DE102015211405A1 - Improvement of the temperature drift compensation by calibration at the component and teaching of the parameter sets - Google Patents

Abstract

The present invention relates to a method for controlling a manipulator. In this case, the method comprises providing a temperature-dependent calibration, on the basis of which a correction parameter set is calculated based on specific reference point residual drift values. Subsequently, the manipulator is controlled taking into account the correction parameter set.

Description

1. Technical area

The present invention relates to a method for controlling a manipulator and a corresponding robot system.

2. Technical background

Robots, and in particular industrial robots, are automatically controlled, freely programmable multipurpose manipulators. They are designed for use in industrial environments and can thus be used in different areas of a production plant. For example, an industrial robot can be used to measure one or more measuring points of a component in a processing station or surveying station, or to run a specific path program on a workpiece while processing the workpiece or component by means of, for example, a folding tool.

Before use, the industrial robot usually needs to be calibrated to obtain a complete kinematic model of the robot. For this purpose, different parameters of the robot mechanics must be determined in order to finally obtain a complete robot model. In addition, the robot must be aligned with respect to the component to be measured or machined: a reference from the coordinate system of the robot to the coordinate system of the workstation or the component must be established. For this example, reference points can be approached, which represent a fixed point in space. A reference point may e.g. be read optically, or be approached manually by a robot. For example, since the position of a reference point is defined in the world coordinate system, by detecting the reference point, the exact position of an end effector of the robot can be determined by corresponding transformations between the robot coordinate system and the world coordinate system.

By friction in the mechanical elements of a robot and by the waste heat of electrical components, the temperature of the robot mechanics changes. This can lead to a change in size of the individual elements, as well as a change in the viscosity of liquids and changes in the elasticity of elements, such as the transmission, bring with it. If the parameters of the robot model are not updated cyclically, there is a so-called drift in the positioning of the robot, which superimposes the repetition and absolute accuracy of the robot. This drift can be measured on a fixed reference body, or reference body, with known or measured positions.

From an in-house method is known to determine an optimized model parameter set during calibration. The model parameter set or the parameterizable mathematical robot model is used for the control of the manipulator. In order to determine the optimized model parameter set, deviations of the positioning of the robot on a temperature-stable calibration body are measured. The optimization then takes place by minimizing the residual error of the positioning. Since the effect of the temperature drift on different poses has different effects, an attempt is made to determine the axis position of the manipulator that varies as widely as possible and covers a large axis range. However, only the effects resulting from the measurements on the calibration body are compensated for. However, effects that are caused by the temperature change of the kinematics can not be mapped by the robot poses on the calibration body and are therefore not compensated.

During a typical measuring cycle, a component is introduced into a measuring cell, measured and then discharged again. This heats the robot. Consequently, the parameters of the robot model determined at the beginning of the cycle are becoming increasingly poorer in the actual positioning of the robot during the course of the cycle. Consequently, a higher residual drift at the measured points is to be expected late in the cycle.

The validity of the optimized parameter set for the entire workspace is based on the assumption that a selection of extreme poses shows all temperature effects. However, this approach involves the risk that parts of the working space are not covered or that temperature-dependent effects are generally not taken into account. This leads to a visible residual drift at the component points. This residual drift on the component is composed of uncompensated, temperature-dependent influences on the positioning as well as a temperature drift due to heating of the kinematics since the last calibration.

The present invention is therefore based on the object to provide a method and a robot system, which minimize the disadvantages listed above - at least partially. It is a further object of the present invention to increase the precision of a robot or manipulator.

These and other objects are achieved by the subject of the main claims.

3. Content of the invention

The present invention includes a method for controlling a manipulator. The manipulator can be differently pronounced and is controlled based on a manipulator or robot parameter set.

The method according to the invention comprises providing a calibration, which is preferably a temperature-dependent calibration. Preferably, the providing may include creating the calibration. The calibration has the following steps: Repeated start of at least one reference point, and in particular two or M reference points until a temperature criterion of the manipulator is met, and determining, after each start, a calibration set comprising at least one comparison reference point residual drift and a manipulator parameter set.

