JP2008217410A - Actuator controller and actuator control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To cooperatively control actuators while control tracking between respective axes without any delay with respect to an instruction value based on a robust acceleration control system when an actuator is saturated. <P>SOLUTION: This actuator controller is provided with an acceleration control system 1 applying robust acceleration control to actuators 2, a tracking control system 20 controlling tracking of the actuators 2, a torque current limiter 95 regulating an acceleration instruction current I<SB>FFi</SB>of a feedfoward component for restricting a current reference value I<SP>ref</SP>to the actuator 2 to be a previously set maximum current value I<SB>MAX</SB>or less, and a delay regulation means 86 correcting an acceleration/deceleration torque current I<SB>FBi</SB>of a feedback component and the acceleration instruction current I<SB>FFi</SB>of the feedforward component according to a position instruction regulation rate γ(n+1)<SB>i</SB>for cooperatively controlling the actuators 2 between the respective axes. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an actuator control apparatus and an actuator control method that can control the tracking of a controlled object on each axis without delay even when torque saturation occurs.

  2. Description of the Related Art In recent years, motion control devices typified by positioning control of various machine tools, semiconductor manufacturing devices, and robots in the industry are demanded for high-speed and high-precision control for the purpose of improving productivity. Furthermore, in order to reduce the cost of the entire system, a feed drive system actuator using an AC servo motor with a small wattage and an inexpensive ball screw is often used for such a positioning mechanism. In the conventional system, the motor causes torque saturation especially when the motor is started and braked. Therefore, in a system in which controlled objects are operated on each axis by multiple actuators, the controlled objects between the axes operate in a coordinated manner. There is a problem that the target locus cannot be tracked. Therefore, in order to solve these problems, many torque saturation countermeasure techniques for satisfying desired control characteristics have been performed.

FIG. 16 shows an example of a system configuration of a servo system to which such measures are taken. Here, the servo motor 501 to be controlled is driven by the current value I given from the control device 502. The control device 502 includes subtracters 503 to 505, a position controller 506, a speed controller 507, a current controller 508, and a current limiter 509, and performs local current control for monitoring the current value I to the servo motor 501. the outer loop 511, the position control loop 512 for monitoring the position theta _m of the servo motor 501, and each comprise a speed control loop 513 which monitors the response speed (angular velocity) omega _m of the servo motor 501.

The control device 502 here calculates a position deviation between the actual position θ _m of the servo motor 501 and the target position reference value θ _ref by the subtractor 503, and this position deviation is calculated by the position controller 506. Convert to speed reference value ω _ref . A deviation between the speed reference value ω _ref and the actual response speed ω _m of the servo motor 501 is calculated by a subtractor 504 connected to the subsequent stage of the position controller 506, and converted into a current reference value I _ref by the speed controller 507. Further, a deviation between the current reference value I _ref and an actual current value I given to the servo motor 501 is calculated by a subtracter 505 connected to a later stage of a current limiter 509 described later, and this is a control for the servo motor 501. A signal, that is, a current value I is output from the current controller 508. As a result, the servo motor 501 is controlled such that the actual position θ _m of the servo motor 501 follows the position reference value θ _ref given to the control device 502.

In addition, a current limiter 509 is connected between the speed controller 507 and the subtractor 505 so that the current reference value I _ref converted by the speed controller 507 does not exceed a certain range in this series of controls. . Thereby, even when the current reference value I _ref that the servo motor 501 is likely to cause torque saturation is output from the speed controller 507, the current value I to the servo motor 501 is suppressed by the current limiter 509 within a certain range. Torque saturation of the servo motor 501 can be prevented.

  Such a torque saturation countermeasure technique is disclosed not only in the example shown in FIG. 16 but also in Patent Document 1, for example.

On the other hand, the applicant of the present application has realized an acceleration control system (acceleration controller) that drives an actuator represented by a motor with an acceleration command, and has proposed the implementation of a high-speed and high-performance servo control system. Although the acceleration control system enables high-speed control, on the other hand, the current command and voltage command of the operation amount to the motor are often limited (saturated). For this reason, the conventional methods as shown in Patent Document 1 and FIG. 16 are all techniques based on position servo and speed servo, so that dynamic delay occurs, the direction of the torque command changes, etc. As a countermeasure against the acceleration control system, it was insufficient.
JP-A-7-62052

  Thus, since the conventional countermeasure against torque saturation is a method based on the speed PI control system, it is often not always suitable for a control system based on a robust acceleration control system.

  The present invention has been made in view of the above-described problems, and its purpose is based on a robust acceleration control system, while performing tracking control between each axis without delay with respect to a command value when the actuator is saturated, while performing coordinated operation. It is an object to provide an actuator control device and an actuator control method capable of performing the above.

  In order to solve the above-described problems, an actuator control device according to the present invention is an actuator control device that individually controls the operation of a plurality of actuators that move on each axis, and includes a disturbance observer that estimates a disturbance to the actuator. The disturbance estimated value obtained by the disturbance observer is converted into a disturbance compensation current, and an acceleration control unit that performs robust acceleration control of the actuator based on a current reference value to the actuator that is feedback-compensated with the disturbance compensation current, and the actuator In order to perform trajectory tracking control, a trajectory tracking control unit that calculates a current reference value of a feedback component and a current reference value of a feedforward component, a current reference value of the feedback component, a current reference value of the feedforward component, and the disturbance Actuator with total compensation current In order to limit the current reference value to the predetermined maximum current value or less, a current limiter that adjusts the current reference value of the feedforward component and the feedforward to coordinately control the actuator between the axes. And an adjustment unit for correcting the current reference value of the feedback component and the current reference value of the feedforward component from the adjustment rate obtained when adjusting the current reference value of the component.

  In the actuator control apparatus, it is preferable that the pole of the disturbance observer is infinite.

  The current limiter for each axis calculates the adjustment rate of the position command so that the actuator does not deviate from the trajectory on all axes when the actuator is saturated with current on any one axis. It is preferable to adjust the current reference value of the feedforward component for each axis using the rate.

Specifically, the current limiter of each axis includes the current reference value of the i-axis feedback component as I _FBi , the current reference value of the i-axis feedforward component as I _FFi , and the i-axis disturbance compensation. The i-axis for maintaining the robust control when the current is I _RBi , the current reference value to the i-th axis actuator is I _i ^ref , and the maximum torque current of the motor constituting the actuator is I _{MAX. Calculate} the maximum acceleration / deceleration torque current I _accMAX (n) _i of the current sample n of the eye using the following formula:

When the sum of the current reference value I _FFi of the feed forward component and the current reference value I _FBi of the feedback component is larger than the maximum acceleration / deceleration torque current I _accMAX, Acceleration θ ^·· _max (n) _i is calculated (where K _t is the torque constant, J _ni is the nominal value of the moment of inertia of the i-axis),

The adjustment rate γ (n + 1) _i of the position command is calculated by the following formula (where Ts is the sampling time from the current sample to the next sample, θ ^{· res} (n) _i is the current sample of the i-axis Speed response value, θ ^cmd (n + 1) _i is a position command value in the next sample of the i-axis, θ ^cmd (n) _i is a position command value in the current sample of the i-axis),

Next, the minimum value γ (n + 1) min is extracted from the adjustment rate γ (n + 1) _i of each axis i,
Values I ^to _FFi (n) obtained by adjusting the current reference value of the feedforward component are calculated from the following equations (where K _tn is the nominal value of the torque constant, and f _i (t) is the position command for the i-th axis. The time function of the value, θ ^· _i ^res (n) is a velocity response value in the current sample on the i-axis, and θ ^cmd _i (n) is a position command value in the current sample on the i-axis).

Further, the adjustment unit of each axis calculates a correction time t ^to (n + 1) in the next sample based on the adjustment rate γ (n + 1) of the position command from the following equation (where t (n) is the current time) ),

  It is preferable to correct the current reference value of the feedback component and the current reference value of the feedforward component based on the correction time.

  The actuator control method according to the present invention is an actuator control method for individually controlling the operations of a plurality of actuators movable on each axis, and includes a disturbance observer for estimating a disturbance to the actuator, and the disturbance obtained by the disturbance observer An estimated value is converted into a disturbance compensation current, and an acceleration control step for robust acceleration control of the actuator based on a current reference value to the actuator feedback-compensated with the disturbance compensation current, and feedback for controlling the locus tracking of the actuator. A trajectory tracking step for calculating a current reference value of a component and a current reference value of a feedforward component; and a current reference value of the feedback component, a current reference value of the feedforward component, and the disturbance compensation current. Current A current limiting step for adjusting the current reference value of the feedforward component to limit the value to a preset maximum current value or less, and a current of the feedforward component for cooperative control of the actuator between the axes. An adjustment step of correcting the current reference value of the feedback component and the current reference value of the feedforward component from the adjustment rate obtained when adjusting the reference value.

