JP6107338B2 - Center of gravity generation method for biped robot - Google Patents

Description

  The present invention relates to a center-of-gravity position generation method for a biped robot.

  Regarding a walking trajectory generation method for a biped robot, for example, Patent Document 1 proposes a technique related to generation of a target center of gravity trajectory. In the technique disclosed in Patent Document 1, a ZMP (zero moment point) trajectory is calculated in advance by looking at a robot as a simple one mass point model. Then, a correction amount of the center of gravity trajectory is calculated for the motion predicted in the detailed model of the robot, and the center of gravity trajectory is corrected using the calculated correction amount. Here, the one-mass model is a model that assumes that the total mass of the robot is concentrated at the center of gravity, and is also referred to as a single mass model. The detailed model is a model that is closer to the actual shape of the robot, and is a multi-mass model in which mass points are arranged with respect to joints and links such as the robot's legs and upper body. By using the detailed model, it is possible to generate a center-of-gravity trajectory that also takes into account the movement of the robot's legs and upper body.

Japanese Patent No. 3834629

  However, in order to calculate the ZMP correction amount related to the future operation using the detailed model, the amount of data is large and a calculation time is required. For this reason, when the robot posture becomes unstable during walking on rough terrain, there is a problem that a trajectory for posture recovery cannot be generated immediately, and the possibility of falling is increased.

  The present invention has been made to solve such a problem, and can calculate a center of gravity position of a biped robot that can calculate a target center of gravity position correction amount at high speed and can realize more stable control. Is intended to provide.

  The center-of-gravity position generation method for a biped robot according to an embodiment calculates a target center-of-gravity position of the robot in a one-mass model that assumes that the total mass of the robot is concentrated on the center-of-gravity position of the robot, A current ZMP is calculated using the current motion of the robot in a detailed model closer to the shape of, and based on the target center-of-gravity position based on the calculated one mass point model and the calculated current ZMP, A target center-of-gravity position correction amount is calculated, a correction amount of the center-of-gravity position correction amount calculation ZMP is calculated using the calculated target center-of-gravity position correction amount, and the calculated correction amount of the center-of-gravity position correction amount calculation ZMP is calculated. And correcting the calculated current ZMP and using the calculated target centroid position correction amount, the target centroid based on the one mass point model. By correcting the location to generate a target barycentric position data of the robot, it is characterized in.

  Thus, when correcting the target center-of-gravity position obtained from the one mass point model, the target center-of-gravity trajectory can be corrected based on the current ZMP in the detailed model without calculating the future ZMP in the detailed model. Therefore, the target center-of-gravity position correction amount can be calculated at high speed with a small calculation amount. Then, by correcting the target center-of-gravity position using such a target center-of-gravity position correction amount, target center-of-gravity position data for bringing the ZMP trajectory closer to the target ZMP trajectory originally targeted in the one mass point model is generated at high speed. can do. As a result, in the walking control, it is possible to generate the target gravity center position data, which is a control parameter for following the target ZMP trajectory, at higher speed, and to realize more stable control.

  In addition, the calculation of the target center-of-gravity position correction amount is performed by performing an inverse conversion process of a predetermined process using the calculated current ZMP, and calculating the result of the inverse conversion process and the calculated one mass point model. The target center-of-gravity position correction amount may be calculated by obtaining a difference from the target center-of-gravity position.

  Further, the correction amount of the ZMP for calculating the center of gravity position correction amount is calculated by multiplying the calculated target center of gravity position correction amount by a predetermined coefficient, thereby calculating the correction amount of the ZMP for calculating the center of gravity position correction amount. You may do it.

  ADVANTAGE OF THE INVENTION According to this invention, the target gravity center position correction amount can be calculated at high-speed, and the gravity center position generation method of the biped walking robot which can implement | achieve more stable control can be provided.

3 is a functional block diagram illustrating a centroid position generator according to Embodiment 1. FIG. 3 is a detailed functional block diagram of the barycentric position generator according to Embodiment 1. FIG. FIG. 6 is a diagram for explaining an effect according to the first embodiment. FIG. 6 is a diagram for explaining an effect according to the first embodiment. FIG. 6 is a diagram for explaining an effect according to the first embodiment. 6 is a flowchart showing a center-of-gravity position generation process according to the first embodiment.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings. For clarity of explanation, duplicate explanation is omitted as necessary.

<Embodiment 1. >
FIG. 1 shows a functional block configuration diagram of the barycentric position generator 100 according to the present embodiment. The center-of-gravity position generator 100 corrects the target center-of-gravity position obtained from the one mass point model, and generates a target center-of-gravity trajectory composed of time series data of the target center-of-gravity position. The center-of-gravity position generator 100 includes processing units 10, 20, and 30 and can execute characteristic processes shown in the following (1) to (5).

