JP2024016491A - Visibility evaluation system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the adaptability of confirmation action simulation for a design element with respect to the situation, thereby improving the quality of visibility evaluation in designing a product using a 3DCAD.

SOLUTION: A visibility evaluation system includes: a biomechanical model for simulating a confirmation action with respect to a visual recognition target point in a virtual 3D space in which an object and the visual recognition target point are defined; a field-of-view determination model for calculating visibility with respect to the visual recognition target point in a field-of-view image in a line of sight of the biomechanical model; and a consolidated control unit which couples the biomechanical model and the field-of-view model. The consolidated control unit applies perturbation to the biomechanical model, executes an operation to generate a visual recognition attitude with behavior having minimum purposeful burden and maximum visibility, the purposeful burden being obtained by multiplying a degree of deviation from a reference attitude set in the biomechanical model by a degree of burden of the biomechanical model due to the perturbation in the attitude, and calculates easiness of visual recognition from an accumulated value of the degree of burden and the final value of the visibility with respect to the visual recognition target point.

SELECTED DRAWING: Figure 1

COPYRIGHT: (C)2024,JPO&INPIT

Description

The present invention relates to a 3D CAD visibility evaluation system using a biomechanical model.

Conventionally, in product design using 3D CAD, evaluations regarding visual effects such as visibility and evaluations regarding practical effects such as operability and ease of handling have been performed separately. Furthermore, evaluations by testers cannot exclude subjective influences, making it difficult to obtain quantitative evaluations.

Therefore, the present inventors used a biomechanical model to simulate the confirmation movement for a visible target point defined in a virtual 3D space, reflected the posture of the biomechanical model on the human body model, and responded to the line of sight of the biomechanical model. In addition to calculating the visibility of the visual target point from the visual field image of the human body model, the degree of burden on the biomechanical model required for the confirmation movement is calculated, and the easy visibility is evaluated from the integrated value of the degree of burden and the final value of visibility. We have created a visibility evaluation system (see Patent Documents 1 to 3).

JP2020-42734A JP2020-107278A Japanese Patent Application Publication No. 2020-107279

This visibility evaluation system is based on a configuration that autonomously simulates the confirmation movement performed by humans by perturbing the biomechanical model and generating a posture with a behavior that provides the minimum burden and maximum visibility. There is. However, depending on the conditions and samples, the posture prediction may not be accurate to the situation if only the assumptions (constraint conditions) of minimum burden and maximum visibility are used. For example, in the case of a confirmation operation by a driver in the cockpit of a vehicle, it is desirable that the driving posture be maintained as much as possible not only while driving but also while the vehicle is stopped. However, depending on the conditions and samples, the confirmation motion in the simulation, especially the final posture, may be excessive (overaction) compared to the confirmation motion normally performed by humans.

The present invention has been made in view of the above-mentioned circumstances, and its purpose is to improve the suitability of the confirmation motion simulation for design elements to the situation in the visibility evaluation in product design using 3D CAD, and to improve the visibility. The aim is to improve the quality of evaluation.

In order to solve the above problems, the present invention
A visibility evaluation system for objects using 3D CAD, the system comprising:
a biomechanical model for simulating a confirmation movement for the visual target point in a virtual 3D space in which the object and its visual target point are defined;
a visual field determination model for calculating the degree of visibility for the visual target point in the visual field image in the line of sight of the biomechanical model;
A perturbation is applied to the biomechanical model, and in that posture, the visibility with respect to the visual target point, the degree of burden on the biomechanical model due to the perturbation, and the degree of deviation from the reference posture set for the biomechanical model. The biomechanical model and a general control unit that couples the visual field judgment model;
The visibility evaluation system is configured to calculate easy visibility from the integrated value of the degree of burden and the final value of the visibility.

As described above, the visibility evaluation system according to the present invention introduces an index called the degree of deviation from the reference posture set in the biomechanical model into the overall control unit that couples the biomechanical model and the visual field judgment model. , an autonomous confirmation movement that attempts to maintain the reference posture as much as possible by generating a posture with a behavior that minimizes the purposeful degree of burden and maximizes visibility considering the degree of deviation in the degree of burden. This is advantageous for improving suitability to the simulation situation and improving the quality of visibility evaluation accordingly.

FIG. 1 is a block diagram showing a visibility evaluation system according to an embodiment of the present invention. It is a typical side view showing simulation by a visibility evaluation system concerning an embodiment of the present invention. FIG. 2 is a schematic plan view illustrating coupling of a biomechanical model and a visual field judgment model in the visibility evaluation system according to an embodiment of the present invention through data exchange. It is an ER diagram showing a data structure used in a visibility evaluation system concerning an embodiment of the present invention. It is a flow chart showing processing of a visibility evaluation system concerning an embodiment of the present invention. It is a flow chart which describes processing of the visibility evaluation system concerning an embodiment of the present invention divided into a biomechanical model and a visual field judgment model.

Embodiments of the present invention will be described in detail below with reference to the drawings.

1. Overview of Evaluation System FIG. 1 is a block diagram showing a visibility evaluation system 1 according to an embodiment of the present invention. The evaluation system 1 of this embodiment uses 3D CAD data of an instrument panel arranged in the front part of the cockpit of a vehicle as evaluation target data 2, and uses a biomechanical model 10 as shown in FIGS. 2 and 3. A confirmation operation for visually confirming the visual recognition target point 20 is simulated, and visibility including ease of the confirmation operation is evaluated.