  In the actuator control method, it is preferable that the pole of the disturbance observer is infinite.

  Further, the current limiting step calculates the adjustment rate of the position command so that the actuator does not deviate from the trajectory on all axes when the actuator causes current saturation on any one axis, and calculates the adjustment rate. Use to adjust the current reference value of the feedforward component for each axis.

Specifically, the current reference value of the i-axis feedback component is I _FBi , the current reference value of the i-axis feedforward component is I _FFi , the i-axis disturbance compensation current is I _RBi , and i When the current reference value to the actuator on the axis is I _i ^ref and the maximum torque current of the motor constituting the actuator is I _MAX , the current limiting step is for maintaining the robust control for each axis. The maximum acceleration / deceleration torque current I _accMAX (n) _i of the current sample n on the i-axis is calculated by the following equation:

When the sum of the current reference value I _FFi of the feed forward component and the current reference value I _FBi of the feedback component is larger than the maximum acceleration / deceleration torque current I _accMAX, Acceleration θ ^·· _max (n) _i is calculated (where K _t is the torque constant, J _ni is the nominal value of the moment of inertia of the i-axis),

The adjustment rate γ (n + 1) _i of the position command is calculated by the following formula (where Ts is the sampling time from the current sample to the next sample, θ ^{· res} (n) _i is the current sample of the i-axis Speed response value, θ ^cmd (n + 1) _i is a position command value in the next sample of the i-axis, θ ^cmd (n) _i is a position command value in the current sample of the i-axis),

Next, the minimum value γ (n + 1) min is extracted from the adjustment rate γ (n + 1) _i of each axis i,
Values I ^to _FFi (n) obtained by adjusting the current reference value of the feedforward component are calculated from the following equations (where K _tn is the nominal value of the torque constant, and f _i (t) is the position command for the i-th axis. The time function of the value, θ ^· _i ^res (n) is a velocity response value in the current sample on the i-axis, and θ ^cmd _i (n) is a position command value in the current sample on the i-axis).

Further, the adjustment step calculates a correction time t ^to (n + 1) in the next sample based on the adjustment rate γ (n + 1) of the position command from the following equation (where t (n) is the current time),

  It is preferable to correct the current reference value of the feedback component and the current reference value of the feedforward component based on the correction time.

  According to the apparatus of claim 1 and the method of claim 6, high-speed motion based on the robust acceleration control is achieved by causing the actuator to track the target value while realizing robust acceleration control that compensates for disturbance to the actuator. Trajectory tracking control is possible. In addition, under such control, when the current reference value to the actuator 2 exceeds the maximum current value and is saturated, the current reference value of the feedforward component that has a great influence on the position in the next sample is appropriate. By adjusting to, it is possible to perform follow-up control so that the actuators are not delayed from the trajectory on all axes including the actuators that have caused current saturation. In addition, the current reference value of the feedback component and the current reference value of the feedforward component calculated by the trajectory tracking control system and the tracking control step are corrected from the adjustment rate obtained when adjusting the current reference value of the feedforward component. The actuator can be coordinated between the axes.

  According to the apparatus of claim 2 and the method of claim 7, the transfer function from the acceleration reference value to the acceleration response value to the actuator can be regarded as 1, and the actuator follows the target command value almost without delay. Therefore, robust trajectory tracking control can be realized more completely.

  According to the apparatus of claim 3 and the method of claim 8, the current reference value of the feedforward component adjusted to the actuator necessary for locus tracking control of the actuator 2 of each axis is obtained from the adjustment rate of the position command. Easy to calculate.

  According to the apparatus of claim 4 and the method of claim 9, the maximum acceleration current of the axis is determined at the time of saturation, and from this determination result, robust acceleration control for the actuator is performed reliably, and the optimum position command time is determined. It is possible to calculate an adjustment rate γ that is a target reduction rate, and thus a current reference value of the adjusted feedforward component.

  According to the apparatus of the fifth aspect and the method of the tenth aspect, it is possible to adjust the optimum trajectory command for cooperatively operating the actuators of all the axes from the adjustment rate of the position command.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. First, an example of a robust acceleration control system using a disturbance observer applied by the actuator control apparatus of the present invention will be described with reference to FIG.

In the figure, reference numeral 1 denotes a control system of a robust acceleration control system. The acceleration control system 1 serving as an acceleration control unit applies a disturbance in an acceleration dimension to a control target actuator 2 by a disturbance observer 3 serving as a disturbance compensation unit. It is configured to perform torque compensation. The movable actuator 2 that converts energy into motive power is provided as a target for estimating the disturbance torque T _dis to the actuator 2, and this operates by inputting a current value i to a drive source such as a servo motor. It is. Actuator 2 has an actual speed response value θ ^· _res (hereinafter, except for figures and formulas, the first-order differential is indicated as “ ^• ” and the second-order differential is indicated as “ ^•• ” for convenience. ) Is converted into an electrical speed response signal and detected and output, a speed sensor 4 as a speed detecting means is attached. The speed sensor 4 includes a position sensor that detects the actual position of the actuator 2 and converts it into an electrical signal, and a pseudo-differentiator that obtains an estimated speed response signal by pseudo-differentiating the detection signal from the position sensor. It may be used instead of.

On the other hand, the disturbance observer 3 corresponds to an estimated value of disturbance torque ^ T _dis (hereinafter, “^” representing this estimated value except for the figures and mathematical expressions) as a command value for removing disturbance to the actuator 2. In particular, the estimated value ^ T _dis here can estimate not only the disturbance torque T _dis but also parameter variation errors such as inertia variation torque. Specifically, the current value i that is the operation amount to the actuator 2 is input, and the speed response value θ ^· _m from the speed sensor 4 is input, and the estimated value of the disturbance torque ^ T _dis from these values. Is calculated and output, and is actually composed of computer software. The disturbance observer 3 incorporates an inverse model equivalent to the actuator 2, converts the current value i into a first signal in a torque (force) unit, and the speed response value detected by the first signal and the speed sensor 4. an inverse model unit 6 for outputting a third signal obtained by comparing the second signal obtained by differentiating θ ^{· res} , and setting a cutoff frequency upon differentiation in the inverse model unit 6; A low-pass filter 7 that extracts a third signal of a low-frequency band component from the inverse model unit 6 and outputs the third signal as a disturbance torque estimation value ^ T _dis . Reference _numeral 8 denotes torque-current conversion means for inversely converting the disturbance torque estimated value ^ T _dis into a disturbance compensation current value I ^cmp in the same unit as the target current reference value I ^ref , and this disturbance compensation current value I ^The current input to the actuator 2 is added by adding the ^cmp and the current reference value I ^ref obtained by converting the reference acceleration reference value θ ^{·· ref} by the acceleration-current conversion means 9 with the adder 10. The value i is calculated. Thereby, in the acceleration control system 1 which consists of each structure of FIG. 1, the influence of the disturbance torque _Tdis added to the actuator 2 from the outside can be removed.

The actuator 2 includes a torque value calculated by the product of the input current value i and the torque constant K _t, the deviation between the disturbance torque T _dis, be the actual output torque values, the moment of inertia of the output torque value Dividing by J is the (angular) acceleration response value θ ^{·· res} that is actually output. Further, the speed response value θ ^{· res} from the actuator 2 is a value obtained by integrating (1 / s) the acceleration response value θ ^{·· res} by the integrator 11. In FIG. 1, the operation of the actuator 2 is shown for convenience in each component including the integrator 11.

Here, symbols in each component shown in FIG. 1 will be explained. J is a moment of inertia of the motor, K _t is a torque constant, and a suffix _n after the symbol is a nominal value. Means. Further, assuming that the cutoff frequency of the disturbance observer 3 is g, the disturbance torque estimated value ^ T _dis is calculated by the following equation.

In this way, in the acceleration control system 1 shown in FIG. 1, the disturbance torque estimated value ^ T ^dis calculated by the disturbance observer 3 is converted into the disturbance compensation I ^cmp by the torque-current conversion means 8 to perform feedback compensation. Based on the conventional current control system, a simple and robust acceleration control system can be realized.

1 is transmitted from the acceleration reference value θ ^{·· ref} to the acceleration response value θ ^{·· res} when the pole (cutoff frequency) g of the disturbance observer 3 is infinite (g = ∞). The function can be regarded as 1. FIG. 2 shows a block diagram of a trajectory tracking control system based on the robust acceleration control system of FIG. 1. The trajectory tracking control system 20 serving as a trajectory tracking control unit here is a target (position) command value. Since θ ^cmd is a time function, the speed command value θ ^{· cmd} is obtained by differentiation of the function of the target command value θ ^cmd by the first-order differentiator 21, and the acceleration command value θ ^{·· cmd} is obtained by the second-order differentiator 22. It has a feedforward system configuration obtained by second-order differentiation of the function of θ ^cmd . The acceleration control system 1 in FIG. 1 includes the main control system unit 23 and the integrator 11.

