DE102015209773B3 - A method for continuously synchronizing a pose of a manipulator and an input device - Google Patents

Abstract

The present invention comprises a manipulator system and method for continuously synchronizing the pose of a manipulator 100 and an input device 200, wherein the input device 200 is arranged to control the manipulator 100, and the method comprises the steps of: determining 12 a deviation of the pose of the input device and the manipulator; Detecting 13 a commanded change of the pose of the input device; Determining 15 at least one correction term; Determining 16 a new pose of the manipulator from a combination of the commanded change in the pose of the input device and the correction term; and moving 17 the manipulator into the previously determined new pose, so that the deviation is at least reduced.

Description

1. Field of the invention

The invention relates to a method and a manipulator system for the continuous synchronization of a pose of a manipulator and a pose of an input device, wherein the input device is arranged to control the manipulator. The method and the manipulator system can be used, for example, in telemanipulation.

2. Background

For many applications in robotics, especially in the field of telemanipulation, input devices are needed that can command translational or rotational changes to a pose of a manipulator. In telemanipulation, movements of an input point of the input device are typically detected and converted into corresponding translatory and / or rotational movements of a manipulator.

The input point of the input device forms the origin of an input coordinate system. When the input point of the input device is moved, the position of the input point and / or the orientation of the input coordinate system change. The position of the input point and the orientation of the input coordinate system determine the pose of the input device. The position is determined by information regarding all three translatory degrees of freedom and determines the location of a point in space. The orientation is determined by information regarding all three rotational degrees of freedom and determines the rotation of a coordinate system in space. Analogously, the manipulator is also assigned a reference point. This reference point can be, for example, the Tool Center Point (TCP). The reference point of the manipulator typically forms the origin of an output coordinate system. The position of the reference point and the orientation of the output coordinate system determine the pose of the manipulator.

To control the manipulator, the input device is connected to the manipulator, so that at least data can be exchanged between the input device and the manipulator. The data exchange can be wired or wireless. Input device and manipulator can be spatially separated from each other. Telemanipulation systems are particularly useful where direct intervention is impossible or dangerous for an operator or in areas where particularly accurate and safe movements are required. This is the case, for example, in nuclear technology, in space technology or in medical technology.

According to DIN EN ISO 8373, a manipulator is an automatically controlled, freely programmable multipurpose manipulator which can be programmed in three or more axes and can be arranged for use, for example, in automation technology, either at a fixed location or in a movable manner. A manipulator carries end effectors, such as grippers or tools, or workpieces. Such manipulators are used for example in the field of medical technology, such as in minimally invasive surgery. There, the manipulator can, for example, perform a laparoscopic instrument and precisely process very small structures, or enable minimally invasive imaging by holding and guiding an endoscope.

Depending on the application, it may be necessary at certain times or due to events such as a dead man's switch signal or leaving the working space of the manipulator to no longer translate user input directly into manipulator motion. If the input device or an arm of the manipulator is moved during this decoupled state, the pose of the input device and the corresponding pose of the manipulator are out of sync, d. H. the manipulator's pose commanded by the input device is different from the actual pose of the manipulator. Likewise, an external force acting on the manipulator in the coupled state can lead to an asynchronous pose of the manipulator.

If the pose of the input device and the corresponding pose of the manipulator do not match, ie if there is a deviation, an input on the input device (master) may not be converted as expected by the manipulator (slave). For example, a deviation in the orientation of the output coordinate system of the manipulator and the input coordinate system of the input device may cause a rotation that the operator would command about the x-axis actually commanded as a rotation about the y-axis. A corresponding rethinking of the operator would be necessary to the usual axes of rotation of the input device, which deviates due to the deviation To assign rotary axes of the manipulator and so command the desired change in the pose on the manipulator properly. However, this rethinking is extremely prone to error and therefore undesirable.

In known methods for synchronizing the pose of an input device and a manipulator, the pose of the input device must be corrected after detecting the deviation in a decoupled state, for example by manually moving the input device to the correct pose. According to this known method, controlling the manipulator by means of the input device during the synchronization is not possible, which delays work processes. Other methods require active input devices that self-adjust to the correct pose.

For example, the US 2014/0 236 565 A1 a robotic teaching apparatus and method using a virtual image of a robot. The US 9,052,710 B1 discloses a system for tracking human gestures used to control a manipulator. Furthermore are out DE 10 2005 047 489 A1 a method and a corresponding system for programming operations and movements of robots for automation with at least one movable actuator and with at least one motion control or a drive control for the actuator known. The DE 10 2011 014 299 A1 discloses methods and means for controlling an automation device, in particular a robot, wherein initially a virtual model of the real automation device is set up and the model has a virtual detection means. In addition are out EP 2 483 039 B1 a system and method for synchronously mimicking the movement of a hand or other end element by a robot.

It is therefore an object of the present invention, in the event of deviations in the pose of the manipulator and the pose of the input device, to continuously synchronize these deviations during the control of the manipulator. Thus, the disadvantages described above can be eliminated.

3. Detailed description of the invention

The above objects are achieved by a method according to the invention for the continuous synchronization of a pose of a manipulator and a pose of the input device according to claim 1, and by a corresponding manipulator system according to claim 14.