The evaluation system 1 includes a biomechanical model 10 and a visual field judgment model 50, and separate programs exist for each operation. Each program and data, as well as an operation program that collectively operates each program, are stored in an external storage device (not shown) of the computer, etc., are read into the main memory (RAM), and arithmetic processing is executed by the CPU. It is configured to function as an evaluation system.

The biomechanical model 10 is composed of a musculoskeletal model 3 and an eye movement model 4, and by integrating these, it generates body movements and eye movements for visual field judgment, and also calculates the degree of burden 13 required for them. The degree of deviation 14 from the reference posture is calculated, and the posture data of the eyeball 5 obtained as a result of each movement is reflected on the camera 55 in the visual field judgment model 50, thereby calculating the degree of visibility 15 for the visibility target point 20. Finally, the evaluation unit 17 evaluates the ease of viewing (easy visibility) including the ease of confirmation operation based on the burden level 13 and the visibility level 15.

The visibility 15 calculated by the visual field judgment model 50 is fed back to the general control unit 11 of the biomechanical model 10, and the biomechanical It becomes part of the motion norm potential that optimizes the checking motion of the model 10. In this way, the biomechanical model 10 and the visual field judgment model 50 operate independently but are coupled by sharing or exchanging data between them, thereby minimizing the burden level 13 on the biomechanical model 10, It is possible to perform a simulation that simultaneously maximizes the visibility 15 in the visual field judgment model 50.

That is, in the two models, a purposeful optimal movement that minimizes the burden level 13 while maintaining the reference posture of the biomechanical model 10 (reducing the degree of deviation 14) improves the visibility level 15 in the visual field judgment model 50. There is a coupled relationship in which an improvement in the visibility 15 in the visual field judgment model 50 brings about an optimal movement that minimizes the burden 13 in the biomechanical model 10.

Data can be shared or exchanged between two models by accessing the database 8 as shown in FIG. 4, but in addition to this, any other method can be used, such as data communication or transferring data files through file operations. It can be implemented with

The database 8 includes simulation data 80 that stores conditions, results, etc. in each simulation (trial), behavior data 81 that stores data of each behavior process in each simulation (trial), and data for acquiring each behavior process. includes perturbation data 82 that stores data of a perturbation process virtually performed in , and there is a one-to-many relationship between the primary key (PK) of each data table and the foreign key (FK) of the corresponding data table. Configured as a configured relational database.

The simulation data 80 stores input conditions and the like for each simulation (trial). For example, in the illustrated example, support structure information (first support point (hip point; HP) on the seat 6, second support point (left and right grip points; coordinates) on the steering wheel 7, corresponding to the CAD data of the cockpit where the simulation is performed, x, y, z)), physique setting information (height, weight) of the person to be simulated, CAD data to be evaluated (name, identification code), visible target point coordinates (coordinates (x, y, z)), etc. are input. The visibility evaluation and rendering information as a result of the simulation process are stored.

Although not shown, basic information and rendering data such as support structure information (cockpit CAD data), physique setting information, and evaluation target CAD data are each registered in separate data tables, and the identification code (foreign key). In addition, when performing a simulation by passing data files, data files of appropriate formats corresponding to simulation data 80, behavior data 81, and perturbation data 82 are configured, and data is additionally updated. After the behavior data 81 in each behavior process is acquired, the perturbation data 82 required for the behavior data 81 can also be set not to be retained.

Regarding the behavior data 81 and the perturbation data 82, as will be described in detail later, the movements of the musculoskeletal model 3 and the eye movement model 4 in the biomechanical model 10 are virtually calculated by the number of state variables in order to determine the optimal behavior. By repeating the perturbation process that calculates the operating burden B, deviation degree D, visibility R, and visibility objective function Fo, and the behavior process of actually operating for a unit time under conditions that minimize the visibility objective function Fo. Implemented.

Therefore, first, the behavior number (ID) in the behavior data 81 is assigned by automatic numbering, etc., and then, regarding the behavior number, the perturbation numbers for the state variables are assigned to the perturbation data 82, and the muscle and bone Perturbations of case model 3 and eye movement model 4 are performed. The skeletal link posture, eyeball posture, burden B, and deviation degree D obtained as a result of each perturbation are stored in the perturbation data 82, and the eyeball posture is reflected on the camera 55 of the visual field judgment model 50 and calculated from the visual field image. The visibility R is stored in the perturbation data 82.

Next, the conditions under which the visibility objective function Fo calculated based on the burden level B, the deviation level D, and the visibility level R are minimized are stored in the behavior data 81, and the behavior process is implemented according to the condition to determine the skeletal link posture. and eyeball posture are registered as key frame information. By accumulating this key frame information, it becomes possible to generate a 3DCG video of a visual field image corresponding to the confirmation operation (and a 3DCG video viewed from an arbitrarily set viewpoint). Further, as the behavior process is executed, the visual recognition objective function Fo acquired in the previous behavior process is updated and stored in the behavior data 81. This visibility objective function Fo becomes basic information for the final easy visibility evaluation in the simulation.