The trajectory tracking control system 20 has a position and speed control loop as a feedback system, and the deviation between the target command value θ ^cmd and the position response value θ ^res obtained by the subtractor 24 is the position -The speed conversion means 25 multiplies the gain Kp of the position control system to obtain a speed deviation value. This speed deviation value and the speed command value θ ^{· cmd} obtained by the first-order differentiator 21 are added by the adder 26, and The subtractor 27 calculates a deviation between the added value and the speed response value θ ^{· res} . The deviation value obtained by the subtractor 27 is multiplied by the speed control system gain Kv by the speed-acceleration conversion means 28, and the acceleration deviation value obtained thereby and the acceleration command value θ obtained by the second-order differentiator 22 are obtained. ^{.. Cmd} is added by the adder 29 to calculate the acceleration reference value θ ^{... Ref} to the acceleration control system 1.

In the locus tracking control system 20 in which the acceleration control system 1 is incorporated, a position control loop is provided in addition to the speed control loop, so that the actuator 2 is provided with a position response from the speed response value θ ^{· res for} convenience. An integrator 30 is provided as means for converting to the value θ ^res, and a position sensor 31 is provided as position detection means for detecting the position response value θ ^res of the actuator 2.

In FIG. 2, the transfer function from the position command value θ ^cmd to the position response value θ ^res can be expressed by the following equation.

From the above equation 22, the position response value θ ^res , speed response value θ ^{· res} , and acceleration response value θ ^{·· res} of the actuator 2 are the target position command value θ ^cmd , speed command value θ ^{· cmd} , and acceleration command value, respectively. Follows θ ^{·· cmd} almost without delay. Therefore, in the system as shown in FIG. 2, when the trajectory tracking control of a multi-axis system (for example, an XY table) is performed using the acceleration control system 1 that realizes robust acceleration control, high-speed and high-precision Motion control becomes possible.

Next, FIG. 3 shows a configuration in which a current limiter 41 is incorporated in the acceleration control system 1 shown in FIGS. 1 and 2 in consideration of the motor current limit (saturation). When the acceleration control system 1 of the robust acceleration control system is constructed using the disturbance observer 3 based on the conventional acceleration control system, an adder 10 is used to avoid torque saturation of the motor incorporated in the actuator 2. A current limiter 41 is provided for limiting the current value i obtained by adding the obtained current reference value I ^ref and the disturbance compensation current value I ^cmp within a certain range. Thereby, for example, when a current value i exceeding the maximum torque of the motor is generated due to a high-speed trajectory command value or a load condition, the current limiter 41 limits the current value i to a value within a certain range. Output to the actuator 2 to avoid torque saturation of the motor. However, when the limiter is considered in the current dimension as before, it is difficult to maintain robust acceleration control for the actuator 2 when the current is saturated.

Therefore, as shown in FIGS. 4 to 5, the disturbance observer 3 in FIG. 3 is equivalently converted. In the equivalent conversion diagram shown in FIG. 4, the acceleration-current conversion means 9 divides the acceleration reference value θ ^{·· ref} by the nominal value K _tn of the torque constant, and multiplies this by the nominal value J _n of the moment of inertia. The current reference value I ^ref is obtained. The current limiter 41 acts on this current reference value I ^ref . The current reference value I ^{ref that} has passed through the current limiter 41 is multiplied by the nominal value K _{tn of the} torque constant by the current-torque conversion means 42, and the torque reference value resulting from this multiplication and an element equivalent to the disturbance torque T _dis A torque response value of the actuator 2 input to the torque-speed conversion means 45 is obtained by subtracting the estimated torque value obtained by multiplying the sensitivity function of the mathematical formula shown in 43 by the equivalent subtractor 44. The torque response value is converted into a speed response value θ ^{· res} by the torque-speed conversion means 45, and this speed response value θ ^{· res} is further converted into a position response value θ ^res by an integrator 30 which is a speed-position conversion means. Is done.

  Note that g ′ shown in the element 43 of FIGS. 4 and 5 can be expressed by the following equation.

  On the other hand, in FIG. 5, the current limiting by the current limiter 41 is equivalently replaced with the acceleration limiting by the acceleration limiter 51. 5 that the current limiter 41 shown in FIG. 3 can be expressed as the acceleration limiter 51 in the robust acceleration control system. Therefore, in this embodiment, an acceleration limiting algorithm is incorporated in the computer constituting the control device for the actuator 2 as a technique that considers the limiter in the acceleration dimension.

The acceleration limiter 51 configures the actuator 2 by limiting the acceleration reference value θ ^{·· ref} to a value within a certain range when an acceleration reference value θ ^{·· ref} that exceeds the maximum torque of the motor occurs. This avoids saturation of the motor in the acceleration dimension. The acceleration reference value θ ^{·· ref} obtained through the acceleration limiter 51 and the estimated acceleration value corresponding to the disturbance calculated by multiplying the disturbance torque T _dis and the sensitivity function of the mathematical formula shown in the element 52 are obtained. Subtracted by the equivalent subtractor 53 is the acceleration response value of the actuator 2 that is input to the integrator 54 that is acceleration-speed conversion means. The acceleration response value is converted into a speed response value θ ^{· res} by the integrator 54, and this speed response value θ ^{· res} is further converted into a position response value θ ^res by an integrator 30 which is a speed-position conversion means.

FIG. 6 shows the configuration of acceleration limiting means 61 for executing the acceleration limiting algorithm. The acceleration limiting means 61 includes disturbance compensation acceleration θ ^·· _RB ^cmp , acceleration reference value θ ^·· _FB ^ref of the feedback system of the acceleration control system 1, and acceleration reference value θ ^·· _FF of the feed forward system of the acceleration control system 1. a ^ref as input, between _{^{·· ref = θ ·· RB cmp +}} θ ·· FB ref + θ ·· FF ref acceleration reference value theta is the sum of each of these acceleration values, the maximum acceleration theta ^{· ·} _MAX of the motor Based on the comparison result, a final acceleration reference value θ ^{··· ˜ref} (according to acceleration saturation) is calculated (hereinafter, the final value is expressed as “ ^˜ ” for convenience except for figures and mathematical expressions, and is also written after the corresponding symbol). To do.

Here, the disturbance compensation acceleration θ ^·· _RB ^cmp for realizing the robust acceleration control is the disturbance torque estimated value ^ T _dis calculated by the disturbance observer 3 and the torque-acceleration converting means 62 uses the nominal value J of the moment of inertia. Calculated by multiplying by _n . Also, the acceleration reference value θ ^·· _FB ^ref of the feedback system is obtained by calculating the deviation between the target feedback reference value and the response value obtained from the actual operation of the actuator 2 by the subtractor 63, Calculated by multiplying the appropriate feedback gain G _FB by element 64. The feedback reference value or response value may be other than acceleration, and the acceleration reference value θ ^·· _FB ^ref may be calculated from the element 64 using the feedback reference value and response value.

The acceleration limiter 51 determines the final acceleration reference value to the actuator 2 when the total acceleration reference value θ ^{·· ref} calculated by the adder 65 exceeds a preset maximum acceleration θ ^·· _MAX . θ ^{·· ~ ref} is ^limited to the maximum acceleration θ ^·· _MAX , but this is actually limited to the disturbance compensation acceleration θ ^·· _RB ^cmp to adjust the disturbance compensation acceleration θ ^·· _RB ^cmp . To adjust the RB term adjustment rate k1 to be multiplied and the acceleration reference value θ ^·· _FB ^ref of the feedback system, the FB term adjustment rate k2 multiplied by the acceleration reference value θ ^·· _FB ^ref and the feed forward system to condition the acceleration reference value theta ^{· ·} _FF ^ref, this acceleration reference value theta ^{· ·} _FF is multiplied ^ref FF term adjustment rate k3 determined, summing the values obtained by the adjustment by the adder 65 Therefore, it has a function to calculate the final acceleration reference value θ ^{·· ~ ref} . In FIG. 6, elements 66 to 68 are shown as means for determining the RB term adjustment rate k1, the FB term adjustment rate k2, and the FF term adjustment rate k3.

The acceleration limiting algorithm realized by the acceleration limiting means 61 configured as described above determines the adjustment rates k1, k2, and k3 by the following procedure, and finally adjusts the acceleration reference value θ ^{·・ Calculate ~ ref} .

