FR3115482A1 - Method for calibrating a cell, in particular a robotic cell, and associated cell - Google Patents

Abstract

Method for calibrating a cell, in particular robotic, and associated cell The invention relates to a method for calibrating a cell (1) comprising at least one machine (R1, R2) defining a fixed reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2) and at least one mobile platform (3). According to the invention, the calibration method comprises a phase of characterization of the relative positioning of the platform (3) with respect to said machine (R1, R2), according to which a measurement unit (7) records the coordinates of at least one standard (REF1, REF2, REF3-1, REF3-2) positioned on the platform (3) in at least one predefined position and the control unit expresses the coordinates recorded in the reference (O1, X1, Y1, Z1 O2, X2, Y2, Z2) of said machine (R1, R2), a phase of identifying deviations in the relative positioning of the platform (3) with respect to said machine (R1, R2) in the real world with respect to to a theoretical virtual model, and a defect correction phase. The invention also relates to an associated cell. Figure for the abstract: Fig. 1

Description

Method for calibrating a cell, in particular a robotic cell, and associated cell

The present invention is in the field of cells comprising at least one machine intended to work on at least one part positioned on a mobile platform. The invention relates to the method for calibrating such a cell. The invention applies in particular to a robotic cell comprising one or more robots.

In general, the particular robotic cell comprises one or more machines.

By machine we mean both a machine tool, a Cartesian machine, and an articulated, poly-articulated industrial robot. A Cartesian machine uses a Cartesian coordinate system and generally evolves with linear movements, while an articulated, poly-articulated robot uses a polar coordinate system, and has at least one arm capable of pivoting around an axis of 'joint.

The machines (robots) make it possible to carry out respectively at least one operation, an automated task, for example machining, by means of a corresponding effector. The effector is also called aggregate in the particular field of woodworking.

Machining can be milling, drilling, multi-drilling, or any other material removal process. It can advantageously be precision machining on parts of complex shapes, for example at least partly concave, convex, or conical, or multi-materials or even of different thicknesses.

Different applications can be envisaged, in particular in mechanical construction.

In particular, in the aeronautical field, an application may be the machining of acoustic plates for aircraft nacelles, for the production of noise reduction devices, such as Helmholtz resonators, making it possible to reduce the noise of turbojet engines housed in the nacelles.

The nacelles are generally tubular in shape, or more specifically have sections of conical or frustoconical shape.

The acoustic plates are generally made of carbon composite, aluminum, titanium, inconel. They have a multitude of small-diameter holes, distributed in such a way as to provide an acoustic absorption effect for the noise generated by the turbojet engines. In order to create Helmholtz resonators with an effective impact on sound absorption, the distribution of holes in the acoustic plate is very important and significantly influences the noise level generated by the turbojet engine.

However, such an acoustic plate can only be perforated after having been shaped beforehand. As a result, the acoustic plate generally has a surface of conical, frustoconical, concave, convex or complex shape (with concave zones and convex zones). To achieve effective, optimized sound absorption, it is necessary that the precise positioning of the holes made be respected within fairly restricted limits, even on such complex shapes.

The cell also comprises a platform intended to support a tool, the part to be worked by the machines, the robots. This platform can be mobile and include one or more devices to support the workpiece and cause it to move. The platform may include, in a non-exhaustive manner, a fixed worktable, a mobile worktable along an axis, a rotary table, a rotator.

The overall architecture of the cell is defined by the architecture of these multiple sub-assemblies, put in relation to each other (the platform devices, mobile or not, the machines / robots and even their possible supports).

A program, for example managed by a control unit of the cell, makes it possible, when it is executed, to control the movement of this entire architecture, to move the machines (the robots), to move one or more mobile devices of the platform carrying the workpiece.

According to a known solution, such a program is generated based on a theoretical virtual model of the architecture of the cell which is perfect.

However, the reality is generally different. The relative positions of the sub-assemblies of the cell, their orientations, often present shifts or defects with respect to the theoretical virtual model.

The implementation of these subsets involves residual defects and therefore inaccuracy. The more complex the architecture of the cell, i.e. the more subsets there are, the more the defects pile up. The inaccuracy is all the more important. This can be particularly critical when dealing with a particularly demanding machining operation in terms of precision, i.e. the machine / robot must move with precision, to reach a precise location on the part to be machined, for example.

In order to remedy this imprecision, a known solution seeks to conform reality to theory. In other words, significant tuning devices are implemented to install the cell so that the real-world installation matches the theoretical virtual model. This involves lengthy and costly adjustment and installation times for the cell. Moreover, it is extremely difficult to obtain a real-world installation that is perfectly identical to the theory. And once the installation is complete, it is complicated to start the process again.

According to another solution, this time we seek to conform the theory to reality. In this case, the cell is installed and measurement devices external to the cell are added, and are implemented to identify the faults between theory and reality. Depending on the identified defects, the theoretical virtual model is deformed to conform to reality. And once this compliance is complete, it is complicated to start the process again.

Mostly, the external measuring devices used are tracker lasers also called pursuit lasers. However, such lasers are complex and expensive to implement.

Moreover, according to one or the other of the solutions of the prior art, the bringing into conformity satisfies only the defects identified at a given moment. These solutions cannot be implemented easily and quickly at any time, to take into account, for example, changes over time of the sub-assemblies, in particular with respect to wear over time, to aspects of variability of the atmosphere, temperature. Indeed, each implementation would be tedious, expensive and time-consuming to implement. As a result, defects may appear and/or evolve over time.

The aim of the invention is to overcome the aforementioned drawbacks in order to guarantee a cell whose machines, in particular the robots in the case of a robotic cell, move with great precision. Another object of the invention is to guarantee stability over time.

To this end, the subject of the invention is a method for calibrating a cell comprising at least one machine defining a fixed reference point and at least one platform mobile with respect to at least one axis, such that the machine comprises an arm configured for carrying at least one tool so as to perform at least one automated operation on a workpiece intended to be positioned on the mobile platform. The calibration method comprises a preliminary step of arranging said at least one machine with respect to the mobile platform in the real world, from a theoretical virtual model of the cell used to program movements of said at least one machine and of the mobile platform, and of the operations to be carried out by said at least one machine on the workpiece according to a program managed by a control unit of the cell.

• une phase de caractérisation du positionnement relatif de la plateforme mobile par rapport à ladite au moins une machine dans le monde réel, selon laquelle une unité de mesure portée par le bras relève les coordonnées d’au moins un étalon positionné sur la plateforme mobile dans au moins une position prédéfinie et l’unité de commande exprime les coordonnées relevées dans le repère de ladite au moins une machine,
• une phase d’identification des écarts du positionnement relatif de la plateforme mobile par rapport à ladite au moins une machine dans le monde réel par rapport au modèle virtuel théorique, selon laquelle l’unité de commande compare des coordonnées relevées par l’unité de mesure et exprimées dans le repère de ladite au moins une machine avec les coordonnées établies selon le modèle virtuel théorique, et
• une phase de correction des défauts, selon laquelle l’unité de commande détermine en fonction des écarts identifiés au moins une consigne de correction à appliquer sur une chaine cinématique virtuelle de ladite au moins une machine, et enregistre la consigne de correction au niveau du programme de la cellule.

According to the invention, the calibration method comprises the following phases:

• a characterization phase of the relative positioning of the mobile platform with respect to said at least one machine in the real world, according to which a measurement unit carried by the arm records the coordinates of at least one standard positioned on the mobile platform in at least at least one predefined position and the control unit expresses the coordinates recorded in the reference frame of said at least one machine,
• a phase of identifying deviations from the relative positioning of the mobile platform with respect to said at least one machine in the real world with respect to the theoretical virtual model, according to which the control unit compares the coordinates recorded by the measurement unit and expressed in the frame of said at least one machine with the coordinates established according to the theoretical virtual model, and
• a fault correction phase, in which the control unit determines, based on the deviations identified, at least one correction setpoint to be applied to a virtual kinematic chain of said at least one machine, and saves the correction setpoint at program level of the cell.

The invention applies in particular to a robotic cell comprising at least one robot defining a fixed frame of reference and at least one platform mobile with respect to at least one axis, such that the robot comprises a robot arm configured to carry at least one tool so as to perform at least one automated operation on a workpiece intended to be positioned on the mobile platform. The calibration method comprises in this case a preliminary step of arranging said at least one robot with respect to the mobile platform in the real world, from a theoretical virtual model of the robotic cell used to program movements of said robot and of the mobile platform, and of the operations to be carried out by said robot on the workpiece according to a program managed by a control unit of the robotic cell.

• une phase de caractérisation du positionnement relatif de la plateforme mobile par rapport audit au moins un robot dans le monde réel, selon laquelle une unité de mesure portée par le bras de robot relève les coordonnées d’au moins un étalon positionné sur la plateforme mobile dans au moins une position prédéfinie et l’unité de commande exprime les coordonnées relevées dans le repère dudit au moins un robot,
• une phase d’identification des écarts du positionnement relatif de la plateforme mobile par rapport audit au moins un robot dans le monde réel par rapport au modèle virtuel théorique, selon laquelle l’unité de commande compare des coordonnées relevées par l’unité de mesure et exprimées dans le repère dudit au moins un robot avec les coordonnées établies selon le modèle virtuel théorique, et
• une phase de correction des défauts, selon laquelle l’unité de commande détermine en fonction des écarts identifiés au moins une consigne de correction à appliquer sur une chaine cinématique virtuelle dudit au moins un robot, et enregistre la consigne de correction au niveau du programme de la cellule robotique.