2. Details of Visual Field Judgment Model The visual field judgment model 50 includes a human body corresponding to evaluation target data 2 having a visual recognition target point 20 and a biomechanical model (musculoskeletal model 3, eye movement model 4) in a 3D coordinate system (virtual data space). The model 53 and camera 55 are defined, and the body movement and eye movement (body posture and eyeball posture) generated by the biomechanical model are reflected on the human body model 53 and camera 55, and the visual target point 20 in that state is reflected. This is to obtain visibility.

This camera 55 is a viewpoint (corresponding to the human eye) set in the 3D coordinate system, and the camera image becomes the field of view of the human (human body model 53). By registering the posture of the human body model 53 in key frames from moment to moment, a moving image can finally be created. If you set an external viewpoint, you can also visually observe your body's behavior.

The visibility calculation process using the visual field judgment model 50 will be described below. Note that all processes are performed automatically by a program. Since it is based on image processing in 3D data space, general-purpose 3DCG applications can also be used.

As mentioned earlier, the evaluation system 1 of this embodiment uses 3D CAD data of the instrument panel arranged in the front part of the cockpit of the vehicle as the evaluation target data 2, and as shown in FIGS. 2 and 3, the evaluation system 1 corresponds to the driver. The biomechanical model 3 is used to simulate the confirmation motion to visually confirm the visual recognition target point 20, and the visibility including the ease of the confirmation motion is evaluated. and behavior), the visibility with respect to the visibility target point 20 at each point in time is calculated.

Humans have a visual field of about 200° in the horizontal direction, but the visual field that has a certain influence on cognitive behavior and can effectively utilize information is about 2° from central vision, so eye movement can change the visual field. Furthermore, although the maximum range of eye movement is about 55 degrees, large eye movements are taxing, so the reflex movement of the eyeballs called saccades is less than 15 degrees. , If further viewpoint movement is required, not only the eyeballs but also the head actively move.

Therefore, even when the visual recognition target point 20 is set within the range of the viewing angle, confirmation motions that involve body movements such as eye movements and eye/head coordination movements are induced. Even if the visible target point 20 is outside the range of the viewing angle, if knowledge of the position of the visible target point 20 is obtained in advance, a confirmation action will be triggered in that direction.

In the illustrated example, the visual target point 20 is assumed to be a switch located at the lower right side of the instrument panel 2. This switch is located in a position hidden behind the steering wheel 7, making it difficult to see it directly from the viewpoint (5) of the driver (3) seated on the seat 6. As shown in FIG. It gradually becomes visible through confirmation movements that involve movement and head rotation.

In order to obtain the visibility of the visible target point 20 even when it is not directly visible at the initial position, the gradation sphere 21 is defined as a three-dimensional object obtained by radially enlarging the visible target point 20. Thereby, visibility can be calculated by at least partially including the gradation sphere 21 in the visual field image.

Furthermore, in order to weight the visibility in the center of the gradation sphere 21, that is, in the vicinity of the visibility target point 20 and in the peripheral area, only specific pixel information is gradually decreased according to the distance from the center of the gradation sphere 21. In this embodiment, the R value indicating the intensity (luminance) of red color is used as specific pixel information, and the gradation sphere 21 has a radial shape such that the red color becomes darker on the center side and the red color becomes lighter on the peripheral side. It has a gradation.

The pixel information used in the visual field determination model 50 is based on the "RGBA" color model, which is the three primary colors of red (R), green (G), and blue (B) plus an alpha channel (A) for transparency. RBGA is a numerical value of 0 to 1.0, and is expressed as [R, B, G, A]. The gradation sphere 21 installed at the visual target point 20 has an R value that gradually decreases between the center R value = 1.0 (maximum) and the surface R value = 0 (minimum), and the G value and B value are 0, the A value is 1.0 because it is based on a completely transparent sphere, and therefore the RGBA of the gradation sphere 21 is:
[R, G, B, A] = [0 to 1.0, 0, 0, 1.0]
It will be expressed in the range of

When calculating the visibility in the visual field judgment model 50, all solids other than the gradation sphere 21 (evaluation target data 2, handle 7, sheet 6, person model 53, etc.) are black and opaque (does not reflect the light source, [R, G, B, A] = [0, 0, 0, 0]), only the gradation sphere 21 partially hidden by the black solid (handle 7, etc.) can be seen in the field of view image of the camera 55. It has become.

In this state, by calculating the sum of the R values of all pixels in the field of view image of the camera 55, it is possible to quantify how visible the red gradation sphere 21 is, and based on this, the visual recognition target point 20 can be visually recognized. degree can be judged.

Since this visibility (total R value) uses the gradation sphere 21, it can be extended and quantified even when the visibility target point 20 is not directly visible, and the state is approaching the state where it can be seen from a state where it is not directly visible. Since the situation can be acquired as an improvement in visibility, it is ideal for evaluating visibility (easiness of visibility) including confirmation movements, and as will be described in detail later, it is fed back to the biomechanical model 10 and the biomechanical model 3 It is also significant in that it is reflected in the motor norm potential that directs the movement (confirmation movement) of the body.

Note that as the specific pixel information of the gradation sphere 21 used in the visual field determination model 50, the brightness of green or blue other than red can be used, and furthermore, information other than brightness can also be used. For example, it is also possible to define red dots inside a transparent gradation sphere with a density depending on the distance from the center and calculate the sum of the R values.

