Actuator control method and actuator control device

Abstract

An actuator control method and an actuator control device that incorporate an element of feedback control in time optimal control, including: a calculation step of calculating a switching time at which an acceleration output is switched to a deceleration output and an end time of the deceleration output expressed by time elapsed from a calculation time at which calculation for control is performed using a maximum acceleration and a maximum deceleration, which are measured in advance, at the time of the maximum output of control force of an actuator; a control output step of setting the control force of the actuator to a maximum acceleration output from the calculation time to the switching time, setting the control force of the actuator to a maximum deceleration output from the switching time to the end time, and ending the output of the control force at the end time, and an update step of calculating and updating the switching time and the end time by repeating the calculation step at each preset time.

Claims

1. An actuator control method using time optimal control, comprising: a calculation step of calculating a switching time t1 at which an acceleration output is switched to a deceleration output and an end time t2 of the deceleration output expressed by time elapsed from a calculation time t0 at which calculation for control is performed using a maximum acceleration p and a maximum deceleration m, which are measured in advance, at a time of a maximum output of a control force of the actuator; and a control output step of setting the control force of the actuator to a maximum acceleration output from the calculation time t0 to the switching time t1, setting the control force of the actuator to a maximum deceleration output from the switching time t1 to the end time t2, and ending the output of the control force at the end time t2, and the method also calculates and updates the switching time t1 and the end time t2 by repeating the calculation step at each preset time, wherein, at the calculation step, in a case where a trajectory from a position at the calculation time t0 to a target position at the end time t2 is represented by a combination of two quadratic curves in contact with each other, initial velocity is taken to be VO, and a deviation between the target position and a controlled variable is taken to be X, the switching time t1 and the end time t2 are calculated using Expression (1) and Expression (2) below $\begin{array}{cc}\underset{\_}{\left(\mathrm{Formula}1\right)}& \\ t1=\frac{\begin{array}{c}-2\left(1-2\right)V0\\ \sqrt{4{\left(1-2\right)}^{2}V{0}^2-4\left(1-2\right)\left(V{0}^2-2X2\right)1}\end{array}}{21\left(1-2\right)}& \mathrm{Expression}\left(1\right)\\ \underset{\_}{\left(\mathrm{Formula}2\right)}& \\ t2=\frac{-\left(1-2\right)t1-V0}{2}& \mathrm{Expression}\left(2\right)\end{array}$ where 1 is the maximum acceleration p or the maximum deceleration m between the calculation time t0 and the switching time t1, and 2 is the maximum deceleration m or the maximum acceleration p between the switching time t1 and the end time t2.

2. An actuator control device using time optimal control, comprising: a calculation unit configured to calculate a switching time t1 at which an acceleration output is switched to a deceleration output and an end time t2 of the deceleration output expressed by time elapsed from a calculation time t0 at which calculation for control is performed using a maximum acceleration p and a maximum deceleration m, which are measured in advance, at a time of a maximum output of a control force of the actuator; a control output unit configured to set the control force of the actuator to a maximum acceleration output from the calculation time t0 to the switching time t1, to set the control force of the actuator to a maximum deceleration output from the switching time t1 to the end time t2, and to end the output of the control force at the end time t2; and an update unit configured to repeatedly calculate and update the switching time t1 and the end time t2 by the calculation unit at each preset time, wherein the calculation unit, in a case where a trajectory from a position at the calculation time t0 to a target position at the end time t2 is represented by a combination of two quadratic curves in contact with each other, initial velocity is taken to be VO, and deviation between the target position and a controlled variable is taken to be X, calculates the switching time t1 and the end time t2 using Expression (1) and Expression (2) below $\begin{array}{cc}\underset{\_}{\left(\mathrm{Formula}1\right)}& \\ t1=\frac{\begin{array}{c}-2\left(1-2\right)V0\\ \sqrt{4{\left(1-2\right)}^{2}V{0}^2-4\left(1-2\right)\left(V{0}^2-2X2\right)1}\end{array}}{21\left(1-2\right)}& \mathrm{Expression}\left(1\right)\\ \underset{\_}{\left(\mathrm{Formula}2\right)}& \\ t2=\frac{-\left(1-2\right)t1-V0}{2}& \mathrm{Expression}\left(2\right)\end{array}$ where 1 is the maximum acceleration p or the maximum deceleration m between the calculation time t0 and the switching time t1, and 2 is the maximum deceleration m or the maximum acceleration p between the switching time t1 and the end time t2.

