JPH10172005A - Operation data connection method and device - Google Patents

Abstract

(57) [Summary] The present invention relates to an operation before and after performing a connection.
It is an object of the present invention to provide an operation data connection method capable of generating a natural connection operation with a short calculation time without adding any condition such as periodicity. In the present invention, connection of front and rear motion data is mainly independent of a first processing block for obtaining a current time position, a second processing block for calculating a combined attitude angle, and a combined joint angle. From the third processing block. In the first processing block, the position at the current time is calculated. In the second processing block, a combined posture angle is calculated. In the third processing block, a composite slide vector is calculated, and a composite joint angle synthesized by a connection function is calculated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the field of computer graphics animation, and more particularly to a method and apparatus for connecting a plurality of time-series motion data provided for moving a rigid articulated object. .

[0002]

2. Description of the Related Art In recent years, in the field of computer graphics and animation, it has become essential to realistically move a living thing such as a human modeled by a rigid versus articulated object. The three-dimensional time-series data representing such movements is generated by people called animators using a method using keyframe interpolation, or by capturing real movements using a three-dimensional motion measurement method called motion capturing. Is done.

[0003] Motion data is generated in a minimum unit having a minimum meaning as motion in consideration of reduction of data processing time and reuse thereof. Therefore, in order to construct a series of motions over a long period of time, it is necessary to connect the fragmented motions composed of these minimum unit data. The most primitive connection method is that a professional person called an animator described above manually estimates connection operation between operations and creates connection operation data. However, the connection operation created is often unnatural because the productivity is low due to the manual operation, and since the animator estimates the operation in the three-dimensional space by personal experience. Under such circumstances, the following two types of operation connection data generation methods have been conventionally proposed.

First, as a method of automatically generating operation connection data for a periodic operation, Munetoshi Unuma, et al.,
"Fourier Principles for Emotion-based Human Figur
e Animation ", SIGGRAPH95 Proceeding, pp91-96.1995.
There is. In this method, the time series data of the operation before and after the connection is subjected to Fourier expansion, the operation in the connection section is extrapolated in the frequency space, and is generated by performing inverse Fourier expansion. In other words, the operation connection data is generated on the premise of the prediction using the periodicity.

Further, as a method of generating operation connection data which does not necessarily assume the periodicity of operation, Charles Roth
se, et al., "Efficient Generation of Motion Transi
tions using Spacetime Constraints ", SIGGRAPH96 Pro
ceeding, pp147-154.1996. This is to estimate and generate an operation in a connection section from time-series data of operations before and after connection, by making full use of inverse kinematics, inverse dynamics, optimization calculation and the like.

[0006]

However, since the former method can be applied only to a periodic operation, the range of use is limited. Furthermore, since Fourier expansion and inverse Fourier expansion are essential, there is a problem that the processing load is large and the calculation time is enormous.

In the latter method, connection operation data is generated by extrapolating and estimating the connection operation from the movements before and after.
This is basically a process that requires processing time. In addition, since it includes various numerical calculations, a longer calculation time is required.
Originally, the connection based on the operation predicted from the previous and next operations is a so-called defective setting problem. Therefore, an appropriate action is generated only when the objective function in the optimization calculation matches the condition of the connection at that time. However, since such connection conditions are determined individually for each operation before and after and for connection conditions, there are no conditions that hold in general situations, and connection operation data for unnatural movements often exist. Is generated.

In this extrapolation estimation method, only torque consumption is considered as an objective function, and optimization is performed to minimize the torque consumption. However, the condition of minimum torque consumption is a necessary condition, not a sufficient condition,
Not able to handle all situations. For example, when a smooth movement or a high-speed movement is required, it is not supported. That is, although the processing time is long, the connected operation often becomes unnatural.

In any of the above methods, the connection operation is obtained by estimating from the operations before and after the connection is made. If the estimation is not successful, the connection operation becomes inappropriate. Therefore, in the former method, in order to ensure the estimation, it is essential to assume the periodicity of the front-back operation. And the latter method often deviates its estimation. Thus, there is a root of the problem in generating a connection operation by estimation.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and generates natural connection operation data having a short calculation time without adding any condition such as periodicity to the operation before and after the connection. An operation data connection method and an operation data connection method are provided.

[0011]

Means for Solving the Problems First time-series motion data representing the motion of a subject having at least one or more joints, and subsequent second time-series motion data overlapping for a predetermined period are overlapped. Data connection method for connecting in a period period, the method includes a method of connecting the first time-series operation data of the first time-series operation data and the second time-series operation data in the overlap period. A first step of obtaining second overlapping section operation data; a second step of obtaining the position of the subject at the current time based on the first and second overlapping section operation data; A third step of obtaining a posture angle of the subject at the current time based on the second overlapping section operation data; and a step of obtaining the current time at the current time based on the first and second overlapping section operation data. The slide vector and joint angle of the subject's joint Of Mel fourth step, said second, wherein in that it has a third, and any one of steps and said first step of the fourth step.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Before describing the present invention in detail, referring to FIG. 4, an object such as a human or an organism modeled by a rigid articulated object in computer graphics animation will be described. The concept of the operation data will be described below. In the figure, a model of a rigid articulated object and a human are employed. As mentioned above, the operation data is
Generally, it is given as time-series data. This includes three-dimensional coordinate values representing the position of the entire object and three posture angles Aa and A used to control the posture of the entire object.
v, Ah, a slide vector representing the amount of parallel movement used for motion control of each joint, and a joint angle A representing the amount of rotation
j.

A point P for representing the entire position of the object O is uniquely fixed in advance for each object. In addition, a traveling direction vector Va, an upward vector Vv, and a lateral vector Vh, which represent the posture of the entire object O, are also defined in advance for each object, and these three vectors Va, Vv, and Vh are defined as posture vectors. Call.

By using a coordinate system introduced from the fixed point P and the posture vectors Va, Vv and Vh, the position of each joint of the object can be uniquely determined. Conversely, in computer graphics, an object has a coordinate system called an object coordinate system, which is unique to each object used to define a shape. Therefore, the origin of the object coordinate system represents the position of the entire object O. And a unit vector defining the object coordinate system is defined as a posture vector.

