JP7398090B2 - Information processing device, calculation method and program - Google Patents

Description

The present invention relates to an information processing device, a calculation method, and a program for dynamically analyzing body motion.

In order to dynamically analyze and evaluate physical movements such as walking, it is usually necessary to measure physical movements and forces acting on the body. However, in vivo mechanical quantities such as joint moments are difficult to directly measure. For this reason, inverse dynamic analysis is performed using a musculoskeletal model that numerically expresses the mechanical characteristics of the body to estimate the mechanical quantities in the living body.

In order to perform inverse dynamics analysis using a musculoskeletal model, in addition to the three-dimensional position of each body part obtained from a motion capture system, external force information obtained from a force sensor such as a floor reaction force meter is required. becomes. However, it is not easy to install a floor reaction force meter for measuring the external force generated on the foot because it requires embedding it in the floor or preparing a stand that matches the height of the device. Furthermore, the installation of a floor reaction force meter may make the subject conscious of the measuring device, which may induce unnatural walking or running.

From the above-mentioned background, various developments have been made to eliminate the need for floor reaction force meters during walking. Techniques that enable analysis that does not require a floor reaction force meter include techniques that estimate ground reaction force based on minimizing the state of muscle activity throughout the body (Non-Patent Document 1, Non-Patent Document 2) and artificial neural networks. For example, there is a technique (Non-Patent Document 3) for estimating the ground reaction force using the following method.

Non-Patent Document 1 discloses a configuration in which 12 contact points each having 5 artificial muscles are provided on the foot of a musculoskeletal model constructed based on body surface marker positions obtained by an optical motion capture system. Disclose. Then, by minimizing the muscular activity states of each part of the body, which would normally be found through static optimization, through optimization calculations, the unknown ground reaction force can be found in an exploratory manner. Here, each artificial muscle is provided to express frictional force, and is defined to solve an indeterminate problem that occurs when supporting both legs.

In addition, Non-Patent Document 2 further increases the number of contact points on the foot, and changes the strength information of the artificial muscle, which had previously been determined by trial and error using function approximation, depending on the speed and position information of each point. Discloses an improved technique to do so. In Non-Patent Document 2, the validity of the estimated value was verified by comparing the estimated value and the actual measured value in the walking movements of 10 male subjects, and it was shown that the estimation was performed with relatively high accuracy. ing. However, since muscle characteristics vary widely among individuals, it may be difficult to estimate with high accuracy the physical movements of female subjects, the elderly, and the disabled.

Non-Patent Document 3 discloses a technique aimed at developing a novel technique that can estimate the floor reaction force during walking on a flat surface with high accuracy without using a floor reaction force meter. In Non-Patent Document 3, an artificial neural network model is applied to both leg phases of walking motion in order to solve an indeterminate problem. Note that the method of Non-Patent Document 3 requires learning data.

However, the above-mentioned conventional techniques were not sufficient because they required time and effort to prepare for applying the techniques, and there was a concern that the estimation accuracy may vary greatly depending on the characteristics of the subjects.

R. Fluit, et al., “Prediction of ground reaction forces and moments during various activities of daily living”, Journal of Biomechanics, Vol. 47, No. 10, pp. 2321-2329(2014). Y. Jung, et al. , “Dynamically adjustable foot ground contact model to estimate ground reaction force during walking and running”, Gait & Posture, Vol. 45, pp. 62-68(2016). E. S. Oh, et al., “Prediction of ground reaction forces during gait based on kinematics and a neural network model”, Journal of Biomechanics, 46(14), 2372-2380 (2013).

The present invention has been made in view of the insufficiencies in the prior art described above, and the present invention provides information processing that is capable of robustly estimating external forces acting on the body during physical exercise based on body movement data. The purpose is to provide devices, calculation methods, and programs.

In order to solve the above problems, the present invention provides an information processing device having the following characteristics for estimating external forces generated during physical exercise. This information processing device includes an acquisition means for acquiring body movement data, a setting means for setting a provisional value of external force, and an inverse dynamics analysis using a body model based on the body movement data and provisional value of external force. an execution means, an evaluation means for evaluating the load generated on the body according to the mechanical quantity generated at each joint of the body model based on the execution results of the inverse dynamic analysis, and an evaluation means for evaluating the provisional value of the external force based on the evaluation result. and updating means for updating.

The present invention also provides a calculation method having the following characteristics for estimating an external force generated during body exercise. This calculation method includes the steps of a computer acquiring body motion data and setting a provisional value of external force. This calculation method further includes the step of the computer performing an inverse dynamics analysis using the body model based on the body motion data and provisional values of external forces, and the step of performing an inverse dynamics analysis using the body model based on the execution results of the inverse dynamics analysis. The method includes a step of evaluating the load generated on the body according to the mechanical amount generated in each joint, and a step of updating a provisional value of the external force based on the evaluation result.

The present invention further provides a program for realizing an information processing device having the above characteristics for estimating the external force generated during the above-described physical exercise.

With the above configuration, it is possible to robustly estimate the external force acting on the body during physical exercise based on the physical exercise data.

