JP3148830B2 - Walking control device for legged mobile robot - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a walking control device for a legged mobile robot, and more particularly, to a walking control device capable of walking stably even when there is unexpected unevenness.

[0002]

2. Description of the Related Art As a legged mobile robot, for example, a bipedal walking robot disclosed in Japanese Patent Application Laid-Open No. 62-97006 is known.

[0003]

A legged mobile robot is
Since the supporting polygon changes greatly as compared with other types of robots such as wheel-type and crawler-type robots, the posture is likely to be unstable, particularly in the case of a bipedal walking robot.

Accordingly, it is a first object of the present invention to provide a legged mobile robot that always walks in a stable posture without being affected by unexpected irregularities or inclinations on the floor on which it walks. It is an object of the present invention to provide a walking control device for a legged mobile robot which can perform the following.

Further, in order for such a legged mobile robot to walk stably, it must satisfy predetermined mechanical stability conditions. Therefore, the robot walks while solving the dynamics problem in real time, or walks after solving the problem in advance. It is sufficient that the robot model used at that time approximates the actual robot with sufficient accuracy, but in reality it contains many elements that are difficult to model, and there are various restrictions such as arithmetic processing. Approximate models are often used in order to reduce the time and labor required for model production.

Accordingly, a second object of the present invention is to keep the posture restoring force of the robot as constant as possible so that the robot can always be approximated by an inverted pendulum having a certain restoring force. Another object of the present invention is to provide a walking control device for a legged mobile robot, which aims to linearize a control system.

Further, in such a legged mobile robot, the robot walks while balancing the reaction force received from the floor with the resultant force of gravity and inertia acting on the floor from the robot side. If the impact force at the time is large, it becomes a cause of losing posture, and stable walking cannot be expected.

Accordingly, a third object of the present invention is to provide a walking control device of a legged mobile robot capable of walking with a stable posture by minimizing a landing impact received by the robot. It is in.

[0009]

The invention to solve the problems described above, there is provided a means for solving] is as shown in item 1 example claims, the body
And connected to it via a joint, at least
In a walking control device of a legged mobile robot having a plurality of legs having one joint, a joint of the leg and the leg
Absorbing mechanism connecting the tip of the part, target posture of the robot
Gait generating means for generating a gait, acting on the robot
Measures floor reaction force and detects ZMP actual measurement position
Detecting means for detecting the measured ZMP measured position as a target.
To calculate the deviation by comparing with the ZMP position
Step, and the target posture according to the obtained deviation.
Rutotomoni comprising a target posture correcting means for correcting the target
The posture correcting means is configured to adjust the robot to the corrected target posture.
To follow the joint displacement of
To reduce the stress caused by the deformation of the shock absorbing mechanism.
It was constructed as Ru to time difference.

[0010]

[Function] Equipped with a shock absorbing mechanism and ZMP measurement position
Is shifted from the ZMP target position,
Because it generates stress due to the deformation of the shock absorbing mechanism,
Even when stepping on unexpected irregularities, the collapse of the posture can be suppressed as quickly as possible, and since it is performed according to the deviation,
The restoring force of the robot can be made constant, that is, the robot can always be approximated by an inverted pendulum having a certain restoring force, and a linear control characteristic can be obtained. as a result,
Not only this stabilization control but also other types of stabilization control can be easily adopted. In addition, since the posture is always maintained as designed, the landing impact can be effectively reduced.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below by taking a bipedal walking robot as an example of a legged mobile robot. FIG.
Is an explanatory skeleton diagram showing the robot 1 as a whole, and has six joints on each of the left and right leg links 2 (for the sake of convenience of understanding, each joint is shown by an electric motor that drives it). The six joints are, in order from the top, joints 10R and 10L for turning the hip leg (R is the right side and L is the left side).
The same applies hereinafter), joint 12 in the roll direction of the waist (around the x axis)
R, 12L, joint 14 in the same pitch direction (around the y axis)
R, 14L, knee joints 16R, 16L in the pitch direction,
The joints 18R and 18L in the pitch direction of the ankle part and the joints 20R and 20L in the same roll direction are attached below the feet 22R and 22L, and the uppermost part is the upper body (housing). 24, in which the control unit 2
6 is stored.

In the above description, the hip joint is joint 10R (L),
12R (L) and 14R (L), and the ankle joint is composed of joints 18R (L) and 20R (L).
In addition, the thigh links 32R, 32 between the waist joint and the knee joint.
L, the lower leg links 34R, 34 between the knee joint and the ankle joint.
L connected. Here, the leg link 2 is given six degrees of freedom for each of the left and right feet, and by driving these 6 × 2 = 12 joints (axes) to appropriate angles during walking, the entire foot To a desired motion, and can be arbitrarily walked in a three-dimensional space. As described above, the above-mentioned joint is formed of an electric motor, and further includes a speed reducer for boosting the output thereof. Details of the joint are described in the application proposed by the present applicant (Japanese Patent Application No.
No. 324218, Japanese Patent Laid-Open No. 3-184787), and the like, which is not the subject of the present invention, and therefore, further description is omitted.