The attitude angles Aa, Av, and Ah are given as rotation amounts around three coordinate axes in the world coordinate system (x, y, z). By rotating the attitude vector in the initial state around each of the coordinate axes x, y, and z with the rotation amount, the attitude vectors Va, Vv, and Vh at that time can be calculated. Thus, the posture of the entire object at each time is controlled. The movement of the joint is controlled by a parallel movement based on a slide vector amount in a local coordinate system defined for each joint and a rotation about each coordinate axis of the local coordinate system. Although the description has been given of the rigid articulated object,
It is not limited to a rigid body or an articulated object, but the same applies to a soft body or a single joint, or a combination thereof.

Hereinafter, an operation data connection method according to an embodiment of the present invention will be described with reference to the drawings. FIG.
The processing flow of the operation data connection method in this embodiment is shown below. The entire process is executed in blocks # 1, # 3, # 5, # 7, # 9, and # 11. As is clear from the figure, block # 5, block # 7, and block # 9 are parallel processes, but not all three blocks need to be processed simultaneously.

The processing in each block configured as described above will be described in detail. If connection of two pieces of operation data can be performed, connection of an arbitrary number of pieces of operation data is possible. Therefore, the flow of processing in connection of two pieces of operation data will be described here.

Before that, two operational connections will be described with reference to FIG. In the figure, the previous operation data Da
A case is considered in which two time-series operation data, i.e., and post-operation data Dp are connected. The post-operation data Dp is operation data having a predetermined time length that starts at time t0 and ends after time t0. On the other hand, the previous operation data Da is operation data Da that starts before time t0 and ends at time te. That is, two operation data Da
And Dp, the period between times t0 and te that overlap is the connection period Sc for connecting the two data. The portion of the previous operation data Da within the connection period SC is referred to as front connection section operation data Dac, and the portion of the subsequent operation data Dp within the connection period SC is referred to as rear connection section operation data Dpc.

Returning to FIG. 1, first, at block # 1, these two front-back operation data Da and Dp and the connection time Sc
Is entered.

In block # 3, until the time t0 is reached, that is, until the connection period Sc is entered, necessary processing unrelated to the connection processing is performed on the previous operation data Da, and then the blocks # 5, # 7,. And # 9 are performed.

First, in block # 5, step S
In step 1, the speed of the previous operation in the connection section Sc is calculated by calculating the difference between the three-dimensional coordinate data representing the entire position P of the object O in the previous connection section operation data Dac. That is, the speed at the time ti is equal to the entire position P of the object O at the time ti + 1.
And the total position P of the object O at the time ti. Similarly, the speed of the post-operation is calculated from the post-connection section operation data Dpc. After calculating the speed, each speed is obtained from the speed. The speed is defined as the magnitude of the speed (vector Vs) as usual. Note that the speed of the previous operation data Da is represented as a three-dimensional vector Vsa,
The speed is represented as | Vsa |. Similarly, the speed of the post-operation data Dp is expressed as a three-dimensional vector Vsp, and the speed is expressed as | Vsp |.

In step S3, a combined speed Vss is calculated by combining the speed Vsa of the previous operation and the speed Vsp of the subsequent operation calculated in step S1 with the connection function Fc. As the connection function Fc, 1 at the initial time of the connection section Sc, 0 at the end of the connection section Sc, monotonically decreasing and differentiable, and the value at the middle point time of the connection section Sc is 180 degrees in the connection section Sc. When the function is a rotationally symmetric function, the composite speed Vss is calculated by the following equation (1).

[0023]

Vss = Vsa × Fc + Vsp × (1−Fc)

Equation 1 shows a proportional distribution equation based on a connection function value. Therefore, when the connection function Fc is nonlinear with respect to time, the distribution ratio also changes nonlinearly with respect to time.

The initial time of the connection section Sc is t 0, the end time of the connection section Sc is te, the current time is T, and t = (T−
When the normalization of (t0) / (te-t0) is performed, the following connection functions Fc having t as an independent variable are as follows.

In the case where the connection function Fc1 (t) is used, since linear synthesis is performed, only continuity within the connection section Sc can be guaranteed. Connection functions Fc2 (t), Fc3 (t) and Fc4
In the case of (t), the synthesis is nonlinear. However, the first-order continuation is guaranteed in Fc2 (t), and the second-order continuation is guaranteed in Fc (3).
In c (4), 4th floor continuation is guaranteed.

That is, regardless of whether the functions of the pre-operation and post-operation are similar or not, in the case of Fc2 (t), the continuity of the speed (corresponding to acceleration) of the composite speed Vss in the connection section Sc is guaranteed. , Fc3 (t), the continuity of the acceleration (corresponding to jerk) of the combined speed Vss is guaranteed, and Fc4 (t) guarantees the continuity of the derivative of the jerk of the combined speed Vss.

As the connection function Fc, the connection section Sc
0 at the initial time, 1 at the end of the connection section Sc, a function that is monotonically increasing and differentiable, and whose value in the connection section Sc is 180 degrees rotationally symmetric with respect to the value at the middle point time of the connection section Sc. In this case, the synthesis speed Vss is calculated by the following equation (2).
Is calculated.

[0029]

Vss = Vsa × (1−Fc) + Vsp × Fc

When the above-mentioned normalization is performed, the following connection functions are available. These connection functions Fc1 '(t), Fc'2 (t), Fc'3 (t) and Fc'4
The properties of (t) are the functions Fc2 (t) and Fc
3 (t) and Fc4 (t).

The speed of the pre-operation | Vsa | and the speed of the post-operation | V
For sp |, the same synthesis as in the calculation of the synthesis speed Vss is performed to calculate the synthesis speed Vsc. However, the connection function Fc does not necessarily need to be the same as in the case of calculating the composite speed Vss. In this embodiment, the connection function F is appropriately determined in advance in consideration of the smoothness of the connection and the calculation time.
Although c is selected, the following automatic selection is also possible. With reference to the flowchart shown in FIG. 7, calculation of the composite speed Vss will be described below as an example in the case of automatically selecting the connection function Fc.

