JPH01296301A - Method for controlling servo loop of industrial robot - Google Patents

Abstract

PURPOSE:To largely set the gain of the servo loop of an industrial robot and, at the same time, to sufficiently suppress the vibration of the robot by estimating state variables by means of an observer to which a disturbance estimating term is introduced and feeding back the estimated state variables. CONSTITUTION:The observer 28 obtains the estimated value X1', X2', and X3' of each state variable by inputting command torque T and output rotating speed X1 of an object to be controlled (industrial robot) 24. Then the observer 28 feeds back the estimated values through amplifiers 30, 32, and 34 respectively having gains K1, K2, and K3. In this case, a state variable of X4'=0 is introduced so that the influence of a disturbance factor can be eliminated and the estimation can be performed easily. Thus the damping characteristics and rigidity of the object to be controlled 24 are apparently increased and the object 24 can be controlled stably.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a drive control method for an industrial robot, and more particularly to a servo loop control method using an observer incorporating a disturbance estimation term.

[Conventional technology]

Generally, each arm of a movable body of an industrial robot, such as an articulated robot, is associated with a driving servo motor. In robot work, each servo motor corresponding to each arm is appropriately driven and controlled to cause the tip to perform a predetermined movement or reach a predetermined position.

Unlike machine tools, industrial robots generally have a cantilever structure, and low-rigidity reduction gears are installed between each movable body part such as each arm and each servo motor that is each drive source. Existing. Therefore, it is a controlled system that has low rigidity as a whole, a low resonance frequency, and very poor damping properties. Therefore, in the past, vibrations of the arm, etc. were prevented by setting the gain of the servo loop small.

[Problem to be solved by the invention] ′ However, it is essential to operate industrial robots quickly and accurately from the perspective of improving product production efficiency, so it is necessary to reduce the gain of the servo loop as in the past. If we had set this, we would not be able to meet such requests.

Accordingly, the present invention provides a servo loop control method for an industrial robot that is capable of setting a large servo loop gain and sufficiently suppressing vibrations.

[Means to solve the problem]

In view of the above object, the present invention provides a movable body part of an industrial robot and a drive source for driving the movable body part, which are arranged in parallel.
A state variable is estimated by an observer that introduces a disturbance estimation ring into a state equation based on a motion analysis model that is connected by a dimensional rigid spring and a one-dimensional vibration absorber, and these estimated state variables are fed back to drive and control the drive source. This invention provides a servo loop control method for an industrial robot.

[For production]

A motion analysis model in which the movable body of an industrial robot and a drive source that drives the movable body are connected by a one-dimensional rigid spring and a one-dimensional vibration absorber installed in parallel can relatively accurately understand the motion state of the industrial robot. Although it is expressed faithfully, the damping performance of the system does not necessarily improve with the estimated values of the state variables by the observer due to disturbance factors such as nonlinear friction that actually exist. Therefore, a disturbance estimation ring is introduced to absorb the influence of the disturbance factors, thereby improving the state variable estimation performance of the observer.

〔Example〕

The present invention will be described in more detail below based on embodiments shown in the accompanying drawings. Figure 2 shows the movable body part of an industrial robot, such as the arm 12 and the rotor l of the servo motor that is the drive source.
This is a kinematic analysis model in which O is connected by a one-dimensional spring 14 and a dashpot 16 arranged in parallel. This one-dimensional spring 14 and dashpot 16 are mainly connected to a reducer 18 interposed between the arm 12 and the rotor lO of the servo motor.
It has been found through simple experiments that this model generally accurately represents the dynamic characteristics of the reducer itself. This analytical model establishes the following equation of motion.

Here, T: Motor output torque, θM: Rotation angle of rotor 10, JM: Moment of inertia of rotor 10, θR: Robot arm rotation angle converted to motor axis, JR:
The moment of inertia of the robot arm in terms of the motor shaft, K: Torsional spring coefficient of the reducer part, C: Damping coefficient of the reducer part, Here, by introducing the new coal numbers θM - θR and θS, we get the following. The formula becomes

here, all=o.

a12=-C/JM.

a13=-/JM.

a21=0, a22=-(C/JM+C/JR), a23=-(K/JM+/JR), bl-1/JM.

b2=1/JM.