FIG. 1 is a functional block diagram showing the overall configuration of an analysis system including a walking motion analysis device according to an embodiment of the present invention. FIG. 2 is a more detailed functional block diagram of the optimization calculation unit in the walking motion analysis device according to the embodiment of the present invention. FIG. 3 is a flowchart showing a walking motion analysis process executed by the walking motion analysis device according to the embodiment of the present invention. FIG. 4 is a diagram illustrating a rigid link model used in the walking motion analysis process according to the embodiment of the present invention. FIG. 5 is a diagram illustrating contact points defined on the foot in the walking motion analysis process according to the embodiment of the present invention. FIG. 6 is a hardware block diagram showing a typical computer for implementing a walking motion analysis device according to an embodiment of the present invention. FIG. 7 is a diagram illustrating an environment for a walking motion measurement experiment. FIG. 8 is a graph showing the actual value of the floor reaction force actually measured in the environment shown in FIG. 7 and the estimated value of the floor reaction force estimated by the walking movement analysis process shown in FIG. 3. FIG. 9 is a graph showing the actual values of the point of action of the ground reaction force measured in the environment shown in FIG. 7 and the estimated value of the point of action of the ground reaction force estimated by the walking motion analysis process shown in FIG. 3. be. Figure 10 shows the actual value of the ground reaction force measured in the environment shown in Figure 7 and the estimated value of the ground reaction force estimated by the walking motion analysis process shown in Figure 3 at each point in the gait cycle during normal walking. It is a graph showing joint moments of the lower limbs.

Hereinafter, embodiments of the present invention will be described, but the embodiments of the present invention are not limited to the embodiments described below. In the embodiments described below, walking motion is targeted as an example of an information processing device for estimating external forces that occur during body exercise, and the system estimates the ground reaction force that occurs during walking motion, and also estimates the ground reaction force that occurs during walking motion and calculates the This will be explained using the walking motion analysis device 110 that calculates the generated mechanical quantity.

FIG. 1 is a functional block diagram showing the overall configuration of an analysis system 100 including a walking motion analysis device 110 according to this embodiment. The analysis system 100 shown in FIG. 1 includes a motion capture system 102 that measures the motion of a person's body during walking motion, and receives measurement data from the motion capture system 102 and analyzes the walking motion. The walking motion analysis device 110 is configured to include a walking motion analysis device 110.

The motion capture system 102 is a system that digitizes body movements during movement based on any type of capture technology, such as optical, inertial sensor, mechanical, magnetic, or video. Here, the optical method is a method in which a reflective marker is attached to the subject's body, a plurality of cameras are installed around it, and the position of the reflective marker is optically detected, and includes an image method and an infrared method. The inertial sensor type is a method in which an inertial sensor including an angular velocity meter or an accelerometer is attached to each part of the body, and the position and orientation are calculated backward from the obtained acceleration and angular velocity information. The mechanical method is a method in which each joint angle is measured using sensors such as encoders and potentiometers that mechanically measure rotation angles and displacements. The magnetic type is a method in which a magnetic sensor is attached to the body, a magnetic field is applied from a magnetic generator, and the position and orientation from the magnetic generator are determined based on magnetic lines of force detected by a coil in the magnetic sensor. The video method is a method that analyzes video images and obtains position information. Note that, for convenience, the case in which the optical method, which is a standard method, is adopted will be mainly described below.

FIG. 1 also shows detailed functional blocks of the walking motion analysis device 110. As shown in FIG. 1, the walking motion analysis device 110 includes a body motion data acquisition unit 112 that acquires body motion data from the motion capture system 102, and a body motion data acquisition unit 112 that acquires body motion data from the motion capture system 102. Below, there is an inverse dynamics analysis section 120 that performs inverse dynamics calculations, and an optimization calculation section that estimates the unknown floor reaction force to be found through optimization calculations based on the results of the calculations by the inverse dynamics analysis section 120. 130.

The body motion data acquisition unit 112 acquires body motion data 114 in a predetermined format that describes body motion. Here, the body motion data directly obtained from the motion capture system 102 is body motion displacement data in the case of the above-mentioned optical system. On the other hand, here, body motion data 114 according to a predetermined body model 116 is acquired in order to perform inverse dynamics calculations to be described later.

Body model 116 is a rigid link model in certain embodiments. A rigid link model is a model that assumes that each body segment of the body can be approximated by rigid segments (links) that do not deform, and represents the body mechanics structure as a rigid link model in which rigid links are connected.

The body motion displacement measurement data obtained from marker measurement is the three-dimensional coordinates of the position of the marker provided on the body surface in an absolute coordinate system (spatial coordinate system). The body motion data acquisition unit 112 acquires motion displacement measurement data from the motion capture system 102, and further uses the three-dimensional coordinates of the marker in the absolute coordinate system (spatial coordinate system) using, for example, a global optimization method or a point cluster method. Based on this, three-dimensional movements of each body segment, such as joint angles, are determined. In addition, in the above-mentioned inverse dynamics calculation, in addition to motion displacement such as joint angle, angular velocity and angular acceleration are also used, so in the case of an optical method, these velocity and acceleration information are used to numerically differentiate motion displacement. Calculate by Note that depending on the motion capture method, it may be possible to omit part of the above-mentioned calculation of joint angles and calculations of velocity and acceleration. The body motion data acquisition unit 112 acquires joint angles, angular velocities, and angular accelerations based on kinematic characteristics such as the length of each body joint of the rigid body link model and joint degrees of freedom from data directly obtained from the motion capture system 102. The kinematic quantities are calculated to obtain body motion data 114 according to a predetermined body model 116.

