JPH0127441B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for generating feedforward signals in servo systems such as industrial robots and NC machine tools.

[Conventional technology]

First, a conventional feed-forward type servo system will be explained.

Figure 1 shows the general configuration of a conventional servo system. The deviation between the position command x _r and the position x is given to the position loop controller G _p , and the deviation between the speed command x〓 _r and the speed x〓 is calculated. The deviation between the acceleration command x¨ _r and the acceleration x¨ is given to _the velocity loop controller _{G v ,} and the deviation between the acceleration command Through this, the velocity x〓 and the position x were obtained. The acceleration x¨ may be approximated by a motor current.

In this system, there are positions, speeds, accelerations, and signals corresponding to position commands, speed commands, and acceleration commands. Therefore, in the servo system of FIG. 1, only position commands can be given from the outside, but generally the minor loops, the speed control loop and acceleration control loop, have a faster response than the position control loop.
Therefore, if the servo system is driven using only a position command, the response will be slower than when the servo system is driven by giving not only a position command but also a speed command and an acceleration command.

From this point of view, as a method to improve the response of the servo system shown in Figure 1, a so-called feed forward method is proposed in which speed commands and acceleration commands (or only speed commands) are added to the corresponding signals of the servo system. control is known.

The following two methods are known as methods for realizing this feedforward control.

(i) Differentiate the position command to obtain the velocity command and acceleration command.

(ii) Simultaneously output position command, speed command, and acceleration command from the command generator.

[Problem to be solved by the invention]

The above-mentioned method of "obtaining velocity commands and acceleration commands by differentiating the position command in (i)" uses a differentiator, but a perfect differentiator cannot be constructed from a circuit, and an approximate differentiator circuit Even if the servo system is used, there is a problem in that it amplifies high frequency noise and tends to make the servo system unstable.

On the other hand, the above-mentioned method of "(ii) simultaneously outputting a position command, velocity command, and acceleration command from a command generator" is realized by the configuration shown in the block diagram of FIG.

In the method shown in the block diagram of FIG. 2, the position command group {X _i } given at each discrete time is
Based on this, a command generator calculates and outputs a position command X _N , a speed command V _N , and an acceleration command A _N . In the case of this _method , there is no problem caused by the differentiator, but the _calculations required to simultaneously output the _{position command} The amount of calculation is considerably large compared to the required calculations. Therefore, the calculation processing speed required of the command generator increases.
However, since there is actually an upper limit to the calculation processing speed, commands cannot be generated completely continuously, but are generated discretely with respect to time. Usually, within one discrete time interval, the position, velocity, and acceleration command values are approximated as constants.
This naturally causes an approximation error.

In addition to these problems, regarding the command generator of the servo system, there are the following: ● Improving the accuracy of command calculations ● Reducing the calculation processing speed ● Reducing the trajectory information to be stored ● Reducing the amount of information transmitted from the command generator to the servo system Due to these demands, it is desired to increase the discrete time interval between command generation.

An object of the present invention is to realize a servo system with higher responsiveness and higher accuracy than conventional ones to commands from the same command generator.

[Means to solve the problem]

To achieve this objective, the present invention provides, in addition to position, velocity and acceleration feedback signals, a set of constants [position, velocity, acceleration] continuous at least up to velocity from a command generator for each discrete time interval. [X _N , V _N , A _N ] (N is a discrete time) In a servo system that gives a feed forward command signal and performs control to follow this command signal, the time at which the command signal is given from the command generator In synchronization _with
It is characterized by V _f and A _f .

Formula A _{f =} A _{N V f} = _{V N +} A _N ^τ Strictly speaking, the command value given from the command generator is given in a discrete time manner, and is approximated by a constant within a discrete time interval.

In a system where the command value is given by a set of constants (position, velocity, acceleration), the position command of the servo system can be approximated by a quadratic power function with respect to time. This is shown as follows.

At discrete times N, the set of command values is expressed as (X _N , V _N ,
A _N ), and if the _discrete time interval is T, then during this time X _f (NT)=X _N X〓 _f (NT)=V _N X¨ _f (NT ₎ =A _N NT)...=0...(1). However, X _f (NT) is the time NT of variable X _f
, and "." represents the time derivative of the order of that number.