In the robot 1 shown in FIG. 1, a known six-axis force sensor 36 is provided at the ankle, and force components Fx, F in the x, y, and z directions transmitted to the robot through the foot.
y, Fz and moment components Mx, M around the direction
By measuring y and Mz, the presence or absence of landing on the foot and the magnitude and direction of the force applied to the supporting leg are detected. In addition, foot 22R
At the four corners of (L), a capacitance type ground switch 38 (FIG. 1)
(Not shown) is provided to detect the presence / absence of the touchdown of the foot. Further, a tilt sensor 40 is provided on the upper body 24, and x
The tilt with respect to the z-axis in the -z plane and its angular velocity, similarly y
-Detect the inclination with respect to the z-axis in the z-plane and its angular velocity. The electric motor of each joint is provided with a rotary encoder for detecting the amount of rotation. Further, although omitted in FIG. 1, an inclination sensor 40 is provided at an appropriate position of the robot 1.
Are provided with an origin switch 42 for correcting the output and a limit switch 44 for fail countermeasures. These outputs are sent to the control unit 26 in the body 24 described above.

FIG. 2 is a block diagram showing details of the control unit 26, which is constituted by a microcomputer. The output of the tilt sensor 40 and the like is converted into a digital value by the A / D converter 50, and the output is
Via the RAM 54. The output of an encoder disposed adjacent to each electric motor is input to the RAM 54 via a counter 56, and the ground switch 3
Outputs of the RAM 8 and the like are similarly sent to the RAM 5
4 is stored. In the control unit, there are provided first and second arithmetic units 60 and 62 each composed of a CPU. The first arithmetic unit 60 stores the characteristics of the trajectory of the waist posture stored in the ROM 64 as described later. The reference gait is generated by reading the parameters to be expressed, and then the target joint angle (joint drive pattern) is calculated and sent to the RAM 54. Further, the second arithmetic unit 62 reads out the target value and the detected actual value from the RAM 54 as described later, calculates a control value necessary for driving each joint, and outputs the control value via the D / A converter 66 and the servo amplifier. Output to the electric motor that drives each joint.

Next, the operation of the control device will be described.

FIG. 3 is a structured flow chart (PAD diagram) showing the operation. Referring to FIG.
First, in S10, the waist posture (that is, the inclination and direction of the waist) is calculated from the parameters representing the characteristics of the waist locus . Next, in S12, a ZMP target position derived from the equation of motion is calculated from the parameters representing the characteristics of the ZMP trajectory. (When the ZMP trajectory is represented by a polygonal line expression, the parameters representing the characteristics are also given by the folding point coordinates. Here, “ZMP” means a point on the floor where the moment due to the floor reaction force becomes zero. Next, in S14, the positions and postures of both feet are calculated from parameters representing the characteristics of the foot trajectory, for example, the landing position and the one-leg support period, and then in S16, the waist height that does not cause an unreasonable posture is determined. Then, the horizontal acceleration and the horizontal position of the waist are determined so that the ZMP becomes the target position . As described above , S10 to S18 show the work of creating the reference gait. In this embodiment, as described above, the parameters representing the characteristics of the trajectory of the waist posture are set in advance as data for each step. Then, the trajectories such as the waist, the ZMP, and the position / posture of the foot are calculated and used as a reference gait. As described later, the target angle of each joint is specifically calculated from the reference gait.

Then, the program proceeds to S20, where a leg compliance control value is calculated. FIG. 4 is a subroutine flowchart showing the operation.

Before starting the description with reference to FIG. 4, the compliance referred to in this embodiment will be described with reference to FIG.

As described above, if the force acting on the floor from the robot (the combined force of the gravity and the inertial force of the robot) and the floor reaction force acting on the robot from the floor are balanced, the robot can walk stably. I do. If this is replaced with a concentrated load system, if the ZMP actual measurement position (the center point of the actual floor reaction force) matches the ZMP target position (the center point of the floor reaction force assumed by design values), the robot Walks while maintaining the posture as designed. Therefore, in the present invention, in the first embodiment, the floor reaction force applied to the rear foot and the floor reaction force applied to the front foot are measured from the six-axis force sensors 36 attached to both legs, and the resultant is the resultant force. A total measured floor reaction force is determined, and a ZMP measured position, which is the point of action, is determined and compared with a ZMP target position. Then, as shown in FIG. 6, if the measured ZMP position is shifted to the front (in this case, a moment is generated around the ZMP target position to tilt the robot backward) as shown in FIG. Lift the foot vertically (or lower the rear foot vertically depending on the amount of displacement, or lower the rear foot vertically while lifting the front foot vertically). By this operation, the floor reaction force applied to the front foot is relatively reduced as compared with the floor reaction force applied to the rear foot, so that the ZMP actual measurement position moves backward and the ZMP
Approach the target position. This control is referred to as “leg compliance control” in this specification. That is, in this specification, the leg compliance control means control for eliminating a deviation between the ZMP target position and the ZMP actual measurement position. Or Z
It can also be said that the control cancels the moment generated around the MP target position.

That is, when the robot walks in accordance with the ideal gait based on the dynamics calculation, it is assumed that the robot slightly leans forward due to encountering unexpected irregularities during the period of supporting both legs. If the leg compliance control described in this specification is not performed, the robot lifts its rear foot,
The entire load is applied to the front foot, and as a result, the floor reaction force action point (ZMP actual measurement position) shifts to the front foot sole. That is, the ZMP changes almost binaryly with respect to the inclination of the robot (FIG. 7). On the other hand, when the leg compliance control is performed, when the actually measured floor reaction force center (ZMP actual measurement position) is shifted forward from the designed floor reaction force center (ZMP target position), the front foot is lifted. Accordingly, the posture is such that both legs touch the ground while the upper body is tilted forward. At this time, the deviation between the measured ZMP position and the ZMP target position is proportional to the lifting height of the foot, and the lifting height of the foot is proportional to the inclination angle of the body. , The inclination angle of the upper body is also proportional. As will be described later, the proportional gradient is inversely proportional to the magnitude of the leg compliance, and the larger the amount of the leg compliance, the larger the proportional region with respect to the inclination angle of the robot (FIG. 8).