3. The actuator control method according to claim 1, wherein the acceleration p between the calculation time t0 and the switching time t1 and the deceleration m between the switching time t1 and the end time t2 are determined from Table 1 TABLE-US-00003 TABLE 1 A1 X > 0 V.sub.0 > 0 V.sub.0.sup.2/2 m > X 1 = m, 2 = p A2 V.sub.0.sup.2/2 m < X 1 = p, 2 = m A3 V.sub.0 < 0 1 = p, 2 = m A4 X < 0 V.sub.0 > 0 1 = m, 2 = p A5 V.sub.0 < 0 V.sub.0.sup.2/2 p > X 1 = m, 2 = p A6 V.sub.0.sup.2/2 p < X 1 = p, 2 = m.

4. The actuator control device according to claim 2, wherein the acceleration p between the calculation time t0 and the switching time t1 and the deceleration m between the switching time t1 and the end time t2 are determined from Table 1 TABLE-US-00004 TABLE 1 A1 X > 0 V.sub.0 > 0 V.sub.0.sup.2/2 m > X 1 = m, 2 = p A2 V.sub.0.sup.2/2 m < X 1 = p, 2 = m A3 V.sub.0 < 0 1 = p, 2 = m A4 X < 0 V.sub.0 > 0 1 = m, 2 = p A5 V.sub.0 < 0 V.sub.0.sup.2/2 p > X 1 = m, 2 = p A6 V.sub.0.sup.2/2 p < X 1 = p, 2 = m.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a diagram showing a time optimal control model for explaining time optimal control used in an actuator control method of an embodiment of the present invention.

(2) FIG. 2 is a diagram for explaining calculation conditions of a control trajectory.

(3) FIG. 3 is a diagram for explaining a change in the control trajectory in recalculation.

(4) FIG. 4 is a diagram showing target trajectories that f (t) and g (t) can take.

(5) FIG. 5 is a diagram showing an example of a control flow of the actuator control method of the embodiment of the present invention.

(6) FIG. 6 is a diagram showing a control result in a case where there is no mechanical damping force of feedback time optimal control of the actuator control method of the embodiment of the present invention.

(7) FIG. 7 is a diagram showing a control result in a case where there is a mechanical damping force of the feedback time optimal control of the actuator control method of the embodiment of the present invention.

(8) FIG. 8 is a diagram showing a simulation result of an example.

(9) FIG. 9 is a diagram showing a simulation result of a comparative example.

(10) FIG. 10 is a diagram showing a control result in a case where there is no mechanical damping force of time optimal control of the prior art.

(11) FIG. 11 is a diagram showing a control result in a case where there is a mechanical damping force of the time optimal control of the prior art.

(12) FIG. 12 is a diagram showing a PID control model for explaining PID control.

DESCRIPTION OF EMBODIMENTS

(13) Hereinafter, an actuator control method and an actuator control device of an embodiment according to the present invention are explained with reference to drawings. Here, in order to clarify time optimal control of the present invention, explanation is given in comparison with PID control.

(14) The actuator control device of the embodiment according to the present invention uses the time optical control and includes a calculation unit, a control force output unit, and an update unit.

(15) The calculation unit calculates a switching time t1 at which an acceleration output is switched to a deceleration output and an end time t2 of the deceleration output expressed by time elapsed from a calculation time t0 at which calculation for control is performed using a maximum acceleration p and a maximum deceleration m, which are measured in advance, at the time of a maximum output of control force of the actuator.