_{In other words :} ^_ _{_ _ _} ).

X _f (t) =X _N +V _N (t-NT)+1/2・A _N ×(t-NT) ²
...(3) From equation (3), the position command, velocity command, and acceleration command when NT≦t<(N+1)T can be expressed as in the following equation (4).

X _f (t)=X _N +V _N (t-NT)+1/2・A _N ×(t-NT) ² V _f (t)=V _N +A _N (t-NT) A _f (t)=A _N (4) The set of commands in equation (4) is in the middle of NT≦t<(N+1)T from the command approximated by the constant expressed in equation (1), and in X _f , V _N (t-NT )+1/2・A _N ×(t-NT) ² At V _f , the accuracy is improved by A _N (t-NT).

The gist of the present invention is to calculate equation (4) from a set of constants (X _N , V _N , A _N ), and add and input the result to the corresponding input signal of the servo system.

In addition, in equation (4), if τ = t-NT, then
0≦τ<T, and when formula (4) is expressed using this τ, it becomes a simple formula as shown below.

X _f (t)=X _N +V _N τ+1/2・A _N ×τ ² V _f (t)=V _N +A _N τ A _f (t)=A _N [Example] The present invention will be described in Fig. 3 below. This will be explained based on the embodiment shown in .

The configuration of the servo system is assumed to have a control loop for position, velocity, and acceleration as shown in FIG.

Feedforward signals (additional input signals) for (position command), (speed command, and acceleration command) of the servo system are assumed to be X _f , V _f , and A _{f ,} respectively.

The command generator gives (position, velocity, acceleration) in a discrete time manner as a set of constants (X _N , V _N , A _N ), A _N is A _f , A _N is the input of integrator 1, and the The output is V _i , the sum of V _N and _Vi is V _f , V _f is the input of integrator 2, its output is X _i , the sum of X _N and X _i is X _f , and the command A reset signal is generated in synchronization with the timing at which (X _N , V _N , A _N ) is applied from the generator to the servo system, and the integrator 1 and integrator 2 are reset.

The integrators 1 and 2 are constructed by cascading the circuit shown in FIG. 4 with a gain circuit for correcting the integral constant (1/RC).

By adopting such a control method, even if the acceleration A is given as a discontinuous value at each discrete time interval, the velocity X becomes linearly continuous, and the position It will continue to the next parabola at the point where the slope of is equal.

〔Effect of the invention〕

As described above, the present invention provides discrete time commands (position, velocity, acceleration) as a set of constants (X _N , V _N , A _N ), and generates position commands and feedforward signals of velocity and acceleration. Compared _to the conventional feed _{forward control method in which X f = X N V f = V N A f = A N ,} where Perform integral correction for X _f = X _N + V _N τ + 1/2・A _N ×τ ² V _f = V _N + A _N τ A _f = A _N.

Therefore, the accuracy of X _f and V _f is improved by V _N τ + 1/2・A _N ×τ ² and A _N τ, respectively, and discontinuity at the time of command generation is eliminated, resulting in the following effects. .

Position commands and speed commands can be given in continuous curves or straight lines, achieving smooth movement.

Since the command does not result in a step change in the value of the function, the timing of data output, that is, the discrete time interval, can be lengthened without reducing accuracy.

Therefore, a highly accurate servo system can be realized even using a conventional digital controller.

If the same accuracy is required, the configuration becomes simpler and an inexpensive control system can be realized.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing the general configuration of a conventional servo system, Fig. 2 is a block diagram showing a feed forward control system, and Fig. 3 is a block diagram showing the general configuration of a conventional servo system.
The figure is a block diagram showing the configuration of the servo system of the present invention, and FIG. 4 is a circuit diagram showing an example of the configuration of an integrator.

Claims (1)

[Claims] 1. In addition to the feedback signals of position, velocity, and acceleration, a set of constants [X _N , V _N , A _N ] (N is discrete time)
A feedforward command signal is given as , and in the servo system that performs control to follow this command signal, the calculation expressed by the following formula is performed sequentially in synchronization with the time when the command signal is given from the command generator. , a servo system feedforward signal generation method, characterized in that the servo system position, velocity, and acceleration feedforward signals X _f , V _f , and A _f are used. Formula _{A f} = A _{N V f = V N} + ^A _N τ