The calibration method according to the invention comprises in this case the following phases:

• a characterization phase of the relative positioning of the mobile platform with respect to said at least one robot in the real world, according to which a measurement unit carried by the robot arm records the coordinates of at least one standard positioned on the mobile platform in at least one predefined position and the control unit expresses the coordinates recorded in the reference frame of said at least one robot,
• a phase of identifying deviations from the relative positioning of the mobile platform with respect to said at least one robot in the real world with respect to the theoretical virtual model, according to which the control unit compares the coordinates recorded by the measurement unit and expressed in the frame of said at least one robot with the coordinates established according to the theoretical virtual model, and
• a fault correction phase, according to which the control unit determines, based on the identified deviations, at least one correction instruction to be applied to a virtual kinematic chain of said at least one robot, and records the correction instruction at the level of the the robotic cell.

The calibration method according to the invention makes it possible to define the mobile platform in the reference frame of the machine, for example of the robot, and thus to identify the real architecture of the cell. This is achieved by using the cell directly, more precisely the unit of measurement carried by the machine (the robot) and the calibration markers, without any other tool, without having recourse to means or additional measuring devices external to the cell. which can be costly and mechanically cumbersome to implement. In other words, the cell can self-identify its architecture, based solely on elements already present.

Such a solution is easier, shorter and less expensive to implement than the solutions of the prior art requiring the installation and the intervention of external measuring devices such as tracker lasers also called tracking lasers.

When running the program to machine a part, for example, the values corrected following the correction phase are taken into account by the machines/robots, and not the theoretical virtual model expressed in the program.

Finally, the calibration process can be implemented at any time. This solution is self-adaptive because it can be implemented at any time by integrating any effect, such as dilation effects, which have an impact on the machine, the robot.

The calibration process may also include one or more of the following characteristics described below, taken separately or in combination.

• au moins une étape d’entrainement de la plateforme mobile dans le monde réel, commandée par l’unité de commande de la cellule, de sorte que ledit au moins un étalon positionné sur la plateforme mobile prenne au moins deux positions distinctes, ledit au moins un étalon étant calibré, et
• au moins deux étapes de mesure réalisées par l’unité de mesure portée par le bras, lorsque ledit au moins un étalon est dans les deux positions distinctes, de façon à relever les coordonnées dudit au moins un étalon respectivement dans les deux positions distinctes, les coordonnées relevées sont transmises à l’unité de commande.

According to one embodiment, the characterization phase of the relative positioning of the mobile platform with respect to said at least one machine comprises:

• at least one step of driving the mobile platform in the real world, controlled by the control unit of the cell, so that said at least one standard positioned on the mobile platform takes at least two distinct positions, said at least a standard being calibrated, and
• at least two measurement steps carried out by the measurement unit carried by the arm, when said at least one standard is in the two distinct positions, so as to record the coordinates of said at least one standard respectively in the two distinct positions, the recorded coordinates are transmitted to the control unit.

During the characterization phase of the relative positioning of the mobile platform with respect to said at least one machine, the control unit can express the coordinates of a direction vector of said at least one axis in the frame of said at least one machine.

The method may include a preliminary step of calibrating the stroke of the interface along the translation axis, for example in a non-limiting way from 0mm to 3000mm.

According to one embodiment, the mobile platform comprises a linearly mobile interface along a translation axis, the stroke of the interface along the translation axis being calibrated. During a first training step, the interface is driven in linear displacement until a standard on the interface reaches a first linear position. In a second drive step, the interface is driven in linear motion until the etalon on the interface reaches a second linear position.

During the measurement steps, the measurement unit can respectively record the coordinates of said standard in the first linear position and in the second linear position.

During a step of characterizing the axis of translation relative to said at least one machine, the control unit expresses the coordinates of said standard in the first linear position and the coordinates in the second linear position in the fixed frame of said at least one machine. least one machine.

During the translation axis characterization step, the coordinates of a direction vector of the translation axis can be expressed in the fixed frame of said at least one machine.

The first linear position can be a start or end position of the interface. As a variant, the first linear position can be close to the start or end position, within the limit of the capacity of the arm carrying the measurement unit to reach the standard on the interface.

The second linear position can be a start or end position of the interface. As a variant, the second linear position can be close to the start or end position, within the limit of the capacity of the arm carrying the measurement unit to reach the standard on the interface.

According to one embodiment, the mobile platform comprises a rotating plate, mobile in rotation around an axis of rotation. According to this embodiment, said method includes a phase of setting an original angular position.

During the phase of setting the original angular position, the rotary plate can be driven in rotation, by the control unit of the cell, until the measuring unit carried by said at least one machine detects a standard on the rotary plate in a position such that said standard coincides with a first predefined axis of the fixed reference of said at least one machine and that the coordinate of said standard, along a second predefined axis of the fixed reference of said at least one machine, either at zero.

The turntable can be rotatably mounted on the linearly movable interface. In this case, the phase of setting the original angular position can be preceded by a step of driving the interface in displacement until the standard on the interface reaches a predefined position according to the translation axis. This is, for example, a mid-stroke position.

• Selon une première étape d’entrainement en rotation du plateau rotatif, l’unité de commande peut entrainer le plateau rotatif depuis la position angulaire d’origine jusqu’à ce que l’étalon sur le plateau rotatif atteigne une deuxième position angulaire.
• Selon une deuxième étape d’entrainement en rotation du plateau rotatif, l’unité de commande peut entrainer le plateau rotatif depuis la position angulaire d’origine jusqu’à ce que l’étalon sur le plateau rotatif atteigne une troisième position angulaire.
• Trois étapes de mesure peuvent être mises en œuvre, dans lesquelles l’unité de mesure relève respectivement les coordonnées dudit étalon dans la position angulaire d’origine, dans la deuxième position angulaire et dans la troisième position angulaire.
• Selon au moins une étape d’identification d’un centre de rotation du plateau rotatif, l’unité de commande peut définir un plan de rotation, à partir des trois positions angulaires, et en déduire le centre de rotation.
• Selon une étape de détermination, l’unité de commande peut déterminer un vecteur directeur de l’axe de rotation normal au plan de rotation défini et passant par le centre de rotation.
• Selon une étape de caractérisation, l’unité de commande peut exprimer les coordonnées du vecteur directeur de l’axe de rotation dans le repère fixe de ladite au moins une machine.

According to an example, the characterization phase of the relative positioning of the mobile platform with respect to said at least one machine comprises the following steps so as to characterize the axis of rotation with respect to said at least one machine.

• According to a first step of driving the rotary plate in rotation, the control unit can drive the rotary plate from the original angular position until the standard on the rotary plate reaches a second angular position.
• According to a second step of driving the rotary plate in rotation, the control unit can drive the rotary plate from the original angular position until the standard on the rotary plate reaches a third angular position.
• Three measurement steps can be implemented, in which the measurement unit respectively records the coordinates of said standard in the original angular position, in the second angular position and in the third angular position.
• According to at least one step of identifying a center of rotation of the rotary plate, the control unit can define a plane of rotation, from the three angular positions, and deduce the center of rotation therefrom.
• According to a determination step, the control unit can determine a directing vector of the axis of rotation normal to the plane of rotation defined and passing through the center of rotation.
• According to a characterization step, the control unit can express the coordinates of the direction vector of the axis of rotation in the fixed frame of said at least one machine.

Preferably, the second and third angular positions form an angle of the same measure with respect to the original angular position. This is for example an angle of 120°.

• au moins trois étapes de mesure, dans lesquelles l’unité de mesure portée par le bras peut relever les coordonnées d’au moins trois points dans un plan de référence défini sur le plan de travail, lorsque le plateau rotatif est dans la position angulaire d’origine,
• au moins une étape de caractérisation, dans laquelle l’unité de commande peut exprimer les coordonnées desdits points du plan de référence dans le repère du plateau rotatif,
• au moins une étape de mesure, dans laquelle l’unité de mesure portée par le bras peut identifier un alésage de référence dans le plan de référence, lorsque le plateau rotatif est dans la position angulaire d’origine,
• au moins une étape d’analyse, dans laquelle l’unité de commande peut déterminer le centre de l’alésage de référence identifié, correspondant au centre du plan de travail, et
• au moins une étape de caractérisation, dans laquelle l’unité de commande peut exprimer les coordonnées du centre du plan de travail dans le repère du plateau rotatif.

According to one embodiment, the turntable includes a work surface on which the workpiece is intended to be placed, the turntable defining a mark. In this case, said method comprises a phase of characterization of the relative positioning of the work plane with respect to the plane of rotation of the rotary plate comprising the following steps:

• at least three measuring steps, in which the measuring unit carried by the arm can take the coordinates of at least three points in a reference plane defined on the work plane, when the rotary table is in the angular position d 'origin,
• at least one characterization step, in which the control unit can express the coordinates of said points of the reference plane in the frame of reference of the rotary plate,
• at least one measuring step, in which the measuring unit carried by the arm can identify a reference bore in the reference plane, when the rotary table is in the original angular position,
• at least one analysis step, in which the control unit can determine the center of the identified reference bore, corresponding to the center of the working plane, and
• at least one characterization step, in which the control unit can express the coordinates of the center of the work plane in the frame of reference of the rotary table.