(16) Moreover, the control output unit sets the control force of the actuator to a maximum acceleration output from the calculation time t0 to the switching time t1, sets the control force of the actuator to a maximum deceleration output from the switching time t1 to the end time t2, and ends the output of the control force at the end time t2.

(17) Further, the update unit is configured to repeatedly calculate and update the switching time t1 and the end time t2 by the calculation unit at each time in a fixed period or preset irregularly. The calculation unit, in a case where a trajectory from a position at the calculation time t0 at which recalculation is performed to a target position is represented by a combination of two quadratic curves in contact with each other, initial velocity is taken to be V0, and deviation between a target amount and a controlled variable is taken to be X, calculates the switching time t1 and the end time t2 using Expression (1) and Expression (2) below.

(18) $\begin{array}{cc}\left(\mathrm{Formula}1\right)& \\ t1=\frac{\begin{array}{c}-2\left(1-2\right)V0\\ \sqrt{4{\left(1-2\right)}^{2}V{0}^2-4\left(1-2\right)\left(V{0}^2-2X2\right)1}\end{array}}{21\left(1-2\right)}& \mathrm{Expression}\left(1\right)\\ \left(\mathrm{Formula}2\right)& \\ t2=\frac{-\left(1-2\right)t1-V0}{2}& \mathrm{Expression}\left(2\right)\end{array}$
Here, 1 is the maximum acceleration p or the maximum deceleration am between the calculation time t0 and the switching time t1 and 2 is the maximum deceleration m or the maximum acceleration p between the switching time t1 and the end time t2.

(19) Furthermore, the time optimal control used in the actuator control method of the embodiment of the present invention is equivalent to a mechanical model in which a weight m shown in FIG. 1 descends and ascends a slope. As in FIG. 1, when the weight is lifted from the slope and then is released, the weight descends the slope at constant acceleration by the gravity and then ascends the slope on the opposite side at constant deceleration. If there is no energy loss due to friction etc., the velocity of the weight m becomes zero when it ascends to the same height as the initial position and stops for a moment. If the gravity is eliminated at this point of time, the weight m continues to stay at the position.

(20) This is considered in a control system in which a controlled object is moved by an actuator in place of the mechanical model. If the weight is pressed by the maximum thrust force of the actuator and then the weight is pulled back by the maximum thrust force of the actuator, when the work input when thrusting the weight becomes equal to the work input when pulling back the weight, the controlled object stops. If the position where the controlled object stops in this way is the target position, the control is ended.

(21) That is, in the time optimal control, since the controlled object is accelerated and decelerated by the maximum thrust force of the actuator and thus is controlled to the target position, it is theoretically possible to perform the control in the shortest time. Further, the control output pattern at this time is determined before the control is started, and therefore, the time optimal control is a feedforward control.

(22) In contrast to this, the PID control of the prior art is based on a damping oscillation model with a basic mass m, spring, and damper system as shown in FIG. 12 and the spring plays a role of the P term and the dashpot plays a role of the D term. The correction term at the zero point is the I term, but the I term does not have so much physical meaning here.

(23) If this control is considered in energy conversion, as shown in the middle of FIG. 12, by inputting the target position, the spring is stretched as a target displacement and if the strain energy of the spring is input to the control system, the weight (mass m) begins to move by being pulled by the spring force. At this time, the conversion from strain energy into kinetic energy is carried out. When the weight m begins to move, the kinetic energy is converted into thermal energy by the dashpot and the energy is discharged to the outside of the system. If all the strain energy of the spring input initially is converted into the thermal energy, as shown at the bottom of FIG. 12, the weight m, which is the controlled object, stays at the target position as a result. In the PID control, the control output is determined from the motion state of the controlled object, and therefore, the PID control is a feedback control.