The approach of a reference point means a detection of the reference point, which can be done, for example, optically or by touch. The reference point has a well-known position. Thus, an assignment to the world coordinate system and the basic coordinate system of the measuring robot is known. Preferably, the reference points are in the vicinity of the measuring points, processing points or component points on the component, and can also be arranged directly on the component. When determining the calibration rate, a residual drift with respect to the approached reference point is determined.

Preferably, at least two reference points are approached repeatedly, and after each approach of the two reference points, a calibration set is determined which comprises at least one comparison reference point residual drift and a manipulator parameter set. In this case, preferably one of the reference points is a component point which is to be approached by the manipulator or robot in the operating mode. Further preferably, one of the reference points is not a component point, and is also approached in normal working mode or during a normal operating cycle of the robot. The calibration set preferably comprises a comparison reference point residual drift, which was determined with respect to this reference point. The comparison reference point residual drifts are used to calibrate or optimize the model parameter set. This model parameter set is then provided as a manipulator parameter set in the calibration set. Thus, each calibration set includes a manipulator parameter set created from the residual drifts with respect to at least one comparison reference point residual drift at the reference point.

Preferably, the temperature criterion is selected such that the starting and determining is repeated until the temperature of the robot is in a stable state and changes only marginally. Thus, the calibration is preferably performed while the robot is heating. In particular, after activating or switching on a robot or manipulator, the temperature of the robot changes rapidly and reaches a constant value after a certain time. During this time, the calibration or temperature-dependent calibration is preferably carried out in order to advantageously detect all temperature-dependent effects.

The method according to the invention for controlling a manipulator also has a start-up of, in particular exactly one or N, reference points (s). This startup is carried out in the normal operating mode of the manipulator, or operating mode, and not during the calibration. Furthermore, the method comprises determining a current reference point residual drift. Consequently, in the productive operation, the instantaneous reference point residual drift with respect to the approached reference point is determined. Preferably, the number of reference points N is smaller than the number of reference points M. Thus, a precise correction for various manipulator actions can be granted.

Furthermore, the inventive method comprises calculating a correction parameter set based on the determined instantaneous reference point residual drift and based on at least one of the calibration sets. Thus, preferably at least the manipulator parameter set from the calibration is used to calculate the correction parameter set. In this case, the calibration set of the calibration from which the manipulator parameter set is to be used is selected based on the instantaneous reference point residual drift determined in the operating mode. Thus, advantageously, a suitable correction parameter set can be determined, wherein a suitable value or manipulator parameter set can be selected from the calibration on the basis of the determined instantaneous reference point residual drift.

Furthermore, the method according to the invention comprises controlling the manipulator based on the calculated correction parameter set. In this case, for example, the current robot model can be updated using the correction parameter set, or even be replaced by this. The person skilled in the art understands that not only the robot model can be optimized by means of the present invention but also that track points can be adapted from a path planning. In this case, the control preferably involves measuring with the aid of the manipulator, taking into account of the correction parameter. For example, a distance measurement can be carried out by means of the manipulator, wherein the manipulator preferably has its suitable device, such as a laser. Furthermore, the control may also include moving the manipulator in consideration of the correction parameter.

The "reference point" can thus be any point in the vicinity of the manipulator, and can be, for example, a component point, that is to say a point on a component or workpiece. The term "comparison reference point residual drift" can describe a residual drift at a reference point, which is determined in the course of the calibration. Analogously, the term "instantaneous reference point residual drift" can describe a residual drift at a reference point, which is determined in the course of the operating mode. In turn, a "calibration set" may be a data set that includes a plurality of values, such as a comparison reference point residual drift and a manipulator parameter set. One, several or all calibration sets determined during a calibration may be provided in a controller of the manipulator or separately. A "residual drift" itself does not have to be corrected, so it can only be a "drift".

Preferably, providing the calibration includes creating the calibration. Preferably, the calibration is carried out before the production operation of the manipulator, so that the calibration already exists during the working operation or measuring operation of the manipulator. Thus, the correction parameter sets can be determined directly based on the calibration, so that the manipulator can be controlled precisely based on the correction parameter set.