The leg compliance control is not limited to the two-leg supporting period, but may be performed during the one-leg supporting period. Also, as shown in FIG. 9, a shock absorbing mechanism such as a spring or rubber is provided below the ankle joint 18, 20R (L) to prevent oscillation of the leg compliance control and mechanically absorb the high frequency component of the load variation. It is desirable to insert it.

The calculation of the leg compliance control value will be described with reference to FIG.
At 00, the detection value of the six-axis force sensor 36 is fetched. Then
In S102, a ZMP actual measurement position is obtained. Referring to FIG. 10, the method will be described. A moment M → around an arbitrary origin is obtained, a force F → is obtained, and then M → =
The distance vector L → which is F → × L → is obtained, the object is translated in parallel by the distance L, and the intersection with the floor is obtained.

Then, the program proceeds to S104, in which the measured ZMP position and the ZMP target position are compared, and the deviation direction, that is, ZMP
If the measured position is shifted from the ZMP target position, it is determined whether the measured position is shifted to the front side or to the rear side, and the difference (shift amount) x is calculated by the distance. Then S1
A predetermined gain Kf and a measured floor reaction force F (or a vertical component (z direction) component Fz thereof) are added to the difference x calculated in step 06.
To obtain the foot posture correction amount (a method that does not multiply the floor reaction force may be used). That is, as described above, if the measured ZMP position is shifted before the ZMP target position, the moment generated around the designed floor reaction force center (ZMP target position) tilts the robot backward as shown in FIG. Raise the anterior foot vertically, lower the posterior foot vertically, or raise the anterior foot vertically and lower the posterior foot vertically As described above, the foot posture correction amount is obtained, and the foot position is corrected according to the posture correction amount in S108, so that a moment in the opposite direction is generated.
The measured MP position can be brought closer to the ZMP target position, the balance of the posture can be restored, and the robot can walk in the posture as designed. At this time, as shown in FIG. 8, the posture correction amount (the height at which the foot is lifted) is determined according to the deviation amount x, and since the deviation amount x is proportional to the inclination angle of the body, the posture correction amount is It is proportional to the angle of inclination of the upper body. That is, assuming that the ratio between the robot's posture restoring force and the body's tilt angle is the robot's posture restoring force coefficient, by determining the amount of posture correction according to the body's tilt angle, the robot's posture restoring force coefficient is calculated. It can be as constant as possible. That is, the robot can always be approximated by an inverted pendulum having a constant restoring force coefficient, and a linear control characteristic can be obtained. Note that the posture correction direction is such that when the measured ZMP is shifted rearward from the design ZMP and a forward moment is acting on the robot, the driving direction of the foot is reversed and the foot of the rear foot is raised. Or lower the foot on the front foot,
Or, the posture correction amount of the foot is determined so as to perform both.

Returning again to the flowchart of FIG. 3, the program proceeds to S22, in which the target angles of the twelve joints are obtained from the position and posture of the foot and the position and posture of the waist. still,
When the posture of the foot is corrected according to the flowchart of FIG. 4 in the calculation of the leg compliance control value in S20, the target angle is obtained based on the corrected posture. Then, the process proceeds to S24, where the inclination of the robot (the actual posture inclination angle and the actual posture inclination angular velocity) is determined from the output of the inclination sensor 40 described above.
Is detected, the deviation from the command value is calculated, and the deviation decreases .
To modify the target attitude so. A detailed description of this correction will be omitted. Then, the process proceeds to S26, in which the joint is controlled to follow the target angle. This is controlled by the second arithmetic unit 62 shown in FIG. 2, but since this control is not related to the gist of the present invention, detailed description is omitted.

[0025] Since this embodiment is configured as described above, Osamu
Make the joint displacement of the robot follow the corrected target posture,
Therefore, the impact is determined so that the determined deviation is reduced.
Generates stress due to deformation of the absorption mechanism (shown in FIG. 9)
As a result, even if there are unexpected irregularities on the floor and the measured ZMP position may deviate from the ZMP target position, the difference is effectively eliminated, and a moment is generated to turn the robot around the ZMP target position. But you can negate it . That is, since to be able to be approximated by an inverted pendulum having a constant restoring force is always the robot, the control characteristic can be linearly becomes easy to design the control system, and other posture stabilization control This facilitates the combination and facilitates stable walking even when unexpected irregularities are present on the floor. In addition, ZMP actual measurement position is Z
Since control is performed so as to match the MP target position, landing impact is also reduced (here, landing impact is a large floor reaction force).

FIG. 11 shows a second embodiment of the present invention.
5 is a flowchart showing a leg compliance control value calculation subroutine similar to FIG. 4. Explaining focusing on the differences from the first embodiment, the shift direction and the shift amount x are obtained (S200 to S204), and then the process proceeds to S206, as shown at the end of FIG. As shown in the figure, the coordinate rotation angle when assuming that the floor is tilted while keeping the constant is obtained as shown in the figure, and the process proceeds to S208, and the position and posture of both feet are rotated around the ZMP target position by the rotation angle described above. To correct. Here, the floor inclination angle θ is determined by θ = shift amount x gain Kf.