First, in the first procedure, the disturbance reference acceleration θ ^·· _RB ^cmp , the acceleration reference value θ ^·· _FB ^ref , and the acceleration reference value θ ^·· _FF ^ref are summed by the adder 65 as they are, and the acceleration reference value θ ^{·· ref} It is determined whether (= θ ^·· _RB ^cmp + θ ^·· _FB ^ref + θ ^·· _FF ^ref ) exceeds the maximum acceleration θ ^·· _MAX . Here, if the acceleration reference value θ ^{·· ref} does not exceed the maximum acceleration θ ^·· _MAX , that is, if θ ^{·· ref} ≤ θ ^·· _MAX , each adjustment rate k1, k2, k3 is assumed to be 1. (K1 = k2 = k3 = 1), the total acceleration reference value θ ^{·· ref} is output as it is as the final acceleration reference value θ ^{··· -ref} .

On the other hand, if the acceleration reference value θ ^{·· ref} exceeds the maximum acceleration θ ^·· _MAX , that is, if θ ^{·· ref} > θ ^·· _MAX , each adjustment rate k1 is satisfied so that the relationship of the following equation is satisfied. , K2, k3 are determined so that the final acceleration reference value θ ^{··· ref} is equal to the maximum acceleration θ ^·· _MAX .

Here, an example of adjusting each of the adjustment rates k1, k2, and k3 when θ ^{·· ref} > θ ^·· _MAX will be described. In the feedforward adjustment type shown in the following equation, a feedforward (FF) term is used. By adjusting a certain k3 · θ ^·· _FB ^ref , the size of other feedback (FB) terms k2, ^···· _FB ^ref and robust disturbance compensation (RB) terms k1, ^···· _RB ^cmp is maintained. .

  In another feedback adjustment type, the size of other FF terms and RB terms is maintained by adjusting the FB terms. This can be expressed as:

  Further, in the feedback + feedforward adjustment type, the size of the RB term is maintained by adjusting the FB term + FF term. This can be expressed as:

  The trajectory tracking control system 20 performs feedback and feedforward control of the actuator 2 based on the adjusted FB term and FF term. In the multi-axis system, the expansion to the multi-axis system can be expanded by considering the values of the three adjustment factors k1, k2, and k3.

  Next, an example of a multi-axis system configuration incorporating the concept of the acceleration limiting algorithm shown in FIG. 5 will be described with reference to FIGS. FIG. 7 shows the external appearance of a multi-axis experimental system. Reference numeral 71 denotes a computer as an actuator control device, which incorporates an algorithm for realizing the locus tracking control system based on the above-described robust acceleration control. Yes. An input / output port of the computer 71 is connected to a PCI bus bridge 73 serving as an input / output interface via a PCI bus 72 serving as connection means. The PCI bus bridge 73 includes an AD (analog-digital conversion) board, a counter board, a DIO (digital input / output) board, and the like.

  On the other hand, the actuator 2 in this experimental system is provided with two actuators 2A and 2B corresponding to the two tables 78 moving in the X direction and the Y direction, respectively. In the X-axis actuator 2A, a ball screw 77A is connected to a rotation shaft of an AC motor 75A that is a drive source by using a joint 76A, and a table 78A that is a controlled object that is movable is screwed to the ball screw 77A. Configured. Similarly, the Y-axis actuator 2B has a configuration in which a ball screw 77B is connected to a rotation shaft of an AC motor 75B as a drive source using a joint 76B, and a movable table 78B is screwed to the ball screw 77B. Is done.

  Then, the control signal output from the computer 71 to the inverter 79A through the PCI bus bridge 73 converts the inverter 79A into an alternating current sufficient to drive the AC motor 75A, and supplies it to the AC motor 75A. The ball screw 77A rotates in the forward direction or the reverse direction, and the table 78A moves along the axial direction of the ball screw 77A. Similarly, a control signal output from the computer 71 to another inverter 79B through the PCI bus bridge 73 converts the inverter 79B into an alternating current sufficient to drive the AC motor 75B, and supplies this to the AC motor 75B. Accordingly, the ball screw 77B rotates in the forward direction or the reverse direction, and the table 78B moves along the axial direction of the ball screw 77B.

  The current supplied from the inverter 79A to the AC motor 75A is detected by the current sensor 81A, and the rotational speed of the AC motor 75A is detected by the encoder 82A and is taken into the computer 71 through the PCI bus bridge 73. Similarly, the current supplied from the inverter 79B to the AC motor 75B is detected by the current sensor 81B, and the rotational speed of the AC motor 75B is detected by the encoder 82B, and is taken into the computer 71 through the PCI bus bridge 73. It has become.

  In this case, an example in which the tables 78A and 78B are moved along the two axes X and Y has been shown. However, a similar actuator 2A, and an accompanying inverter 79A, current sensor 81A, and encoder 82A are added. If so, the present invention can be applied to a multi-axis system configuration having three or more axes.

By the way, in the multi-axis system as shown in FIG. 7, when locus tracking control based on robust acceleration control is performed, for example, when the actuator 2A on one axis causes acceleration saturation, an acceleration limiter as shown in FIG. Even if 51 is used, a trajectory error occurs. The main reason for this is that the torque current reference value I ^{ref of the} shaft that has not caused acceleration saturation needs to be adjusted so as not to deviate from the target locus.

  FIG. 8 shows an example of a trajectory tracking system that takes into account torque saturation based on multi-axis robust acceleration control according to a conventional method, and each component other than the actuator 2 in this figure is programmed in the computer 71. Built in as The tracking control system shown here is shown only for one-axis actuator 2, and a similar tracking control system is provided in a multi-axis system according to the number of axes.

Reference numeral 85 denotes a torque current limiter based on the acceleration limiting means 61, which includes a feedforward acceleration reference value θ ^·· _FF ^ref and a feedback acceleration reference value θ ^·· _FB ^ref supplied from the trajectory tracking control system 20. And an adder 65 for adding the disturbance compensation acceleration θ ^·· _RB ^cmp obtained by converting the estimated disturbance torque ^ T ^dis from the disturbance observer 3 by the torque-acceleration converting means 62, and adding by the adder 65 Acceleration-current conversion means 62 for converting the resulting acceleration reference value θ ^{·· ref} to a torque current reference value I ^ref , a current limiter 41 for limiting the torque current reference value I ^ref to a maximum torque current I _MAX or less, The final current reference values I to ^ref obtained by the current limiter 41 are supplied to the actuator 2.

Reference numeral 86 denotes a delay adjustment means including a motion command adjustment algorithm (Adjustment algorithm of motion command) for adjusting a dynamic delay generated by torque saturation. This is because when torque saturation occurs in the trajectory tracking control, if the torque current reference value I ^ref of each axis is adjusted by the torque current limiter 85, the torque becomes insufficient at the current sample, and the target trajectory cannot be tracked. In order to avoid an increase in the trajectory error, a position command value θ ^cmd (t), which is a target trajectory command value at the original time t, based on the position command adjustment rate γ (n + 1) obtained by the torque current limiter 85. The speed command value θ ^{· cmd} (t) and the acceleration command value θ ^{·· cmd} (t) are ^converted into a position command value θ ^cmd (t ^to ), a speed command value θ ^{· cmd} (t ^To ), acceleration correction command value θ ^{·· cmd} (t ^to ). The acceleration correction command value θ ^{·· cmd} (t ^to ) is directly used as the feedforward system acceleration reference value θ ^·· _FF ^ref , and the position correction command value θ ^cmd (t ^to ) and the speed correction command value θ ^{· cmd.} and (t ^~), on the basis of the deviation between the feedback position response value theta ^res and velocity response value theta ^{· res,} acceleration reference value theta ^{· ·} _FB ^ref of the feedback system are calculated.

FIG. 9 is a block diagram showing torque saturation of the i-th axis in robust acceleration control when the number of axes of the multi-axis system is N (i = 1 to N). The torque current limiter 85 here is a rewrite of the configuration shown in FIG. 8 to sum the current components, and the i-axis acceleration command value θ ^{·· cmd} _i (the acceleration reference value θ ^·· _FF). (corresponding to ^ref ) is converted into an acceleration command current I _FFi of the feedforward component of the acceleration command value by the acceleration-current conversion means 88 as in the following equation.

Further, a value obtained by converting the deviation between the _i- th axis speed reference value ω ^ref _i calculated by the subtractor 89 and the speed response value ω ^res _i into an acceleration reference value by the speed-acceleration conversion means 28 (the acceleration reference value). (corresponding to θ ^·· _FB ^ref ) is converted into an acceleration / deceleration torque current I _FBi as a feedback component that compensates for the speed deviation by another acceleration-current conversion means 90 as in the following equation.

Further disturbance torque estimated value ^ T _dis, as in the following equation, the torque - is converted to the disturbance compensation torque current I _RBi to achieve a robust acceleration control by the current converting unit 91.

The adder 65 calculates an i-axis torque current reference value I _refi (= I _FFi + I _FBi + I _RBi ) obtained by adding the acceleration command current I _FFi , acceleration / deceleration torque current I _FBi , and disturbance compensation torque current I _RBi. calculated, the current limiter 41, the torque current reference value I _refi an AC motor 75 for each axis (75A, 75B, ...) maximum torque current to be equal to or less than I _MAX of the final current reference values ^{I ～Ref} Limit.