According to the example with the rotary plate comprising the work surface, said method further comprises a phase of identifying the deviations of the relative positioning of the work surface with respect to the rotary plate in the real world and according to the theoretical virtual model.

The three points in the reference plane make it possible to characterize the work plane (support plane for the part to be worked) in the reference frame of the rotary table. This makes it possible to identify any defects in the parallelism of the work plane with respect to the rotation plane.

These steps make it possible to identify a possible rotational offset of the work surface relative to the rotary plate as well as a possible misalignment of the center of the work surface relative to the axis of rotation according to the component according to the vertical axis of the center of the worktop.

Advantageously, the reference plane is defined by the bottom of the reference bore, so as to avoid introducing a lack of flatness.

In particular, the cell can comprise at least two machines respectively defining a benchmark, the machines being configured to respectively carry at least one tool so as to perform automated tasks on the same part intended to be positioned on the mobile platform.

The theoretical setpoint value of said at least one axis of the mobile platform is equal to the actual setpoint value of said at least one axis.

According to this example, at least the phases of characterization of the relative positioning of the mobile platform and identification of deviations are implemented at least in part with respect to each machine.

In particular, the characterization of the relative positioning of the translation axis can be implemented with respect to each machine.

According to the embodiment with a rotary plate, the phase of setting the original angular position is implemented with respect to a machine. Then, the characterization of the relative positioning of the axis of rotation can be implemented with respect to each machine.

During the correction phase, the control unit can determine, based on the identified differences in relative positioning of the mobile platform with respect to a first machine, at least one correction setpoint to be applied to a setpoint value of the virtual kinematic chain of the first machine as a function of the theoretical and actual setpoint value of said at least one axis of the mobile platform. Similarly, the control unit can determine, based on the identified differences in relative positioning of the mobile platform with respect to a second machine, at least one correction setpoint to be applied to a setpoint value of the virtual kinematic chain of the second machine as a function of the theoretical and actual setpoint value of said at least one axis of the mobile platform.

The cell such as a robotic cell is installed at an initial time. The calibration process is implemented at this initial instant in order to identify the faults and integrate them at the program level. The calibration process can be repeated at any time. This reiteration at any time is possible because the calibration process does not use other materials external to the cell, in particular robotics. The calibration process can be easily implemented, restarted at any time.

We can provide a preliminary phase of calibration of the machine, of the robot, so as to obtain the real kinematic chain of the machine, of the robot.

It is also possible to provide a preliminary phase of identifying the position of the center of the tool, for example with respect to an output flange of the carrier axis of said robot.

These preliminary phases can be carried out independently or be integrated into the calibration process as described.

Preferably, at the initial instant, the calibration process is implemented at a predetermined reference temperature, for example at 20°C. The predetermined reference temperature corresponds to the temperature at which said machine, said robot has been calibrated.

The invention also relates to an associated cell. The cell comprises at least one machine, a platform movable relative to said at least one machine, and a control unit, said at least one machine comprising at least one arm configured to carry at least one tool so as to perform at least one automated operation on a part intended to be positioned on the mobile platform. The cell is configured to implement at least certain steps of a calibration method as defined above, the arm carrying a measurement unit configured to implement at least one measurement step and the mobile platform having at least one standard.

In particular, the cell is a robotic cell comprising at least one robot comprising a robot arm carrying the measurement unit, and the platform is mobile relative to said at least one robot.

The measurement unit is for example a measurement unit by contact such as a 3D feeler or other identification means of the feeler type, or without contact, for example of the optical type. The contact measurement unit such as a 3D probe makes it possible to probe a marker, a reference element, a reference plane or even a reference bore, a reference groove, and the control unit can integrate the coordinates with respect to a given reference frame.

At least one standard is chosen from a sphere, a ring. According to one embodiment, the mobile platform has at least one other standard from among a reference bore, a reference groove, a reference plane.

Other advantages and characteristics of the invention will appear more clearly on reading the following description given by way of illustrative and non-limiting example, and the appended drawings, among which:

FIG. 1 is an overall view of a robotic cell for machining a part comprising robots and a mobile platform.

Figure 2 shows an embodiment of a linearly movable interface of the mobile platform of Figure 1.

Figure 3 is a DC cross-sectional view of a standard sphere on the linearly moving interface of Figure 2.

Figure 4 shows one embodiment of a turntable of the mobile platform of Figure 1.

Figure 5 is a sectional view AA of the turntable of Figure 4.

Figure 6 is a detail view of Figure 5 showing a standard sphere, a reference bore and a reference plane on the rotary plate.

Figure 7 is a schematic representation of the reference frames of the subassemblies of the robotic cell of Figure 1.

FIG. 8 is a table of values of theoretical and real axes of the robotic cell for examples.

In these figures, identical elements bear the same reference numbers.

The following achievements are examples. Although the description refers to one or more embodiments, this does not necessarily mean that each reference is to the same embodiment, or that the features apply only to a single embodiment. Simple features of different embodiments can also be combined or interchanged to provide other embodiments.

In the description, certain elements can be indexed, such as for example first, second element. It can be a simple indexing to differentiate and name close but not identical elements. This indexing does not necessarily imply a priority of one element over another and it is easy to interchange such denominations without departing from the scope of the present description. Nor does this indexing necessarily imply an order in time.

CELL
The invention relates to a cell, or operational cell comprising one or more machines and at least one mobile platform 3 shaped to carry at least one part on which the machine or machines must work.

In the present, the term “machine” refers both to a robot R1, R2 as illustrated in the example of figure 1, and to a machine tool, a Cartesian machine.

The cell is in particular a robotic cell 1 comprising at least one robot R1, R2, in particular an industrial robot. This robot R1, R2 is articulated, poly-articulated.

The rest of the description refers to such a robotic cell 1.

The robotic cell 1 represented schematically in FIG. 1 makes it possible to carry out one or more operations, in particular machining, on one or more parts. This may involve precision machining, for example, in a non-limiting way, the machining of acoustic plates for the production of Helmholtz resonators. For example, without limitation, the robotic cell 1 can carry out milling, drilling, multi-drilling, and in particular multi-drilling operations with a number of spindles or electro-spindles for different drilling.

In the example shown, robotic cell 1 has two robots R1, R2. Of course, this number is not limiting. The two robots R1, R2 are for example intended to face each other.

The robots R1, R2 respectively define a fixed reference. According to the example described with two robots, a first robot R1 defines a first fixed frame O1, X1, Y1, Z1, and a second robot R2 defines a second fixed frame O2, X2, Y2, Z2. With reference to the orientation in figure 1, the axes X1, Y1, respectively X2, Y2 define a horizontal plane and the axis Z1, respectively Z2, is vertical.

The robots R1, R2 each comprise at least one robot arm 5 configured to carry at least one tool so as to perform at least one automated operation on the workpiece.

The robots R1, R2 are articulated, poly-articulated. These are, for example, six-axis robots.

In particular, the robots R1, R2 respectively comprise a base from which extends the robot arm 5. The base can be fixed.

The robot arm 5 is configured to carry at least one tool, by means of an effector, so as to be able to perform at least one automated operation on the workpiece. For example for machining, the robot arm 5 can carry a machining spindle.

The robot arm 5 carries a measurement unit 7. This is preferably a measurement unit 7 fitted by default to the robots R1, R2.

The measurement unit 7 is advantageously a contact measurement unit such as a 3D feeler, as shown schematically in the zoomed part of FIG. 1. Such a 3D feeler is advantageous because in a large majority of robots are equipped with it by default. .

The contact measurement unit such as a 3D probe makes it possible to probe a standard, a marker, a reference element, a reference plane or even a reference bore.

The measurement unit 7, such as a 3D probe, is carried by the last axis of the robot R1 or R2.

Other identification devices or means may be envisaged, for example of the feeler type, or without contact, for example of the optical type.

The mobile platform 3 for its part is shaped to carry at least one part on which the robot or robots R1, R2 must work.

The mobile platform 3 can comprise one or more mobile elements among, an interface, a tray, a work table, a work plan. The platform 3 may optionally further comprise one or more fixed elements. The mobile platform 3 can comprise several devices.

In addition, the platform 3 is mobile with respect to at least one axis V, C. Generally, one or more devices of the mobile platform 3 can move linearly and/or be rotary.

The mobile platform 3 comprises for example an interface 9 mobile in a linear fashion along a translation axis V.

In theory, the translation axis V is assumed to be parallel to the axis Y1, respectively Y2, of the robot R1, respectively R2. It evolves for example in the middle, at an equal distance from the two robots R1, R2.

The mobile platform 3 also comprises a rotary plate 11 around an axis of rotation C. The rotary plate 11 defines another mark O′, Xc, Yc, Zc.

In theory, the axis of rotation C is for example orthogonal to the axis of translation V.