In the case of this embodiment, when the measured ZMP position is shifted forward as shown at the end of FIG. 6, the foot on the rear side strongly kicks the actual floor (shown by the solid line). A floor reaction force is generated, and the measured ZMP position can be brought closer to the ZMP target position. That is, also in this case, a moment in the opposite direction to the moment shown in the upper part of FIG. 6 can be generated, and the posture can be prevented from being collapsed. Also in this embodiment, a restoring force proportional to the amount of displacement can be provided. That is, a linear control characteristic can be obtained with a constant restoring force coefficient.

FIG. 12 is a flow chart similar to FIG. 3 but showing a third embodiment of the present invention. The following description focuses on the differences from the first embodiment (or the second embodiment). In this embodiment, after calculating the leg compliance control values in S300 to S306, the process proceeds to S308, and the posture is not excessive. The height of the waist in the vertical direction (z direction) is obtained, and then the process proceeds to S310, where the leg compliance control is performed to correct the foot position and posture, so that the horizontal acceleration and the horizontal position of the waist are set so that the ZMP is at the target position. And S
Proceeding to 312, a target joint angle is determined based on the foot position / posture and the waist position / posture (correction values when corrected in the subroutine of S306). The remaining configuration including S314 and S316 is not different from the first embodiment (or the second embodiment).

That is, the leg compliance control is intended to prevent the actual measured ZMP position from deviating from the ZMP target position even if there is disturbance due to unevenness or inclination of the floor. It is also desirable that the ZMP target position itself does not shift. That is, only by correcting the foot posture by leg compliance control, Z
The MP measurement position can be returned to the ZMP target position,
As a result, the posture of the body may change, and the ZMP target position itself may deviate from the expected position. Here, if the mass of the legs, especially the mass of the tip, is sufficiently smaller than that of the upper body, the deviation of the design ZMP position can be ignored, and the horizontal position of the upper body can be the same as the reference gait. In the configuration shown in FIG. 1, it is difficult to say that the mass of the foot 22 is sufficiently smaller than that of the upper body 24, but the first embodiment (or the second embodiment)
In the case of Example), the effect on the upper body caused by performing the leg compliance control was considered to be substantially very small, and was ignored. On the other hand, in the third embodiment, as shown, the horizontal position and acceleration of the body are corrected. Therefore, in addition to the effects of the first embodiment (or the second embodiment), there is an effect that the posture can be stabilized more precisely. In the third embodiment, the correction of the waist height in the vertical direction in S308 may not be performed.

[0030] Incidentally, the smaller the leg compliance operation, may be transferred to the leg compliance control value calculation subroutine gait generator 2 during or subsequently. In doing so, the calculations of the gait generator 1 and the gait generator 2 can be performed off-line in advance, which is convenient when using a computer with low computational power.

FIG. 13 shows a fourth embodiment of the present invention. Similar to FIGS. 4 (first embodiment) and FIG. 11 (second embodiment), a subroutine flow chart relating to leg compliance control value calculation is shown. It is. Referring to the figure, after the values detected by the six-axis force sensor are taken in S400, the flow proceeds to S402 to obtain the moment actually generated around the ZMP target position, and proceeds to S404 to obtain the actual moment and the moment command. The difference from the value (usually set to zero) is obtained, and the process proceeds to S406 and S408, where the deviation is multiplied by the gain as in the second embodiment to obtain the coordinate rotation angle. I asked for.

The fourth embodiment differs from the previous embodiment in the method of decomposing the measured floor reaction force. That is, while the previous embodiment focuses on the action point position when the floor reaction force is decomposed by a vector of only a force without a moment, in the fourth embodiment, the value applied to the ZMP target position is defined as the force and moment. It focuses on the moment when it is decomposed into, and in that sense, it is not essentially different from the conventional detection method.

After calculating the leg compliance control value in accordance with the subroutine flowchart of FIG. 13 in the fourth embodiment, the target joint angle is calculated. In that case, FIG. 3 of the first embodiment is used. As shown, the position / posture of the body may not be corrected, or the position / acceleration of the body in the horizontal direction may be corrected as shown in FIG. 12 of the third embodiment. As an effect of the fourth embodiment, since the leg compliance control value is determined by directly detecting the moment around the ZMP target position, a more linear and stable posture control is realized as compared with the previous embodiment. Can be.

The fourth embodiment will be described with reference to the block diagram of FIG.
If the gait is on the slope of 1, the following description will show Δθ and Δθ
comm may be replaced by θ1 + Δθ and θ1 + Δθcomm). In the figure, a so-called reverse kinematics calculation unit that calculates the displacement of each joint from the set position and posture calculates the posture when the floor is inclined by Δθcomm in the reference gait. Fig. 1
The robot joint displacement shown in (1) follows the posture command issued from the inverse kinematics calculation unit. The robot illustrated in FIG. 1 assumes that perfect rigid body, and have One the attitude obtained from the actual joint displacement, robots and foot ground surface tangent (A in FIG.
(A dash line) is defined as Δθ. Is sufficiently high follow-up performance of the displacement controller, is Δθ Δ θ com
matches m. In this case, the transfer function G from the robot and the relative angle of the floor, to a floor reaction force actual moment M about the ZMP target position is as shown in Equation 1.