(1) The center-of-gravity position generator 100 uses the “target center-of-gravity position (COM: Center Of Mass) obtained from one mass point model” and the “ZMP (current value) calculated by the detailed model” to “correct the target center-of-gravity position” Amount (a correction amount of the center of gravity trajectory) "is calculated. In this processing, “target center of gravity position (COM) obtained from one mass point model” and “ZMP (current value) of the detailed model” calculated in the processing unit 10 are input to the processing unit 20, and the processing unit 20 Corresponds to a process of calculating a “target center-of-gravity position correction amount (center-of-gravity trajectory correction amount)” based on these input data.

(2) The center-of-gravity position generator 100 generates a “target center-of-gravity position correction amount (computation amount) calculated by the processing unit 20 from“ target center-of-gravity position (COM) obtained from one mass point model ”regarding the characteristic processing of (1) above. Subtract center of gravity trajectory correction). Then, the processing unit 20 calculates a “target center-of-gravity position correction amount (center-of-gravity trajectory correction amount)” using the “target center-of-gravity position obtained from one mass point model” after the subtraction.

(3) The center-of-gravity position generator 100 uses the “target center-of-gravity position correction amount (center-of-gravity trajectory correction amount)” calculated by the processing unit 20 in relation to the characteristic processing of (2) above. The center of gravity position correction amount calculation ZMP correction amount "is calculated. Using the calculated “correction amount of the ZMP for calculating the center of gravity position correction amount”, the “ZMP (current value) of the detailed model” calculated in the processing unit 10 is corrected.

(4) The center-of-gravity position generator 100 uses the processing unit 20 to calculate the “target center-of-gravity position correction amount (center-of-gravity trajectory correction amount)” by the processing unit 20 regarding the characteristic processes (1) to (3) above. 20 uses the “detailed model ZMP (current value)” calculated by the processing unit 10 to execute an inverse conversion process of “convert the center of gravity position in one mass point model to ZMP”, and perform the inverse conversion A “target centroid position correction amount” is calculated by obtaining a difference between the result of the processing and “target centroid position obtained from one mass point model”.

(5) With respect to the characteristic processing of (3) above, the center-of-gravity position generator 100 uses the processing unit 30 to calculate “target center-of-gravity, By multiplying the “position correction amount (the correction amount of the center of gravity trajectory)” by a predetermined coefficient K, the “correction amount of the ZMP for calculating the center of gravity position correction amount” is calculated.

Hereinafter, with reference to FIG. 2, the calculation processing of “target center-of-gravity position correction amount (center-of-gravity trajectory correction amount)” by the processing unit 20 will be described in more detail.
Regarding the calculation process of “target center-of-gravity position correction amount (center-of-gravity trajectory correction amount)” in the processing unit 20, as described in the feature (4), the processing unit 20 calculates the “detailed model calculated by the processing unit 10. ZMP (current value) ”is used to execute an inverse conversion process of“ convert the center of gravity position in one mass point model to ZMP ”, the result of the inverse conversion process, and the target obtained from the one mass point model By calculating the difference from the “center of gravity position”, the “target center of gravity position correction amount” is calculated. The inverse conversion process of the process of “converting the center-of-gravity position in one mass point model to ZMP” here can be performed based on the following formula. In Equation (1), p represents ZMP, x represents the position of the center of gravity, g represents the acceleration of gravity, and z _c represents the height position of the center of gravity.

ZMP in the case where the position of the center of gravity and the acceleration are determined can be calculated by Expression (1). Then, Formula (2) can be obtained by performing Laplace transform on Formula (1).

Furthermore, Formula (3) can be obtained by converting Formula (2).

  The center-of-gravity position when the ZMP is located at a predetermined position can be calculated by Expression (3). Thereby, by substituting ZMP calculated based on the detailed model into the equation (3), it is possible to calculate the position of the center of gravity necessary for that purpose. By calculating the difference between the calculated center of gravity position and the “target center of gravity position determined from the one mass point model”, the correction amount of the “target center of gravity position determined from the one mass point model” (that is, the target center of gravity position correction amount) is obtained. Can be calculated.

  On the other hand, as can be seen from Equation (1), since the ZMP is determined by the position of the center of gravity and the acceleration, the center of gravity trajectory that satisfies the ZMP cannot be uniquely determined. That is, (3) is only seeking one solution. Further, even when ZMP is satisfied, the center of gravity may diverge greatly (that is, the target center-of-gravity position correction amount may diverge). In the present embodiment, this can be prevented by adding a function that does not diverge the target center-of-gravity position correction amount.

  Specifically, divergence is prevented by correcting “ZMP (current value) of detailed model”. In the present embodiment, as shown in FIG. 2, this is calculated by multiplying the “target center-of-gravity position correction amount” by a predetermined coefficient K. According to this method, the prevention of complicated divergence can be realized most easily.

  Note that the biped robot according to the present embodiment incorporates an actuator for controlling the joint angle and an encoder for detecting the joint angle in each joint of the leg. Further, the robot includes a control controller including the above-described center-of-gravity position generator 100. The control controller controls each joint based on the input user command. That is, the control controller controls the actuator so as to follow the target angle of each joint obtained based on the user command.