Preferably, the repeated starting comprises a repeated approach of two reference points, wherein preferably one of the reference points is a component point. Thus, by means of the calibration set, the model parameters or robot parameters can be optimized efficiently with respect to the points approached in the production mode. Preferably, after each repeated startup during calibration, a parameter set is stored for a comparison reference point residual drift and recorded in the calibration set.

Preferably, the determined calibration sets comprise two comparison reference point residual drifts and one manipulator parameter set.

Preferably, calculating the calibration set comprises the steps of: selecting a calibration set based on the determined instantaneous reference point residual drift such that the comparative reference point residual drift of the selected calibration set is closest to the determined instantaneous reference point residual drift, and generating the correction parameter set based on the selected calibration set. Thus, a best fit calibration set is selected such that the difference between the determined instantaneous reference point residual drift and the comparative reference point residual drift of the selected calibration set is minimal. This ensures that the calibration set from the calibration best matches the instantaneous state of the manipulator.

The creation of the correction parameter set based on the selected calibration set may include, for example, creating a correction parameter set based on the manipulator parameter set of the selected calibration set. Preferably, a calibration set is selected, wherein the comparison reference point residual drift from the calibration with the instantaneous reference point residual drift from the productive operation is closest, so that the manipulator parameter set determined in the calibration optimally matches the instantaneous state of the manipulator and thus can be used for the calculation of the correction parameter set and thus for controlling the manipulator, such as the start or measurement of a component point.

Preferably, calculating the correction parameter set comprises the steps of: selecting at least two different calibration sets based on the determined instantaneous reference point residual drift such that the comparison reference point residual drifts of the selected calibration sets are closest to the determined instantaneous reference point residual drift and creating the correction parameter set based on an interpolation of the selected calibration sets. Thus, two calibration sets are used, which best reflect the current state of the manipulator. The correction parameter set is then created based on an interpolation of these two calibration sets or based on an interpolation of the manipulator parameter sets of the selected calibration sets. It is thus possible to obtain a correction parameter set from the calibration that optimally matches the current state of the manipulator.

The person skilled in the art understands that more than two calibration sets can also be selected based on the instantaneous reference point residual drift determined in the production mode and can be used for the calculation of the correction parameter set. Thus, for example, if three calibration sets have been selected, a correction parameter set can be calculated based on a spline interpolation of the calibration sets or the corresponding manipulator parameter sets. Further the skilled person understands that the method according to the invention can also be combined with other methods for the compensation of different drifts. In particular, the method according to the invention is suitable for minimizing residual drift or the effect of residual drift.

With the present invention, a calibration is thus carried out in which a great many (temperature-dependent) states of the robot are detected. By comparing a momentary reference point residual drift with a comparison reference point residual drift from the calibration, a suitable manipulator parameter set from the calibration can thus be used for the instantaneous state of the manipulator. In this case, the present invention allows to improve the positioning accuracy of a robot or manipulator. In this case, for example, a high accuracy of better than +0.15 mm can be achieved.

The present invention further comprises a robot system having means for carrying out the method according to the invention. The means include in particular a robot controller. Further, the present invention includes a computer-readable medium containing instructions that, when executed by a processing system, perform steps to control a manipulator according to the inventive method for controlling a manipulator.

4th embodiment (e)

In the following, the present invention will be described in more detail with reference to the accompanying figures. It shows:

1 the sequence of the calibration according to the present invention, and

2 an inventive method for controlling a manipulator.

In 1 is schematically the process or process 10 a (temperature-dependent) calibration according to the present invention. The process 10 starts in step 11 , In step 12 a reference point is approached by the manipulator. This reference point is preferably not on the component, but is provided independently of the component. In step 13 is a comparison reference point residual drift with respect to in step 12 approached reference point determined.

In step 14 a component point is approached, ie a point on the component to be measured or machined is approached by the manipulator. In the following step 15 a comparison component residual drift is determined, ie the residual drift between the manipulator and the approached component point.