[0035]

(Equation 1)

This is because, as shown in FIG.
It is equivalent to a bending spring of (1 / Kleg + Kf).

In the fourth embodiment, when a moment is generated around the target ZMP position due to the robot stepping on unexpected irregularities or the like, the moment is directly detected, and the coordinate rotation is proportional to the magnitude of the moment. The angle is virtually obtained, and the position / posture of the foot is set to the target ZM by the virtual angle.
Correct to rotate around P position, ie, target Z
Directly detecting the moment generated around the MP position,
Since a moment in the opposite direction is generated around the same position so as to cancel this, as described above, the linearization of the control characteristic and the posture are further achieved as compared with the first to third embodiments. Can be realized. Furthermore,
When correcting the position / acceleration of the upper body in the horizontal direction in the same manner as shown in FIG. 12, the linearization of the control characteristics and the stabilization of the posture can be achieved even more accurately.

Since the ZMP target position moves continuously or intermittently, for a gait in which the ZMP target position moves rapidly, the center of moment calculation and the rotation center of the leg compliance operation are set to the ZMP target position. In such a case, a sudden change in the behavior is likely to occur, and there is a possibility that the user cannot stably walk due to bouncing or the like. Therefore,
For such a gait, the moment calculation center and the rotation center of the leg compliance operation are close to the ZMP target position but move more gently, for example, a point where the ZMP target position is filtered and smoothed. Is good.

FIG. 16 is a flowchart showing a leg compliance control value calculation subroutine similar to that of FIG. 13 of the fourth embodiment, showing a fifth embodiment of the present invention. Instead of the ZMP target position, a reference point,
For example, ankles of supporting legs (ankle joints 18, 2 in FIG. 1)
The moment around the point of projection of 0R (L) on the floor is determined, and the moment is rotated around that point (S5).
02, S508). The rest of the configuration including the other steps is the same as in the fourth embodiment, and whether or not the body is modified is also the same as in the fourth embodiment. The floor reaction force moment is slightly easier to obtain, but the effect is the same as that of the fourth embodiment except that the control value is slightly inferior to that of the fourth embodiment. Note that the reference point may be a moving point.

FIG. 17 shows the sixth embodiment. The difference from the previous embodiment is that the legs may oscillate if the foot far from the ZMP target position is moved sharply. As shown in FIG. 18, the movement angle θ1 of the foot on the far side on the rotational coordinate is smaller than that of the foot on the near side θ2. Although the driving method of the foot is in accordance with the second embodiment, the vertical movement of the first embodiment is also appropriate, and the upper body may or may not be modified.

FIG. 19 shows a seventh embodiment of the present invention, which deals with a case where a lateral overturning force is applied in the traveling direction. That is, in the previous embodiment, the control value was calculated by detecting the traveling direction, that is, the moment My acting around the y-axis, but in this embodiment, the moment in the lateral direction (around the x-axis) was calculated. Mx is also detected to calculate the control value (S700 to S708). The second
Although described according to the embodiment, this technique is also applicable to other embodiments. Although the moment was used, the lateral force F
y may be used.

FIG. 20 shows an eighth embodiment of the present invention. In the block diagram of FIG. 14 showing the fourth embodiment, a filter 80 is interposed between the robot main body and the floor. . For example, the transfer function of this filter is 1 /
(1 + TS) (where T is a time constant). Assuming that the followability of the displacement controller is sufficiently high, Δθ is Δθco
mm. Accordingly, the transfer function G from the relative angle between the robot and the floor to the actual floor reaction force moment M around the ZMP target position is as shown in Expression 2.

[0043]

(Equation 2)

If the leg rigidity Kleg is sufficiently high, 1 / Kl
eg can be ignored, and the equation becomes as shown in Equation 3.

[0045]

(Equation 3)

This is equivalent to a mechanism in which a torsion spring and a torsion damper are assembled in parallel as shown in FIG. That is,
A damping effect equivalent to inserting a mechanical damper between the robot body and the floor can be obtained, and jumping when the free leg lands can be prevented. Further, since such a low-pass filter is inserted into the feedback loop of the compliance control, as a secondary effect, the loop gain for high frequencies can be reduced, the stability of the compliance control system can be increased, and oscillation can be prevented. In addition, high frequency noise entering from the six-axis force sensor 36 can be removed.

FIGS. 22 to 28 show a ninth embodiment of the present invention.

In the legged mobile robot, when the joint is driven by displacement control, if the above-described leg compliance control is not performed, and if the posture is slightly inclined forward during the support period of both legs while walking, the rear leg is moved. Since the flat is off the floor and the entire load is on the front foot, an extremely large restoring force is generated to return to the rear. That is, the above-mentioned restoring force coefficient becomes extremely large during the two-leg supporting period. Therefore, if the posture stabilization control that feeds back the inclination of the body or the shift of the center of gravity to the movement of the two legs in such a state is further performed, the restoring force coefficient is further increased, and a sufficient damping effect corresponding thereto can be provided. become unable. As a result, the posture becomes rather unstable. Therefore, in this embodiment, the posture stabilization control is performed according to the inclination of the body while performing the leg compliance control described above. In addition, instead of performing the leg compliance described above, compliance can be given to the leg by driving the joint by torque control.However, the compliance characteristic corresponding to the relative position between the robot and the floor greatly depends on the posture. Therefore, the disturbance suppression characteristics for unevenness and inclination of the floor do not match.