If the output of the control system is the speed of the motor, that is, the rotational speed bM of the rotor 10, then the following equation holds true.

Two (100) When introduced, V = (10003 summers. By the observer based on this state equation, (
xi x2 x3 x4), in which △
/\ △ xl, x2. The estimated value xi of x3, x2. Servo loop control is performed by feeding x3.

The entire control loop is shown in Figure 1, in which a command signal for the position to which each arm of the robot should move is input into a controller (not shown), and the position loop gain K
The commanded velocity is converted into a velocity through a position/velocity converter 20 having an integral gain of 4 and a torque T through a velocity/torque converter 22 having an integral gain of 4 and a proportional gain of 5. A predetermined servo motor of the industrial robot, which is the controlled object 24, is driven by current control according to the command torque T, and outputs a rotational speed x1. This output rotation speed xi
is actually measured by a pulse encoder or the like, and the measured rotational speed x1 is fed back, and the difference between it and the commanded rotational speed output from the position/speed converter 20 is calculated and inputted to the speed/torque converter 22. Further, the output rotation speed xi of the controlled servo motor 24 is passed through the integrator 26 to calculate the position, the calculated position is fed back and the difference from the command position is calculated,
It is input to the position/velocity converter 20.

The motion state equation A of the industrial robot (controlled object) 24 and the variable X used to express the mouth are xi, x as shown in Equation 2.
2, and x3, that is, θM, θS, and θS. The fact that there are three independent variables corresponds to the fact that the characteristic matrix of equation (c) has three rows and three columns, and is based on the fact that the characteristic equation is a cubic equation. In controlling this controlled object 24, an observer 28 is used for asymptotic stabilization. The observer 28 feedbanks through amplifiers 30 , 32 , 34 with command torque T and output rotation to the controlled object 24 . When estimating this state variable, the influence of nonlinear disturbance factors such as friction is taken into consideration in the state of motion equation (A) used for estimation. Speed x which is the output
Because i is input, estimation of each state variable may not go well.

Therefore, the previously mentioned state variable of ``Yakure=0'' is introduced, which makes it possible to remove the influence of disturbance factors and facilitate estimation.

Note that the characteristic matrix of the state equation C described above is its component a
Since both ll and a21 are zero, feeding back the estimated value of the state variable x4, which is indicated by the apparent deflection -0, as shown in Figure 1 becomes the following equation by moving it to the right side of the motion equation A. .

T=JM・θM+C・(θ unit − δR)+K・(θ unit − θ
R) As can be seen from this equation, the damping coefficient C has apparently increased by 2 minutes, and the torsion spring coefficient K has increased by 3 minutes. That is, the damping performance and rigidity of the controlled object 24 are apparently increased, and stable control becomes possible.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, since the controlled object can be stabilized, the forward gain of the position loop gain KP, the integral gain of 4, and the proportional gain of 5 can be set large, and as a result, the controlled object can be stabilized. Able to operate things quickly and accurately.

[Brief explanation of the drawing] FIG. 1 is a control loop diagram according to the present invention, Figure 2 is a motion analysis model diagram of an industrial robot. lO...servo motor rotor, l2...Arm, 18...Reducer.

Claims (1)

[Claims] 1. When making an industrial robot perform a desired operation, the movable body of the industrial robot and the drive source that drives the movable body are connected by a one-dimensional rigid spring and a one-dimensional vibration absorber arranged in parallel. An industrial robot characterized in that a state variable is estimated by an observer in which a disturbance estimation term is introduced into a state equation based on a kinematic analysis model connected to each other, and the estimated state variables are fed back to drive and control the drive source. Servo loop control method.