3. Details of the biomechanical model The biomechanical model 10 is composed of a musculoskeletal model 3 and an eye movement model 4. Basically, body movement is generated in the musculoskeletal model 3, and eye movement is coupled to the body movement. evoked. Therefore, first, a flow of generating a body movement for confirming the visual target point 20 in the musculoskeletal model 3 will be explained.

3.1 Musculoskeletal model (body mechanics model)
Musculoskeletal model 3 is a rigid link model in which human body segments correspond to three-dimensional rigid links, and has 17 to 20 links and 39 to 43 joint degrees of freedom in the whole body.

That is, the musculoskeletal model 3 includes a head 31, a cervical vertebrae 32, a thorax (thoracic vertebrae) 33, a lumbar vertebrae, a pelvis 35, thighs, lower legs, feet 36c constituting the left and right legs 36, left and right collarbones, and left and right arm parts 37. It consists of the upper arm, forearm, and hand. For example, the elbow joint between the upper arm and the forearm has one degree of freedom for bending, and the wrist joint between the forearm and the hand 37d has three degrees of freedom for forward and backward bending, internal and external bending, and twisting.

In the example, the values of body parameters such as the mass and moment of inertia of each body joint are assumed to be a typical adult male, and the values of body parameters such as the mass and moment of inertia of each body joint are set based on data such as height and weight in input data (simulation data). Scale parameters. Furthermore, passive resistance due to joint structures such as joint soft tissues acts on each joint.

3.2 Definition of motion representative points It is preferable that the musculoskeletal model 3 is biomechanically appropriate for reproducing various body motions, and that it is simple and has a low calculation cost. Therefore, representative motion points are set for each body part, and the motion trajectory of each representative motion point can be defined. In the embodiment, the eyeball coordinates of the head 31 (the intermediate coordinates of the left and right eyeballs 5 or the central coordinates of the head 31 may be used) and By defining two points on the chest 33 (also a reference part) as the motion representative point p _rep (3 degrees of freedom) and as the motion representative angle q _rep (3 degrees of freedom) on the neck 32, the body can be calculated using a total of 9 variables. prescribed exercise.

The three-dimensional position of the motion representative point is described by a spatial coordinate system, and body motion is defined as movement of the motion representative point from an initial position to a target (terminal) position. This coordinate position can also be defined as a relative coordinate value with respect to another motion representative point. For example, if the center coordinates of the head 31 are used as a representative motion point, the confirmation motion can be defined as the movement of this point, and the eyeball coordinates can also be defined based on this representative motion point.

3.3 Posture generation using quasi-static motion norm potential The joint degrees of freedom of the musculoskeletal model 3 are redundant for the nine variables of motion representative points and motion representative angles that define body motion. In order to determine these redundant joint angles in a manner that is autonomous and physiologically valid, a mechanical calculation algorithm based on an inverse model using a quasi-static motion norm potential is used as described below.

First, the joint angle q of each joint at one point in time in discrete time is used as a state variable, and the posture and motion are realized by inverse dynamics calculation for the equation of motion of the musculoskeletal model of the body 3 based on angular velocity, angular acceleration, and external force. The virtual joint drive torque n _v is defined as follows.
Here, d _inv (・) is a function for performing inverse dynamics calculation, and f _ext indicates an external force acting on the musculoskeletal model 3, which is the contact force between the musculoskeletal model 3 and the mechanical environment such as the seat 6. , is determined as a function of joint displacement and joint displacement velocity.

Next, once the current joint angle q is determined, the motion representative point position p _rep (q) defined on the musculoskeletal model 3 determined by it, the visual target point position obtained as setting information, and the optimization described later. The magnitude of the difference from the target motion representative point position p _ser determined by calculation is expressed as the motion norm potential UΔp for the motion representative point, and similarly, the motion representative angle q _rep (q) on the current musculoskeletal model 3 and the target motion representative angle Letting the difference from q _ser be the motion norm potential UΔq for the motion representative angle, the motion norm potential UΔp and the motion norm potential UΔq are as follows.

In addition to these motion norm potentials, considering the motion norm potential for the body motion load defined by the sum of the squares of the virtual joint torque n _v by the inverse dynamics calculation of equation (3.1), the weighted linear sum of these potentials is expressed as Define the general motion norm potential U(q).
Here, k ₁ , k ₂ , and k ₃ are weighting coefficients. Since it is difficult to analytically obtain this comprehensive motion norm potential, we calculate the potential U (q + Δq) when perturbation Δq is given to the joint angle q as a state variable, and calculate the approximate motion norm potential by the difference. Find the gradient ∇U.

By adding the influence of the gradient ∇U of the motion norm potential to the virtual joint drive torque n _v obtained by the above inverse dynamics calculation, the actual joint drive torque n _real becomes as shown in the following equation.
Here, k ₄ and k ₅ are weighting coefficients. By considering the gradient ∇U of the motion norm potential in this way, a posture is generated that reduces both the body motion load and the difference between the current position and the target position of the motion representative point.

3.4 Body motion generation by forward dynamics calculation In addition to the joint drive torque n _real determined by the motion norm potential as described above, by performing forward dynamics calculation, the body motion corresponding to the given motion representative point position can be calculated. generate. Here, we assume a state variable v corresponding to the joint drive torque, define a motion norm potential U _fwd (v) based on this, and define dynamics for reducing this as shown in the following equation.
Here, n(q) is a joint drive torque obtained by inverse dynamic calculation, and k ₆ and k ₇ are coefficients.