(24) As described above, it can be said that the time optimal control and the PID control are fundamentally different, but if the P gain and D gain of the PID control are made very large and the upper limit of the output is cut at the maximum output of the actuator, the control output waveform of the PID control becomes close to the control output waveform of the time optimal control, and therefore, it is also possible to regard the time optimal control as the PID control in which the PD gains are increased to the limit. However, in the PID control, if the PD gains are made too large, the control diverges normally because of a delay in calculation and control. The reason for this is that the PID control is a feedback control and the control output is determined with a delay from a phenomenon at all times, and therefore, the control diverges if the delay becomes too large.

(25) In contrast to this, the time optimal control is a feedforward control and the control output that has taken into consideration the motion from the start to its end is determined before a phenomenon at all times, and therefore, the control is stable and even if the control output equivalent to that in the PID control in which the PD gains are increased to the limit is output in the time optimal control, the control does not diverge.

(26) The simplest example of the time optimal control is shown in FIG. 10 and FIG. 11. FIG. 10 shows a case where there is no mechanical damping force and FIG. 11 shows a case where there is a mechanical damping force. In FIG. 10 and FIG. 11, the actuator is operated toward a target by the maximum thrust force from the calculation time t0 and then, the actuator is operated at the maximum deceleration at the switching time t1, and then the thrust force of the actuator is reduced to zero at the end time t2. In this manner, in the time optimal control, control can be performed by determining the switching time t1 and the end time t2. It is possible to determine the switching time t1 and the end time t2 based on the calculation conditions below.

(27) Under the calculation conditions, it is assumed that the controlled object reaches a target through two quadratic curves in contact with each other. Then, the maximum acceleration p that the actuator can generate, the maximum deceleration m that the actuator can generate, the deviation X between a target amount Tx and a controlled variable x (=target amountcontrolled variable: difference between a target position and a position at the time of control) at the calculation time t0, and the velocity V0 of the controlled object at the calculation time t0 are assumed to be already-known values, and under restriction conditions that the two quadratic curves come into contact at the switching time t1, the velocity of the controlled object is reduced to zero (V=0) at the end time t2, and the deviation X is reduced to zero (X=0) at the end time t2, then, the switching time t1 and the end time t2 are found.

(28) It is assumed that the control trajectory is configured by two quadratic curves f (t) and g (t) as shown in FIG. 2, and based on the conditions below, the switching time t1 at which the control outputs are switched and the end time t2 at which the control output is ended are found. Further, the accelerations 1 and 2 are determined from Table 1. Moreover, V0 is a first-order derivative value (or difference value) of the controlled variable obtained during the control.

(29) Calculation conditions are (1) to (7) below.

(30) (1) The maximum acceleration p and the maximum deceleration m that the actuator can generate at the time of the maximum output are already known, that is, are obtained by accelerations measured in advance.

(31) (2) The velocity V0 at the calculation time t0 is already known, that is, is obtained from the first-order derivative value (or difference) of the measured value.

(32) (3) The value of the first quadratic curve f (t0) at the calculation time t0 is zero.

(33) (4) The first-order derivative value of the first quadratic curve f (t0) at the calculation time t0 is the velocity (initial velocity) at the time t0.

(34) (5) The second quadratic curve g (t) comes into contact with the first quadratic curve f (t) at the switching time t1.

(35) (6) The value of the second quadratic curve g (t2) at the end time t2 is the target value.

(36) (7) The first-order derivative value of the second quadratic curve g (t) at the end time t2 is zero.

(37) From the above conditions, Expression (3) to Expression (13) below are obtained. By putting them into simultaneous equations, the switching time t1 and the end time t2 of the output are found and the accelerations 1 and 2 are found. Here, 1 is the acceleration from the calculation time t0 to the switching time t1 and 2 is the acceleration from the switching time t1 to the end time t2, and when 1=p, 2=m, and when 1=m, 2=p.