In the method of the track following system shown in FIG. 8, when the actuator 2 generates acceleration saturation, the magnitude of the robust disturbance compensation (RB) term described above is maintained, and the feedforward (FF) term and the feedback (FB) term are maintained. As a term to be emphasized, the respective axes are coordinated with a minimum allowable rate α _min of torque current. Specifically, the torque current limiter 85 determines the acceleration torque current adjustment rate α for each axis as follows.

First, _when the absolute value of the i-axis disturbance compensation torque current I _RBi exceeds the maximum torque current I _MAX (I _MAX <| I _RBi |), the i-axis adjustment rate α _{i is set} to zero. (Α _i = 0).

Next, _{even if} the absolute value of the i-axis disturbance compensation torque current I _RBi does not exceed the maximum torque current I _MAX , the absolute value of the i-axis torque current reference value I _refi is the maximum torque. When the current I _MAX is exceeded (I _MAX <| I _FBi + I _FFi + I _RBi |), the adjustment rate α _i for the i-axis is calculated by the following equation.

That is, if I _refi exceeds I _MAX , the adjustment rate α _i is calculated using the upper equation, and if I _refi is less than −I _MAX , the adjustment rate α _i is calculated using the lower equation.

Further, _when the absolute value of the i-th axis torque current reference value I _refi is equal to or less than the maximum torque current I _MAX (I _MAX ≧ | I _FBi + I _FFi + I _RBi |), the i-th axis adjustment rate α _i Is set to 1 (α _i = 1).

When considering the coordination of the torque current reference value I _refi of each axis, the torque current limiter 85 determines the minimum value (minimum allowable ratio) α from among the adjustment rates α _i (i = 1 to N) determined for each axis. _Min is extracted, and the torque current reference value I _refi of each axis is adjusted using the minimum value α _min as _shown in the following equation. Here, the vector direction of the acceleration / deceleration torque current component of each axis is maintained, and only the magnitude thereof is adjusted.

Then, the actuator 2 between the axes is cooperatively controlled by the final torque current reference values I ^to _refi thus obtained.

  10A to 10D show experimental results when the conventional locus tracking system shown in FIG. 8 is incorporated in the computer 71 shown in FIG. 7 and a circular locus is drawn. Here, a load of 8.9 kg is applied to the Y axis, and a circular locus with a radius of 100 mm is drawn. FIG. 10A traces the position θ, (angular) speed ω, and command value and response value of the current iq in the X-axis actuator 2A, and FIG. 10B shows the position θ, (angle) in the Y-axis actuator 2B. The command value and response value of speed ω and current iq are traced. In addition, FIG. 10C shows the tracking error and the radius error of the X axis and the Y axis at this time. Further, FIG. 10D is an enlarged view of the trajectory response at the time of starting and stopping.

In the conventional trajectory tracking system, as shown in FIG. 10C, when the torque saturation occurs, the tracking error of the X axis is particularly large. As a result, as shown in FIG. 10D, the target sinusoidal XY The trajectory response value ^pres deviates greatly from the trajectory reference value p ^ref , and the trajectory error is increased.

The reason for this is that, as shown in the following equation, if the acceleration α, the position θ, and their correction values are α ^˜ , θ ^˜ , the acceleration and position allowances expressed as the ratio of the original value and the correction value This is because the trajectories cannot be matched even if the axes are coordinated with the minimum allowable rate α _min of torque current as in the conventional method. Therefore, when performing trajectory tracking control of a multi-axis system, each axis must be coordinated with the position dimension instead of the current dimension.

Therefore, a system that can coordinately control each axis in the dimension of such a position and its method will be described in detail below. FIG. 11 shows acceleration saturation of the i-axis in the newly proposed robust acceleration control. Here, a new torque current limiter 95 having a maximum acceleration current limiter 94 is incorporated in place of the torque current limiter 85 described above. ing. In FIG. 11, I _FBi is an acceleration / deceleration torque current as a feedback component that compensates for a speed deviation, I _FFi is an acceleration command current as a feedforward component of an acceleration command value, and I _RBi realizes robust acceleration control. This is a disturbance compensation torque current. The acceleration command current I _FFi obtained by conversion by the acceleration-current conversion means 88 and the acceleration / deceleration torque current I _FBi obtained by conversion by the acceleration-current conversion means 90 are added by an adder 65A, and this adder 65A And the disturbance compensation torque current I _RBi obtained by the conversion by the torque-current conversion means 91 are added by the adder 65B configured in the maximum acceleration current limiter 94. That is, in the locus tracking control system based on the robust acceleration control system, unless the torque current saturation occurs, the torque current reference value I _refi of the i-axis AC motor 75 is added by the adders 65A and 65B to the acceleration command current I _FFi. , _Acceleration / deceleration torque current I _FBi , and disturbance compensation torque current I _RBi .

Here, since the disturbance compensation torque current I _RBi is an essential component for robust control, it is necessary to maintain the magnitude for each axis. Further, since the acceleration / deceleration torque current I _FBi is necessary to compensate for a deviation caused by a modeling error that differs depending on each axis, the magnitude of the acceleration / deceleration torque current I _FBi must be maintained for each axis. Furthermore, the acceleration command current _IFFi is also a term necessary for following the target locus, and is a term that greatly affects the position at the next sample (n + 1). Thus, it can be understood that each current reference value has an important role in the trajectory tracking control system based on the robust acceleration control.

As described above, in the trajectory tracking control system based on the robust acceleration control, if the magnitudes of the acceleration command current I _FFi , the acceleration / deceleration torque current I _FBi , and the disturbance compensation torque current I _RBi do not maintain the magnitude, the trajectory tracking is performed. I can't. Therefore, here, attention is focused on the acceleration command current I _FFi necessary for following the target locus. The FF term related to the acceleration command current _IFFi greatly affects the position in the next sample. If torque current saturation occurs, the position command in the next sample is adjusted for each axis so that no trajectory error occurs, so that the acceleration command current I _FFi of the current sample is the optimum acceleration command current I ^~ Need to adjust to _FFi . Therefore, instead of the conventional method of coordinating the current reference value with the minimum allowable rate α _min , when torque current saturation occurs even in one of a plurality of axes, the optimum acceleration command current for each axis I ^to _FFi are calculated by back-calculating from the position command adjusted so as not to deviate from the locus, and the preferred method unique to this embodiment for adjusting the torque current reference value for each axis is improved.

Here, the new torque current limiter 95 shown in FIG. 11 is applied to the tables 78A and 78B that move in the two-axis XY directions as shown in FIG. As an example, when the torque of the Y-axis actuator 2B is saturated, the position command value θ _x ^cmd (n + 1) in the next sample (n + 1) for the X-axis actuator 2A that does not cause torque saturation is The torque current limiter 95 can be calculated from the equation.

Where T _s is the sampling time of the control system, and θ ^{· avg} (n) is the average speed from the current sample (n) to the next sample (n + 1). However, since the Y-axis actuator 2B actually has torque saturation, the position command value θ _x ^cmd (n + 1) at the next sample is corrected by the delay adjusting means 86 as shown in the following equation. it is necessary to adjust the ^{_{θ ~ x cmd (n + 1}} ).

_{Incidentally, f} x (t) is the time function of the position command value θ _x ^cmd (n) of x-axis, also γ (n + 1) is a position command adjustment rate which will be described later.

Therefore, if I ^to _FFx (n) is calculated such that the above Expression 35 and Expression 36 are equal as shown in the following expression, this is the optimum acceleration command current in the current sample.

The optimum acceleration command currents I ^to _FFx (n) can be easily calculated by the torque current limiter 95 from the following equation.

Therefore, the final x-axis corrected torque current reference values I ^to _refx obtained by the torque current limiter 95 are _expressed by the following equations.

  From the above, in the torque current limiter 95 provided with the newly proposed feed forward current correction type torque limiting algorithm, the torque saturation countermeasure is executed in the following procedure.

First, in the first step, the position command adjustment rate γ (n + 1) _i is calculated for each axis as follows. The position command adjustment rate γ (n + 1) _i is an allowable rate in the position dimension. Here, I _MAX is the maximum torque current of the AC motor 75. In the robust acceleration control system, there is a limiter based on the maximum acceleration / deceleration torque current I _accMAX for the purpose of maintaining robust control. The maximum acceleration current limiter 94 _derives the maximum acceleration / deceleration torque current I _accMAX by the following equation in consideration of the direction of the disturbance compensation torque current I _RBi .