The inverse dynamics analysis unit 120 performs an inverse dynamics calculation of the equation of motion of the body model 116 under a predetermined ground reaction force based on the acquired body motion data (joint angles, angular velocities, angular accelerations, etc.) 114. Then, the mechanical quantities of each body part, more specifically, the joint moment (also referred to as joint torque, muscle torque, etc., which are synonymous) and joint reaction force generated at each joint of the body model 116 are calculated. Generally, external force information such as ground reaction force is used in inverse dynamics calculations, but in the embodiment of the present invention, a provisional value 118 of the ground reaction force is assumed, and a value below that provisional value 118 is assumed. Inverse dynamics calculations are performed. By executing the inverse dynamics calculation, an estimated value 122 of each joint moment is calculated for the assumed provisional value of the ground reaction force.

Here, more specifically, the floor reaction force used in the inverse dynamics calculation includes a floor reaction force vector representing the magnitude and direction of the floor reaction force, and a point of action of the floor reaction force. Furthermore, the walking movement has a single-leg support phase in which either the left or right foot is in contact with the floor, and a double-leg support phase in which both the left and right feet are in contact with the floor. In the single-leg support phase, the ground reaction force is that of the foot that touches the ground, and in the double-leg support phase, the ground reaction force is typically set separately for each of the left and right feet.

The optimization calculation unit 130 collaborates with the inverse dynamics analysis unit 120 to perform optimization calculations on the unknown floor reaction force 118 (floor reaction force vector and center of pressure (COP)) to be determined. Estimated by The optimization calculation unit 130 gives a provisional value 118 of a predetermined floor reaction force, causes the inverse dynamics analysis unit 120 to perform an inverse dynamics analysis, evaluates the resulting analysis result 122, and performs the next optimization. Repeat updating the provisional value 118 of the ground reaction force used in calculations and search for a ground reaction force that minimizes (if defined as a penalty) or maximizes (if defines as a reward) a predetermined objective function. do. Based on the obtained analysis results, the direction in which the objective function is improved in the search space is obtained. By repeating this inverse dynamics analysis, evaluation, and updating of the provisional value of the floor reaction force, the estimated value of the floor reaction force is optimized and obtained as an analysis result. At the same time, estimated values 122 of mechanical quantities such as joint moments and joint reaction forces of the body model 116 are also provided as analysis results when the optimized estimated values of the ground reaction forces are provided.

Once the optimal values of mechanical quantities such as joint moments and joint reaction forces of the body model 116 are obtained, they are combined with a muscular strength model and a musculoskeletal geometry model, and an optimization calculation method is further applied to minimize muscle load. By applying this method, it is possible to calculate the in-vivo load on the musculoskeletal system, such as muscle strength. Here, the muscle strength model is a model that expresses the mechanical characteristics associated with the force exertion of individual muscles, and the musculoskeletal geometry model is a model that expresses the mechanical characteristics associated with the force exertion of individual muscles. This is a model that represents A musculoskeletal model is constructed by combining a rigid link model, a muscle strength model, and a musculoskeletal geometry model. Note that there are no particular limitations on how the mechanical quantities of the obtained body model 116 (rigid link model) are used. Furthermore, not only the mechanical quantities of the body model 116 (rigid link model) but also the finally obtained floor reaction force vector and point of application are useful information. Note that in the embodiment of the present invention, the explanation will focus on estimating external forces such as ground reaction force used in inverse dynamics calculations, so the processing after estimating the ground reaction force vector and the point of application will not be described further. Do not enter.

FIG. 2 shows more detailed functional blocks of the optimization calculation unit 130 shown in FIG. 1. As shown in FIG. 2, the optimization calculation unit 130 includes a floor reaction force initialization unit 132, an analysis execution unit 134, a total load evaluation unit 136, an end determination unit 138, and a floor reaction force update unit 140. Contains and consists of.

The floor reaction force initial setting unit 132 first sets a provisional value of the floor reaction force as an initial value. Although the initial value is not particularly limited, any fixed value may be set, or a random number generated within a predetermined range using a random number function may be set. Here, the ground reaction force given to the inverse dynamics analysis includes the ground reaction force vector (direction and magnitude) and the point of application of the ground reaction force, and the provisional value of the ground reaction force includes the ground reaction force vector (direction and magnitude) and the point of application of the ground reaction force. and provisional values of the coordinates of the point of application.

The analysis execution unit 134 provides the acquired body motion data and the provisional value of the ground reaction force set at the current time to the inverse dynamics analysis unit 120, executes an inverse dynamics analysis using the body model 116, and performs an inverse dynamics analysis using the body model 116. Obtain the results of the dynamic analysis. The execution results of the inverse dynamics analysis include estimated values of the mechanical quantities occurring in each joint. The total load evaluation unit 136 calculates the sum of the mechanical quantities occurring in each joint of the body model based on the result of the inverse dynamic analysis performed by the inverse dynamic analysis unit 120, that is, based on the estimated value of this mechanical quantity. , evaluate the load that occurs on the body. The floor reaction force updating unit 140 updates the provisional value of the floor reaction force based on the result of the evaluation by the total load evaluation unit 136.

By repeating execution of inverse dynamics analysis by the inverse dynamics analysis section 120, evaluation by the total load evaluation section 136, and updating of the provisional value of the floor reaction force by the floor reaction force updating section 140 under predetermined conditions, a predetermined result is obtained. The objective function is minimized (if defined as a penalty) or maximized (if defined as a reward) and the ground reaction force is optimized.