(38) $\begin{array}{cc}\left(\mathrm{Formula}3\right)& \\ f\left(t\right)=t^2+t+& \mathrm{Expression}\left(3\right)\\ g\left(t\right)=t^2+\mathrm{.Math.}t+& \mathrm{Expression}\left(4\right)\\ f\left(0\right)=0& \mathrm{Expression}\left(5\right)\\ \frac{f\left(0\right)}{t}=V0& \mathrm{Expression}\left(6\right)\\ \frac{{}^{2}f\left(0\right)}{t^2}=1& \mathrm{Expression}\left(7\right)\\ f\left(t1\right)=g\left(t1\right)& \mathrm{Expression}\left(8\right)\\ \frac{f\left(t1\right)}{t}=\frac{g\left(t1\right)}{t}& \mathrm{Expression}\left(9\right)\\ g\left(t2\right)=X& \mathrm{Expression}\left(11\right)\\ \frac{g\left(t2\right)}{t}=0& \mathrm{Expression}\left(12\right)\\ \frac{{}^{2}g\left(t2\right)}{t^2}=2& \mathrm{Expression}\left(13\right)\end{array}$
From Expression (3) and Expression (5),
(Formula 4)
=0Expression (13)
from Expression (3), Expression (6), and Expression (13a),

(39) $\begin{array}{cc}\left(\mathrm{Formula}5\right)& \\ \frac{f\left(0\right)}{t}=2.Math.0+==V0& \mathrm{Expression}\left(14\right)\end{array}$
from Expression (3) and Expression (7),

(40) $\begin{array}{cc}\left(\mathrm{Formula}6\right)& \\ \frac{f\left(0\right)}{t}=2=1=\frac{1}{2}& \mathrm{Expression}\left(15\right)\end{array}$
from Expression (4) and Expression (12),

(41) $\begin{array}{cc}\left(\mathrm{Formula}7\right)& \\ =\frac{2}{2}& \mathrm{Expression}\left(16\right)\end{array}$
from Expression (4), Expression (11), and Expression (16),

(42) $\begin{array}{cc}\left(\mathrm{Formula}8\right)& \\ \frac{g\left(t2\right)}{t}=2.Math.t2+\mathrm{.Math.}=2.Math.t2+\mathrm{.Math.}=0\mathrm{.Math.}=-2.Math.t2& \mathrm{Expression}\left(17\right)\end{array}$
from Expression (4), Expression (10), Expression (16), and Expression (17),

(43) $\begin{array}{cc}\left(\mathrm{Formula}9\right)& \\ \begin{array}{c}g\left(t2\right)=.Math.t{2}^2+\mathrm{.Math.}.Math.t2+\\ =\frac{2}{2}t{2}^2-2.Math.t{2}^2+\\ =-\frac{2}{2}t{2}^2+=X\end{array}& \mathrm{Expression}\left(18\right)\end{array}$
by modifying Expression (18),

(44) 0$\begin{array}{cc}\left(\mathrm{Formula}10\right)& \\ =X+\frac{2}{2}t{2}^2& \mathrm{Expression}\left(19\right)\end{array}$
from Expression (3), Expression (5), Expression (9), Expression (15), Expression (16), Expression (17), and Expression (19),

(45) $\begin{array}{cc}\left(\mathrm{Formula}11\right)& \\ \frac{f\left(t1\right)}{t}=\frac{g\left(t1\right)}{t}1.Math.t1+V0=2.Math.t1-2.Math.t2\underset{\_}{t2=\frac{-\left(1-2\right)t1-V0}{2}}& \mathrm{Expression}\left(20\right)\end{array}$
from Expression (3), Expression (5), Expression (8), Expression (15), Expression (16), Expression (17), Expression (19), and Expression (20),