That is, if the torque current reference value I _i ^{ref on} the i-th axis exceeds I _MAX , the maximum acceleration / deceleration torque current I _accMAX (n) _i of the current sample is calculated by the above equation, and I _i ^ref is −I _MAX If it is less, I _accMAX (n) _i is calculated by the lower equation. Therefore, when the sum of the acceleration command current I _FFi and the acceleration / deceleration torque current I _FBi is larger than the maximum acceleration / deceleration torque current I _accMAX (I _FFi + I _RBi > I _accMAX ), that is, the torque current reference value I _refi When the absolute value exceeds the maximum torque current I _MAX , the maximum acceleration current limiter 94 calculates the maximum acceleration θ ^·· _max (n) _i for each axis by the following equation.

Then, a position command adjustment rate γ (n + 1) _i is calculated for each axis from the maximum acceleration θ ^·· _max (n) _i .

In the above equation, Ts is the sampling time from the current sample to the next sample, θ ^{· res} (n) _i is the velocity response value in the current sample on the i-axis, and θ ^cmd (n + 1) _i is the next on the i-axis. The position command value in the current sample, θ ^cmd (n) _i is the position command value in the current sample on the i-axis.

On the other hand, _when the absolute value of the torque current reference value I _refi is less than the maximum torque current I _MAX , the maximum acceleration current limiter 94 sets the position command adjustment rate γ (n + 1) _i to 1.

In the next second step, the torque current limiter 95 extracts the minimum position command adjustment rate γ (n + 1) min from the position command adjustment rate γ (n + 1) _i of each axis i (i = 1 to N). This is the position command adjustment rate γ (n + 1). That is, γ (n + 1) = γ (n + 1) min, so that it can be matched in the dimension of the position to the axis having the largest allowable amount when the torque current is saturated, and the respective axes are coordinated.

Further, in the third step, the torque current limiter 95 calculates the feedforward component acceleration command currents I ^to _FFi (n) corrected for all the axes i by the following equations.

In the above equation, θ ^· _i ^res (n) is a velocity response value in the current sample on the i-axis, and θ ^cmd _i (n) is a position command value in the current sample on the i-axis.

Then, by using the corrected acceleration command currents I ^to _FFi (n), final torque current reference values I ^to _refi for the actuators 2 are calculated.

The procedure in the third step is when the position command adjustment rate γ (n + 1) on the axis _i is different from the position command adjustment rate γ (n + 1) _i obtained in the second step (ie, γ (n + 1) ) ≠ γ (n + 1) i), the otherwise (γ (n + 1) = γ (n + 1) i), the ^{I ~} _refi = ^I _MAX.

By the above procedure, when the torque saturation of the multi-axis system is calculated, the optimum acceleration command currents I ^to _FFi are calculated, and each actuator 2 is controlled by the final torque current reference values I ^to _refi calculated based on the acceleration command currents I ^to _FFi . Trajectory errors can be suppressed.

  Next, a description will be given of the delay adjusting means 86 that enables a coordinated operation between the respective axes by outputting the command value corrected for each axis i by using the position command adjustment rate γ (n + 1) as an input.

FIG. 12 illustrates the adjustment of the position command value in the next sample based on the position command vector adjustment rate γ (n + 1) obtained by the equation 42. Here, θ ^vector (n + 1) shows the position command vector at the next sample, which is the next sample position command value θ ^cmd (n + 1), the current sample position command value θ ^cmd (n) It is obtained by the subtracted value (θ ^vector (n + 1) = θ ^cmd (n + 1) −θ ^cmd (n)). In addition, θ 1 to ^vector (n + 1) indicates a corrected position command vector in the next sample, which is ^{calculated from} the corrected position command value θ 1 to ^cmd (n + 1) of the next sample and the current command position command value θ ^cmd ( n) is obtained by subtracting (θ ^vector (n + 1) = θ to ^cmd (n + 1) −θ ^cmd (n)). Position command adjustment rate γ (n + 1) can be calculated by dividing the corrected position command vector ^θ ~vector (n + 1) the position command vector θ ^vector (n + 1).

In FIG. 12, the correction time t ^to (n + 1) in the next sample can be calculated from the time function f (t) of the position command by the following equation.

In the figure, t ^to '(n + 1) is a correction time in the current sampling when the position command between each sampling is linearly approximated. If the sampling time Ts≈0, t ^to (n + 1) = t ^to ' Since there is no problem even if (n + 1), this expression can be replaced as follows.

Therefore, in the motion command adjustment algorithm executed by the delay adjustment means 86, the corrected position command value θ ^{~ cmd} (n + 1) _i in the next sample of each axis _i is generated as shown in the following equation using the above equation. . Further, a corrected speed command value θ ^{~ · cmd} (n + 1) _i and a corrected acceleration command value θ ^{·· cmd} (n + 1) _i in the next sample of each axis _i are generated. By outputting these values to the trajectory tracking control system 20, the actuator 2 of each axis i is cooperatively operated.

Note that f _i (t) is a time function of the position command value in each axis, g _i (t) is a time function of the speed command value in each axis, and h _i (t) is an acceleration command value in each axis. Is the time function of Ts is the sampling time of the control system. In the derivation of the corrected speed command value θ ^{~ · cmd} (n + 1) _i and the corrected acceleration command value θ ^{·· cmd} (n + 1), the upper equation is the adjustment rate γ (n + 1) _i <1 and γ (n + 1) When min = γ (n + 1) _i , and when the adjustment rate γ (n + 1) _i <1, and γ (n + 1) min ≠ γ (n + 1) _i , the lower equation is applied.

  FIGS. 13A to 13D show experimental results when the preferred locus tracking system of the present embodiment shown in FIG. 11 is incorporated in the computer 71 shown in FIG. 7 and a circular locus is drawn. That is, each of these figures is an experimental result of trajectory tracking control based on robust acceleration control in consideration of torque saturation to which the method of this embodiment is applied. Note that the experimental condition is that a load of 8.9 kg is applied to the Y-axis and a circular locus with a radius of 100 mm is drawn as in the conventional locus tracking system.

  13A traces the position θ, (angular) speed ω, and command value and response value of the current iq in the X-axis actuator 2A, and FIG. 13B shows the position θ, (corner) in the Y-axis actuator 2B. The command value and response value of speed ω and current iq are traced. In addition, FIG. 13C shows the tracking error and the radius error of the X axis and the Y axis at this time. Further, FIG. 13D is an enlarged view of the trajectory response at the time of starting / stopping.

  Here, it is recognized that torque saturation occurs on the Y axis, but the torque current limiter 95 including the maximum acceleration current limiter 94 is corrected to an optimal command value, so that a good trajectory response is obtained. be able to.

  By the way, when the torque saturation of the actuator 2 occurs in the sudden stop operation or the reversing operation in the moving direction, the deceleration torque becomes insufficient. As a result, the vehicle cannot stop at the target stop position or deviates from the target trajectory. In order to avoid such a phenomenon, it is necessary to brake in advance. Then, the brake mode algorithm (Braking Mode Algorithm) proposed in the present embodiment will be described next.

FIG. 14 shows an ideal deceleration stop response when the actuator 2 is decelerated with the maximum torque. In the figure, the target stop position (target position) must be maintained. However, since the speed response ω0 changes depending on the load state, the target stop position (target position) changes when the brake is started from the brake start time t _brake . Therefore, to calculate the moving distance .DELTA..theta _err to stop time t _stop the brake start time t _brake that varies with load conditions, to estimate the deviation brake start time from t _brake and the stop position. Thus, by applying the _brake at the estimated brake start time t _brake , the target stop position (target position) can be stopped without being excessively performed.

The deviation Θ _err from the stop position can be calculated by the following equation from the relationship between the target stop position (target position) and the position response value θ ^res .

  The following equation is a deceleration start threshold value obtained from the optimum profile during deceleration shown in FIG. 14 in consideration of Coulomb friction.

In the proposed brake mode algorithm, when the deviation Θ _{err from the} stop position is within the movement distance ΔΘ _err , it is determined that the brake start time t _brake has been reached, and the stop operation is performed with the maximum deceleration torque.

  FIG. 15 is an example of a trajectory tracking control system that takes into account torque saturation by applying the concept of FIG. 11. Here, disturbance compensation is performed by the maximum acceleration current limiter 94 that executes the torque limiting algorithm described above and the disturbance observer 3. A switch 98 is interposed between the actuator 2 and a switch control means 99 capable of executing the brake mode algorithm, so that the switching operation of the switch 98 is appropriately controlled.

Here, the corrected acceleration command value theta ^{· · cmd} from delay adjusting unit 86 (t ^~), speed - acceleration reference value theta ^{· ·} _FB ^ref from the acceleration conversion means 28 are added by the adder 101, the The added value is converted into a current value i _acc (corresponding to the above-mentioned I _FFi + I _FBi ) by the acceleration-current conversion means 90 and the current value i _cmp (the above-mentioned I _cmp converted by the torque-current conversion means 91). and corresponding) to _RBi, the acceleration - current value i _acc and the adder torque current reference value obtained by adding at 65B I _ref from the current converter 90 is output to the maximum acceleration current limiter 94. A speed limiter 102 is connected between the adders 26 and 27.