The ninth embodiment will be described below with reference to the flow charts of FIGS.

FIG. 22 is an illustration of FIG. 1 for the third embodiment.
2 is the same main routine flow chart as that of FIG. 2 except that in S810, not only the leg compliance control value but also the stabilization control value is calculated. FIG. 23 is a subroutine flow diagram showing the operation.
After the difference between the actual moment and the command value is obtained (S900 to S904), the deviation between the actual tilt angle, the tilt angular velocity and the command value of the body is determined in S906 as shown in the figure.
(Second deviation described above) is multiplied by a predetermined gain to obtain a stabilization control value, which is added to the leg compliance control value to obtain a coordinate rotation angle, which is corrected to that value in S908 .
It was in the jar. In this way, it has to be directed to the same joints stabilization control value and the leg compliance control value, and controls the easily.

FIG. 24 is a block diagram showing this.
As shown, the posture stabilization control of the upper body is performed using PD control. Here, as described above, Δθ is Δθcom
m, and assuming that the measured ZMP position matches the ZMP target position, FIG. 24 can be modified as shown in FIG. As is clear from FIG. 25, the total system including the robot and the control system is linear, and
Various linear control theories such as classical control theory, optimal control theory, and robust control theory can be applied to the attitude tilt stabilization control. FIG. 26 shows an example using the state feedback control.

[0052] Since the ninth embodiment constructed as described above, Osamu
Corrected target posture (more specifically, the detected floor reaction force mode
Deviation from the desired floor reaction force moment (or
Is the detected ZMP actual measurement position and the target ZMP position.
Deviation corrected according to the above-mentioned second deviation)
Target posture) to follow the joint displacement of the robot.
The shock absorber so that the determined deviation is reduced.
Causing stress due to deformation of the structure (shown in FIG. 9)
Therefore, the posture stabilization control can be realized without introducing the joint torque control which is difficult to be put into practical use due to the influence of the wear and inertia of the joint. That is, if the displacement of both feet is deliberately shifted from the design value, the leg compliance control
A posture restoring force corresponding to the shift amount is generated. Therefore, in the posture stabilization control based on the body inclination feedback, the amount of displacement of the foot can be used as the operation amount for generating the restoring force. In addition, since the posture restoring force coefficient of the robot can be kept substantially constant regardless of the one-leg supporting period or the two-leg supporting period, the robot can be approximated to an inverted pendulum having a constant restoring force at all times. As a result, the control system can be linearly approximated, and the attitude stabilization control law can be easily designed. Even if there is unexpected unevenness or inclination on the floor, it is possible to walk while maintaining the inclination of the upper body with respect to the vertical direction almost as designed, without being affected so much.

Here, the same configuration as that described in the eighth embodiment can be applied to the filter 800 shown in FIG. 25, as will be described later. Add consideration. That is, if attention is paid to the dynamic characteristics of the control system and the compliance control is summarized by regarding the floor reaction force moment command as 0, FIG. 25 can be modified as shown in FIG. FIG. 27 is equivalent to the inverted pendulum having the spring and the actuator shown in FIG. This simplification enables not only the application of various linear control theories, but also the inference of robot attitude control from the behavior of a simple model. The best combination of properties can be easily found without much experimentation and simulation. Needless to say, the filter 800 may have a damping effect equivalent to a mechanical damper, as in FIG. That is, in the posture control for feeding back the inclination angle of the body, FIG.
As shown in, the inclination angle velocity of the body (or the inclination angle of the straight line connecting the ground point and the center of gravity) is also fed back in order to enhance stability. However, due to insufficient link system rigidity and insufficient gait, etc. When the body vibrates at a high frequency, a large high-frequency fluctuation occurs in the tilt angular velocity. Therefore, when the tilt angular velocity feedback gain is increased, the legs may vibrate or oscillate. Therefore, if the same filter as that used in the eighth embodiment is used, the stability of the attitude control system can be increased by supplementing the tilt angular velocity feedback. Stability can be ensured. The effect of preventing jumping when the free leg lands is the eighth effect.
This is the same as the embodiment.

Although various examples have been shown in the above embodiment, the modifications are not limited thereto. That is, in this control, the following are arranged. Detection target a. Amount of deviation between ZMP target position and ZMP actual measurement position b. Moment of force generated by floor reaction force around ZMP target position c. 1. Moment of force generated by floor reaction force around reference point Foot movement a. Up and down one leg b. Upper and lower legs c. Foot rotation d. 2. Reduce the amount of movement of the foot farther from the ZMP target position than that of the other foot. Horizontal movement of upper body a. Maintain reference gait b. 3. Correct position and acceleration in the horizontal direction. Vertical movement of upper body a. Maintain reference gait b. 4. Recalculate waist height Upper body posture stabilization control a. No b. However, all of these can be combined, and the examples are only examples.

In the above description, an example in which the present invention is applied to the case where the walking data is set in advance has been described. However, the present invention is not limited to this. The present invention may be applied to a technique in which walking data is obtained in real time.

In the above description, a legged mobile robot that walks on two legs has been described as an example. However, the present invention is not limited to this, and is applicable to a legged mobile robot having three or more legs.