(46) $\begin{array}{cc}\left(\mathrm{Formula}12\right)& \\ \mathrm{Expression}\text{(21)}& \\ \begin{array}{c}\frac{1}{2}t{1}^2+V0.Math.t1=\frac{2}{2}t{1}^2-2.Math.t2.Math.t1+X+\frac{2}{2}t{2}^2\\ =\frac{\left(1-2\right)}{2}t{1}^2+\left(V0+2.Math.t2\right)t1-X-\frac{2}{2}t{2}^2\\ =\frac{\left(1-2\right)}{2}t{1}^2+V0+\left[2\frac{-\left(1-2\right)t1-V0}{2}\right]t1-\\ X-{\frac{2}{2}\left[\frac{-\left(1-2\right)t1-V0}{2}\right]}^2\\ =-\frac{\left(1-2\right)}{2}t{1}^2-X-\frac{{\left(\left(1-2\right)t1+V0\right)}^2}{22}\\ =-\frac{\left(1-2\right)}{2}t{1}^2-\\ X-\frac{{\left(1-2\right)}^{2}t{1}^2+2V0\left(1-2\right)t1+V{0}^2}{2}\\ =-2{\left(1-2\right)}^{2}t{1}^2-22.Math.X-{\left(1-2\right)}^{2}t{1}^2-\\ 2V0\left(1-2\right)t1-V{0}^2\\ =-1{\left(1-2\right)}^{2}t{1}^2-2V0\left(1-2\right)t1-22.Math.X-V{0}^2=0\end{array}& \\ 1{\left(1-2\right)}^{2}t{1}^2+2\left(1-2\right)V0.Math.t1+22.Math.X+V{0}^2=0& \end{array}$
by modifying Expression (21),
(Formula 13)
(12)1.Math.t1.sup.2+2(12)V0.Math.t1+V0.sup.2+2X.Math.2=0Expression (22)
by applying formula for solution of the quadratic formula to Expression (22),

(47) $\begin{array}{cc}\left(\mathrm{Formula}14\right)& \\ \underset{\_}{t1=\frac{\begin{array}{c}-2\left(1-2\right)V0\\ \sqrt{4{\left(1-2\right)}^{2}V{0}^2-4\left(1-2\right)\left(V{0}^2-2X2\right)1}\end{array}}{21\left(1-2\right)}}& \mathrm{Expression}\left(23\right)\end{array}$
and by substituting Expression (23) in Expression (20), t2 is obtained.

(48) Here, the trajectories that f (t) and g (t) can take will be six trajectories A1 to A6 shown in FIG. 4. Here, the six trajectories are classified according to the state.

(49) A1 corresponds to a state where even if deceleration is performed at the maximum deceleration m generated by the actuator when X>0 and V0>0, the target is overshot. If the time taken for the velocity to reduce to zero by performing deceleration at the maximum deceleration m is taken to be t3, t3=V0/m, and therefore, the condition under which overshoot occurs at the time t3 is V0t3/2=V0.sup.2/2m>X.

(50) A2 corresponds to a state where if deceleration is performed at the maximum deceleration m generated by the actuator when X>0 and V0>0, the target is not overshot. By the same calculation as that for A1, the condition under which overshoot does not occur at the time t3 is V0.sup.2/2m<X.

(51) A3 is the trajectory under the condition that X>0 and V0<0 and A4 under the condition that X<0 and V0>0. Moreover, A5 corresponds to a state where if deceleration is performed at the maximum deceleration p generated by the actuator when X<0 and V0<0, the target is not overshot. By the same calculation as that for A1, the condition under which overshoot does not occur at the time t3 is V0.sup.2/2p>X.

(52) A6 corresponds to a state where even if deceleration is performed at the maximum deceleration p generated by the actuator when X<0 and V0<0, the target is overshot. By the same calculation as that for A1, the condition under which overshoot occurs at the time t3 is V0.sup.2/2p<X.

(53) A1, A2, and A3 show the change from the upwardly convex shape into the downwardly convex shape and 1=m<0 and 2=p>0 hold, and A4, A5, and A6 show the change from the downwardly convex shape into the upwardly convex shape and 1=p>0 and 2=m<0. From the above, 1 and 2 are determined.

(54) This classification according to the state is shown in Table 1.