The termination determination unit 138 determines whether the termination conditions for the inverse dynamics analysis by the inverse dynamics analysis unit 120 are satisfied. For example, a predetermined number of aborts or conditions for detecting convergence of the objective function are used. When the termination determination unit 138 determines that the termination condition is satisfied, the analysis is terminated and the final value of the mechanical quantity and the final value of the ground reaction force generated in each joint of the body model 116 are obtained.

A preferred embodiment is also shown in FIG. 2, and as shown in FIG. . In a preferred embodiment, the predetermined conditions for the above-mentioned optimization include a constraint to keep the point of application of the ground reaction force within the ground contact surface of the foot. By introducing this constraint, it is possible to reduce the possibility that the position of the point of action of the ground reaction force is estimated to be outside the sole of the foot.

The movement displacement measurement data holds the three-dimensional coordinate position of the marker provided on the body surface, as described above, but in a preferred embodiment, the movement displacement measurement data further includes a plurality of coordinates to interpolate the plurality of markers forming the foot. contact points are defined. The ground contact determination unit 142 performs a ground contact determination for each contact point defined on the foot based on the height of the contact point and the speed of the contact point. Then, the contact surface is given as a convex hull composed of the contact points determined to be in contact with the ground.

The inside/outside determining unit 144 determines whether the provisional value of the point of action is located inside the ground contact surface obtained based on the ground contact determination of the contact point. The constraint for keeping the point of application of this ground reaction force within the contact surface of the foot may be incorporated as a constraint expression (constraint condition) in the optimization calculation, or may be incorporated in the objective function of the optimization calculation. good. In certain embodiments, as a penalty term given when the provisional value of the point of action is located outside the ground plane, or as a reward term given when the provisional value of the point of action is located inside the ground plane. Incorporated into the objective function of optimization calculation.

Hereinafter, with reference to FIG. 3, the walking motion analysis process including estimation of the floor reaction force during walking motion according to the present embodiment will be described in more detail. FIG. 3 is a flowchart showing a walking motion analysis process executed by the walking motion analysis device 110 according to the embodiment of the present invention. The walking motion analysis process shown in FIG. 3 is executed by a processor such as a CPU included in the walking motion analysis device 110. Furthermore, the processing shown in Fig. 3 is based on body movement data at a predetermined point in time during walking motion, and in order to analyze the step cycle, the process from one foot's heel contact to heel separation and toe separation is necessary. Similar processing is performed at each point in time until the heel contacts the ground again. In this case, the initial value for calculation at the next time point may be determined based on the result obtained at the previous time point.

The walking motion analysis process shown in FIG. 3 starts from step S100, for example, in response to an instruction from an operator. In step S101, the processor reads the body model 116.

FIG. 4 is a diagram illustrating a rigid link model 200 as the body model 116 used in the walking motion analysis process according to the embodiment of the present invention. As shown in FIG. 4, the rigid link model 200 includes a plurality of links 202 and a joint section 204 that connects the links. Degrees of freedom are set. For example, basic motion constraints are set assuming that the hip joint has three degrees of freedom: flexion/extension, internal/external rotation, and rotation, and the knee joint and ankle joint have three degrees of freedom: rotation.

Referring again to FIG. 3, in step S102, the processor acquires measurement data from the motion capture system 102, and acquires body movement data at a predetermined time point during walking movement. The body motion data is converted into a format that includes information such as joint angles, angular velocities, and angular accelerations of each joint of the body model 116. In step S103, the processor sets provisional values of the floor reaction force vector and the point of action of the floor reaction force as initial values. The initial value is set as a fixed value, a random value, or an optimal value at a past point in time. Thereafter, in steps S104 to S107, optimization calculations are performed using a predetermined objective function and constraint expression.

In step S104, the processor performs an inverse dynamics analysis of the body model 116, giving the acquired body motion data and the current ground reaction force vector and provisional values of the point of action of the ground reaction force. In step S105, the processor evaluates the load generated on the body according to the mechanical amount generated at each joint of the body model, based on the execution result of the inverse dynamic analysis. More specifically, in step S105, the following objective function I is evaluated. Note that the objective function I is configured with the load as a penalty.

Here, f _pelvis indicates the joint reaction force generated at the pelvic joint corresponding to the degree of freedom with respect to the spatial coordinates during inverse dynamics calculation, n _i indicates the joint moment generated at joint i, and f _grf is the body represents the ground reaction force, which is an external force that acts on the Note that the hat attached to each symbol indicates that it has been normalized based on Hof's method. In this normalization, the force is normalized by the body weight given as the physique condition, and the moment is the body weight given as the physique condition and the length of each part such as leg length (foot length etc.).These are also physique conditions. (obtained from the height given as ). Also, in certain embodiments, normalization processing for speed changes may be performed. In the case of walking motion, the number of external forces is 2 for the left and right feet. In the above formula, a ₁ , a ₂ , and a ₃ are optimization weighting coefficients. Parameters a ₁ , a ₂ , and a ₃ are typically set to arbitrary values that satisfy predetermined conditions.

The above objective function I is based on the assumption that ``a person walks so that the load generated on each part of the human body is minimized,'' and it evaluates the sum of mechanical quantities obtained by inverse dynamics calculation. be. The optimization calculation is formulated as shown in equations (1) and (2) below.