[0057]

According to the first aspect, the upper body 24 and the upper body 24 are provided.
And connected via joints (10-14R, L)
Has at least one joint (16-20R, L)
In the walking control device of the legged mobile robot 1 having a plurality of legs (leg links 2) , the joint of the leg and the leg
A shock absorbing mechanism for connecting the ends of the legs (the spring
Or rubber), the desired gait to generate the desired posture of the robot
Means (control unit 26, S10 to S18, S300
To S304, S308, S310),
Measure the floor reaction force to be used and measure the ZMP actual measurement position
(6-axis force sensor 36, control unit
26, S20, S306, S100, S102, S20
0, S202), target the detected ZMP actual measurement position
And the deviation (shift direction, shift)
Deviation calculating means (control unit 26, S2
0, S306, S106, S108, S206, S20
8) and the target posture according to the obtained deviation.
Posture correcting means (control unit 26, S2
0, S306, S106, S108, S206, S20
8), includes a Rutotomoni, the target posture correcting means, wherein
The joint displacement of the robot follows the corrected target posture.
So that the determined deviation is reduced.
Ru cause stress due to the deformation of the shock absorbing mechanism (electric motor
Data, control unit 26, S22 to S26, S312 to S
316), the posture can be minimized even when unexpected irregularities are encountered,
Since the target posture is corrected according to the deviation, the restoring force of the posture can be made constant, the robot can be approximated with an inverted pendulum, and linear control characteristics can be obtained.
In addition, the impact at the time of landing can be absorbed.

[0058] In the second aspect, wherein, the body 24, it
When connected to joints via joints (10-14R, L)
Has at least one joint (16-20R, L)
Legged mobile robot with multiple legs (leg links)
In the walking control device of 1, the joint of the leg and the leg
Shock absorbing mechanism connecting the ends of the parts (springs and
Is rubber), the desired gait to generate the desired posture of the robot
Generation means (control unit 26, S10 to S18, S30
0 to S304, S308, S310, S800 to S80
8), around a predetermined reference point acting on the robot
Floor reaction force moment detection means for detecting floor reaction force moment
(6-axis force sensor 36, control unit 26, S20, S3
06, S400, S402, S500, S502, S6
00, S602, S700, S702, S810, S9
00, S902), the detected floor reaction force moment
Compare with the desired floor reaction force moment and calculate the deviation
Deviation calculating means (control unit 26, S20, S30
6, S404, S504, S604, S704, S81
0, S904), and according to the obtained deviation,
Target posture correcting means (control unit for correcting the target posture)
26, S20, S306, S406, S408, S50
6, S508, S606, S608, S706, S70
8, S810, S908), and the target
The posture correcting means is configured to adjust the robot to the corrected target posture.
To follow the joint displacement of
To reduce the stress caused by the deformation of the shock absorbing mechanism.
Ru was time difference (electric motor, the control unit 26, S22～S
26, S312 to S316, S812 to S816), so that a stable posture can be ensured at all times.
And a linear control characteristic can be obtained,
Can it to absorb the shock of landing.

[0059] In the claim 3, wherein said predetermined reference point
Since There was configured Ru ZM P position location der a target, can be modified even better when you move the results ZMP target position and drives the legs, it is possible to obtain the effects described above even more accurately. According to claim 4, further comprising the upper body
Detecting means for detecting the tilt angle and / or tilt angular velocity of the vehicle
Step (control unit 26, tilt sensor 40),
Command the tilt angle and / or tilt angular velocity of the upper body
And the deviation is determined as the second deviation.
Calculation means (control unit 26, S24, S314, S8
10, S814, S906), and the eye
The target posture correcting means is obtained by the deviation calculating means.
Calculated by the deviation and inclination deviation calculating means.
Correcting the target attitude according to the deviation of the control unit (control unit
26, S24, S314, S810, S814, S90
6, S908), the upper body tilt feed
Generates restoring force in attitude stabilization control by back
Therefore, the displacement of the leg joint should be used as the operation amount.
Practical use due to the effects of joint wear and inertia
Wire the control system without introducing difficult joint torque control
Shape stabilization control can be easily realized.
Can be

[Brief description of the drawings]

FIG. 1 is a schematic view showing the entire walking control device of a legged mobile robot according to the present invention.

FIG. 2 is an explanatory block diagram of a control unit shown in FIG.

FIG. 3 is a main diagram showing the operation of the control unit shown in FIG. 2;
It is a flow chart.

4 is a flowchart showing a leg compliance control value calculation subroutine in the flowchart of FIG. 3;

FIG. 5 is an explanatory diagram illustrating leg compliance control of FIG. 4;

FIG. 6 is a diagram for explaining the leg compliance control of FIG. 4 and is an explanatory diagram showing a driving method of the leg.

7 is a graph showing a relationship between a deviation amount between a ZMP target position and a ZMP actual measurement position in the leg compliance control of FIG. 4 and a tilt angle of the robot.

8 is a graph showing a relationship between a deviation amount between a ZMP target position and a ZMP actual measurement position in the leg compliance control of FIG. 4 and a magnitude of a leg compliance control value.

FIG. 9 is an explanatory diagram showing a foot structure of a robot suitable for the leg compliance control of FIG. 4;

FIG. 10 is an explanatory diagram showing a method of detecting a ZMP actually measured position in the leg compliance control of FIG. 4;

FIG. 11 is a flowchart showing another example of leg compliance control value calculation according to the second embodiment of the present invention.