The motion norm potential U _fwd is determined by forward dynamics calculation as follows. First, when a state variable v equivalent to a joint drive torque is given, the acceleration at the present moment (time t) can be estimated as shown in the following equation.

Here, M is the inertia matrix, h is the Coriolis force, and f _ext is the external force. From this acceleration, the displacement and velocity of the body at time t+Δt can be estimated by time integration as shown in the following equation.

From these, the speed and position of the motion representative point at time t+Δt can be estimated as shown in the following equation.

The above f ₁ and f ₂ are functions for determining the motion representative point position and velocity from the joint displacement and velocity by forward kinematic calculation of the body structure. The motion reference potential U _fwd is defined by the following equation based on the difference between the predicted motion representative point position P _pre obtained by forward dynamics calculation and the target motion representative point position Pser. A ₂ and A ₃ are weighting coefficient matrices.

The motion norm potential U _fwd defined by the above equation does not take into account coupling with eye movement. The coupling between the body movement and the eye movement is realized by optimization calculation that minimizes the visual recognition objective function Fo, which will be described later.

3.5 Eye movement model The eyeball achieves eye movement through the contraction of six types of muscles called extraocular muscles.The greater the movement, the more fatigue the extraocular muscles cause, and humans feel strain on their eyes. Therefore, the degree of burden on the eyeballs is defined as follows based on the eyeball angular velocity and eyeball posture during eyeball movement.
Here, ω _eye is an eyeball angular velocity vector, θ _eye is an eyeball posture, and k ₁₁ , k ₁₂ , and k ₁₃ are coefficients. Since this is the eye movement model 4 for evaluating visibility, it is assumed that visual recognition is performed through central vision, and the line of sight is defined by a vector connecting the position of the eyeball 5 determined by the mechanical calculation of the musculoskeletal model 3 and the position of the visual target point 20. Ru. The eyeball posture θ _eye is the angle of the inner product of the line of sight when viewing the target point and the neutral posture of the eyeball 5 .

The eyeball/head coordination movement has different characteristics depending on the static environment and the running environment, and the degree of burden on the eyeballs also changes depending on these environments. Specifically, in a static environment, head movement occurs when the horizontal angle of the target point of eye movement exceeds 10°, and the head angle (rotation angle of the neck 32) becomes larger than the eyeball angle (eyeball angle). On the other hand, in a driving environment, eye movement takes priority and active head movement does not occur until the horizontal angle of the target point is about 30°.

Therefore, different values were set for the coefficient _k13 in the static environment and in the driving environment so that the degree of eye strain increases rapidly from θ _eye = 10° in a static environment and from θ _eye = 30° in a driving environment. By setting this, it becomes possible to perform simulations and visibility evaluations that take into account the visibility environment.

Furthermore, regarding the neutral posture of the eyeballs, in addition to the geometric neutral posture with zero elevation and depression angles, there is also a posture in which the eyeball load is minimal regardless of visual field information, that is, there is no contraction of the extraocular muscles and the eyeballs are not moving. It is now possible to set the eyeballs to a neutral posture. We conducted a verification experiment with subjects to determine how much the neutral posture of the eyeballs is tilted in the vertical direction with respect to the head coordinate system (31a). It was confirmed that it was pointing downward with an angle of depression of ~10°. Therefore, in addition to the horizontal angular characteristics of eye movement under the driving environment and static environment mentioned above, by considering the characteristics of the neutral posture of the eyeball in the vertical direction, we can improve the simulation and visibility that reflect each viewing environment. Evaluation can be carried out.

3.6 Optimization of visual confirmation movement As mentioned earlier, the body movement of the musculoskeletal model 3 (rigid link model) is based on the motion representative points of the head 31 and chest 33, which play a central role in the confirmation movement, and the movement representative points of the neck 32. A quasi-static motion posture is generated by defining the confirmation motion and eye coordinates using the motion representative angle as a variable, giving a perturbation Δq using the joint angle q as a state variable, and finding the joint drive torque. The motion representative point position and motion representative angle after unit time Δt are calculated using the motion norm potential U _fwd , but coupling with eye movement is not taken into consideration.

Therefore, in the visual recognition operation, a visual recognition objective function that takes into account the degree of deviation D from the reference posture and defines the relationship between the degree of burden B of physical exercise including eyeball load and body load and the degree of visibility R output from the visual field judgment model. By defining Fo as the following formula, the body movement and eye movement are coupled through an optimization calculation that calculates the body movement and eye movement to minimize Fo, and the optimal visual recognition movement is generated. .
Here, L _eye is the eyeball load, L _neck is the mechanical load generated in the neck, L _body is the mechanical load on the body other than the neck, R is visibility, and k ₁₄ , k ₁₅ , k ₁₆ , and k ₁₇ are The weighting coefficient C _pos is the displacement (distance) of the reference part from the reference posture, and the degree of deviation from the reference posture D (1+k ₁₄ C _pos ) is 1 when the displacement C _pos from the reference posture is the minimum of zero. It has been adjusted so that

The eyeball load L _eye is given by the above-mentioned equation (3.14), and the neck load L _neck and body load L _body are given by the following equation.
Here, ξ is a number corresponding to each body part, and neck load L _neck corresponds to the neck 32, body load L _body corresponds to body parts other than the neck 32, n _ξ is joint drive torque, and τ _ξ is It is joint passive resistance, and the burden on the body due to muscle activity is calculated as the square root of the sum of the squares of the joint drive torque, which corresponds to the force exerted by the muscles. In addition, at joint angles near the limit of joint range of motion, a reaction occurs due to joint passive resistance, which also places a load on the joint, so the root sum of the squares of joint passive resistance is also added, but the deviation from the reference posture On the premise that the degree D is minimized, it can be assumed that the joint angle will remain within a relatively small range and may be omitted.