Regarding the first condition of equation (2) above, the reaction force f _pelvis of the pelvic joint is ideally zero, but in reality it is not zero due to modeling and measurement errors, so it is weighted more heavily than other terms. By doing so (a ₁ >a ₂ >a ₃ ), the influence is minimized. Further, each term in the above equation (1) is expressed as the sum of cubes of absolute values in order to clearly show the characteristics of each term. Although not limited to the sum of cubes, it is preferable to use a nonlinear monotonically increasing function.

The second condition of the above equation (2) is a constraint for keeping the action point position p _grf (x, y) of the ground reaction force within the ground contact area C of the foot. The foot ground contact area C is determined by defining contact points based on the marker positions of the foot and performing a ground contact determination for each contact point.

FIG. 5 is a diagram illustrating contact points defined on the foot in the walking motion analysis process according to the embodiment of the present invention. As shown in FIG. 5, a plurality of contact points 214 are defined to interpolate the plurality of markers 212 that make up the foot 210. Then, the grounding discriminant shown in equation (3) below is applied to each defined contact point 214, and the foot grounding area C is obtained as a convex hull composed of the contact points determined to be in contact with the ground. It will be done.

In the above equations (3) and (4), the subscript grf is the external force number, j is the contact point number, p is the position of the contact point, v is the speed of the contact point, and the subscript added to p and v is The letter thres represents a threshold value for the contact point position and contact point velocity, and x, y, and z appended to p and v represent components in the traveling direction, left-right direction, and vertical direction, respectively. f represents the estimated ground reaction force vector. In (3) above, a contact point whose height is below a predetermined standard and whose horizontal and vertical velocities are smaller than the standard is determined to be in contact with the ground. A contact area can be defined when there are three or more contact points per foot. Therefore, according to the above formula (4), the number of contact points is counted, and when the number of contact points reaches three or more, a floor reaction force is applied.

In a particular embodiment, the constraint for keeping the point of action position p _frg (x, y) of the ground reaction force within the foot ground contact area C is such that the point of action position p _frg (x, y) is outside the ground contact area C. In some cases, the penalty shown in equation (5) below can be added to the objective function of equation (1) as the fourth term to incorporate it into the optimization calculation.

In the above formula (5), a is a weighting coefficient, and d represents the distance from the center of the foot ground contact area C to the point of action p _grf (x, y). Divergence of values can be prevented by adding a penalty depending on the distance of the point of action from the center.

The predetermined conditions in the optimization calculation described above may further include a constraint for minimizing the distance between the point of application of external force and the zero moment point (ZMP). Referring again to the above equation (2), the third condition of the above equation (2) represents a constraint that minimizes the distance between the estimated point of action position of the floor reaction force and the zero moment point. ZMP is a point on the floor surface where the moment caused by inertia force and gravity at the center of gravity of the body becomes zero, and when walking on a flat surface, this point coincides with the position of the point of action of the floor reaction force. In addition, during the single-leg support period, the distance between the point of action of the ground reaction force on one supported leg and the ZMP is evaluated, and during the double-leg support period, the distance between the point of application of the ground reaction force of the left and right feet and the ZMP is evaluated. distance can be evaluated. The zero moment point is determined by recursively determining the center of gravity position of each body segment based on the measurement data from the motion capture system 102, and calculating the center of gravity acceleration of each body segment from the center of gravity position. It is derived from the acceleration of the center of gravity of each body segment.

Referring again to FIG. 3, in step S106, the processor determines whether a termination condition is satisfied. When a convergence condition is satisfied, such as reaching a predetermined upper limit number of aborts or no further improvement of the objective function is expected, it is determined that the termination condition is satisfied. If it is determined in step S106 that the optimization has not been completed yet, the process branches to step S107. In step S107, the processor updates the provisional value of the floor reaction force, preferably in the direction of improvement in which the objective function decreases, based on the results of the above-described evaluation.

If it is determined in step S106 that the termination condition is satisfied, the process branches to step S108. The execution of the inverse dynamics calculation in step S104, the evaluation of the objective function in step S105, and the updating of the provisional value of the floor reaction force in step S107 are repeated until the termination condition is satisfied in step S106. force vector and point of application) are optimized. In step S108, the processor sets the current provisional value of the ground reaction force as the final value, and calculates the mechanical quantities such as joint moment and joint reaction force as a result of the inverse dynamics calculation when the ground reaction force is given. Output the value as the final value. In step S109, the process ends.

The hardware configuration of the walking motion analysis device 110 will be described below. Gait motion analysis device 110 is implemented as a general-purpose computer in certain embodiments. FIG. 6 is a hardware configuration diagram of the walking motion analysis device 110 according to this embodiment. The gait motion analysis device 110 includes a single-core or multi-core microprocessor unit (MPU) 12 on a board 10, a nonvolatile memory 14 that stores a BIOS (Basic Input Output System), and a program that can be processed by the MPU 12. and a memory 16 that provides execution storage space for the operations.

The MPU 12 is connected to the storage control interface 18 via an internal bus 22, and the hard disk 20 writes or reads data in response to input/output requests from the MPU 12. MPU 12 controls a serial or parallel interface 24, such as a USB, via an internal bus 22 to communicate with input/output devices 26, such as a keyboard, mouse, printer, etc., and to receive input from a user.

