 or 3, further comprising the step of:   The skeletal model is a multi-joint rigid body model that expresses a plurality of parts constituting the human body, which is the human body, as a plurality of rigid body segments, so that the plurality of rigid body segments can rotate around a joint. The method for estimating muscle activity according to claim 4.   Further, by using a musculoskeletal model that physically represents the human on at least the plurality of muscles and the plurality of bones on the computer, and by giving a predetermined operation to the musculoskeletal model, The muscle activity state estimation method according to any one of claims 1 to 5, further comprising a muscle motion state amount calculation step of calculating the muscle motion state amount.   The musculoskeletal model expresses a plurality of parts constituting the human body, which is the human body, as a plurality of rigid body segments so that the plurality of rigid body segments can rotate around a joint, The muscle activity estimation method according to claim 6, wherein each muscle is expressed as a plurality of wire elements. The joint motion state quantity includes at least one of an angle, an angular velocity, and an angular acceleration of the joint,
The muscle activity state estimation method according to claim 1, wherein the muscle movement state quantity includes at least one of a length of the muscle and a change speed thereof.   The human is mainly composed of a plurality of muscle element models and a plurality of bone element models, and a physical characteristic that is at least one of stiffness and viscosity assigned to each muscle element model is assigned to the muscle element model. The muscle activity estimation method according to any one of claims 1 to 8, wherein the muscle activity is represented on the computer by a human body model having a characteristic that changes in accordance with the activity. The human being is an occupant riding a vehicle,
The object is a structure of the vehicle where the occupant is expected to collide when the vehicle collides with an obstacle,
The muscle activity estimation method according to any one of claims 1 to 9, wherein the muscle activity estimated by the muscle activity estimation method is used to analyze a response behavior of the occupant with respect to a collision with the structure. . The human being is an occupant riding a vehicle,
The object is an operating device that is assumed to be operated by the occupant in the vehicle,
The muscle activity estimation according to any one of claims 1 to 9, wherein the muscle activity estimated by the muscle activity estimation method is used to analyze a response behavior of the occupant responding to an operation of the operating device. Method.   The muscle activity estimation method according to claim 11, wherein the operation device includes at least one of a brake operation device, an accelerator operation device, a steering operation device, and a shift operation device. The occupant is a driver of the vehicle,
In the vehicle, at least one of its running state and behavior is automatically controlled regardless of the driver's intention,
13. The muscle activity estimation method is used to analyze the response behavior in consideration of an automatic control state of at least one of a running state and a behavior of the vehicle. Muscle activity estimation method. A person composed mainly of a plurality of muscles and a plurality of bones and having physical characteristics that change at least one of rigidity and viscosity of each muscle in accordance with the activity of the muscles constitutes the person. The computer displays the muscle strength of each of the plurality of muscles in the selected portion selected as the analysis target among the plurality of portions in a state where at least one of the plurality of portions to be contacted is in contact with the object at the contact portion. The muscle strength estimation method estimated by
A muscle activity estimation method according to any one of claims 1 to 13,
A muscle strength estimation method including: a muscle strength calculation step of calculating muscle strength according to a predetermined relationship between muscle activity and strength based on the muscle activity estimated by the muscle activity estimation method.   The muscle strength estimation method according to claim 14, wherein the predetermined relationship includes a predetermined relationship among muscle activity, maximum muscle strength, muscle length, and muscle length change rate for each muscle. A person composed mainly of a plurality of muscles and a plurality of bones and having physical characteristics that change at least one of rigidity and viscosity of each muscle in accordance with the activity of the muscles constitutes the person. Selection selected as an analysis target among the plurality of parts in an analysis target period that is at least one of immediately before and immediately after colliding with an object at the contact part. A human behavior estimation method for estimating a behavior shown in a part by a computer,
A muscle activity estimation method according to any one of claims 1 to 13,
An input step of inputting an external force and a displacement acting on the person in the analysis target period;
Based on the input external force and displacement and the muscle activity estimated by the muscle activity estimation method, the person is physically represented on the computer with respect to at least the plurality of muscles and the plurality of bones. A human behavior estimation method including: a behavior calculation step of calculating a behavior exhibited by the human in the period to be analyzed by using a human physical model. By using a finite element human body model that expresses a human body, which is a human body, by a plurality of finite elements, in which belonging relationships between a plurality of parts constituting the human and the plurality of finite elements are set in advance , A human behavior analysis method for analyzing human behavior,