First, in step S110, the connection section Sc
Then, the difference amount (corresponding to acceleration) dA (t) between the speed Vsa of the previous operation and the difference amount dB (t) between the speed Vsp of the subsequent operation is calculated. Next, in step S102, | dA (t) -dB (t) |
A value S is calculated by taking the sum of the raised values over the connection section. In step S104, if S is equal to or smaller than α based on thresholds α, β, and γ determined in advance by a previous experiment or the like (step S104), the connection function Fc1 (t) is automatically selected (step S106). If S is larger than α and equal to or smaller than β (step S108), the connection function Fc2 (t) is automatically selected (S110). If S is greater than β and equal to or less than γ (S112), the connection function Fc3 (t) is automatically selected (S1).
14). If S is larger than γ (S112), the connection function Fc4 (t) is automatically selected (step S116).

In this automatic selection process, the value S indicates the degree of similarity of the functions, that is, the closeness of the differential coefficients. However, the smaller the value, the greater the degree of similarity. According to the above-described processing, when the operation of the previous connection section and the operation of the rear connection section are similar, the connection function Fc1 having a short calculation time is used.
Is used. Therefore, as the degree of similarity decreases, the calculation time increases, but a connection function Fc that enables smoother connection can be selected. As described above, according to the present invention, automatic selection of the connection function Fc for optimizing the calculation time and the processing result can be realized. Also,
When processing is performed based on equation (2), Fc1, Fc
Needless to say, connection functions Fc1 ′, Fc′2, Fc′3 and Fc′4 may be selected instead of Fc3 and Fc4. In step S5, the position P (ti) at the current time ti is calculated by the following equation (3).

[0034]

[Equation 3] P (ti) = P (ti-1) + Vss (ti-1) × | V
sc (ti-1) | ÷ | Vss (ti-1) |

Note that ti-1 is the previous time and Vss (ti-1)
Is the combined speed at the previous time ti-1, | Vsc (ti-1) | is the combined speed at the previous time ti-1, and | Vss (ti-1) | is the magnitude of the combined speed at the previous time ti-1. It is. In this way, after the process of block # 5 is completed, the process proceeds to block # 11.

On the other hand, in block # 7, in step S7, the posture angles Aaaa,
From Ava and Aha, the front posture vectors Vaa, Vv
a and Vha are calculated, and from the posture angles Aaac, Avac, and Ahac of the connection section operation data Dac and the posture angles Aapc, Avpc, and Ahpc of the connection section operation data Dpc, the rear posture vectors Vap, Vvp, and V.
hp is calculated.

In step 9, the front posture vector Vaa,
Vva and Vha and the rear posture vectors Vap, Vvp and Vhp are synthesized in the same manner as in the case of the calculation of the synthesized speed Vss, and the synthesized posture vectors VaS, VvS and V
Calculate hS. However, the connection function is not necessarily the composite speed V
It need not be the same as in the case of calculating ss. An appropriate selection is made in consideration of smoothness of connection, calculation time, and the like.

In step S11, the combined posture vector V
The combined attitude angle AvS is calculated from aS, VvS, and VhS. Since the correspondence between the combined posture vectors VaS, VvS and VhS and the combined posture angle AvS is generally one-to-many, it cannot be uniquely determined without any conditions. But,
A unique decision can be made by adding a condition for the rotation angles of the three direction vectors constituting the posture vectors Va, Vv and Vh. For example, the rotation angle Avv of the upward vector Vv is solved by adding a condition such as −90 ° or more and 90 ° or less. Such a restriction does not affect the calculation of the combined posture vectors VaS, VvS and VhS from the combined posture angle AvS. After finishing the process of block # 7 in this way, the process proceeds to block # 11.

Further, in block # 9, step S13
Then, the front slide direction DsVac and the front slide amount | Svac | are calculated from the joint slide vector Svac of the front connection section motion data Dac, and the rear connection section motion data D is calculated.
The rear slide direction DsVpc and the rear slide amount | Svpc | are calculated from the pc slide vector Svpc. However, the slide direction DsV is the slide vector Sv itself or a vector normalized to the size 1, and the slide amount | Sv | is the size of the slide vector Sv.

In step S15, the front slide direction Ds
Vac and the rear slide direction DsVpc, and the front slide amount | Svac | and the rear slide amount | Svpc | are synthesized in the same manner as in the case of the above-described calculation of the synthesis speed Vss, and the synthesized slide direction DsVS and the synthesized slide amount SVc are calculated. calculate. However, the connection function Fc does not necessarily need to be the same as in the case of calculating the composite speed Vss. An appropriate selection is made in consideration of smoothness of connection, calculation time, and the like. Next, a combined slide vector SvS is calculated by the following equation (4) from the combined slide direction DsvS and the combined slide amount SVc.

[0041]

SvS = DsvS × SVc ÷ | DsVS || D
sVS | is the size of the composite slide direction DsvS.

In step S17, the joint angle Ajac of the front connection section motion data Dac and the rear connection section motion data Dp
The joint angle AjS of c is calculated by synthesizing the joint angle Ajpc in the same manner as in the case of calculating the synthetic speed Vss described above. However, the connection function Fc does not necessarily need to be the same as in the case of calculating the composite speed Vss. An appropriate selection is made in consideration of smoothness of connection, calculation time, and the like. Automatic selection is also possible in the same manner as described above. Thus, after the processing of block # 9 is completed, the process proceeds to block # 11.

After the processing of any of blocks # 5, # 7, and # 9, in block # 11, for the post-motion data Dp, the global global of the entire object O required for the motion data after the connection section Sc A conversion process is performed to make the movement correspond to the connection operation. As can be seen from the above description, in the present invention, the combination of the data of the preceding connection section operation and the data of the subsequent connection section operation is equivalent to generation of the connection operation. In FIG.
A connection state by the connection function Fc described above is schematically shown. In the connection period Sc, the end of the pre-operation data Da (time t0) and the start of the post-operation data Dp (time te) are determined by the connection function F selected in the above processing.
The connection is smooth by c. Hereinafter, a modified example of the above-described operation data connection method will be described with reference to the drawings. The processing flow of the operation data connection method in the present example shown in FIG. 2 is basically the same as the processing flow shown in FIG. However, step S2 is added to block # 5 in FIG. 1 to change to block # 5R, and step S8 is added to block # 7 to change to block # 7R. Therefore, only the differences from the method shown in FIG. 1 will be described below.