Gait motion analysis device 110 may further include VRAM 28 and graphics chip 30. The graphics chip 30 processes video signals in response to commands from the MPU 12 and displays them on the display device 32. The MPU 12 is also connected to a network interface card (NIC) 34 via an internal bus 22. This allows the walking motion analysis device 110 to communicate with an external device such as a motion capture system through the network.

The walking motion analysis device 110 is stored in a storage device such as a non-volatile memory 14, a hard disk 20 (or SSD (solid state drive)), an NV-RAM (not shown), an SD card (not shown), etc. A program (not shown) is read and expanded in the memory area of the memory 16. Thereby, each functional means and each process described above is realized under an appropriate operating system (OS). The above OS may be Windows (registered trademark), UNIX (registered trademark), LINUX (registered trademark), android (registered trademark), iOS (registered trademark), MacOS (registered trademark), iPadOS (registered trademark), etc. An OS having an architecture can be adopted.

The above functional unit can be realized by a computer executable program written in a legacy programming language such as assembler, C, C++, C#, Java (registered trademark), or an object-oriented programming language, and can be implemented in a ROM, EEPROM, Stored in a device-readable recording medium such as EPROM, flash memory, flexible disk, CD-ROM, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, Blu-ray disc, SD card, MO, or connected to a telecommunications line. It can be distributed through.

As described above, the embodiments of the present invention provide an information processing device, a calculation method, and a program that can robustly estimate the ground reaction force acting on the body during walking motion based on body motion data. can do.

The algorithm for estimating the ground reaction force described above is based on the assumption that "people walk so as to minimize the load on their bodies," and uses a body model 116 that expresses the kinematic and inertial characteristics of the body. By minimizing the sum of the mechanical quantities obtained from the inverse dynamics calculation based on this method, the unknown ground reaction force and the location of its point of action can be found in an exploratory manner.

The movement information required for mechanical analysis is only body movement data and external force information is not required, so if the exercise is in an environment where a motion capture camera can be installed, it is recommended to install a force sensor such as a floor reaction force meter. At least it can be analyzed. Furthermore, since the estimation method uses inverse dynamic calculation of a body model constructed based on marker positions, uniform estimation is possible regardless of the subject. As a result, uniform estimation can be expected regardless of the measuring device or subject.

In certain embodiments, multiple contact points are provided within the sole of the foot from the foot marker location to which a ground contact formula consisting of height and velocity is applied. Then, a contact surface is formed by only the contact points that touch the ground, and a determination is made as to whether or not the estimated point of action is on this surface. If the point of action is determined to be outside, by adding a constraint amount corresponding to that amount to the optimization formula, it becomes possible to estimate the position of the point of action with the minimum in-vivo load within the sole of the foot.

In addition, in the above description, human walking was the subject of analysis, but the subject of analysis is not limited to humans. Even objects such as animals and robots can be analyzed.

Verification Experiment Below, we implemented the above-mentioned floor reaction force estimation algorithm and applied it to actually measured physical movement data during walking to perform a verification experiment. The verification experiment will be explained below.

In order to verify the validity of the ground reaction force and the position of the point of action estimated by the above-mentioned ground reaction force estimation algorithm, an experiment was conducted to measure actual walking motion. The subjects were six healthy adult males, and the measurement devices used were an optical motion capture system (Optitrack, Natural Point) and a floor reaction force meter (TF-406-D, Tech Gihan). It was installed as shown in 7. A floor reaction force meter was used to obtain measured values for comparison with estimated values. In addition, since it is difficult to measure the end point of a single step cycle with a single floor reaction force meter, we used a mat switch (OM-1074, Osaka Jidodenki Co., Ltd.) to measure the point of heel contact, which is the end point of one walk cycle. I measured it. In FIG. 7, the camera of the optical motion capture system is indicated at 308 and the floor reaction force meter is indicated at 304. 306 is a mat switch, and 302 is a stand for configuring a surface that is flush with the location where the floor reaction force meter is installed.

The measurement frequency of each device was unified to 100 Hz, and the measurement signal was passed through a low-pass filter for smoothing, and its cutoff frequency was 10 Hz. In the verification experiment, each subject's normal walking, fast walking, and slow walking on the walking path were measured three times each. The degree of slowness compared to normal walking at this time was determined by each subject. As indicators for comparing estimated values and actual values, we used the coefficient of determination, which indicates the degree of similarity, and the mean squared error, which indicates the size of the error.

Furthermore, based on the three-dimensional positions of each body part obtained by the optical motion capture system, we defined the positions and coordinate systems of each joint of the body, and constructed a body model with 18 whole body joints and 47 rotational degrees of freedom. . In addition to the placement of the VICON (registered trademark) Plug-in-Gait, markers for measurement were attached to a total of 51 locations on various parts of the body, with reference to the VICON (registered trademark) Oxford Foot Model, which is specialized for feet. Further, as shown in FIG. 5, the contact points were arranged so that each contact point was located at the midpoint of the marker and other contact points, for a total of 31 contact points.