(55) TABLE-US-00002 TABLE 1 A1 X > 0 V.sub.0 > 0 V.sub.0.sup.2/2 m > X 1 = m, 2 = p A2 V.sub.0.sup.2/2 m < X 1 = p, 2 = m A3 V.sub.0 < 0 1 = p, 2 = m A4 X < 0 V.sub.0 > 0 1 = m, 2 = p A5 V.sub.0 < 0 V.sub.0.sup.2/2 p > X 1 = m, 2 = p A6 V.sub.0.sup.2/2 p < X 1 = p, 2 = m

(56) By multiplying 1 and 2 obtained as described above by a virtual mass m, the actuator thrust force is found. That is, between the calculation time t0 and the switching time t1, an actuator thrust force 1 (=1virtual mass) is obtained and between the switching time t1 and the end time t2, an actuator thrust force 2 (=2virtual mass) is obtained.

(57) It is possible to perform the time optimal control used in the actuator control method of the embodiment of the present invention in accordance with a control flow as shown in FIG. 5. When the control flow starts, at step S11, data of the maximum acceleration p and the maximum deceleration m is read. At the next step S12, data of the target amount (target value) Tx and the controlled variable (controlled value) x is read. At the same time, counting of an elapsed time t and an elapsed time for recalculation tc is started.

(58) At step S13, whether or not the period of trajectory recalculation is reached, that is, whether or not the elapsed time for recalculation tc after the trajectory is calculated becomes equal to or more than a period of trajectory recalculation tcr is determined. In the case where the period of trajectory recalculation tcr is reached (YES) at step S13, the procedure proceeds to step S15 after calculating the trajectory again at step S14 and in the case where the period of trajectory recalculation is not reached (NO) at step S13, the procedure proceeds to step 15, bypassing the trajectory calculation at step S14. It is preferable to set the period of trajectory recalculation tcr to a period about 1/10 of the control period, but there arise no problem even if the period of trajectory recalculation tcr is set the same as the control period.

(59) In the trajectory calculation at step S14, the initial velocity V0 is calculated by V0=(xx.sub.1)/tcr (x.sub.1 is the controlled variable before the calculation period) and kinetic energy E is calculated by E=V0.sup.2/2, and 1 and 2 are determined from Table 1, and by Expression (23) and Expression (20), the switching time t1 and the end time t2 are calculated. Further, the elapsed time for recalculation tc is reset to zero (Tc=0).

(60) At step S15, whether or not the elapsed time t is smaller than the switching time t1 is determined and in the case where the elapsed time t is smaller than the switching time t1 (YES), the procedure proceeds to step S16 and then proceeds to step S20 after setting the output acceleration to 1. In the case where the elapsed time t is not smaller than the switching time t1 (NO) at step S15, the procedure proceeds to step S17 and determining whether or not the elapsed time t is smaller than the end time t2 and in the case where the elapsed time t is smaller than the end time t2 (YES), the procedure proceeds to step S18 and then to step S20 after setting the output acceleration to 2. In the case where the elapsed time t is not smaller than the end time t2 (NO) at step S17, the procedure proceeds to step S19 and then to step S20 after setting the output acceleration to zero.

(61) At step S20, the actuator thrust force corresponding to the output acceleration is kept being generated during a preset time (time in relation to the interval of various kinds of determination) and the controlled object is controlled. Further, the elapsed time t and the elapsed time for recalculation tc are counted. After that, the procedure returns to step S12 and repeats step S12 to step S20. Due to this, it is possible to control the actuator thrust force while calculating the trajectory again at step S14 each time the elapsed time for recalculation tc reaches the period of trajectory recalculation tcr.

(62) If an event by which the control in FIG. 5 should be ended occurs, such as the elapsed time t exceeds a preset time and a switch signal to end the control is input, an interrupt occurs even when any step of the control flow is performed and the procedure proceeds to return and returns to the upper control flow, and when the upper control flow ends, the control flow in FIG. 5 also ends.