The V, C axes are common for the R1, R2 robots. They are named auxiliary axes in order to differentiate them from the axes of the robots R1, R2.

As in the example shown, the turntable 11 can be mounted on the interface 9 which can move linearly along the translation axis V.

The mobile platform 3 further comprises a work table or a work surface 13 on which, for example, a tool can be arranged with the workpiece. In the example described, the work plane 13 is defined on the rotary plate 11. More precisely, the work plane 13 is defined by the upper face of the rotary plate 11 along the vertical axis Z1 of the reference mark of the first robot R1, respectively Z2 of the reference frame of the second robot R2. In theory, this work plane 13 is perfectly coaxial with the rotary plate 11. Thus, the normal N to the work plane 13 corresponds in theory to the axis of rotation C.

The mobile platform 3 also has at least one standard REF1, REF2, REF3-1, REF3-2. The quality of the standard is important so as not to introduce defects.

At least one standard REF1, REF2 is chosen from a sphere, a ring. Such a standard sphere or ring is chosen, dimensioned according to a compromise to optimize the approach distances of the measuring unit 7 and to minimize the amplitude of movement to carry out the measurements (as described later) at the level of the sphere. or standard ring. In a non-limiting manner, the diameter of the standard sphere or ring can for example be of the order of 10mm to 200mm.

The mobile platform 3 can also present at least one other standard from a reference bore REF3-2, a reference plane REF3-1.

At least one standard REF1 can be provided on the mobile interface 9 in a linear fashion. It can be for example a sphere or a standard ring. An exemplary embodiment is shown in Figures 2 and 3.

At least one other standard REF2, REF3-1, REF3-2 can be provided on the rotary plate 11. An example embodiment is shown in FIGS. 4 to 6.

In the example described, the rotary plate 11 can comprise a standard REF2 such as a sphere or a standard ring. In order to differentiate the standards, such as spheres or standard rings, the standard REF1 on the interface 9 (figures 2-3) is named first standard REF1, and the standard REF2 on the rotary plate 11 (figures 4-6 ) is named second stallion REF2. However, this does not imply priority of one over the other. The second standard REF2 can be similar to the first standard provided on the mobile interface in a linear fashion.

In addition, the rotary plate 11 may have other standards, for example a bore or a reference groove REF3-2, visible in Figures 4 and 6. The rotary plate 11 may still have a reference plane REF3-1. The reference plane REF3-1 is for example defined by the bottom of the reference bore REF3-2.

Furthermore, a theoretical virtual model of the particular robotic cell 1 (FIG. 1) is generated by a modeling system. The modeling system is for example a CAD system for computer-aided design and/or CAM for computer-aided manufacturing.

The theoretical virtual model is used to program movements of the robots R1, R2 and of the mobile platform 3, and operations to be carried out by the robots R1, R2 on the workpiece, according to a program, which can be managed, executed by a control unit of the robotic cell 1.

The control unit may include one or more processing means. These processing means are configured in particular to manage the execution of the program, for example machining. In general, the processing means may comprise one or more of the means from one or more means of communication, at least one comparator, at least one calculator or interpolator, a processor and/or any other hardware making it possible to execute a program or software, at least one memory, or other conventional or custom hardware.

The program can express, from the theoretical virtual model, the theoretical virtual kinematic chains of the robot or robots R1, R2 with a theoretical positioning of all the joints, of the length of the robot arm 5, of the length of the axes. According to a particular non-limiting example, the program can express the coordinates of all the axes, for example of the six axes of the robots R1, R2. The program can also express the positioning of the so-called auxiliary axes along which the platform 3 is mobile, i.e. in the example written the axis of translation V and the axis of rotation C.

Other programming examples otherwise known to those skilled in the art of robotic programming are possible. In a non-limiting way, a variant can be to use a programming, which is in particular standardized ISO for “International Organization for Standardization” in English, to express the end of the tool and the normal of the tool carried by the robot. R1, R2. According to this programming, the coordinates of the points to be reached by the robot R1, R2 can be given with the orientation of the normal to the points and bits which impose the posture of the robot R1, R2 (depending on whether these bits take the value zero or a) to achieve this normal at this point. In principle, a six-axis robot is able to reach the same point with its normal from eight different postures.

The preceding description refers to a robotic cell 1 comprising at least one robot R1, R2, in particular an industrial robot. Of course, this description can also apply to a machine tool, a conventional Cartesian machine, in particular this machine can include an arm carrying the measuring unit 7 previously described, and whose movements are programmed according to the theoretical virtual model.

The particular robotic cell 1 is configured to implement at least certain steps of a calibration method described below. In particular, the measurement unit 7 is configured to implement at least one measurement step. The control unit, and in particular one or more of these processing means, can at least partially implement the calibration method.

CALIBRATION PROCEDURE
The calibration process makes it possible to identify the real architecture of the cell, in particular robotics 1 compared to the machine(s) / robot(s) R1, R2 and to integrate the faults detected in the program.

An embodiment of the calibration method described below relates more particularly to the calibration of a robotic cell 1 as described above.

In a preliminary step, the robotic cell 1 is installed in the real world. This involves arranging the robot or robots R1, R2 with respect to the mobile platform 3, based on the theoretical virtual model of the robotic cell 1. After installation in the real world, in a non-exhaustive manner, it is the robots R1, R2 may have relative positions with respect to the mobile platform 3 different from the theory, different orientations, the axis of rotation C of the rotary plate 11 may be tilted, may not be perfectly vertical to the axis translation V, or other orientation defect, misalignment, alignment.

Absolute calibration of machines or robots R1, R2
Before starting the calibration, the machine(s), in particular the robot(s) R1, R2 must be calibrated, so as not to introduce defects in the event that the machines, in particular the robots R1, R2 operate a theoretical virtual kinematic chain does not correspond to reality.

This is an absolute calibration to identify any defects intrinsic to robots R1, R2 compared to the theoretical virtual model. For example, purely illustrative, according to theory, robot arm 5 is supposed to extend to 1500mm but in the real world extends to 1503mm.

The absolute calibration of robots is generally carried out at a predefined temperature, for example in a non-limiting manner at 20°C +/-1°C or 2°C.

The calibration of the robots R1, R2 can be done independently, for example by the robot manufacturer.

The calibration process may optionally include one or more preliminary phases for calibrating the robots R1, R2.

According to one embodiment, the calibration of the robot R1, R2, comprises at least one step of measuring the dimensions of the robot R1, R2, so as to obtain the real kinematic chain of the robot R1, R2. This can be done for each robot R1, R2.

The step of measuring the dimensions of the robots R1, R2 can be implemented by a measuring instrument, for example such as a laser, of the laser tracker type also called tracking laser.

In a next step, the real kinematic chain is compared to the theoretical virtual kinematic chain of the robot R1, R2 in order to identify the deviations. These identified deviations are then integrated into the program by the control unit.

Another preliminary phase can be the identification of the position of the center of the tool, or TCP for tool center point in English, carried by the robot arm 5. The position of the center of the tool is in particular defined with respect to an output flange of the carrier axis of the robot R1, R2, generally the last axis of the robot R1, R2. Identifying the position of the tool center, or TCP, is a process known to those skilled in the art of robotics, and is not described in more detail herein.

Once the position of the center of the tool has been identified, the real coordinates can be integrated at the program level, so as to integrate any residual defects of positioning, of orientation with respect to the end of the robot arm 5.

Also, in a preliminary manner, the standards REF1, REF2, REF3-1, REF3-2 positioned on the mobile platform 3 are preferably calibrated.

• une phase de caractérisation du positionnement relatif de la plateforme 3 mobile par rapport à au moins un robot R1, R2 dans le monde réel,
• une phase d’identification des écarts de positionnement relatif dans le monde réel par rapport au modèle virtuel théorique, et
• une phase de correction des défauts.

In general, the calibration process includes in particular the following phases:

• a characterization phase of the relative positioning of the mobile platform 3 with respect to at least one robot R1, R2 in the real world,
• a phase of identifying deviations in relative positioning in the real world compared to the theoretical virtual model, and
• a fault correction phase.

Characterization phase of the relative positioning of the mobile platform 3 with respect to the robot R1, R2
The characterization phase of the relative positioning of the mobile platform 3 with respect to the robot R1, R2 in the real world, aims to identify and express the relative positioning between subassemblies of the robotic cell 1.

In particular, the positioning of one or more devices of the mobile platform 3 is identified by using at least one standard REF1, REF2 or reference element. The measurement unit 7 carried by the robot arm 5 records the coordinates of at least one standard REF1, REF2 positioned on the mobile platform 3 in at least one predefined position.

To this end, the characterization phase of the relative positioning of the mobile platform 3 with respect to the robot R1, R2 may include at least one training stage of the mobile platform 3. This training can be controlled by a control unit of the robotic cell 1.

In particular, the mobile platform 3 or at least one of its devices can assume at least two positions. In each of these positions, the measuring unit 7 carried by the robot arm 5 can record the coordinates of the standard REF1, REF2 positioned on the mobile platform 3. The coordinates recorded can then be transmitted to the control unit.

The control unit can then express the coordinates recorded in the frame O1, X1, Y1, Z1, respectively O2, X2, Y2, Z2, of the robot R1, respectively R2. This is at least a characterization step.