Results of verification experiment Compare the estimated and measured values of the ground reaction force and its point of action. FIG. 8 is a graph showing the actual value of the floor reaction force actually measured in the environment shown in FIG. 7 and the estimated value of the floor reaction force estimated by the walking movement analysis process shown in FIG. 3. FIG. 9 is a graph showing the actual values of the point of action of the ground reaction force measured in the environment shown in FIG. 7 and the estimated value of the point of action of the ground reaction force estimated by the walking motion analysis process shown in FIG. 3. be. FIG. 8 and FIG. 9 show estimated values and actual measured values of the ground reaction force and the position of the point of action for one subject at each walking speed. On the horizontal axis, one gait cycle was taken from the time when the right foot touched the floor reaction force meter until the ipsilateral foot touched the mat switch. Furthermore, the position of the point of action indicates the locus during the stance period. Table 1 below summarizes the comparison results between estimated values and actual measured values using evaluation indicators.

The upper part of Table 1 above shows the evaluation of this experiment. If the coefficient of determination is close to 1 and the mean squared error is close to 0, it can be said that the estimated value is close to the measured value.

From the above results, first of all, regarding the ground reaction force, the longitudinal and vertical components were estimated to be close to the actually measured values when the behavior was similar. Regarding the size of the error, the longitudinal and vertical directions showed relatively small errors, but the error in the vertical direction was larger than other components, and the error increased as the walking speed increased. Regarding the position of the point of action, the error was larger in the front-rear direction than in the left-right direction, where there is less movement. Regarding the difference in estimation accuracy between subjects, no major difference was observed in ground reaction force. Regarding the position of the point of action, there were variations between subjects, but the subjects shown in FIGS. 8 and 9 had a coefficient of determination of 0.7 or more in the anteroposterior direction, indicating relatively high estimation accuracy. There were also variations between subjects in terms of the magnitude of the error, but in the examples shown in FIGS. 8 and 9, the error was approximately 0.03 m.

The reason why the estimation accuracy of both the ground reaction force and the position of the point of action decreases as the walking speed increases is that normalization processing for speed changes is not performed in each term of the objective function of the optimization calculation. Conceivable. Since each mechanical quantity of the objective function is normalized based on Hof's method, the optimization calculation is less susceptible to differences in the body physique information of the subjects. However, when the same subject walks at a speed higher than normal walking speed, the burden on each joint of the body becomes greater, so the value of the objective function increases compared to when walking at a low speed. Therefore, it is thought that the higher the walking speed, the lower the estimation accuracy.

Regarding the position of the point of action, compared to the ground reaction force and the position of the point of action in the longitudinal direction, the value of the error variation in the horizontal direction of the position of the point of action is very close to the average miscalculation value, and this is due to the ZMP This is thought to be due to the estimation error. ZMP is derived from the position of the center of gravity of the body segment, which is recursively determined based on motion capture data. Therefore, it is thought that the error at that time had an influence. Furthermore, since there is less change in the position of the point of action in the left-right direction than in the front-back direction, it is thought that the error in the center of gravity acceleration may have a greater influence on the results.

The joint moment in the bending/extending direction at each joint of the lower limb is calculated using the estimated ground reaction force and the position of the point of action, and compared with the joint moment calculated from the actually measured ground reaction force.

FIG. 10 shows joint moments based on estimated values and actual measured values of ground reaction force during normal walking of the same subject as shown in FIGS. 8 and 9, which showed the highest estimation accuracy. From FIG. 10, it can be seen that the error that occurs in the ankle joint in the latter half of the stance phase affects the knee joint and hip joint. It is expected that the estimation accuracy will be improved by normalizing the joint moments of the objective function according to each joint and body segment. The current normalization method divides each joint moment by a formula using body weight and leg length, so by normalizing by the length and weight of the body segment connected to the joint, the priority of minimization can be improved. This can be unified for each joint, which is thought to have the potential to improve estimation accuracy.

With measurement data from a motion capture system, it is possible to mechanically evaluate body movements without using information from force sensors, making it ideal for facilities where elderly people or people with disabilities who have difficulty in moving live. It is now possible to perform mechanical analysis of body movements in places where it has been difficult to install floor reaction force meters, such as factories with a large number of machine and production machines.

Although the embodiments of the present invention have been described so far, the embodiments of the present invention are not limited to the above-described embodiments, and those skilled in the art may come up with other embodiments, additions, changes, deletions, etc. Modifications may be made within the range that is possible, and any embodiment is included within the scope of the present invention as long as the effects and effects of the present invention are achieved.

DESCRIPTION OF SYMBOLS 100... Analysis system, 102... Motion capture system, 110... Walking motion analysis device, 112... Body motion data acquisition part, 114... Body motion data, 116... Body model, 118... Provisional value of ground reaction force, 120... Inverse dynamics analysis section, 122... Estimated value of mechanical quantity, 130... Optimization calculation section, 132... Floor reaction force initial setting section, 134... Analysis execution section, 136... Total load evaluation section, 138... Convergence determination section, 140 ... Floor reaction force updating section, 142 ... Ground contact determination section, 144 ... Internal/external determination section, 200 ... Rigid body link model, 202 ... Link, 204 ... Joint section, 210 ... Foot section, 212 ... Foot section marker, 214 ... Contact point, 302...stand, 304...floor reaction force meter, 306...mat switch, 308...camera, 310...body, 10...board, 12...MPU, 14...BIOS, 16...memory, 18...memory control interface, 20...hard disk Drive (HDD), 22...Internal bus, 24...Interface, 26...I/O device, 28...VRAM, 30...Graphic chip, 32...Display device, 34...NIC

Claims (12)