Further, the maximum acceleration current limiter 94 here uses the speed-current conversion means 104 to convert the deviation between the speed reference value ω ^ref = 0 and the speed response value θ ^{· res} obtained by the subtractor 103 by switching the switch 98. Based on the converted current value, the actuator 2 has a function of speed control.

The switch control means 99 switches the switch 98 with the axis that has reached the deceleration start time as the high-speed stop motion axis, and performs speed control with the speed reference value ω ′ ^ref = 0. At this time, the shaft generates a maximum deceleration torque and tries to stop, so torque saturation occurs, and the torque current limiter 95 executes a “torque limitation (limitation) algorithm”, which is a correction maximum opposite to that during acceleration. The acceleration command θ ^·· _max (n) is calculated by Equation 41. As a result, a position command vector adjustment rate γ (n + 1) is generated. Therefore, the correction time t ^to (n + 1) of the next sample is obtained by Equation 45, and the command value of the other axis of the next sample is adjusted by the delay adjusting means 86. In this way, it is possible to perform a stop operation while cooperating with the stop operation axis, and control so as not to deviate from the target locus.

Thereafter, when the position deviation theta _err becomes equal to or less than .DELTA..theta ^~ _err is sufficiently small value, it is determined that the positioning completion time, the switch control means 99 for performing a braking mode algorithm, normal position control switch 98 The positioning is completed by returning to the system.

As described above, the present embodiment includes the disturbance observer 3 for estimating the disturbance to the actuator 2 in the actuator control device and the control method for individually controlling the operations of the plurality of actuators 2 movable on each axis. The disturbance estimated value obtained in 3, that is, the disturbance torque estimated value ^ T _dis is converted into a disturbance compensating torque current I _RBi which is a disturbance compensating current, and the current to the actuator 2 obtained by feedback compensation with the disturbance compensating torque current I _RBi is obtained. An acceleration control system 1 that executes an acceleration control step for performing robust acceleration control of the actuator 2 based on the reference value I ^ref , and an acceleration / deceleration torque current I _FBi that is a current reference value of a feedback component in order to cause the actuator 2 to follow the locus. calculates an acceleration command current I _FFi is a current reference value of the feed-forward component A trajectory tracking control system 20 for executing a trace tracking step, feedback component of acceleration and deceleration torque current I _FBi, acceleration command current I _FFi feedforward component, and the disturbance compensation torque current I current reference values _RBi to the actuator 2 which is the sum of A torque current limiter 95 as a current limiter for executing a current limiting step for adjusting the acceleration command current I _FFi of the feedforward component in order to limit I ^ref to a preset maximum current value I _MAX or less; In order to coordinately control the actuator 2 between the axes, the trajectory tracking control system 20 calculates from the adjustment rate obtained when adjusting the acceleration command current _IFFi of the feedforward component, that is, the position command adjustment rate γ (n + 1) _i. Acceleration / deceleration torque current I _{FBi of} feedback component and acceleration of feedforward component And a delay adjusting unit 86 as an adjusting unit that executes an adjusting step for correcting the command current I _FFi .

In this way, the acceleration control system 1 realizes robust acceleration control that compensates for the disturbance to the actuator 2, and the locus tracking control system 20 causes the actuator 2 to follow the locus to the target value, thereby achieving high speed based on the robust acceleration control. Enables motion tracking control. In addition, when the current reference value I ^ref to the actuator 2 is saturated beyond the maximum current value I _MAX under such control, the acceleration of the feedforward component that has a great influence on the position in the next sample, in particular. By appropriately adjusting the command current I _FFi , it is possible to perform follow-up control so that the actuator 2 is not delayed from the trajectory on all axes including the actuator 2 that has caused current saturation. Further, from the adjustment rate obtained when adjusting the feedforward component acceleration command current I _FFi , the feedback component acceleration / deceleration torque current I _FBi calculated by the locus tracking control system 20 and the feedforward component acceleration command current I _FFi are calculated. Can be corrected, and the actuator 2 can be operated cooperatively between the axes.

Further, the disturbance observer 3 in the present embodiment is configured so that the pole g is infinite when the disturbance torque estimated value ^ T _dis is calculated by the relational expression as shown in the above equation (21). In this way, the transfer function from the acceleration reference value θ ^{·· ref} to the acceleration response value θ ^{·· res} to the actuator 2 can be regarded as 1, and the actuator 2 follows the target command value almost without delay. Therefore, robust trajectory tracking control can be realized more completely.

Further, the torque current limiter 95 for each axis in the present embodiment, when executing the current limiting step, if the actuator 2 causes current saturation on any one axis, the actuator 2 does not deviate from the locus on all axes. The position command adjustment rate γ (n + 1) _i is calculated, and the final acceleration command current adjusted from the feed-forward component acceleration command current _{IFFi for} each axis using this adjustment rate γ (n + 1) _i. Since I ^to _FFi are calculated, the adjusted acceleration command currents I ^to _FFi necessary for the trajectory tracking control of the actuator 2 of each axis can be easily calculated from the adjustment rate γ (n + 1) _i of the position command.

Specifically, the torque current limiter 95 for each axis calculates the maximum acceleration / deceleration torque current I _accMAX (n) _i in the current sample n of the i-th axis for maintaining robust control by the above formula 40, and feedforward component When the sum of the acceleration command current I _FFi and the feedback component acceleration / deceleration torque current I _FBi becomes larger than the maximum acceleration / deceleration torque current I _accMAX , the maximum acceleration θ ^·· _max of the i-axis from the above _equation 41 (N) _i is calculated, and the position command adjustment rate γ (n + 1) _i is calculated by the above formula 42. Next, a minimum value γ (n + 1) min is extracted from the position command adjustment rate γ (n + 1) _i of each axis i, and values I ^to _FFi (n) obtained by adjusting the acceleration command current of the feedforward component are Calculated from Equation 43.

In such a procedure, the maximum acceleration / deceleration torque current I _accMAX (n) _i of the axis is determined at the time of saturation, and from this determination result, the robust acceleration control for the actuator 2 is reliably performed and the time of the optimum position command is determined. It is possible to calculate the adjustment rate γ (n + 1) _i that is the reduction rate, and thus the acceleration command current adjustment values I ^to _FFi (n).

Further, the delay adjusting means 86 for each axis calculates the correction time t ^to (n + 1) in the next sample based on the position command adjustment rate γ (n + 1) from the above formula 45 in executing the adjustment step. based on corrected time, and fixes the acceleration command current I _FFi the acceleration and deceleration torque current I _FBi and feedforward component of the feedback component.

As a result, it is possible to adjust the optimum trajectory command for cooperatively operating the actuators 2 of all the axes from the position command adjustment rate γ (n + 1) _i .

  In addition, this invention is not limited to the said Example, A various deformation | transformation implementation is possible in the range of the summary of this invention. The concept of the present invention is also effective for trajectory tracking control (high-speed motion control) of various mechatronic devices other than the experimental apparatus shown in FIG.

  Since conventional torque saturation countermeasures are basically based on the speed PI control system, they are not always suitable for motion control based on a robust acceleration control system. However, in the above embodiment, torque saturation based on robust acceleration control is newly defined and the optimum feedforward torque current command is corrected, so that the conventional problem can be solved. Furthermore, by considering the emphasis operation in the other axis system at the same time, it is theoretically possible to follow the locus without generating a locus error in the control of the multi-axis system having infinite axes.

  In this way, by taking measures against torque saturation based on robust acceleration control, it is possible to perform highly accurate trajectory tracking control that far exceeds the barriers of the prior art, such as various machine tools and robots currently being realized in the industry. It is expected to renew motion control in production technology.