In step 16 a correction parameter set is created. For this purpose, at least in step 13 determined comparison reference point residual drift and / or in step 15 determined comparison component residual drift a manipulator parameter set, or model parameter set determined. In the correction parameter set, this manipulator parameter set is replaced with that in step 13 linked comparison reference point residual drift linked. Subsequently, this correction parameter set is stored in a corresponding calibration database.

At the decision 17 it is checked whether the temperature of the manipulator has exceeded a predefined limit. Alternatively, in the decision 17 Also be checked whether the temperature of the manipulator has reached a constant value. If the decision 17 is negative, a new calibration set is determined by following the procedure in step 12 is continued. If the decision 17 positive, the calibration ends in step 18 ,

Thus, several calibration sets are determined with the calibration, wherein in each calibration set a comparison reference point residual drift is linked to a manipulator parameter set. Each calibration set was determined at a different temperature of the manipulator. Only when the temperature of the manipulator is preferably approximately stable, the calibration is terminated.

In 2 is schematically the process 20 of a method according to the invention for controlling a manipulator. This method is preferably performed in the working or operating mode. The procedure begins in step 21 , In step 22 a (temperature-dependent) calibration is provided. Preferably, the calibration has been performed according to the in 1 performed process and provides appropriate calibration kits.

In step 23 a reference point is approached by the manipulator. This reference point is identical to the reference point which was approached or measured during the calibration. In step 24 an instantaneous reference point residual drift is determined, ie the residual drift between the manipulator and the approached reference point.

In step 25 two calibration sets are selected from the calibration. The calibration sets are compared by comparing in step 24 selected instantaneous reference point residual drift and the comparison reference point residual drift of the calibration sets of the calibration. The comparison determines which comparison reference point residual drifts of the calibration sets of calibration are closest to that in step 24 certain instantaneous reference point residual drift. Correspondingly, two calibration sets with the closest comparison reference point residual drifts with respect to the instantaneous reference point residual drift are then selected.

In step 26 For example, the manipulator parameter sets of the selected calibration sets associated with the two comparison reference point residual drifts are interpolated for the calibration, and a correction parameter set is calculated on the basis of the interpolation. In the following step 27 the manipulator is controlled based on at least the correction parameter set. Then the process ends in step 28 ,

Claims (9)

A method of controlling a manipulator, comprising: a) providing a calibration, comprising the following steps: - Repeated start of at least one, in particular at least two or M, reference points (s) until a temperature criterion of the manipulator is met, and Determining, after each startup, a calibration set comprising at least one comparison reference point residual drift and a manipulator parameter set; b) approaching one, in particular exactly one or N, reference points (s); c) determining a current reference point residual drift; d) calculating a correction parameter set based on the determined instantaneous reference point residual drift and based on at least one of the calibration sets, and e) controlling the manipulator based on the correction parameter set. The method of claim 1, wherein controlling the manipulator comprises measuring with the aid of the manipulator taking into account the correction parameter set. Method according to claim 1 or 2, wherein the number of reference points N from step b) is smaller than the number of reference points M from step a). The method of any one of the preceding claims, wherein providing the calibration comprises creating the calibration. Method according to one of the preceding claims, wherein the repeated start comprises a repeated approach of two reference points, and wherein preferably one of the reference points is a component point. Method according to one of the preceding claims, wherein the determined calibration sets comprise two comparison reference point residual drifts and a manipulator parameter set. Method according to one of the preceding claims, wherein the calculating comprises the following steps: Selecting a calibration set based on the determined instantaneous reference point residual drift such that the comparative reference point residual drift of the selected calibration set is closest to the determined instantaneous reference point residual drift, and - Create the correction parameter set based on the selected calibration set. Method according to one of the preceding claims, wherein the calculating comprises the following steps: Selecting at least two different calibration sets based on the determined instantaneous reference point residual drift such that the comparison reference point residual drifts of the selected calibration sets are closest to the determined instantaneous reference point residual drift, and - Creating the correction parameter set based on an interpolation of the selected calibration sets. Robot system comprising means for performing a method according to any one of claims 1-8.

Images