In particular, during this characterization step, the control unit can express the coordinates of a direction vector , of at least one of the auxiliary axes, in the frame O1, X1, Y1, Z1, respectively O2, X2, Y2, Z2 of the robot R1, respectively R2.

More precisely, the characterization phase of the relative positioning of the mobile platform 3 makes it possible to define the positioning of the so-called auxiliary axis or axes, in particular the axis of translation V and/or the axis of rotation C, with respect to a robot or to each robot R1, R2.

Characterization of the relative positioning of the translation axis V with respect to a robot
According to the particular embodiment previously described with the interface 9 moving linearly along the axis of translation V, it is a question of characterizing this axis of translation V in the reference frame of the robot R1, R2.

To this end, the stroke of the interface 9 moving linearly along the translation axis V must be calibrated. This can be done in a preliminary phase.

By way of illustrative and non-limiting example, the race is limited between 0mm and 3000mm, that is to say from a starting point or first position of end of race V0 to a point of arrival or second position of limit switch V3000. In theory, the starting point or first end position V0 can be set 1500mm from one side of the robot R1, R2 and the end point or second end position V3000 is 1500mm from the robot. other side. If an offset is detected in the real world, for example if the starting point V0 is located at 1503mm on one side and the ending point V3000 at 1497mm on the other side, this defect is integrated at the program level.

The objective is to identify and position in space, the translation axis V with respect to the robot R1, R2, and more precisely its direction vector .

Referring also to FIG. 7, the characterization phase of the relative positioning of the mobile platform 3 with respect to the robot R1, R2 generally makes it possible to identify two positions P1 and P2 along the stroke along the axis of translation V, using the first standard REF1 on the interface 9, and to express the coordinates of the first standard REF1 in each position P1, P2, in the reference frame of the robot R1, R2.

To do this, during a first training step, the interface 9 can be driven in linear displacement until the first standard REF1 reaches a first linear position P1. During a second training step, the interface 9 can be driven in linear displacement until the first standard REF1 on the interface 9 reaches a second linear position P2.

Two measurement steps can be implemented when the first standard REF1 on the interface 9 is in each of these linear positions P1, P2. The measurement unit respectively records the coordinates X _P1 , Y _P1 , Z _P1 of the first standard REF1 in the first linear position P1 and the coordinates X _P2 , Y _P2 , Z _P2 in the second linear position P2.

In order to improve the precision of the definition of the translation axis V, the two positions P1 and P2 are chosen as far apart as possible according to the stroke of the interface 9. The positions P1, P2 are also dependent on the capacity of the robot arm 5 carrying the measuring unit 7 to reach the first standard REF1, to feel it for example.

The linear positions P1, P2 can be start or end of travel positions or close to such start or end of travel positions, if the span of the robot arm 5 allows it.

Thus, the first linear position P1 can be a start position V0 if the span of the robot arm 5 allows it. As a variant, the first linear position P1 can be close to the first end-of-travel position V0, within the limit of the capacity of the robot arm carrying the measurement unit to reach the standard REF1 on the interface 9.

Similarly, the second linear position P2 can be the limit position V3000 of the interface 9 if the span of the robot arm 5 allows it. As a variant, the second linear position P2 can be close to the end-of-travel position V3000, within the limit of the capacity of the robot arm 5 carrying the measurement unit to reach the standard REF1 on the interface 9.

The coordinates recorded can be transmitted to the control unit. The latter can, during a stage of characterization of the axis of translation V with respect to the robot R1, R2, express the coordinates X _P1 , Y _P1 , Z _P1 of the first standard REF1 in the first linear position P1 and the coordinates X _P2 , Y _P2 , Z _P2 of the first standard REF1 in the second linear position P2 in the fixed frame O1, X1, Y1, Z1, respectively O2, X2, Y2, Z2 of the robot R1, respectively R2.

In particular, the control unit can deduce therefrom a direction vector of the translation axis and express its coordinates Xv, Yv, Zv in the fixed frame O1, X1, Y1, Z1 respectively O2, X2, Y2, Z2 of the robot R1, respectively R2.

The coordinates Xv, Yv, Zv of the direction vector of the translation axis have the form of the following equations:

The translation axis V is thus characterized in the frame O1, X1, Y1, Z1 respectively O2, X2, Y2, Z2 of the robot R1, respectively R2.

When the robotic cell 1 comprises several robots R1, R2, all these steps to characterize the translation axis V can be compared to the reference frame of each robot. According to the particular embodiment described, these steps are implemented with respect to the first robot R1 so as to express the translation axis V in the reference frame of the latter, and with respect to the second robot R2 so as to express the translation axis V in the reference frame of the latter.

Setting phase of an original angular position C0
Furthermore, according to the particular embodiment described with a rotary plate 11, mobile in rotation around an axis of rotation C, in order to be able to characterize this axis of rotation C in the reference frame of the robot or of one of the robots R1, R2, the calibration method comprises a phase of setting an original angular position C0 from which the positioning of the rotary plate 11 and of its axis of rotation C in space can be determined.

The setting phase makes it possible to position the rotary plate 11 in a predefined position such as the original angular position C0 by exploiting the second standard REF2.

This calibration phase is carried out with respect to a single robot. Moreover, when the robotic cell 1 comprises several robots, for example two robots R1, R2, intended to work on the same part, for example intended to make each of the holes on this part, a first robot R1 is defined as a master robot . The phase of setting the original angular position C0 is carried out only with respect to this first robot R1, known as the master robot.

This original angular position C0 is for example defined as the position in which the second standard REF2 is opposite the robot R1, and coincides with a first axis X1, of the reference defined by the robot R1, and the coordinate of the second standard REF2 along a second axis Y1 of the reference defined by the robot R1, is at zero.

To do this, the wedging phase includes a step of driving the rotary plate 11 in rotation, until the measuring unit 7 carried by the robot R1 detects that the second standard REF2 positioned on the rotary plate 11 reaches a position in which it coincides with the X1 axis, of the robot R1, and that its coordinate along the Y1 axis of the robot R1, is at zero. An angular position of the rotary plate 11 is thus associated with the original angular position C0.

As in the example shown, the turntable 11 can be mounted on the interface 9 which can be moved in a linear fashion. In this case, the phase of setting the original angular position C0 is implemented after a predefined positioning of the mobile interface 9 in a linear fashion, along the translation axis V. This positioning uses the first standard REF1.

During a training step, the mobile interface 9 in a linear fashion is moved until the first standard REF1 reaches the predefined position along the translation axis V. This training is for example controlled by the control unit.

The predefined position is for example the mid-travel V1500 along the translation axis V. The position of the interface in the mid-travel V1500 corresponds for example, when there are two robots R1, R2, to a position in which these two robots R1, R2 face each other, which makes it easier to define the relative positions of the subassemblies of the robotic cell 1. Alternatively, any other predefined position can be envisaged, such as for example at the start of V0 or at the limit switch V3000.

Of course, according to an alternative in which the rotary plate 11 would not be mounted on such a linearly mobile interface 9, there would be no need to associate the original angular position C0 with a particular positioning (for example in the middle race V1500) of this interface 9.

When the robotic cell 1 comprises several robots R1, R2, the original angular position C0 is not calibrated again with respect to the other or to the other robots R2. At the following iterations of the calibration process, the following robot or robots R2 come to identify, for example by coming to feel, the second standard REF2 at the original angular position C0 adopted previously. He(they) use(s) the value calibrated with respect to the first robot R1, called the master robot.

Characterization of the relative positioning of the axis of rotation C with respect to a robot
Once the original angular position C0 has been set, the rotary plate 11 is driven from this original angular position C0 to at least two other angular positions, using the second standard REF2 so as to be able to record and express its coordinates in each angular position in the frame of the robot R1, R2.

Advantageously, the two other angular positions form an angle of the same measure with respect to the original angular position C0. The angle is for example +120°, -120°.

For this, the characterization phase includes a first step of rotating the rotary plate 11 until the second standard REF2 reaches a second angular position C+120. The first rotational drive step is performed in a first direction of rotation, for example +120°.

The characterization phase includes a second step of driving the rotary plate in rotation until the second standard REF2 reaches a third angular position C-120. The second rotational drive step is carried out according to a second direction of rotation opposite to the first direction of rotation, for example -120°. The rotational drive of the rotary plate 11 can be controlled by the control unit of the robotic cell 1.

Three measurement steps can be implemented. The measuring unit 7 records the coordinates of the second standard REF2 in the original angular position C0 or first angular position, in the second angular position C+120 and also in the third angular position C-120.

By palpating the second standard REF2, in three positions: the original angular position C0, the second position C+120 and the third position C-120, this makes it possible to define a plane of rotation.

The control unit can, from the plane of rotation, deduce the center of rotation O' of the rotary plate 11.

The plane of rotation defined and the center of rotation O' identified, the control unit can determine a direction vector of the axis of rotation normal to the plane of rotation and passing through the center of rotation O'.

There follows a characterization step in which the control unit expresses the coordinates of the direction vector of the axis of rotation in the fixed frame O1, X1, Y1, Z1, respectively O2, X2, Y2, Z2, of the robot R1, respectively R2.