The degree of deviation D (1+k ₁₄ C _pos ) from the standard posture is an index for maintaining the standard posture if the reciprocal is taken, and when the visual recognition movement in the sitting posture is targeted, the standard posture can be maintained by setting the chest 33 as the standard part. The degree of retention of can be suitably reflected. The displacement C _pos of the reference part is given by the following equation, where the current position p _pre and the initial position p _int of the chest 33 are expressed.

If the displacement C _pos of the reference part set on the chest 33 is zero, this means that the visual recognition movement is performed only by eye movement, or by eye movement and head rotation, and the head 31 must be tilted. In this case, the displacement C _pos becomes a significant value. Therefore, the reference region can also be set to the head 31.

In addition, even in cases other than the seated posture, for example, when the target is a visual recognition movement in a standing posture, the degree of maintenance of the reference posture is reflected in the simulation by considering the degree of deviation D with the chest 33 and the head 31 as reference parts. be able to.

In the above equation (3.14), since the eyeball posture θ _eye is taken into account in the eyeball load L _eye , for example, even if the visual target point 20 is set within the viewing angle, the eyeball angle will be large due to eyeball movement alone. In such a case, by minimizing the visual recognition objective function Fo, the characteristics of eyeball/head coordination movement, which can be substituted for head movement, can be reproduced while avoiding an increase in eyeball load.

Furthermore, since the degree of deviation D from the basic posture is taken into account in the above equation (3.15), eye movements are first induced so that the basic posture is maintained as much as possible, and the eyeballs and head are The natural confirmation movement that transitions from coordinated movement to physical movement is autonomously reproduced, and it is expected to improve the suitability to the simulation situation and improve the quality of visibility evaluation accordingly.

In particular, the degree of deviation D is adjusted so that it is 1 when the displacement C _pos of the reference part is a minimum of zero, so that it is possible to avoid confirmation movements by eye movement or eye/head coordination movement that maintains the basic posture. This is advantageous in continuously evaluating visibility through confirmation movements that involve physical movement.

4. Processing Flow of Visibility Evaluation System As described above, the evaluation system 1 according to the present invention uses the biomechanical model 10 to simulate the confirmation movement of the visibility target point 20 in the evaluation target data 2. The visibility is evaluated in consideration of the ease (degree of burden) and the degree of maintenance of the reference posture, and FIG. 5 is a flowchart showing the processing flow of the evaluation system 1 according to the embodiment.

Furthermore, as described above, the evaluation system 1 is composed of the biomechanical model 10 and the visual field judgment model 50 that operate independently and in conjunction with each other, and FIG. 6 shows the processing flow of the evaluation system 1 according to the embodiment. is described separately into processing in the biomechanical model 10 (confirmation motion simulation) and processing in the visual field judgment model 50 (visibility judgment). Further, in FIG. 3, reference numerals for some of the processes in FIG. 6 are also shown, and data exchange between the biomechanical model 10 and the visual field judgment model 50 is shown.

The last two digits of the code of each step shown in the 100s in FIG. 5 correspond to the last two digits of the code of each step shown in the 200s and 300s in FIG. The processing flow of the evaluation system 1 will be explained with reference to FIG.

When evaluating visibility, first, the computer that constitutes the evaluation system 1 is started, and the biomechanical model 10 and the visual field judgment model 50 are started while the main program (1) that controls the simulation and the database 8 are running. (steps 100, 200, 300).

Next, according to the input information of the simulation data 80, the predetermined evaluation target data 2 is read into the 3D data space of the visual field judgment model 50, and the visual recognition target point 20 of the evaluation target data 2 is acquired by the biomechanical model 10 (step 101 , 201, 301).

Next, in the visual field judgment model 50, the gradation sphere 21 is set at the visual recognition target point 20 of the evaluation target data 2 (steps 102, 302). The data of the gradation sphere 21 itself is prepared in advance in the visual field determination model 50, and the center coordinates of the gradation sphere 21 are set as the visual recognition target point 20.

Next, posture information is read from the simulation data 80, an initial posture (reference posture) is generated in the musculoskeletal model 3 and the eye movement model 4 of the biomechanical model 10, and the initial posture is reflected in the human body model 53 of the visual field judgment model 50. At the same time, the initial position of the camera 55 is acquired (steps 103, 203, 303). In this state, a visual field image is acquired by the camera 55, and an initial value R ₀ of visibility is calculated.

When the perturbation/behavior process is started after the preparations described above, first, the motion generation module 12 calculates the state variables of the inverse dynamics calculation routine and the forward dynamics calculation routine (steps 110, 210, 310).