More specifically, in the present embodiment, the conversion amount of the motion direction of the pre-motion data Da given at each time is defined as the pre-motion conversion amount MAa, and the conversion amount of the motion direction of the post-motion data Dp is defined as the post-motion data. The difference from the method shown in FIG. 1 is that the conversion amount MAp is used.

In block # 5R, step S2
Then, the speed Vsa of the previous operation calculated in step S1 is converted into the pre-motion conversion amount MAa at the time. Since this transformation can be given by a rotation matrix, it is performed by the product of the rotation matrix and the velocity. Similarly, the post-motion speed Vsp is changed to the post-motion conversion amount MAp at the time.
Convert to

In step S4, a combined speed MVsS is calculated by combining the speed MAa of the pre-operation and the speed MAp of the post-operation converted in step S2 by a connection function. This calculation is performed using the same function as the connection function Fc used for calculating the composite speed shown in FIG. For the speed of the pre-operation and the speed of the post-operation, the same synthesis as in the calculation of the synthesis speed is performed, and the synthesis speed MVsc is calculated. However,
The connection function does not necessarily have to be the same as in the case of calculating the composite speed. An appropriate selection is made in consideration of smoothness of connection, calculation time, and the like.

In block # 7R, step S8
Then, the front posture vectors Vaa, Vva, and Vha are changed to the front movement posture vectors MVaa, MVva, M at the time.
Convert to Vha. Similarly, the rear posture vectors Vap, V
vp and Vhp are the post-motion vectors MV at the time
ap, MVvp, MVhp. This conversion can be performed by the product of the rotation matrix and the attitude vector, as in step S2.

In step S10, the motion / posture vectors MVaa, MVva, MVh obtained in step S8 are obtained.
a, MVap, MVvp, and MVhp are synthesized in the same manner as in the case of the above-described calculation of the synthesized speed, and the synthesized motion posture vectors MVaS, MVvS, and MVhS are calculated. In other words, with such a configuration, as an example, when the motion directions of the motion data before and after the connection are different from each other, the direction of the connection data is adjusted, and a smooth motion connection is enabled. . However, the connection function does not necessarily have to be the same as in the case of calculating the composite speed. An appropriate selection is made in consideration of smoothness of connection, calculation time, and the like.

Hereinafter, an apparatus for realizing the above-described operation data connection method according to the present invention will be described with reference to FIG. The operation data connection device 100 includes an input / output unit 1 connected to the bus 3 for exchanging data with each other, a storage unit 2, a pre-operation data connection time interval processor 4, a speed and speed calculator 5, a synthesized speed and a synthesized speed. Calculator 6, current time position calculator 7, attitude vector calculator 8, combined attitude vector calculator 9, combined attitude angle calculator 10, slide direction and slide amount calculator 1
1. Synthetic slide direction amount vector calculator 12, synthesized joint calculators 13, post-operation data processor 14 after connection section
Having.

The input / output unit 1 connects the pre-operation data Da and the post-operation data Dp input from the outside with the connection times t0 and te.
Is stored in the storage device 2 via the bus 3. The pre-operation data connection time inter-processor 4 compares the pre-operation data Da stored in the storage unit 2 via the bus 3 with the pre-operation data Da up to the connection section Sc and has no relation to the connection processing. After performing any necessary processing, it is stored in the storage device 2 via the bus 3.

The speed and speed calculator 5 calculates the previous connection section operation data Dac in the connection section stored in the storage 2.
Of the object O and the subsequent connection section motion data Dp
The position P of the entire object c is read out via the bus 3. Next, the speed Vsa of the previous operation is changed to the previous connection section operation data Dac.
Is calculated by calculating the difference between the three-dimensional coordinate data representing the position P of the entire object O.

That is, the speed Vsa at the time ti is
It is calculated from the difference between the position of the entire object at time ti + 1 and the position P of the entire object at time ti. This velocity is a three-dimensional vector. Similarly, the speed Vsp of the post-operation is calculated from the post-connection section operation data Dpc. After calculating the speed, each speed is obtained from the speed. The speed is normally defined as the magnitude of the speed (vector). These calculated speeds and speeds are stored in the storage device 2 via the bus 3.

The composite speed and composite speed calculator 6 is composed of a memory 2
Speed Vsa of the previous operation and speed Vsp of the subsequent operation stored in
Is read out via the bus 3. Next, step 3 shown in FIG.
Synthesized speed Vss synthesized by the connection function Fc described in
Is calculated. This is in accordance with step S3 in FIG. 1 of the present invention when the connection function Fc is automatically selected. Regarding the speed Vsa of the pre-operation and the speed Vsp of the post-operation, the combined speed Vs
s is calculated, and the synthesis speed V is calculated.
Calculate sc. However, the connection function Fc does not necessarily need to be the same as in the case of calculating the composite speed. The calculated composite speed Vss and composite speed Vsc are stored in the storage device 2 via the bus 3.

The current time position calculator 7 calculates the whole object O in the previous connection section operation data Dac stored in the storage 2.
The position P of the entire object, the composite speed Vss, and the composite speed Vsc of the post-connection section operation data Dpc are read out via the bus 3, and the position P of the entire object in the connection operation is calculated by the above-described third formula Can be calculated. The calculated position P of the whole object of the connection operation is stored in the storage device 2 via the bus 3.

The posture vector calculator 8 calculates the posture angle Aaaa of the previous connection section operation data Dac stored in the storage device 2,
Ava and Aha are read out via the bus 3. Next, the posture angles Aaaa, A of the previous connection section operation data Dac
va and Aha, the front posture vectors Vsa, Vv
a and Vha are calculated, and the subsequent connection section operation data Dpc is calculated.
Are calculated from the posture angles of (1) and (2), and the calculated front and rear posture vectors are stored in the storage unit 2 via the bus 3.