It is a block diagram which shows an example of the system of the robust acceleration control system using the disturbance observer which the actuator control apparatus of this invention applies. FIG. 2 is a block diagram of a trajectory tracking control system based on the robust acceleration control system of FIG. 1. It is a block diagram which shows the structure which incorporated the current limiter in the control system shown in FIG.1 and FIG.2. It is the block diagram which carried out equivalent conversion of the system shown in FIG. FIG. 4 is a block diagram obtained by further equivalently converting the system shown in FIG. 3. It is a block diagram which shows the structure of the acceleration limiting means which performs the acceleration limiting algorithm in this invention. It is the schematic explanatory drawing which showed the external appearance of the experiment system of a multi-axis system It is a block diagram which shows an example of the locus | trajectory tracking system which considered the torque saturation based on the robust acceleration control of the multi-axis system by a conventional method. It is the block diagram which showed the torque saturation of the i-axis in robust acceleration control. It is the wave form diagram which showed the experimental result when drawing a circular locus | trajectory using the conventional locus | trajectory tracking system, and traced the command value and response value of the position of the X-axis, speed, and electric current. It is the wave form diagram which showed the experimental result when drawing a circular locus using the conventional locus tracking system, and traced the command value and response value of the position of Y axis, velocity, and current. It is an experimental result when a circular locus is drawn using the conventional locus tracking system, and is a waveform diagram of each tracking error and radius error of the X axis and the Y axis. It is an enlarged waveform diagram of the trajectory response at the time of starting and stopping, showing experimental results when a circular trajectory is drawn using a conventional trajectory tracking system. It is a block diagram which shows an example of the locus | trajectory tracking system which considered the torque saturation based on the robust acceleration control of a multi-axis system in the preferable method of a present Example. It is a graph which shows the adjustment image of the position command value in the next sample by the calculated position command adjustment rate same as the above. It is a wave form diagram which showed the experimental result when drawing a circular locus using the desirable locus tracking system of this example, and traced the command value and response value of the position of the X-axis, speed, and current. It is a wave form diagram which showed the experimental result when a circular locus was drawn using the desirable locus tracking system of this example, and traced the command value and response value of the position of Y axis, velocity, and current. It is a waveform figure of each tracking error of a X-axis and a Y-axis, and a radius error, showing an experimental result when a circular locus is drawn using a desirable locus tracking system of this example. It is an enlarged waveform diagram of a trajectory response at the time of starting and stopping, showing experimental results when a circular trajectory is drawn using the preferred trajectory tracking system of the present embodiment. It is each graph of a position, speed, and current showing an ideal deceleration stop response when decelerating at the maximum torque. It is a block diagram which shows an example of the locus | trajectory tracking control system which considered the torque saturation which applied the concept of FIG. It is a block diagram which shows the system configuration | structure of a general servo type.

Explanation of symbols

1 Acceleration control system (acceleration control unit)
2 Actuator 3 Disturbance observer 20 Trajectory tracking control system (trajectory tracking control unit)
86 Delay adjustment means (adjustment part)
95 Torque current limiter (current limiter)

Claims (10)

An actuator control device for individually controlling the operation of a plurality of actuators movable on each axis,
A disturbance observer for estimating disturbance to the actuator is provided, the disturbance estimated value obtained by the disturbance observer is converted into a disturbance compensation current, and the actuator is converted by a current reference value to the actuator that is feedback-compensated with the disturbance compensation current. An acceleration controller for robust acceleration control;
In order to control the trajectory tracking of the actuator, a trajectory tracking control unit that calculates a current reference value of a feedback component and a current reference value of a feedforward component;
In order to limit the current reference value of the feedback component, the current reference value of the feedforward component, and the current reference value to the actuator, which is the sum of the disturbance compensation currents, to a predetermined maximum current value or less, the feedforward A current limiter for adjusting the current reference value of the component;
In order to coordinately control the actuator between the axes, the current reference value of the feedback component and the current reference value of the feedforward component are corrected from the adjustment rate obtained when the current reference value of the feedforward component is adjusted. An actuator control device comprising an adjustment unit.   2. The actuator control device according to claim 1, wherein the disturbance observer has an infinite pole.   The current limiter for each axis calculates the adjustment rate of the position command so that the actuator does not deviate from the trajectory on all axes when the actuator is saturated with current on any one axis, and this adjustment rate is calculated. 3. The actuator control device according to claim 1, wherein the current reference value of the feedforward component is adjusted for each axis. The i-axis feedback component current reference value is I _FBi , the i-axis feedforward component current reference value is I _FFi , the i-axis disturbance compensation current is I _RBi , and the i-axis actuator is supplied to the i-axis actuator. When I _i ^ref is the current reference value of I and the maximum torque current of the motor constituting the actuator is I _MAX ,
The current limiter for each axis calculates the maximum acceleration / deceleration torque current I _accMAX (n) _i of the current sample n of the i-axis for maintaining the robust control by the following equation:

When the sum of the current reference value I _FFi of the feed forward component and the current reference value I _FBi of the feedback component is larger than the maximum acceleration / deceleration torque current I _accMAX, Acceleration θ ^·· _max (n) _i is calculated (where K _t is the torque constant, J _ni is the nominal value of the moment of inertia of the i-axis),

The adjustment rate γ (n + 1) _i of the position command is calculated by the following formula (where Ts is the sampling time from the current sample to the next sample, θ ^{· res} (n) _i is the current sample of the i-axis Speed response value, θ ^cmd (n + 1) _i is a position command value in the next sample of the i-axis, θ ^cmd (n) _i is a position command value in the current sample of the i-axis),

Next, the minimum value γ (n + 1) min is extracted from the adjustment rate γ (n + 1) _i of each axis i,
Values I ^to _FFi (n) obtained by adjusting the current reference value of the feedforward component are calculated from the following equations (where K _tn is the nominal value of the torque constant, and f _i (t) is the position command for the i-th axis. Value time function, θ ^· _i ^res (n) is the velocity response value in the current sample on the i-axis, and θ ^cmd _i (n) is the position command value in the current sample on the i-axis)

The actuator control device according to claim 3. The adjustment unit for each axis calculates a correction time t ^to (n + 1) in the next sample based on the adjustment rate γ (n + 1) of the position command from the following equation (where t (n) is the current time),

5. The actuator control apparatus according to claim 4, wherein the current reference value of the feedback component and the current reference value of the feedforward component are corrected based on the correction time. An actuator control method for individually controlling the operation of a plurality of actuators movable on each axis,
A disturbance observer for estimating disturbance to the actuator is provided, the disturbance estimated value obtained by the disturbance observer is converted into a disturbance compensation current, and the actuator is converted by a current reference value to the actuator that is feedback-compensated with the disturbance compensation current. An acceleration control step for robust acceleration control;
A trajectory tracking step for calculating a current reference value of a feedback component and a current reference value of a feedforward component in order to perform trajectory tracking control of the actuator,
In order to limit the current reference value of the feedback component, the current reference value of the feedforward component, and the current reference value to the actuator, which is the sum of the disturbance compensation currents, to a predetermined maximum current value or less, the feedforward A current limiting step for adjusting the current reference value of the component;
In order to coordinately control the actuator between the axes, the current reference value of the feedback component and the current reference value of the feedforward component are corrected from the adjustment rate obtained when the current reference value of the feedforward component is adjusted. An actuator control method comprising: an adjusting step.   The actuator control method according to claim 6, wherein a pole of the disturbance observer is made infinite.   The current limiting step calculates the adjustment rate of the position command so that the actuator does not deviate from the locus on all axes when the actuator causes current saturation on any one axis, and uses this adjustment rate. The actuator control method according to claim 6 or 7, wherein a current reference value of the feedforward component is adjusted for each axis. The i-axis feedback component current reference value is I _FBi , the i-axis feedforward component current reference value is I _FFi , the i-axis disturbance compensation current is I _RBi , and the i-axis actuator is supplied to the i-axis actuator. Current reference value of I _i ^ref and the maximum torque current of the motor constituting the actuator as I _MAX ,
The current limiting step calculates the maximum acceleration / deceleration torque current I _accMAX (n) _i of the current sample n of the i-axis for maintaining the robust control for each axis by the following formula:

When the sum of the current reference value I _FFi of the feed forward component and the current reference value I _FBi of the feedback component is larger than the maximum acceleration / deceleration torque current I _accMAX, Acceleration θ ^·· _max (n) _i is calculated (where K _t is the torque constant, J _ni is the nominal value of the moment of inertia of the i-axis),

The adjustment rate γ (n + 1) _i of the position command is calculated by the following formula (where Ts is the sampling time from the current sample to the next sample, θ ^{· res} (n) _i is the current sample of the i-axis Speed response value, θ ^cmd (n + 1) _i is a position command value in the next sample of the i-axis, θ ^cmd (n) _i is a position command value in the current sample of the i-axis),

Next, the minimum value γ (n + 1) min is extracted from the adjustment rate γ (n + 1) _i of each axis i,
Values I ^to _FFi (n) obtained by adjusting the current reference value of the feedforward component are calculated from the following equations (where K _tn is the nominal value of the torque constant, and f _i (t) is the position command for the i-th axis. Value time function, θ ^· _i ^res (n) is the velocity response value in the current sample on the i-axis, and θ ^cmd _i (n) is the position command value in the current sample on the i-axis)

The actuator control method according to claim 8. The adjustment step calculates a correction time t ^to (n + 1) in the next sample based on the adjustment rate γ (n + 1) of the position command from the following equation (where t (n) is the current time),

The actuator control method according to claim 9, wherein the current reference value of the feedback component and the current reference value of the feedforward component are corrected based on the correction time.