Next, the joint drive torque is calculated by inverse dynamic calculation, and the posture after a unit time is estimated by time-integrating the acceleration by forward dynamic calculation. In this way, the biomechanical model 10 is perturbed, and the obtained posture information (skeletal link posture, eyeball posture, all quaternions), eyeball load, and joint drive torque (cervical load, etc.) that is applied to the body mechanics system. body load) is obtained as the degree of burden B, and the amount of displacement of the reference part 33 is obtained as the degree of deviation D from the reference posture, and is stored in the perturbation data 82 together with the state variable (steps 111, 211).

Furthermore, the posture information obtained by the above perturbation is reflected on the human body model 53 of the visual field judgment model 50 and the camera 55 (FIG. 6, step 311), and the visibility R is expressed as the sum of the R values of the visual field images acquired by the camera 55. is calculated (steps 112, 312) and stored in the perturbation data 82.

The perturbation process as described above is repeated for the number of state variables, and when all state variables have been perturbed (steps 113 and 213), the process moves to the behavior process.

First, the visibility objective function Fo is minimized from among the state variables for which the burden degree B, the deviation degree D, and the visibility R are obtained in the perturbation process (minimum purposeful burden degree B and D and maximum visibility R). ) are calculated and the corresponding joint drive torques are determined (steps 120, 220).

Next, the biomechanical model 10 (musculoskeletal model 3 and eye movement model 4) is actually driven by the determined state variables and joint drive torque, a posture based on the behavior is generated, and posture information (skeletal link posture) is added to the behavior data 81. , eyeball posture (all quaternions) are stored (steps 121, 221).

Furthermore, the posture information based on the above behavior is reflected on the human body model 53 of the visual field judgment model 50 and the camera 55 (FIG. 6, step 321), and the visibility R is calculated as the sum of the R values of the visual field images acquired by the camera 55. (Steps 122, 322) and is stored in the behavior data 81. Additionally, posture information of the human body model 53 is registered in the key frame (FIG. 6, step 324).

In parallel with this, in the database 8, the integrated value of the burden level B including the behavioral process is calculated, and with reference to the visibility R, the easy visibility is calculated taking into account the burden level B (integrated value). (step 222).

The visibility R is then compared to a preset threshold (steps 130, 230, 330) and if it is determined that the threshold has not been reached, the perturbation/behavior process as described above is repeated.

When the visibility R exceeds a preset threshold, the simulation by the biomechanical model 10 is terminated, and easy visibility is determined based on the visibility R and the burden B (integrated value) calculated in the final behavior process. The performance evaluation is recorded in the simulation data 80 as a result of the simulation (trial No.) (FIG. 6, step 231).

At the same time, in the visual field judgment model 50, rendering is executed based on the data of key frames (0.1 second interval, 10 fps in the example) registered up to the final behavior process, and an evaluation video is created (FIG. 6, Step 331).

Note that during rendering, body movements between key frames are interpolated by a video creation function. Further, the surface information of the evaluation target data 2 (instrument panel) and the human body model 53 may be reflected, or the position of the visual target point 20 may be made clear by using a gray scale or the like. In addition to the field-of-view image captured by the camera 55 (human body model 53), an image from an external viewpoint set at an appropriate position such as obliquely above the human body model 53 can be created.

In the perturbation process (steps 111-112, 211, 311-312) of the above embodiment, the posture information obtained in the biomechanical model 10 is reflected in the human body model 53 and camera 55 of the visual field judgment model 50 for each perturbation. However, instead of calculating the visibility R from the visual field image, the biomechanical model 10 acquires posture information for all (or some) state variables and then transfers it all at once to the visual field judgment model 50. It is also possible to perform visual field image acquisition and visibility R calculation.

In addition, in the perturbation process (steps 111-112, 211, 311-312) of the above embodiment, a case has been described in which perturbations are given to the biomechanical model 10 by the number of state variables. By applying perturbations, it is also possible to estimate state variables that minimize the visibility objective function Fo and determine joint drive torques for the behavior.

Further, in the above embodiment, a case has been described in which the perturbation/behavior process is repeated until the visibility R becomes equal to or higher than a preset threshold, but an upper limit is set on the number of repetitions (or repetition time) of the perturbation/behavior process, The simulation by the biomechanical model 10 may be terminated when a predetermined number of times (predetermined time) is reached.

On the other hand, as the visibility R increases due to repetition of the perturbation/behavior process, the joint drive torque calculated by the motion generation module 12 gradually decreases, and the acceleration gradually decreases, causing any state value (or its rate of change) to change. When the value falls below the setting, it may be considered that the confirmation operation by the biomechanical model 10 has autonomously converged.

Further, in the above step 100 (200), a plurality of biomechanical models 10 are activated (multiple activation) for one visual field judgment model 50, and in step 111 (211), a state variable is set for each biomechanical model 10. The calculation of the perturbation and the degree of burden B (and the calculation of the degree of deviation D from the reference posture) may be performed by parallel calculation. Since the biomechanical model 10 has a larger calculation load than the visual field judgment model 50, parallel calculation is advantageous in reducing processing time and leveling resources.

In the above embodiment, the biomechanical model 10 simulates the movement of checking one visible target point 20 from the initial posture, but it is possible to simulate the movement of sequentially checking a plurality of visible target points to comprehensively It is also possible to evaluate visibility.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above embodiments, and various modifications and changes can be made based on the technical idea of the present invention.