By classifying and combining the plurality of finite elements constituting the finite element human body model according to the belonging relationship, the plurality of portions are respectively defined as a plurality of rigid body segments, and the plurality of rigid body segments rotate around the joint. A skeletal model creation process for creating a skeletal model to be expressed as possible;
In accordance with what expresses the plurality of muscles among the plurality of finite elements in the finite element model, on the created skeletal model, wire elements that express the plurality of muscles as wires are respectively represented by the rigid bodies of the muscles. A musculoskeletal model creation process for creating a musculoskeletal model by defining it with the position of the attachment point to the segment and the waypoint in the middle;
An information acquisition step for acquiring information necessary for analyzing the behavior exhibited by the human in a specified environment using the finite element human body model by using the created skeletal model and musculoskeletal model;
A finite element human body model restoring step of restoring the finite element human body model from the created musculoskeletal model so as to represent the acquired information;
A human behavior analysis method including: a behavior calculation step of calculating a behavior exhibited by the human in the designated environment by using the restored finite element human body model.   A program executed by a computer to implement the method according to any of claims 1 to 17.   The recording medium which recorded the program of Claim 18 so that computer reading was possible. A person composed mainly of a plurality of muscles and a plurality of bones and having physical characteristics that change at least one of rigidity and viscosity of each muscle in accordance with the activity of the muscles constitutes the person. The degree of activity exhibited by each of a plurality of muscles in an analysis site selected as an analysis target among the plurality of sites in a state where at least one of the plurality of sites is a contact site and is in contact with an object at the contact site Is a muscle activity estimation system that estimates a muscle activity as a muscle activity,
The joint state and muscle motion state quantities at the analysis site are input as joint motion state quantities and muscle motion state quantities, and the rigidity of the object in a state where the contact site and the object are in physical contact with each other. And an input means for inputting an increment of the physical property that is at least one of viscosity and viscosity from the physical property of the object alone as an equivalent physical property that is equivalent to the physical property of the human alone,
Based on the input joint motion state amount, muscle motion state amount, and equivalent physical property, the muscle activity is calculated based on the difference between the input value of the equivalent physical property and the estimated value in all the motion directions of interest in the object. The muscle activity calculating means for calculating the total value so as to be substantially minimized, wherein the estimated value includes the muscle activity, the muscle motion state quantity, and the joint motion state quantity as variables, respectively. A muscle activity estimation system including: The input means further inputs a contact force that acts at the contact point between the contact site and the object,
The muscle activity estimation system further includes:
A joint torque calculating means for calculating a joint torque acting around the joint at the analysis site by an inverse dynamics analysis method based on the input joint motion state quantity and the contact force;
The muscle activity calculation means calculates the muscle activity so that the total value is substantially minimized and a force balance is substantially established between the joint torque and the muscle strength of the muscle. The muscle activity estimation system according to claim 20. A person composed mainly of a plurality of muscles and a plurality of bones and having physical characteristics that change at least one of rigidity and viscosity of each muscle in accordance with the activity of the muscles constitutes the person. The degree of activity exhibited by each of a plurality of muscles in an analysis site selected as an analysis target among the plurality of sites in a state where at least one of the plurality of sites is a contact site and is in contact with an object at the contact site Is a muscle activity database that is read as a muscle activity by executing a search program by a computer,
(A) a joint motion state quantity which is a joint motion state quantity at the analysis site;
(B) Equivalent physical characteristics that are equivalent to the physical characteristics of the person alone, and are at least one of rigidity and viscosity of the object in a state where the contact portion and the object are in physical contact with each other. A physical property equal to the increment from the physical property of the object alone;
(C) a contact force acting at the contact point between the contact site and the object;
(D) A plurality of physical quantities including the muscle activity are recorded in advance in association with each other as variables, and a combination of the joint motion state quantity, the equivalent physical characteristic, and the contact force by executing the search program If a muscle activity is identified, a muscle activity database associated with the combination is identified and output by executing the search program.