JPH01304501A - Servo loop control method for industrial robot - Google Patents

Abstract

PURPOSE:To realize the stable control of an industrial robot by calculating the inertial moment of a rotor, the inertial moment of a robot arm and the angular speed in an on-line method based on the working attitude and the working speed of the robot. CONSTITUTION:When the servo loop of an industrial robot, i.e., a control subject 22 is controlled, a command rotational speed is transmitted from an external controller and converted into a command torque T via a speed torque converter 20 having the integration gain K4 and the proportion gain K5. A prescribed servo motor set in the robot is driven by the current control corresponding to the torque T and the output rotational speed x1 of the motor is measured by a pulse encoder, etc., and fed back. The difference between the speed x1 and the command rotational speed is supplied to the converter 20. An observer 24 is added to the control of the servo loop for improvement of the system stability. Then a prescribed motion state equation is used when the state variable of the observer 24 is estimated. Thus it is possible to ensure the control where the variation of the dynamic characteristics is reflected in accordance with the working attitude and the working speed of the robot.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a servo loop control method for an industrial robot, and in particular a method for changing the feedback gain of a state variable estimate by an observer online in consideration of the robot's posture. Regarding control method.

[Conventional technology]

Generally, industrial robots have a cantilevered structure, and a low-rigidity reduction gear or the like is present between each movable body part such as each arm and each servo motor that is each drive source. Therefore, it is a controlled system that has low rigidity as a whole, a low resonance frequency, and very poor damping properties. When driving and controlling this low-rigidity, low-damping robot using a servo loop, an observer is used to estimate the state variable to be controlled, and a fixed feedback gain is used when feeding back the estimated value of the state variable. They were trying to perform feedback control with good damping performance.

[Problem to be solved by the invention]

However, as mentioned above, industrial robots have a cantilevered structure, and their dynamic characteristics vary greatly with the posture and operating speed that change due to operation. Therefore, fixing the feed-hack gain of the state variable estimate by the observer is not preferable in order to quickly dampen the controlled system.

Therefore, an object of the present invention is to provide a servo loop control method for an industrial robot that reflects changes in the dynamic characteristics of the industrial robot depending on its operating posture.

[Means to solve the problem]

In view of the above object, the present invention estimates state variables using an observer based on a kinematic analysis model of the industrial robot when causing an industrial robot to perform a desired operation, and when feeding back each state variable, the feedback Provided is a servo loop control method for an industrial robot in which each feedback gain is changed online based on conditions for a transfer function in a servo loop control system of an industrial robot including a predetermined ideal transfer function.

[For production]

Regarding the feedback gain of the state variable estimated value by the observer, the transfer function of the industrial robot that includes the feedback in the servo loop that is the overall control system is an ideal transfer function that exhibits ideal damping characteristics for, for example, human step force. Each feedback gain is calculated online from the conditions for the calculation, and each estimated value is fed back. Therefore, control can be performed that reflects changes in dynamic characteristics depending on the operating posture and operating speed of the industrial robot.

〔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 of the industrial robot Nodo.
For example, this is a motion analysis model in which an arm 12 and a rotor 10 of a servo motor serving as a drive source are 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 the arm 12.
and the rotor 10 of the servo motor.
8, and 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 two floater 10 moment of inertia, θR: robot arm rotation angle converted to motor axis, JR:
The moment of inertia of the robot arm in terms of the motor shaft, KC: tear torsional spring coefficient of the reducer section, BK: damping coefficient of the reducer section. In this analytical model, a transfer function G whose input is the output torque T and whose output is the rotational angular velocity jM of the rotor in the motor is expressed by the following equation.

-Jω'-BK・(1/JM+1/JR)・ω”+KC
・(1/JM+1/JR)・Jω−三〇ノλ Gain IGI is expressed by the following formula.

...2 This gain ICI has two extreme values IGII and 1G21, and assuming that they become extreme values in the Bode diagram at ω=ω1 or ω=ω2, respectively, the following equation holds true.

KC Han JR・ω1! ...Here, A=1/JM+17JR B=1/JM C=1/(JM-JR). In the case of robots, the fluctuation range of ω, which has a small change, is determined experimentally in advance, and it is divided into equal parts with the desired accuracy.
Using that division width, move ω starting from the previous estimated value of ω1, determine the position where the minimum gain is achieved as the current ω, and substitute that value ω1 into the above formula E to calculate KC. Then, by substituting this ω-ω1 condition into equation N, the gain l
Calculate G11 and then calculate BK. Note that JM and JR can be calculated from the current robot posture.

FIG. 1 shows servo loop control of an industrial robot, which is a controlled object 22. A command rotational speed is transmitted from an external controller (not shown) and converted into a command torque T through a speed-torque converter 20 having an integral gain of 4 and a proportional gain of 5. A predetermined servo motor within the robot is driven by current control according to the command torque T, and outputs a rotational speed xi (-/7M). This output rotational speed xl is actually measured by a pulse encoder etc.
The measured rotational speed x1 is fed back and the difference between it and the commanded rotational speed is input to the speed-torque converter 20.

In this servo loop control 11, an observer 24 is provided to improve the stability of the system. This observer 24
The state variables of
-θR) estimated values x2 and x3 are set to gains 2 and 3, respectively.
The command torque T is fed back through an amplifier 26. Partial molecule S S”+(BK/JM+BK/
JR-2)・S+ (KC/JM+KC/JR, 3)
) This transfer function is compared with Kitamori's ideal transfer function, which responds as shown in FIG. 3 in response to a step input.

Kitamori's ideal transfer function is expressed by the following equation in the quadratic case.

0.5・δ2・S2+δ・S+1 Here, δ is a time constant indicating the time until 63% of the response target value is reached. The following equation is obtained by comparing the coefficients of the characteristic equations of both transfer functions.

To 8/JM+BK/JR-K 2 = 1 / 0.5
6... KC/JM+KC/JR-K 3 = 1
/0.5δ2...C By substituting the above-mentioned equations E and E into each of these equations G and J, it is possible to calculate a feedback gain of 2 and 3. The time constant δ is set to 1/4 to I15 of the time constant of the speed loop shown in FIG.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, JM, JR1, and ω1 can be calculated online from the ever-changing operating posture and operating speed of the robot. By changing the feedback gains of x2 and x3 by 2 and 3 online, optimal servo loop control can be performed. In other words, stable control with high damping performance is possible.

[Brief explanation of the drawing]

Fig. 1 is a servo loop control diagram of the controlled object, Fig. 2 is an analytical model diagram used for an observer of the controlled object (robot), and Fig. 3 is an explanatory response diagram of the ideal transfer function. IO...Rotor of servo motor, 12...Arm, 18...Reduction gear.

Claims (1)

[Claims] 1. When causing an industrial robot to perform a desired operation, state variables are estimated by an observer based on the motion analysis model of the industrial robot, and when each state variable is fed back, the servo of the industrial robot including the feedback is A servo loop control method for an industrial robot, characterized in that each feedback gain is changed online based on conditions for a transfer function in a loop control system to be a predetermined ideal transfer function.