When the robotic cell 1 comprises several robots R1, R2, the steps for characterizing the axis of rotation C, with the exception of the phase of setting the original angular position C0, can be implemented with respect to the reference frame of each robot. According to the particular embodiment described, these steps are implemented with respect to the first robot R1 so as to express the axis of rotation C in the reference frame of the latter, and with respect to the second robot R2 so as to express the axis of rotation C in the reference frame of the latter.

Characterization phase of the relative positioning of the work surface in relation to the turntable
Furthermore, according to the particular embodiment described with a work table or a work plane 13 on the rotary plate 11, the calibration method also comprises a phase of characterization of the relative positioning of the work plane 13 with respect to the plane rotation of the turntable 11.

This phase contributes to the definition of the positioning in space of the mobile platform 3. The objective is to identify any misalignment and/or angular offset of this work surface 13 relative to the turntable 11.

To do this, the calibration process comprises one or more steps making it possible to define and identify the work surface 13 and its positioning in relation to the turntable 11.

The rotary plate 11 is brought or remains in the original angular position C0.

The measurement unit 7 carried by the robot arm 5 can carry out at least three measurement steps, so as to record the coordinates of at least three points in the reference plane REF3-1 defined on the work plane 13. The three points can be arranged at 120° from each other so as to obtain the most representative reference plane possible. Advantageously, the reference plane is defined by the bottom of the reference bore, so as to avoid introducing a lack of flatness.

The coordinates recorded can be transmitted to the control unit. From these measurements, the control unit can express the coordinates of these three points in the reference O′, Xc, Yc, Zc of the rotary plate 11, so as to define the support or laying plane for the tool, the part to be worked by the robots R1, R2. This is a characterization step of the reference plane REF3-1 in the O', Xc, Yc, Zc frame of the rotary plate 11.

The calibration process also includes one or more steps to identify the center O'' of the work plane. For this, the rotary plate 11 is brought or remains in the original angular position C0. The measurement unit 7 carried by the robot arm 5, comes for example to feel the work plane 13, during at least one measurement step, to identify the reference bore REF3- in the reference plane REF3- 1. The coordinates of the reference bore REF3-2 recorded by the measuring unit 7 can be transmitted to the control unit.

The control unit can, during at least one analysis step, determine the center of the reference bore REF3-2. This center corresponds to the center O'' of the work plane 13. This makes it possible to identify the coordinates Xi, Yi, of the center O''.

The component Zi of the center O'' can be determined from measurements, from the feeling of the three points in the reference plane REF3-1 making it possible to identify a possible lack of parallelism of the work plane 13 with respect to the plane of rotation and an offset of the center O'' along the component Zc of the plane of rotation of the rotary plate 11 previously identified. The control unit can express the coordinates Xi, Yi, Zi of the center O'' of the work plane 13 in the reference O', Xc, Yc, Zc of the rotary plate 11, during at least one characterization step.

As before, according to the embodiment with the rotary plate 11 mounted on the interface 9 movable in a linear manner, the latter is driven into a predefined position along the translation axis V, such as the middle of the stroke, prior to the steps measurement and characterization to identify the translational and rotational shifts of the work surface 13 relative to the turntable 11.

Characterization phase of the tool reference and/or the part to be worked
Furthermore, the calibration method may comprise a phase of characterization of an isostatic reference frame of a tool bearing the part to be worked and/or of the part to be worked.

The tool or the workpiece is intended to be placed on the mobile platform 3, for example on the work surface 13 on the rotary plate 11 as described above, or alternatively on a work surface on the interface 9 linearly movable.

To do this, the calibration method comprises at least one step of measurement by the measuring unit 7, to feel for example a standard such as a standard ring on the tool or the part, so as to define a reference plane (which gives the support plane for the isostatic reference). The control unit can, during at least one characterization step, express the coordinates of this reference plane in the reference of the last identified stage of the architecture, for example in the reference of the rotary plate 11 or of the interface 9.

The calibration process can include at least one other step of measurement by the measuring unit 7, to feel the standard such as the standard ring on the tool or the part, so as to identify its center and define the pivot of the isostatic frame of reference. The control unit can, during at least one characterization step, express the coordinates of this center in the reference of the last identified stage of the architecture, for example in the reference of the rotary plate 11 or of the interface 9 .

Finally, the measuring unit 7 can come and feel, for example, another standard, so as to identify the stoppage in rotation of the isostatic reference frame. The control unit can, during at least one characterization step, express the coordinates of this other standard in the reference of the last identified stage of the architecture, for example in the reference of the rotary plate 11 or of the interface 9.

Gap identification phase
From the coordinates recorded during the various measurement steps, the control unit can analyze these coordinates to identify defects, shifts of the robotic cell 1 in the real world compared to the theoretical virtual model.

Phase of identification of deviations from the relative positioning of the mobile platform 3 with respect to the robot R1, R2
In particular, the control unit can compare the coordinates of the mobile platform 3 detected by the measurement unit and expressed in the reference O1, X1, Y1, Z1, respectively O2, X2, Y2, Z2, of the robot R1, respectively R2, with the coordinates established according to the theoretical virtual model.

It is more precisely a question of comparing the coordinates of the axis of translation V, in particular of its directing vector , for the mobile interface 9 in a linear fashion with the theoretical coordinates.

Similarly, the coordinates of the axis of rotation C, in particular of its direction vector , of the turntable 11 are compared with the theoretical coordinates. Identified deviations can be recorded.

When the robotic cell 1 comprises several robots R1, R2, the steps for identifying the deviations can be repeated as many times as there are robots R1, R2, intended to work on the same part intended to be placed on a common support such as the work surface 13. According to the particular embodiment described with two robots R1, R2, these steps are implemented with respect to the first robot R1 and to the second robot R2.

Phase of identification of deviations from the relative positioning of the work surface 13 in relation to the turntable 11
Furthermore, the control unit can also analyze and compare the coordinates of the center O'' of the work plane 13 with respect to the axis of rotation C and with respect to the theoretical virtual model, so as to identify any misalignment of the center O'' of the work plane 13 with respect to the axis of rotation C.

Also, the control unit can analyze and compare the coordinates of the three points defining the work plane 13 (more precisely the support plane for the part) with respect to the coordinates of the rotary plate 11 and with respect to the theoretical virtual model, so as to identify any rotational offset of the work surface 13 relative to the turntable 11.

Thus, the calibration process also makes it possible to identify the center O'' of the work plane 13, any translational shifts and rotary shifts, for example if the work plane is not normal to the axis of rotation C that it has an angular offset relative to the plane of rotation defined during a previous step. Identified deviations can be recorded.

The control unit can also compare the coordinates of the tool and/or of the part to be worked noted by the measuring unit with the coordinates established according to the theoretical virtual model.

Defect correction phase
Furthermore, the calibration process makes it possible to take into account and correct the faults identified.

For this purpose, the control unit can determine at least one correction instruction to be applied to the theoretical virtual kinematic chain of the robot R1, R2, as a function of the positioning deviations identified with respect to at least one of the so-called auxiliary axes V, C, between the real world and theory. This correction setpoint can be recorded, integrated into the program, so that during its execution it is indeed data corrected by this correction setpoint which is taken into account and not theoretical data according to the theoretical virtual model.

In general, the robot(s) R1, R2 will conform with respect to their positions with respect to the so-called auxiliary axes V, C which are common for all the robots R1, R2 when the robotic cell 1 comprises several of them. According to the embodiment described, the program expresses setpoint values of the so-called auxiliary axes V and C, and no corrective action is applied to these setpoint values. The same setpoint for auxiliary axes V, C applies to all robots. On the other hand, corrections will apply to position instructions for the axes of the or each robot R1, R2.

A correction instruction to be applied to the virtual kinematic chain of the first robot R1 can be determined according to the differences identified between the actual relative positioning of the first robot R1 with respect to the mobile platform 3, and the relative positioning according to the model theoretical virtual. The setpoint values of the axes of the first robot R1 can be calculated according to the identified deviations. This can be done through a mathematical law, in particular a kinematic transform which integrates the setpoint values of virtual auxiliary axes. The virtual auxiliary axes are the virtual images of the V and C auxiliary axes in the real world.

Similarly, an instruction to be applied to the virtual kinematic chain of the second robot R2 is determined according to the identified deviations of relative positioning of the second robot R2 with respect to the mobile platform 3, between the real world and the theoretical model. This can be done through a kinematic transform which integrates the setpoint values of the virtual auxiliary axes.

The kinematic transform is a mathematical law. The kinematic transform consists in applying a calculation model making it possible to position an end device (the tool in the example described) according to the configuration of the mechanism and vice versa. In particular, matrix products make it possible to express each robot joint transition and more generally of the overall architecture. In robotics, one of the most exploited calculation models is the Denavit-Hartenberg convention or notation. To integrate reality (delta versus theory), we then speak of modified Denavit-Hartenberg parameters. Of course, other conventions can be used.

Each robot R1, R2 thus conforms to its real position relative to the common auxiliary axes V, C. The corrections can therefore differ at the level of the robots R1, R2, because their respective positions relative to the common auxiliary axes V and C are different. .