An information processing device for estimating a ground reaction force generated in a foot during physical exercise, the information processing device comprising:
an acquisition means for acquiring physical movement data;
Setting means for setting a provisional value of the ground reaction force;
Execution means for executing an inverse dynamics analysis using a body model based on the body motion data and the provisional value of the ground reaction force;
Evaluation means for evaluating the load generated on the body according to the mechanical amount generated in each joint of the body model, based on the execution result of the inverse dynamics analysis;
updating means for updating the provisional value of the ground reaction force based on the result of the evaluation;
Grounding determination means for determining grounding of a plurality of contact points defined on the foot;
Determining whether or not the provisional value of the point of action on which the ground reaction force acts is located inside a ground contact surface configured based on a convex hull of three or more contact points determined to touch the ground in the ground contact determination. and an inside/outside determining means for determining whether or not the provisional value of the point of action is located inside the ground contact surface of the foot is reflected in the evaluation by the evaluation means. The ground reaction force includes the direction and magnitude of the ground reaction force, and includes the point of action on which the resultant force of the external forces acting on both the left and right feet acts during the two-leg support phase when both the left and right feet are in contact with the floor. 2. The information processing apparatus according to claim 1 , wherein a constraint that minimizes a distance from a zero moment point is taken into consideration in the evaluation by the evaluation means . The execution result of the inverse dynamic analysis includes an estimated value of a mechanical quantity occurring in each joint, and the evaluation by the evaluation means is performed based on the estimated value of the mechanical quantity. Information processing device. Under predetermined conditions, the value of the floor reaction force is optimized by repeating execution of the inverse dynamics analysis by the execution means, evaluation by the evaluation means, and updating of the provisional value of the floor reaction force by the updating means. The information processing device according to any one of claims 1 to 3, wherein the predetermined condition includes a constraint for keeping the point of action within the ground contact surface of the foot. The contact points include points defined to interpolate a plurality of measured contact points constituting the foot, and the ground contact determination includes, for each contact point, a height of the contact point and a point of the contact point. The information processing device according to any one of claims 1 to 4, wherein the information processing device performs the processing based on speed. The information processing device includes:
It further includes a termination determination means for determining the termination of the inverse dynamics analysis, and when the termination determination means determines that the analysis has been completed, the final value of the mechanical quantity occurring in each joint of the body model and the The information processing device according to any one of claims 1 to 5, wherein a final value of the floor reaction force is obtained. Any one of claims 1 to 6, wherein the body movement is a walking movement, the body model includes a rigid link model representing a body mechanical structure as a model in which rigid links are connected, and the mechanical quantity includes a joint moment. The information processing device according to item 1. A calculation method for estimating the ground reaction force generated in the foot during physical exercise, the computer
obtaining physical movement data;
setting a provisional value of the ground reaction force;
performing an inverse dynamics analysis using a body model based on the body motion data and the provisional value of the ground reaction force;
Evaluating the load generated on the body according to the mechanical amount generated at each joint of the body model, based on the execution result of the inverse dynamics analysis;
and updating the provisional value of the ground reaction force based on the result of the evaluation, and the calculation method further includes the step of:
a step of determining ground contact of a plurality of contact points defined on the foot;
Whether or not the provisional value of the point of action on which the ground reaction force acts is located inside a ground contact surface configured based on a convex hull of three or more contact points determined to be in contact with the ground in the step of determining ground contact. and determining whether or not the provisional value of the point of action is located inside the ground contact surface of the foot is reflected in the evaluation in the evaluating step. The ground reaction force includes the direction and magnitude of the ground reaction force, and includes the point of action on which the resultant force of the external forces acting on both the left and right feet acts during the two-leg support phase when both the left and right feet are in contact with the floor. , a constraint that minimizes the distance to the zero moment point is taken into account in the evaluation in the evaluating step . The calculation method includes a step in which the computer optimizes the value of the ground reaction force by repeating the executing step, the evaluating step, and the updating step under predetermined conditions; The calculation method according to claim 8 or 9, wherein the conditions include a constraint for keeping the point of action within the ground contact surface of the foot. A program for realizing an information processing device for estimating the ground reaction force generated in the foot during physical exercise, which uses a computer to
acquisition means for acquiring physical movement data;
a setting means for setting a provisional value of the ground reaction force;
Execution means for executing an inverse dynamics analysis using a body model based on the body motion data and the provisional value of the ground reaction force;
evaluation means for evaluating the load generated on the body according to the mechanical amount generated at each joint of the body model, based on the execution result of the inverse dynamics analysis;
updating means for updating the provisional value of the ground reaction force based on the result of the evaluation;
a ground contact determination means for determining ground contact of a plurality of contact points defined on the foot; and a provisional value of the point of action on which the ground reaction force acts is determined to be a contact point of three or more contact points determined to be in contact with the ground in the ground contact determination. This is a program for functioning as an inside/outside determining means for determining whether or not the foot is located inside a ground contact surface constructed based on a convex hull, and the provisional value of the point of action is determined to be within the ground contact surface of the foot. A program in which the evaluation by the evaluation means reflects whether or not the program is located in the . The ground reaction force includes the direction and magnitude of the ground reaction force, and includes the point of action on which the resultant force of the external forces acting on both the left and right feet acts during the two-leg support phase when both the left and right feet are in contact with the floor. 12. The program according to claim 11, wherein a constraint that minimizes the distance from the zero moment point is taken into consideration in the evaluation by the evaluation means .

Images