  The control controller includes a CPU (Central Processing Unit) that performs control processing, arithmetic processing, and the like as a main hardware configuration, a control program executed by the CPU, and a ROM (Read Only Memory) that stores the arithmetic program and the like. ) And a RAM (Random Access Memory) that temporarily stores processing data and the like. The CPU, ROM, and RAM are connected to each other by a data bus.

Next, the center-of-gravity position generation processing according to the present embodiment will be described with reference to FIG.
S101: The controller determines the landing position of each foot of the robot, and generates a target ZMP trajectory within the determined landing region of each foot.
S102: The centroid position generator 100 generates a target centroid trajectory for generating a target ZMP trajectory.
S103: The controller drives each joint so as to follow the generated target center-of-gravity trajectory.

Hereinafter, the process in S102 will be specifically described.
S201: The center-of-gravity position generator 100 generates a target center-of-gravity trajectory for realizing the target ZMP trajectory generated in S101 using the one mass point model.
S202: The center-of-gravity position generator 100 calculates the current ZMP using the current motion of the robot in the detailed model, and converts the calculated current ZMP into the center-of-gravity position of the robot according to a predetermined conversion process. Then, the target center-of-gravity position correction amount is calculated using the target center-of-gravity trajectory based on the one mass point model generated in S201 and the center-of-gravity position based on the converted current ZMP. Then, the target center-of-gravity position based on the one mass point model generated in S201 is corrected using the calculated center-of-gravity position correction amount. Thereby, target gravity center position data of the robot is generated.

Next, effects of the present embodiment will be described with reference to FIGS.
First, an example of a target ZMP and a center of gravity trajectory that satisfies the target ZMP calculated using a one-mass point model will be described with reference to FIG. In the figure, the horizontal axis indicates time, and the vertical axis indicates distance (m) in the horizontal direction. In the example shown in FIG. 3, the target ZMP changes by 0.1 m in a step shape at 5 sec.

  For example, it is assumed that the ZMP of the detailed model changes due to the movement of the free leg or the movement of the upper body. In FIG. 4, it is assumed that a disturbance of 0.01 m is applied in terms of ZMP. When the method shown in the present embodiment is not applied and the conventional method is not applied, the influence of the disturbance appears as it is in the ZMP trajectory (ZMP trajectory shown in FIG. 4). However, originally, it is necessary to match the original target ZMP trajectory. Therefore, the target center-of-gravity trajectory is corrected by applying the method shown in the present embodiment.

  FIG. 5 shows a trajectory when the method shown in the present embodiment is applied and the target center-of-gravity trajectory is corrected. As shown in the figure, it is understood that the ZMP trajectory (ZMP trajectory) considering the detailed model approaches the ZMP target trajectory (target ZMP trajectory) of the original one-mass model by correcting the target center-of-gravity trajectory. .

  As described above, according to the method shown in the present embodiment, when correcting the target center-of-gravity position obtained from the one mass point model, the current model in the detailed model is not calculated without calculating the future ZMP in the detailed model. The target center of gravity trajectory is corrected based on the ZMP. As a result, the target center-of-gravity position correction amount can be calculated at high speed with a small amount of calculation. Then, by correcting the target centroid position using such a target centroid position correction amount, the target centroid position data for bringing the ZMP trajectory closer to the target ZMP trajectory originally targeted in the one-mass model is faster. Can be generated. As a result, in the walking control, it is possible to generate the target gravity center position data, which is a control parameter for following the target ZMP trajectory, at higher speed, and to realize more stable control.

  The present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the gist of the present invention described above.

10, 20, 30 processing unit,
100 Center of gravity position generator

Claims (2)

Calculating the target center of gravity position of the robot in a one-mass model assuming that the total mass of the robot is concentrated on the center of gravity position of the robot;
Calculating the current ZMP using the current motion of the robot in a detailed model closer to the actual shape of the robot;
Calculating a target center-of-gravity position correction amount based on the target center-of-gravity position based on the calculated one mass point model and the calculated current ZMP;
Using the calculated target center-of-gravity position correction amount, the correction amount of the center-of-gravity position correction amount calculation ZMP is calculated, and using the calculated correction amount of the center-of-gravity position correction amount calculation ZMP, the calculated current Correct ZMP,
By correcting the target centroid position based on the one mass point model using the calculated target centroid position correction amount, target centroid position data of the robot is generated ,
The target centroid position correction amount is calculated as follows:
By executing the inverse transformation process of a predetermined process using the calculated current ZMP, and obtaining the difference between the result of the inverse transformation process and the target center-of-gravity position based on the calculated one mass point model Calculating the target center-of-gravity position correction amount ;
A method for generating a center of gravity trajectory of a biped robot. The calculation of the correction amount of the ZMP for calculating the center of gravity position correction amount is as follows:
Calculating a correction amount of the center-of-gravity position correction amount calculation ZMP by multiplying the calculated target center-of-gravity position correction amount by a predetermined coefficient;
The center of gravity trajectory generation method for a biped robot according to claim 1 .

Images