The combined posture vector calculator 9 calculates the front posture vector and the rear posture vector Vsa, Vsa stored in the storage unit 2.
va, Vha, Vsp, Vvp, and Vhp are read out via the bus 3, and the same synthesis as in the case of the above-described calculation of the synthesis speed is performed to calculate the synthesized posture vectors VaS, VvS, and VhS. However, the connection function does not necessarily have to be the same as in the case of calculating the composite speed. The calculated combined posture vector is stored in the storage device 2 via the bus 3.

The combined posture angle calculator 10 reads out the combined posture vectors VaS, VvS and VhS stored in the storage unit 2 via the bus 3, and calculates the combined posture angle AvS from the combined posture vectors according to the method of step S11. Calculate similarly. The calculated combined attitude angle is stored in the storage device 2 via the bus 3.

The slide direction and slide amount calculator 11 includes:
The slide vector Sva of the joint of the previous connection section operation data and the slide vector Svp of the rear connection section operation data stored in the storage device 2 are read out via the bus 3. next,
From the slide vector SvP of the joint of the previous connection section motion data Dac, the front slide direction DsVa and the front slide amount |
DsVa | is calculated, and the rear slide direction DsVp and the rear slide amount | DsVp | are calculated from the slide vector Svp of the rear connection section operation data Dpc. The calculated slide direction DsV and slide amount | DsV | are stored in the storage device 2 via the bus 3.

Synthetic slide direction amount vector calculator 12
The bus 3 stores the front slide direction and the rear slide direction and the front slide amount and the rear slide amount stored in the storage unit 2 in the bus 3.
And the same synthesis as in the case of the above-described calculation of the synthesis speed is performed to calculate the synthesized slide direction DsVS and the synthesized slide amount SVc. Next, the synthetic slide direction D
A composite slide vector SvS is calculated from svS and the composite slide amount | SvS | based on the fourth equation. The calculated composite slide vector SvS is stored in the storage device 2 via the bus 3.

The composite joint angle calculator 13 calculates the joint angle Aja of the front connection section operation data Dac and the joint angle Ajp of the rear connection section operation data Dpc stored in the storage device 2 using the bus 3.
, And performs the same synthesis as in the case of the above-described calculation of the synthesis speed to calculate the synthesized joint angle AjS. The calculated combined joint angle AjS is stored in the storage device 2 via the bus 3. The post-operation data processor 14 after the connection section reads the post-operation data after the connection section stored in the storage device 2 via the bus 3 and performs necessary conversion processing on the operation data after the connection section, The data is stored in the storage device 2 via the bus 3.

Finally, the pre- and post-connection operation data Dc connecting the pre-operation data and the post-operation data stored in the storage unit 2 by the input / output unit 1 is read out via the bus 3 and output to the outside. Each processing in each stage in the embodiment shown in FIGS. 1 and 2 can be performed in units of operation time. However, in this case, each process is executed in a loop for each operation time.

Further, referring to FIG. 8, the automatic selection of the connection function Fc when performing the real-time processing will be described below. First, a calculation processing performance evaluation value PA of a computer or a processing processor (MPU) that will be used is obtained in advance, and a table that can be searched for by an identifier PI such as a computer or an MPU name is created. The required calculation processing evaluation value PA is calculated for each connection function Fc, and for example, "a" for connection function Fc, "b" for Fc2, "c" for Fc3, and "c" for Fc4. If “d” is set, there is a relationship of a <b <c <d.

Before entering the continuation processing, the above-described identifier PI is input and a search is performed in a table (S200). If the calculation processing performance evaluation value PA of the search result is d or more (S20)
2) Automatically select a connection function Fc4 (S204). When the calculation processing performance evaluation value PA is equal to or larger than c and smaller than d (S206), the connection function Fc3 is automatically selected (S20).
8) If the value is equal to or larger than b and smaller than c (S210), the connection function Fc2 is automatically selected; otherwise, the connection function Fc2 is selected (S21).
0) automatically selects the connection function Fc1 (S214). With this configuration, the best connection function Fc suitable for the processing capacity of the device to be used can be automatically selected.

[0064]

As described in the above embodiments, in the operation data connection method and apparatus according to the present invention, the combination of the data of the previous connection section operation and the data of the subsequent connection section operation is equivalent to the generation of the connection operation. That is, the connection operation is determined from the existing operation, instead of estimating the operation at the non-existing time as in the conventional method. Therefore, for the operation before and after the connection according to the present invention, a natural connection operation can be generated with a short calculation time without adding any condition such as periodicity. When the operation data connection processing according to the present invention is applied in units of time, computer graphics / animation can be performed while connecting in real time by combining with the drawing processing. When combined with autonomous control system processing, living things such as artificial life can be realized. It is a technology that is the basis of technologies that will be important in the future, such as games and agents, and the significance of the present invention is great.

[Brief description of the drawings]

FIG. 1 is a processing configuration diagram showing an operation data connection method in an embodiment of the present invention.

FIG. 2 is a processing configuration diagram showing another example of the operation data connection method shown in FIG.

FIG. 3 is a configuration diagram of an operation data connection device according to the present invention.

FIG. 4 is an explanatory diagram of a posture vector.

FIG. 5 is an explanatory diagram of two operation data for making a connection.

FIG. 6 is an explanatory diagram of operation data synthesis using a connection function.

FIG. 7 is a flowchart illustrating a synthetic speed calculation operation in a case where the connection function Fc is automatically selected according to the present invention.

FIG. 8 is a flowchart showing an automatic selection operation of a connection function Fc when performing real-time processing in the present invention.