EXAMPLE OF REALIZATION
In order to facilitate understanding, a purely illustrative particular example for the calibration of a robotic one 1 is described below using the simplified table in figure 8 and figure 1.

• pour le premier robot R1 : JTA1_1, JTA2_1, JTA3_1, JTA4_1, JTA5_1, JTA6_1, BROT_1
• pour le deuxième robot R2 : JTA1_2, JTA2_2, JTA3_2, JTA4_2, JTA5_2, JTA6_2, BROT_2.

The robots R1 and R2 are six-axis robots (A1 to A6) and each carry a spindle for example BRO machining. The theoretical setpoint values of the axes according to the theoretical virtual model are:

• for the first robot R1: JTA1_1, JTA2_1, JTA3_1, JTA4_1, JTA5_1, JTA6_1, BROT_1
• for the second robot R2: JTA1_2, JTA2_2, JTA3_2, JTA4_2, JTA5_2, JTA6_2, BROT_2.

A theoretical setpoint value for a predefined position of the interface 9 moving linearly along the translation axis V is: VT. This theoretical set point value VT corresponds to the actual set point value. The same theoretical setpoint value VT applies for the first robot R1 or the second robot R2.

A theoretical setpoint value for a predefined angular position of the rotary plate 11 around the axis of rotation C with respect to the original angular position C0 is: CT. Here again, the theoretical setpoint value CT corresponds to the actual setpoint value. The same theoretical setpoint value CT applies for the first robot R1 or the second robot R2.

The setpoint values of the virtual auxiliary axes VV_1, CV_1 are the virtual images of the setpoints of the real auxiliary axes V and C (which correspond to the theoretical values) and are intended to be integrated into the kinematic transform of the first robot R1. Similarly, the setpoint values of the virtual auxiliary axes VV_2, CV_2 are the virtual images of the setpoints of the real axes V and C, and are intended to be integrated into the kinematic transform of the second robot R2. The setpoint values of the virtual auxiliary axes VV_1, CV_1 for the first robot R1, respectively VV_2, CV_2 for the second robot R2, correspond to the theoretical setpoint values VT and CT and therefore real. In other words, the theoretical values VT and CT are input data, which apply for all robots R1, R2.

Robots R1, R2 are calibrated beforehand. The real reference values of the axes of the first robot R1 are: JA1_1, JA2_1, JA3_1, JA4_1, JA5_1, JA6_1, BRO_1. The real reference values of the axes of the second robot R2 are: JA1_2, JA2_2, JA3_2, JA4_2, JA5_2, JA6_2, BRO_2. One or more real setpoint values of the axes are different from the theoretical values.

As described above, the calibration method can be implemented first with respect to the first robot R1 (arbitrarily defined as the master robot for the implementation of the phase of setting the original angular position C0 ).

In particular, the characterization phase of the relative positioning of the mobile platform 3 with respect to the first robot R1 is implemented so as to identify and characterize using a series of measurements by the measuring unit 7 in s' pressing on the standards or reference elements REF1, REF2, REF3-1, REF3-2, the relative positions of the mobile platform 3 with respect to the first robot R1.

The first robot R1 identifies after measurement, for example by probing, its positioning with respect to the interface 9 which moves linearly along the translation axis V.

In addition, the first robot R1 makes it possible to set the original angular position C0 of the rotary plate 11 as previously described. In addition, the first robot R1 identifies its position relative to the rotary plate 11 around the axis of rotation C.

The characterization phase of the relative positioning of the work surface 13 with respect to the rotary plate 11 can then be implemented. As an alternative or in addition, the characterization phase of the tool reference and/or the part to be worked can be implemented.

The gap identification phase is implemented to identify the defects of relative positioning of the mobile platform 3 compared to the first robot R1, that is to say the gaps between theory and reality. It may be that in the real world, the first robot R1 is not positioned with respect to the interface 9 in a manner consistent with the theoretical model. Also, it may be that in the real world, the first robot R1 is not positioned with respect to the rotary plate 11 in a way that conforms to the theory. There may be a difference between one or more theoretical and real position setpoint values for the axes of the first robot R1.

The control unit deduces therefrom a correction setpoint to be applied to one or more of the setpoint values of the axes of the first robot R1 to take these deviations into account. During the correction phase, the corrective setpoints are applied to the setpoint values of the axes of the first robot R1, to integrate the positioning defects with respect to the common auxiliary axes V, C, in the real world with respect to the theory. These include joint correction instructions.

In other words, according to the table in FIG. 8, the setpoints for the axes of the first robot R1 can be corrected at program level according to the differences between the actual setpoint values for the axes JA1,.., JA6, BRO_1, and the values theoretical setpoints for axes JTA1_1, .., JTA6_1, BROT_1. When reading the program, these correction instructions are taken into account and not the values initially expressed according to the theoretical virtual model.

When executing the program, a common setpoint will be required for the auxiliary axis(s) V, C, which are common to all the robots R1, R2. For this, the calibration process includes a step of association or coupling through the virtual auxiliary axes for all the robots R1, R2. In other words, the correction instructions are a function of the common auxiliary axes V and C.

More precisely, the virtual auxiliary axes are coupled in the kinematic transform of each robot by integrating these deviations. The setpoint values of the auxiliary axes are integrated into the kinematic transform of the robots and will make it possible, in relation to the identified deviations, to change the setpoint values for each robot to be taken into account during the execution of the program. When the calibration process is iterated with respect to the first robot R1, the correction has the effect of modifying the instructions on the axes of the first robot R1. It is the first R1 robot that self-corrects against the real architecture of the mobile platform assembly.

Then, these phases can be carried out this time with respect to the second robot R2, and so on with the following robots, if the robotic cell 1 comprises more than two robots R1, R2 intended to work on the same part, with the exception of the phase of setting the original angular position C0 which is carried out only with respect to the first robot R1.

Similarly, the positioning deviations of the second robot R2 relative to the mobile platform 3, more precisely relative to the interface 9 and to the rotary plate 13. The control unit deduces for the second robot R2 according to the axes common auxiliaries V, C, and deviations from the relative positions that had been identified, the correction instructions to be applied to the axis instruction values of the second robot R2.

When executing the program for machining, for example, the correction instructions are taken into account by the robots R1, R2, and not the values according to the theoretical virtual model initially expressed in the program. The program remains unchanged, it is the instructions sent to the axes of robots R1, R2 by the control unit which are corrected values.

Finally, the calibration process can be implemented at any time.

In particular, the particular robotic cell 1 is installed at an initial instant. The calibration process is implemented at this initial instant in order to identify the faults and integrate them into the program, as described previously.

Preferably, at the initial instant, the calibration process is implemented at a predetermined reference temperature. This predetermined reference temperature generally corresponds to the temperature at which the robot R1, R2 has been calibrated. In a non-limiting way, the absolute calibration of the robots is for example carried out at 20°C +/-1°C or 2°C.

However, what was valid at the initial moment may no longer be valid at a later moment. This may be the consequence of an expansion phenomenon, for example due to a higher temperature, for example 40°C or 60°C, or intrinsic wear.

The calibration process can be repeated at any time, simply and quickly, in particular if it is estimated that the architecture of the cell, in particular robotic cell 1, has been able to evolve. This reiteration at any time is possible because the calibration process does not have recourse to other devices or materials external to the cell, in particular robotics 1. It can be easily implemented, restarted at any time.

The cell, in particular robotic 1, will therefore again identify itself and integrate the faults which will generally be different from the faults initially identified during the installation of the cell, in particular robotic 1. The correction instructions to be applied are updated according to new defects identified.

One or more of the steps of the described calibration process can be repeated. The order of certain steps of the calibration process may possibly be reversed.

The preceding description refers to the method for calibrating a robotic cell 1 comprising at least one robot R1, R2, in particular an industrial robot. Of course, this description can also apply to a machine tool, a conventional Cartesian machine, and in particular to a cell comprising several machines, one of which can be defined as a leader machine, master similarly to the first robot R1 defined as the master robot, to perform the phase of setting the original angular position C0.

Thus, the calibration method makes it possible to identify the defects inherent in the relative positioning of the various subassemblies of the cell, such as a robotic cell 1. The calibration method is carried out by directly using the machines or robots R1 , R2, without requiring any element external to the cell, in particular robotics 1.

Such a calibration process can be carried out over a calibration time of less than half an hour, for example of the order of 20 min to 25 min. A significant time saving is obtained compared to the solutions of the prior art using external measuring devices such as scanners or laser trackers, generally implemented over half a day or even a day.

The particular robotic cell 1 can self-identify and allow calibration so as to guarantee precision, for example of machining when the program of cell 1 is implemented. The identified offset components are integrated, and correction instructions are applied and taken into account by the machines/robots R1, R2 when the program is executed.

In addition, the calibration can be repeated in a simple way to identify the reality and integrate it at any time in the program, providing stability over time of the particular robotic cell 1.