For example, in the above embodiment, a case has been described in which the instrument panel of the vehicle is the evaluation target data 2 and the switch on the lower right side thereof is the visual recognition target point 20, but the present invention is not limited to this. The visual target point may be at another position, and data other than the instrument panel may be the data to be evaluated. In addition, it can be applied to the visibility evaluation and design support of various items that are expected to be seen in conjunction with confirmation movements, such as standing postures other than sitting postures, walking postures, and confirmation of visual target points when getting in and out of a vehicle. .

1 Evaluation system 2 Evaluation target data (instrument panel)
3 Musculoskeletal model (body model)
4 Eye movement model 5 Eyeball 6 Support structure (sheet)
7 Support structure (handle)
8 Database 10 Biomechanical model 11 General control unit 12 Movement generation module 13 Load level 14 Discrepancy level 15 Visibility 16 Visibility objective function 17 Easy visibility comprehensive evaluation unit 20 Visibility target point 21 Gradation sphere 30 Body 31 Head 32 Neck 33 Chest (reference site)
50 Visual field judgment model 53 Human body model 55 Camera (viewpoint)

Claims (7)

A visibility evaluation system for objects using 3D CAD, the system comprising:
a biomechanical model for simulating a confirmation movement for the visual target point in a virtual 3D space in which the object and its visual target point are defined;
a visual field determination model for calculating the degree of visibility for the visual target point in the visual field image in the line of sight of the biomechanical model;
A perturbation is applied to the biomechanical model, and in that posture, the visibility with respect to the visual target point, the degree of burden on the biomechanical model due to the perturbation, and the degree of deviation from the reference posture set for the biomechanical model. The biomechanical model and a general control unit that couples the visual field judgment model;
A visibility evaluation system, comprising: a visibility evaluation system configured to calculate easy visibility from the integrated value of the degree of burden and the final value of the visibility. The degree of deviation (D) is a weighting coefficient k ₁₄ that takes into account the degree of burden, and a minimum value of 1 when the amount of displacement Cpos is zero, as the amount of displacement Cpos of the reference part of the biomechanical model from the reference posture. The visibility evaluation system according to claim 1, wherein the visibility evaluation system is calculated by the formula: D=1+k ₁₄ Cpos. The degree of burden (B) is calculated using the following formula: the eyeball load Leye of the biomechanical model, the mechanical load Lneck on the neck, the mechanical load Lbody on the body other than the neck, and the respective weighting coefficients k ₁₅ , k ₁₆ , k ₁₇ :B=k1 ₅ Leye+k ₁₆ Lneck+k ₁₇ Lbody,
The visibility evaluation system according to claim 2, wherein the visibility posture is generated by a behavior that minimizes a visibility objective function Fo=D·B/R with respect to the visibility (R). The object is 3D CAD data of an instrument panel arranged in front of the driver's seat of the vehicle defined as the virtual 3D space, and the biomechanical model is a sitting posture in which the driver is seated and looks straight ahead. is defined as the reference posture, a reference part serving as a reference for calculating the deviation degree is set to the chest or head of the biomechanical model, and the biomechanical model and the visual field judgment model are installed on the instrument panel. The visibility evaluation system according to claim 1, wherein the visibility evaluation system is adapted to perform a simulation of a confirmation operation for the visibility target point. The perturbation of the biomechanical model is defined using the positions of the motion representative points set on the head and the chest and the angle of the motion representative angle set on the neck as variables, and is defined as the index of the degree of burden. The visibility evaluation system according to claim 4, wherein the mechanical load is calculated based on a joint drive torque of the biomechanical model. A visibility evaluation system for objects using 3D CAD, the system comprising:
a biomechanical model for simulating a confirmation movement for the visual target point in a virtual 3D space in which the visual target point of the object is defined;
a visual field for calculating the visibility of the visual target point from a visual field image of the human body model corresponding to the line of sight of the biomechanical model in a virtual 3D space in which a human body model corresponding to the biomechanical model and the object are defined; judgment model,
An overall control unit that couples the biomechanical model and the visual field judgment model,
perturbing the biomechanical model;
reflecting the posture of the biomechanical model due to the perturbation on the human body model;
a step of obtaining from the visual field judgment model a degree of visibility for the visual target point calculated from a visual field image of the human body model reflecting the posture; a step of calculating a degree of burden on the biomechanical model due to the perturbation;
A perturbation process including a step of calculating a deviation degree of the biomechanical model from a reference posture due to the perturbation is performed, and the purposeful load degree obtained by multiplying the load degree by the deviation degree is the minimum and the visibility is the maximum. a general control unit that repeatedly operates a perturbation/behavior process that generates a visual posture of the biomechanical model with a behavior that becomes;
A visibility evaluation system configured to calculate easy visibility from the integrated value of the degree of burden and the final value of the visibility. activating a plurality of the biomechanical models for one of the visual field judgment models;
A step of applying a perturbation equivalent to a state variable to the biomechanical model, a step of calculating a degree of burden on the biomechanical model due to the perturbation, and a step of calculating a degree of deviation of the biomechanical model from a reference posture due to the perturbation. The visibility evaluation system according to claim 6, wherein the steps are executed by parallel calculation.

Images