Claims (27)

[Claims] 1. First time-series motion data (Da) representing a motion of a subject (O) having at least one or more joints.
And a subsequent second time-series operation data (Da) that overlaps for a predetermined period (Sc) in the overlap period (Sc). In the overlapping period (Sc), the first overlapping section operation data (Dac) of the first time series operation data (Da) and the second overlapping section operation data (Dpc) of the second time series operation data (Dp) ), And the first and second overlapping section operation data (Dac, Dp)
c) obtaining a position (Pti) of the subject at the current time (ti) based on c), and the first and second overlapping section operation data (Dac, Dp)
Third step (# 7) of obtaining the posture angle (AvS) of the subject (O) at the current time (ti) based on c).
And the first and second overlapping section operation data (Dac, Dp
c) of the second, third, and fourth steps of the fourth step for calculating the slide vector and the joint angle of the joint of the subject (O) at the current time (ti) at the current time. A method comprising any one of the steps and the first step. 2. The operation data connection method according to claim 1, wherein said second step (# 5) comprises:
Further, based on the first overlapping section operation data (Dac),
A step (S1) of obtaining a first speed (Vsa) and a first speed (| Vsa |) of the subject (O) in the overlapping section (Sc); and the second overlapping section operation data (Dpc). )On the basis of,
Determining a second speed (Vsp) and a second speed (| Vsp |) of the subject (O) during the overlap period (Sc) (S1); Vsa, Vsp) based on a predetermined connection function (Fc), and the overlapping section (Sc)
(S3) determining a third speed (Vss) of the subject (0) in the step (b), and converting the first and second speeds (| Vsa |, | Vsp |) into the connection function (Fc). (S3) obtaining a third speed (Vsc) of the subject (0) in the overlapping section (Sc) based on the third speed (Vss) and the third speed (Vss). Determining a position (P (ti)) of the subject (O) at the current time (ti) based on Vsc). 3. The operation data connection method according to claim 1, wherein said third step (# 7) comprises:
Further, based on the first overlapping section operation data (Dac),
A step (S7) of obtaining a first posture vector (Vaa, Vva, Vha) of the subject (O) in the overlapping section (Sc), and based on the second overlapping section operation data (Dpc),
Obtaining a second attitude vector (Vap, Vvp, Vhp) of the subject (O) in the overlapping section (Sc) (S7); and determining the first attitude vector (Vaa, Vva, Vha). The second posture vector (Vap, Vvp, Vhp) is synthesized based on a predetermined connection function (Fc), and the third posture vector (VaS, VaS, V) of the subject (O) in the overlapping section (Sc). VvS, VhS) (S9)
Calculating a third posture vector (AvS) representing the posture of the subject (O) in the overlapping section (Sc) using the third posture vector (VaS, VvS, VhS) (S1).
A method characterized by having 1). 4. The operation data connection method according to claim 1, wherein said fourth step (# 9) comprises:
Further, based on the first overlapping section operation data (Dac),
Calculating a first slide direction (DSva) and a first slide amount (| Sva |) of the joint of the subject (O) in the overlap section (Sc) (S13); Based on the section operation data (Dap),
Calculating a second slide direction (DSvp) and a second slide amount (| Svp |) of the joint of the subject (O) in the overlap section (Sc) (S13); Direction (DSva), first slide amount (| Sva |), second slide direction (DSv
p) and the second slide amount (| Svp |) are synthesized based on a predetermined connection function (Fc) to obtain the overlap section (Sc).
The direction of the third slide of the subject (O) (DS)
vS), a third slide amount (SVc), and a slide vector (SvS) are obtained (S15). The first joint angle of the first overlap section motion data (Dac) and the second overlap section The second joint angle of the motion data (Dpc) is synthesized based on the connection function (Fc), and the slide vector (Sv) of the joint of the subject (O) and the third joint angle in the overlap section (Sc) are calculated. A step (S17) of obtaining a joint angle (AjS) of the step (c). 5. Two time-series motion data representing the entire position and posture angle of the multi-joint object at each time, and the slide vector and joint angle of the joint are defined as pre-motion data and post-motion data. When the operation data for the connection time from the end of the operation data is the previous connection section operation data, and the operation data for the connection time from the beginning of the post operation data is the rear connection section operation data, The speed and speed are calculated from the entire position, the speed and speed are calculated from the entire position of the rear connection section operation data, and the speed and the speed calculated from the previous connection section operation data and the rear connection section data are calculated. A combined speed is calculated by combining the speeds with a speed connection function, and the speed calculated from the preceding connection section operation data and the speed calculated from the rear connection section data are combined by a speed connection function. The resultant speed is calculated, and the traveling speed of the direction and the speed is added to the entire position of the rigid articulated object at the time immediately before the current time, thereby obtaining the current speed. A first step of calculating the entire position of the rigid-to-articulated object; calculating a posture vector from the entire posture angle of the front connection section motion data; and calculating a posture vector from the whole posture angle of the rear connection section motion data. Calculating a combined posture vector by combining a posture vector calculated from the preceding connection section motion data and a posture vector calculated from the rear connection section data by a posture vector connection function, and converting the combined posture vector into a combined posture angle. By doing so, the second step of calculating the overall posture angle of the rigid-to-articulated object at the current time, and the slide direction and slide from the joint slide vector of the previous connection section motion data. Calculate the amount, calculate the slide direction and the slide amount from the slide vector of the joint of the rear connection section motion data, and calculate the slide direction and the slide amount of the joint of the previous connection section motion data, and the rear connection section motion data. Calculating the combined slide direction and the combined slide amount by combining the slide direction and the slide amount of the joint by the slide direction connection function and the slide amount connection function, and calculating a combined slide vector from the combined slide direction and the combined slide amount; By calculating a combined joint angle obtained by combining the joint angle of the previous connection section motion data and the joint angle of the rear connection section data by a joint angle connection function, the slide vector and the joint of the joint of the rigid versus multi-joint object at the current time are calculated And a third step of calculating an angle. 