(63) By switching the operations of the actuator at the switching time t1 and the end time t2 found as described above, an ideal control result as shown in FIG. 10 is obtained. However, this result is a result under ideal circumstances where there is no friction, damping, or errors. If a mechanical damping force exists in the controlled object, the control result will no longer agree with the target as shown in FIG. 11.

(64) In order to solve this problem, in the actuator control method of the present invention, the target trajectory is modified at each time in a fixed period or preset irregularly. During the period of time of reset for modification, such as the fixed period, it is necessary to set the reset signal period to T/2 or less relative to a fluctuation period T within tolerance. The way the control trajectory changes due to the recalculation is shown in FIG. 3. In FIG. 3, in the case where the displacement that should be X1 according to the control trajectory (dotted line) by the initial calculation is actually X2, a new control trajectory (solid line) is calculated by recalculation and the switching time t1 and the end time t2 change to new values, and the control force is controlled based on the new switching time t1 and the end time t2 as a result.

(65) The control results when the switching time t1 and the end time t2 are calculated again for each fixed period are shown in FIG. 6 and FIG. 7. Recalculation is performed with a timing of the rise of the signal pulse. In order to distinguish this time optimal control method from the time optimal control method of the prior art, this time optimal control method is called a feedback (FB) time optimal control method here. FIG. 6 shows the case where there is no mechanical damping force and FIG. 7 shows the case where there is a mechanical damping force, but it is known that it is possible to cause the controlled object to agree with the target position by substantially an ideal trajectory even in the case where there is a mechanical damping force in the controlled object as shown in FIG. 7 as a result.

(66) FIG. 8 shows a simulation result of an example of the feedback time optimal control and FIG. 9 shows a simulation result of a comparative example of the PID control. The comparison of these results indicates that the control result changes significantly depending on the load fluctuation of the control system in the comparative example of the PID control shown in FIG. 9, but the control result is not disturbed and the stable control result is obtained at all times in the example of the feedback time optimal control shown in FIG. 8. That is, in the comparative example in FIG. 9, the control result changes significantly depending on the load fluctuation of the control system, but in the example in FIG. 8, the control result is not disturbed and the stable control result is obtained at all times.

(67) Consequently, according to the actuator control method and the actuator control device described above, the control speed becomes high because of the time optimal control, the maximum acceleration and the maximum deceleration can be set by measurement, and it is not necessary to adjust the control gain because there is no item to be adjusted other than this. Further, it is not necessary to produce an intermediate output because of the ON/OFF control, and therefore, it is possible to simplify the controller and driver.

(68) Further, the element of feedback is incorporated, in which the switching time t1 and the end time t2 are updated by inputting the deviation X between the target amount and the controlled variable at each time of the control at each preset time, and therefore, even if the external force changes or without shortening the control period, it is possible to obtain a stable control result at all times. As a result of that, it is possible to satisfy both of control speed and stability,which have been big conflicting issues in the conventional control law.

(69) According to the actuator control method and the actuator control device of the present invention, the control speed becomes high, it is not necessary to adjust the control gain, and it is not necessary to produce an intermediate output because of the ON/OFF control, and therefore, it is possible to simplify the controller and driver, and further, the element of feedback is incorporated, in which the switching time t1 and the end time t2 are updated by inputting the deviation X between the target amount and the controlled variable at each time of the control at each preset time, and therefore, even if the external force changes or without shortening the control period, it is possible to obtain a stable control result at all times, and therefore, it is possible to utilize the actuator control method and the actuator control device of the present invention, as the actuator control method and the actuator control device for the position control using, for example, electric power, hydraulic power, and pneumatic power, of equipment etc. mounted on automobiles etc., and other controlled objects to which the PID control is applied.

(70) In the case of two degrees of freedom, control with two degrees of freedom is enabled by applying the control of the present invention independently in the X direction and in the Y direction, respectively, and therefore, it is possible to apply the control of the present invention to the control with multiple degrees of freedom, not only to the control with one degree of freedom.