Claims (10)

Method for calibrating a cell (1) comprising at least one machine (R1, R2) defining a fixed reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2) and at least one platform (3) movable with respect to at least one axis (V, C), such that the machine (R1, R2) comprises an arm (5) configured to carry at least one tool so as to perform at least one automated operation on a workpiece intended to be positioned on the mobile platform (3),
the calibration method comprising a preliminary step of arranging said at least one machine (R1, R2) with respect to the mobile platform (3) in the real world, from a theoretical virtual model of the cell (1 ) used to program movements of said at least one machine (R1, R2) and of the mobile platform (3), and operations to be carried out by said at least one machine (R1, R2) on the workpiece according to a program managed by a cell control unit (1),characterized in thatthe calibration process comprises the following phases:

• a characterization phase of the relative positioning of the mobile platform (3) with respect to said at least one machine (R1, R2) in the real world, according to which a measurement unit (7) carried by the arm (5) records the coordinates of at least one standard (REF1, REF2, REF3-1, REF3-2) positioned on the mobile platform (3) in at least one predefined position and the control unit expresses the coordinates recorded in the reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2) of said at least one machine (R1, R2),
• a phase of identifying deviations from the relative positioning of the mobile platform (3) with respect to said at least one machine (R1, R2) in the real world with respect to the theoretical virtual model, according to which the control unit compares the coordinates recorded by the measuring unit (7) and expressed in the coordinate system (O1, X1, Y1, Z1; O2, X2, Y2, Z2) of the said at least one machine (R1, R2) with the established coordinates according to the theoretical virtual model, and
• a fault correction phase, in which the control unit determines, based on the identified discrepancies, at least one correction instruction to be applied to a virtual kinematic chain of said at least one machine (R1, R2), and records the correction at cell program level (1).

Calibration method according to the preceding claim, in which the cell is a robotic cell (1) comprising at least one robot (R1, R2) defining the fixed reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2 ) and said at least one platform (3) movable with respect to at least one axis (V, C), said at least one robot (R1, R2) comprising a robot arm (5) configured to carry said at least one tool so as to perform at least one automated operation on the workpiece intended to be positioned on the mobile platform (3).
Calibration method according to one of the preceding claims, in which the characterization phase of the relative positioning of the mobile platform (3) with respect to said at least one machine (R1, R2) comprises:

• at least one step of driving the platform (3) mobile in the real world, controlled by the control unit of the cell (1), so that said at least one standard (REF1, REF2) positioned on the platform (3) mobile takes at least two distinct positions, said at least one standard (REF1, REF2) being calibrated, and
• at least two measurement steps performed by the measurement unit (7) carried by the arm (5), when said at least one standard (REF1, REF2) is in the two distinct positions, so as to record the coordinates of said at least least one standard (REF1, REF2) respectively in the two distinct positions, the coordinates recorded are transmitted to the control unit.

Calibration method according to the preceding claim, in which during the characterization phase of the relative positioning of the mobile platform (3) with respect to said at least one machine (R1, R2), the control unit expresses the coordinates of a direction vector ( , ) of said at least one axis (V, C) in the frame of said at least one machine (O1, X1, Y1, Z1; O2, X2, Y2, Z2). Calibration method according to one of Claims 3 or 4, in which:

• the mobile platform (3) comprises an interface (9) mobile in a linear manner along a translation axis (V), the travel of the interface (9) along the translation axis (V) being calibrated, and in which
• during a first drive step, the interface (9) is driven in linear displacement until a standard (REF1) on the interface (9) reaches a first linear position (P1),
• during a second drive step, the interface (9) is driven in linear displacement until the standard (REF1) on the interface (9) reaches a second linear position (P2),
• during the measurement steps, the measurement unit (7) respectively records the coordinates (X _P1 , Y _P1 , Z _P1 ) of said standard (REF1) in the first linear position (P1) and in the second linear position (P2), and
• during a step of characterizing the translation axis (V) with respect to said at least one machine (R1, R2), the control unit expresses the coordinates (X _P1 , Y _P1 , Z _P1 ) of said standard (REF1) in the first linear position (P1) and the coordinates (X _P2 , Y _P2 , Z _P2 ) in the second linear position (P2) in the fixed reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2 ) of said at least one machine (R1, R2).

Calibration method according to one of the preceding claims, in which the mobile platform (3) comprises a rotary plate (11), mobile in rotation about an axis of rotation (C), the said method comprising a phase of setting an original angular position (C0), during which the rotary plate (11) is driven in rotation, by the control unit of the cell (1),until the measuring unit (7) carried by said at least one machine (R1, R2) detects a standard (REF2) on the rotary plate (11) in a position such that:

• said standard (REF2) coincides with a first predefined axis (X1, X2) of the fixed reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2) of said at least one machine (R1, R2) and that
• the coordinate of said standard (REF2), along a second predefined axis (Y1, Y2) of the fixed reference (O1, X1, Y1, Z1; O2, X2, Y2, Z2) of said at least one machine (R1, R2), either at zero.

Calibration method according to the preceding claim, in which the characterization phase of the relative positioning of the mobile platform (3) with respect to said at least one machine (R1, R2) comprises the following steps so as to characterize the axis of rotation (C) relative to said at least one machine (R1, R2):

• a first step of driving the rotary plate (11) in rotation, in which the control unit drives the rotary plate from the original angular position (C0) until the standard (REF2) on the plate rotary (11) reaches a second angular position (C+120), and
• a second step of driving the rotary plate (11) in rotation, in which the control unit drives the rotary plate from the original angular position (C0) until the standard (REF2) on the plate rotary (11) reaches a third angular position (C-120),
• three measurement steps, in which the measurement unit (7) respectively records the coordinates of said standard (REF2) in the original angular position (C0), in the second angular position (C+120) and in the third position angular (C+120),
• at least one step of identifying a center of rotation (O') of the turntable (11), in which the control unit defines a plane of rotation, from the three angular positions (C0, C+120, C-120), and deduces the center of rotation (O'),
• a determining step, in which the control unit determines a direction vector of the axis of rotation ( ) normal to the defined plane of rotation and passing through the center of rotation (O'), and
• a characterization step, in which the control unit expresses the coordinates of the direction vector of the axis of rotation ( ) in the fixed frame (O1, X1, Y1, Z1; O2, X2, Y2, Z2) of said at least one machine (R1, R2).

Calibration method according to the preceding claim, in which the rotary plate (11) comprises a work surface (13) on which the workpiece is intended to be placed, the rotary plate (11) defining a mark (O', X_VS, Y_VS, Z_VS), said method comprising:

• a characterization phase of the relative positioning of the work plane (13) with respect to the plane of rotation of the rotary plate (11) comprising the following steps:
  1. at least three measurement steps, in which the measurement unit (7) carried by the arm (5) records the coordinates of at least three points in a reference plane (REF3-1) defined on the work plane ( 13), when the turntable (11) is in the original angular position (C0),
  2. at least one characterization step, in which the control unit expresses the coordinates of said points of the reference plane (REF3-1) in the reference (O', Xc, Yc, Zc) of the rotary plate (11),
  3. at least one measuring step, in which the measuring unit (7) carried by the arm (5), identifies a reference bore (REF3-2) in the reference plane (REF3-1), when the rotary table (11) is in the original angular position (C0),
  4. at least one analysis step, in which the control unit determines the center of the reference bore (REF3-2) identified, corresponding to the center (O'') of the work plane (13), and
  5. at least one characterization step, in which the control unit expresses the coordinates of the center (O'') of the work surface (13) in the reference (O', Xc, Yc, Zc) of the rotary plate (11),
• a phase of identification of the differences in the relative positioning of the work surface (13) with respect to the rotary plate (11) in the real world and according to the theoretical virtual model.

Calibration method according to one of the preceding claims, the cell (1) comprising at least two machines (R1, R2) respectively defining a mark (O1, X1, Y1, Z1; O2, X2, Y2, Z2), the machines (R1, R2) being configured to respectively carry at least one tool so as to perform automated tasks on the same part intended to be positioned on the mobile platform (3), the theoretical setpoint value of said at least one axis (V , C) of the mobile platform (3) being equal to the actual setpoint value of said at least one axis (V, C), in which:

• at least the phases of characterization of the relative positioning of the mobile platform (3) and of identification of the deviations are implemented at least in part with respect to each machine (R1, R2), and in which
• during the correction phase, the control unit:
  1. determines according to the identified differences in relative positioning of the mobile platform (3) with respect to a first machine (R1), at least one correction setpoint to be applied to a setpoint value of the virtual kinematic chain of the first machine (R1 ) as a function of the theoretical and actual setpoint value of said at least one axis (V, C) of the mobile platform (3), and
  2. determines according to the identified differences in relative positioning of the mobile platform (3) with respect to a second machine (R2), at least one correction setpoint to be applied to a setpoint value of the virtual kinematic chain of the second machine (R2 ) as a function of the theoretical and actual setpoint value of said at least one axis (V, C) of the mobile platform (3).

Cell (1) comprising at least one machine (R1, R2), a platform (3) movable relative to said at least one machine (R1, R2), and a control unit, said at least one machine (R1, R2 ) comprising at least one arm (5) configured to carry at least one tool so as to perform at least one automated operation on a part intended to be positioned on the mobile platform (3), characterized in that the cell (1) is configured to implement at least certain steps of a calibration method according to one of the preceding claims, the arm (5) carrying a measurement unit (7) configured to implement at least one measurement step and the mobile platform (3) presenting at least one standard (REF1, REF2, REF3-1, REF3-2).