6. Two time-series motion data representing the overall position and posture angle of the articulated object at each time, and a slide vector and a joint angle of a joint are defined as pre-motion data and post-motion data. The conversion amount of the motion direction of the previous motion data given for each time is the pre-motion conversion amount, the conversion amount of the motion direction of the post-motion data is the post-motion conversion amount, and the motion data for the connection time from the end of the previous motion data. As the previous connection section operation data, when the operation data for the connection time from the beginning of the post operation data as the rear connection section operation data, calculate the speed and speed from the entire position of the previous connection section operation data, The speed and speed are calculated from the entire position of the post-connection section operation data, and the speed calculated from the pre-connection section operation data by the pre-motion conversion amount at the time and the post-connection section data. A calculated speed is calculated by the connection function after converting the calculated speed by the post-motion conversion amount at the time, and a synthesized speed is calculated, and a speed calculated from the previous connection section operation data and a speed calculated from the rear connection section data. Is calculated by the connection function, and the progression speed of the synthesized speed and the synthesized speed is added to the entire position of the rigid articulated object at the time immediately before the current time. Thus, a first step of calculating the entire position of the rigid-joint articulated object at the current time, and calculating a posture vector from the whole posture angle of the previous connection section motion data, and calculating the whole of the rear connection section motion data A posture vector is calculated from the posture angle, a posture vector calculated from the previous connection section motion data by the pre-motion conversion amount at the time, and a posture vector calculated from the rear connection section data. By calculating the combined posture vector by combining the posture vector obtained by converting the torque with the post-motion conversion amount at the time by the connection function, and converting the combined posture vector into the combined posture angle, the rigid-to-multipoint at the current time is calculated. A second step of calculating the overall posture angle of the joint object; calculating a slide direction and a slide amount from the slide vector of the joint in the previous connection section motion data; and performing a slide from the slide vector of the joint in the rear connection section motion data. And the slide amount of the joint of the previous connection section motion data and the joint slide direction and the slide quantity of the rear connection section motion data are synthesized by the connection function. Calculating a direction and a combined slide amount, calculating a combined slide vector from the combined slide direction and the combined slide amount, By calculating a composite joint angle obtained by combining the joint angle of the previous connection section motion data and the joint angle of the rear connection section data by the connection function, the slide vector and the joint angle of the joint of the rigid versus multi-joint object at the current time are calculated. And a third step of calculating the operation data. 7. Two time-series motion data representing the entire position and posture angle of the articulated object at each time, and the slide vector and joint angle of the joint are defined as pre-motion data and post-motion data. The conversion amount of the motion direction of the previous motion data given for each time is the pre-motion conversion amount, the conversion amount of the motion direction of the post-motion data is the post-motion conversion amount, and the motion data for the connection time from the end of the previous motion data. As the previous connection section operation data, when the operation data for the connection time from the beginning of the post operation data as the rear connection section operation data, calculate the speed and speed from the entire position of the previous connection section operation data, The speed and speed are calculated from the entire position of the post-connection section operation data, and the speed calculated from the pre-connection section operation data by the pre-motion conversion amount at the time and the post-connection section data. A speed obtained by converting the calculated speed by the post-motion conversion amount at the time is calculated by a speed connection function to calculate a synthesized speed, and the speed calculated from the previous connection section operation data and the speed calculated from the rear connection section data. The combined speed is calculated by the speed connection function, and the direction of the combined position of the rigid-articulated object at the time immediately before the current time is the combined speed, and the size is the progress speed of the combined speed. A first step of calculating the overall position of the rigid-joint articulated object at the current time by adding, and calculating a posture vector from the whole posture angle of the previous connection section motion data, A posture vector is calculated from the entire posture angle, and a posture vector calculated from the previous connection section motion data by a pre-conversion amount at the time and a posture vector calculated from the rear connection section data. By calculating the combined posture vector by combining the posture vector obtained by converting the torque by the post-conversion amount at the time with the posture vector connection function, and converting the combined posture vector into the combined posture angle, the rigid-to-multipoint at the current time is calculated. A second step of calculating the overall posture angle of the joint object; calculating a slide direction and a slide amount from the slide vector of the joint in the previous connection section motion data; and performing a slide from the slide vector of the joint in the rear connection section motion data. Of the joint and the slide amount of the joint in the preceding connection section operation data, and the joint direction and the slide amount of the joint in the rear connection section operation data. The direction of the synthesized slide and the amount of the synthesized slide synthesized by the function are calculated. By calculating the composite slide vector from the joint angle of the previous connection section motion data and the joint angle of the rear connection section data by the connection function, the rigid joint articulated object at the current time is calculated. And a third step of calculating a joint angle and a joint angle of the joint. 8. A value of 1 at the initial time of the connecting section, 0 at the end of the connecting section, monotonically decreasing and differentiable, and a value at the middle point time of the connecting section has a rotational symmetry of 180 degrees in the connecting section. 7. The operation data connection method according to claim 1, wherein such a function is used as a connection function. 9. The initial value of the connection section is 1; the end of the connection section is 0; the value is monotonically decreasing and differentiable; and the value in the connection section is 180 degrees rotationally symmetric with respect to the value at the middle point time of the connection section. 8. The motion data connection method according to claim 5, wherein such functions are used as a speed connection function, a speed connection function, a posture vector connection function, a slide direction connection function, and a slide amount connection function. 10. The operation data connection according to claim 1, wherein (1 + cos (πt)) / 2 is used as a connection function when the connection section is normalized to the section [0, 1]. Method. 11. When a connection section is normalized to a section [0,1], ((6t + 3) t + 1) (1-t) (1-t) (1-t) is used as a connection function. 7. The operation data connection method according to claim 1, wherein: 12. When the connection section is normalized to the section [0,1], ((((70t + 35) t + 15) t + 5) t + 1) (1-t) (1-t) (1-t) (1-t) (1
-t) is used as a connection function.