FIG. 12 is a flowchart showing an example in which the body position is re-corrected together with leg compliance control according to the third embodiment of the present invention.

FIG. 13 is a flowchart showing another example of leg compliance control value calculation according to the fourth embodiment of the present invention.

FIG. 14 is a block diagram illustrating a fourth embodiment.

FIG. 15 is a diagram schematically illustrating the robot shown in FIG. 14;

FIG. 16 is a flowchart showing another example of leg compliance control value calculation according to the fifth embodiment of the present invention.

FIG. 17 is a flowchart showing another example of leg compliance control value calculation according to the sixth embodiment of the present invention.

FIG. 18 is an explanatory diagram illustrating control of a sixth embodiment.

FIG. 19 is a flowchart showing another example of leg compliance control value calculation according to the seventh embodiment of the present invention.

FIG. 20 is a block diagram showing a state where a filter is interposed in a control loop according to an eighth embodiment of the present invention.

FIG. 21 is an explanatory diagram for mechanically explaining the characteristics of the filter of FIG. 20;

FIG. 22 is a flowchart showing control in which another stabilization control is combined with leg compliance control according to a ninth embodiment of the present invention.

FIG. 23 is a flow chart showing a control value calculation subroutine of leg compliance control and stabilization control in the flow chart of FIG. 22;

FIG. 24 is a block diagram illustrating a ninth embodiment.

FIG. 25 is a block diagram showing a simplified modification of the block diagram of FIG. 24;

FIG. 26 is a block diagram showing an example in which the ninth embodiment is implemented by a state feedback control method.

27 is a block diagram showing an example in which the block diagram of FIG. 25 is modified by setting the transfer characteristic of the filter to 1;

FIG. 28 is an explanatory diagram showing FIG. 27 by replacing it with a mechanical configuration.

[Description of Signs] 1 Legged mobile robot (biped robot) 2 Leg link 10R, 10L Leg rotation joint 12R, 12L Crotch roll direction joint 14R, 14L Crotch pitch direction joint 16R , 16L Joints in the pitch direction of the knee 18R, 18L Joints in the pitch direction of the ankle 20R, 20L Joints in the roll direction of the ankle 22R, 22L Foot 24 Upper body 26 Control unit 36 6-axis force sensor 80, 800 Filter

Continuation of the front page (56) References JP-A-3-161290 (JP, A) JP-A-61-211177 (JP, A) JP-B-48-39425 (JP, B1) Miomir Vukobratovic, translated by Ichiro Kato , "Walking Robots and Artificial Feet", Nikkan Kogyo Shimbun, March 31, 1975, p. 156-178 (58) Field surveyed (Int. Cl. ⁷ , DB name) B25J 5/00 B25J 13/00, 13/08

Claims (4)

(57) [Claims] 1. An upper body connected to the upper body via a joint.
And a walking control device for a legged mobile robot having a plurality of legs having at least one joint , comprising: a . Shock absorbing connecting the joint of the leg and the tip of the leg
Collection mechanism, b . A desired gait generator for generating a desired posture of the robot; c . Detecting means for measuring a floor reaction force acting on the robot and detecting a ZMP actual measurement position as an action point thereof ; d . ZMP targeting the detected ZMP actual measurement position
A deviation calculating means for comparing the position with the position to determine the deviation, and e . Correcting the target posture according to the obtained deviation.
That target attitude correction means, equipped with a Rutotomoni, the target posture Osamu
The corrective means includes the joint of the robot in the corrected target posture.
Follow the displacement, thus reducing the determined deviation
As described above, the stress caused by the deformation of the shock absorbing mechanism is generated.
Walk controller of a legged mobile robot, characterized in that that. 2. The upper body is connected to the upper body via a joint.
Together with a plurality of legs having at least one joint
A walking control device for a legged mobile robot comprising: a . Shock absorbing connecting the joint of the leg and the tip of the leg
Collection mechanism, b . A desired gait generator for generating a desired posture of the robot
Steps, c . Floor around a predetermined reference point acting on the robot
You detect a reaction force moment floor reaction force moment detection means, d. The floor reaction force which is the target of the detected floor reaction force moment.
Deviation calculator for calculating the deviation by comparing with the force moment
Steps, and e . Correcting the target posture according to the obtained deviation.
Means for correcting the desired posture.
The corrective means includes the joint of the robot in the corrected target posture.
Follow the displacement, thus reducing the determined deviation
Let so, cause stress due to the deformation of the shock absorbing mechanism
Walk controller of a legged mobile robot, characterized in that that. Wherein the predetermined walk controller of a legged mobile robot according to claim 2 wherein wherein the reference point is ZM P position <br/> location of a target. 4. The method of claim 1, further comprising: f . Detecting the tilt angle and / or tilt angular velocity of the body
Outgoing detection means, g . The detected tilt angle and / or tilt of the upper body
Compare the angular velocity with the command value and use the deviation as the second deviation
Calculating means for calculating the inclination deviation, and the target posture correcting means includes a calculating means for calculating the deviation.
Calculating means for calculating the deviation and inclination deviation obtained by the output means
The target posture according to the second deviation obtained by
4. The method according to claim 1, wherein the correction is performed.
A walking control device for a legged mobile robot according to any of the first to third aspects.