And the operation data connection method described in 3 + 3. 13. When the connection section is normalized to the section [0,1], (1 + cos (πt)) / 2 is a speed connection function, a speed connection function, a posture vector connection function, a slide direction connection function, 8. A method according to claim 5, wherein the function is used as a slide amount connection function.
The operation data connection method described. 14. When the connection section is normalized to the section [0,1], ((6t + 3) t + 1) (1-t) (1-t) (1-t) is a velocity connection function, 8. The motion data connection method according to claim 5, wherein the motion data connection method is used as a speed connection function, a posture vector connection function, a slide direction connection function, or a slide amount connection function. 15. When the connection section is normalized to the section [0,1], ((((70t + 35) t + 15) t + 5) t + 1) (1-t) (1-t) (1-t) (1-t) (1
8. The motion data connection method according to claim 5, wherein -t) is used as a speed connection function, a speed connection function, a posture vector connection function, a slide direction connection function, and a slide amount connection function. 16. The initial value of the connection section is 0, the end of the connection section is 1, the monotonically increasing value is differentiable, and the value at the middle point time of the connection section has a rotational symmetry of 180 degrees in the connection section. 7. The operation data connection method according to claim 1, wherein such a function is used as a connection function. 17. The initial value of the connection section is 0, the end of the connection section is 1, the differentiation is monotonically increasing, and the value in the connection section is 180 degrees rotationally symmetric with respect to the value at the middle point time of the connection section. 8. The motion data connection method according to claim 5, wherein such functions are used as a speed connection function, a speed connection function, a posture vector connection function, a slide direction connection function, and a slide amount connection function. 18. The operation data connection according to claim 1, wherein (1-cos (πt)) / 2 is used as a connection function when the connection section is normalized to the section [0, 1]. Method. 19. When the connection section is normalized to the section [0,1], 1-((6t + 3) t + 1) (1-t) (1-t) (1-t) is used as a connection function. 7. The operation data connection method according to claim 1, wherein the operation data connection method is used. 20. When the connection section is normalized to the section [0,1], 1-((((70t + 35) t + 15) t + 5) t + 1) (1-t) (1- t) (1-t) (1-t)
7. The operation data connection method according to claim 1, wherein (1-t) is used as a connection function. 21. When the connection section is normalized to the section [0,1], (1-cos (πt)) / 2 is a speed connection function, a speed connection function, a posture vector connection function, a slide direction connection function, 8. A method according to claim 5, wherein the function is used as a slide amount connection function.
The operation data connection method described. 22. When the connection section is normalized to the section [0, 1], 1-((6t + 3) t + 1) (1-t) (1-t) (1-t) is speed-connected. 8. The motion data connection method according to claim 5, wherein the motion data connection method is used as a function, a speed connection function, a posture vector connection function, a slide direction connection function, and a slide amount connection function. 23. When the connection section is normalized to the section [0,1], 1-(((((70t + 35) t + 15) t + 5) t + 1) (1-t) (1- t) (1-t) (1-t)
8. The motion data connection method according to claim 5, wherein (1-t) is used as a speed connection function, a speed connection function, a posture vector connection function, a slide direction connection function, and a slide amount connection function. 24. First time-series motion data (Da) representing the motion of a subject (O) having at least one joint.
And the subsequent second time-series operation data (Da) that overlaps for a predetermined period (Sc) in the overlap period (Sc). In the overlapping period (Sc), the first overlapping section operation data (Dac) of the first time series operation data (Da) and the second overlapping section operation data (Dpc) of the second time series operation data (Dp) ), And the first and second overlapping section operation data (Dac, Dp)
c) determining a position (Pti) of the subject at the current time (ti) based on the first and second overlapping section operation data (Dac, Dp).
c) determining the posture angle (AjS) of the subject (O) at the current time (ti) based on c), and the first and second overlapping section motion data (Dac, Dp
and c) determining a slide vector and a joint angle of the joint of the subject (O) at the current time (ti) at the current time. An apparatus comprising: one means and the first means. 25. The operation data connection device according to claim 25, wherein said second means (# 5) comprises:
Further, based on the first overlapping section operation data (Dac),
Means (S1) for determining a first speed (Vsa) and a first speed (| Vsa |) of the subject (O) in the overlapping section (Sc); and the second overlapping section operation data (Dpc )On the basis of,
Means (S1) for determining a second speed (Vsp) and a second speed (| Vsp |) of the subject (O) during the overlap period (Sc); and the first and second speeds (| Vsa, Vsp) based on a predetermined connection function (Fc), and the overlapping section (Sc)
Means (S3) for determining a third speed (Vss) of the subject (0) in the step (a), and connecting the first and second speeds (| Vsa |, | Vsp |) to the connection function (Fc). Means (S3) for obtaining a third speed (Vsc) of the subject (0) in the overlapping section (Sc) based on the third speed (Vss) and the third speed (Vss). A device (S5) for determining the position (P (ti)) of the subject (O) at the current time (ti) based on Vsc). 26. The operation data connection device according to claim 25, wherein the third means (# 7) comprises:
Further, based on the first overlapping section operation data (Dac),
A means (S7) for obtaining a first posture vector (Vaa, Vva, Vha) of the subject (O) in the overlapping section (Sc); and a second overlapping section motion data (Dpc).
Means (S7) for obtaining a second posture vector (Vap, Vvp, Vhp) of the subject (O) in the overlapping section (Sc); a first posture vector (Vaa, Vva, Vha); The second posture vector (Vap, Vvp, Vhp) is synthesized based on a predetermined connection function (Fc), and the third posture vector (VaS, VaS, V) of the subject (O) in the overlapping section (Sc). VvS, VhS) and a third posture vector (VaS, VvS, VhS) representing the posture of the subject (O) in the overlapping section (Sc). Means for determining AvS) (S11)
An apparatus comprising: 27. The operation data connection device according to claim 25, wherein said fourth means (# 9) comprises:
Further, based on the first overlapping section operation data (Dac),
Means (S13) for determining a first slide direction (DSva) and a first slide amount (| Sva |) of the joint of the subject (O) in the overlap section (Sc); Based on the section operation data (Dap),
Means (S13) for determining a second slide direction (DSvp) and a second slide amount (| Svp |) of the joint of the subject (O) in the overlap section (Sc); and the first slide. Direction (DSva), first slide amount (| Sva |), second slide direction (DSv
p) and the second slide amount (| Svp |) are synthesized based on a predetermined connection function (Fc) to obtain the overlap section (Sc).
The direction of the third slide of the subject (O) (DS)
vS), means (S15) for calculating a third slide amount (SVc), and a slide vector (SvS); a first joint angle of the first overlap section motion data (Dac) and the second overlap section The second joint angle of the motion data (Dpc) is synthesized based on the connection function (Fc), and the slide vector (Sv) of the joint of the subject (O) and the third joint angle in the overlap section (Sc) are calculated. A means (S17) for determining a joint angle (AjS).