• a) Bestimmen der Abweichung in der Pose des Manipulators und der Eingabevorrichtung zu einem ersten Zeitpunkt t_k;
• b) Erfassen einer kommandierten Änderung der Pose der Eingabevorrichtung, und Konvertieren der kommandierten Änderung der Pose der Eingabevorrichtung in eine kommandierte Änderung der Pose des Manipulators;
• c) Bestimmen zumindest eines Korrekturterms, wenn zumindest eine Komponente der kommandierten Änderung der Pose der Eingabevorrichtung in der Richtung der Abweichung verläuft, wobei der Korrekturterm so bestimmt wird, dass der Korrekturterm in Richtung der genannten zumindest einen Komponente verläuft;
• d) Bestimmen einer neuen Pose des Manipulators, wobei sich die neue Pose des Manipulators aus der Kombination der kommandierten Änderung der Pose der Eingabevorrichtung und des zuvor bestimmten Korrekturterms ergibt, wobei sich die Abweichung zumindest verringert; und
• e) Bewegen des Manipulators in die zuvor bestimmte, neue Pose des Manipulators, sodass die neue Pose bis zu einem zweiten Zeitpunkt t_k+1 erreicht wird.

In particular, the above-mentioned objects are achieved by a method for continuously synchronizing the pose of a manipulator and an input device, wherein the input device is arranged to control the manipulator, and the method comprises the following method steps:

• a) determining the deviation in the pose of the manipulator and the input device at a first time t _k ;
• b) detecting a commanded change in the pose of the input device, and converting the commanded change in the pose of the input device into a commanded change in the pose of the manipulator;
• c) determining at least one correction term when at least one component of the commanded change in pose of the input device is in the direction of the deviation, the correction term being determined such that the correction term extends toward said at least one component;
• d) determining a new pose of the manipulator, wherein the new pose of the manipulator results from the combination of the commanded change of the pose of the input device and the previously determined correction term, wherein the deviation is at least reduced; and
• e) moving the manipulator into the previously determined, new pose of the manipulator so that the new pose is reached until a second time t _{k + 1} .

The determination of the deviation in method step a) can take place, for example, by a comparison of the absolute pose of the input device and the absolute pose of the manipulator. In this case, it is preferable to use a common reference system and to determine the amount and the direction of the deviation in the position of the input point and reference point and the orientation of the input coordinate system and the output coordinate system.

For example, the deviation is determined by the poses (orientation and position) of the input device (master) TM / M and the manipulator (slave) TS / S related to each other. For this purpose, the poses TM / M and TS / S transformed into a common reference system. The necessary transformations are denoted by _M T ^R and _S T ^R. In the trivial case these are equal to the unit matrix. The pose of the manipulator and the pose of the input device in the common reference system thus result in:

Wherein R _M , R _S denotes the orientation of the input coordinate system (index M) or the output coordinate system (index S) and p → _M , p → _S corresponding to the position of the input point or the reference point. Consequently, the deviation in the translation p → _r can be determined: p → _r = p → _S - p → _M

The deviation in the orientation R _{r is} determined, for example, by: R _r = R _s ^-1 .R _M

A common reference system is advantageous because it allows the determination of the deviation to be accelerated. Particularly preferably, the determination of the deviation is merely in retrieving and / or using already known deviation data. For carrying out the method, it is sufficient that the deviation is known to the manipulator system; an explicit calculation is not required in this case. In particular, the deviation in the pose may be a translational deviation between reference point and input point in at least one of the three translatory degrees of freedom, and / or a rotational deviation between output manipulator system of the manipulator and input coordinate system of the input device in at least one of the three rotational degrees of freedom.

The entry point is a point associated with the input device. If the input device is moved, for example, by an operator, the input point is also moved, with the position of the input point and / or the orientation of the input coordinate system changing. Consequently, the pose of the input device, which is determined by the position of the input point and the orientation of the input coordinate system, changes. The entry point is the origin of the input coordinate system. The thus commanded change in the pose of the input device is detected by the manipulator system, and converted into a corresponding change in the pose of the manipulator. To determine the pose of the manipulator, the manipulator is assigned a reference point, which forms the origin of an output coordinate system. The position of the reference point and the orientation of the output coordinate system determine the pose of the manipulator. Typically, the reference point is the Tool Center Point (TCP), the wrist-man point of the manipulator (HWP), or another point characteristic of manipulator motion. When the manipulator is moved in accordance with the commanded change in pose of the manipulator, the operator can control the manipulator by means of the input device. Input devices are known and may include, for example, 6D mice, haptic input devices, such as Force Dimension omega input devices, or sensitive manipulators, such as KUKA AG's LBR iiwa.

In order to correct the previously determined deviation and thus to synchronize the pose of the input device and the pose of the manipulator, it is advantageous to determine a correction term by which the existing deviation during the execution of movements according to the commanded change of the pose of the manipulator, respectively Commanded changes in the pose of the input device, can be reduced or even eliminated. Thus, the deviation can be synchronized while the manipulator is already being controlled by the input device. This is done almost imperceptibly for the operator, so that the synchronization takes place quasi by itself in the course of use.

Preferably, the correction term is determined when at least one component of the commanded change in the pose of the input device is in the direction (translational direction and direction of rotation) of the deviation. In this case, the commanded change in the pose of the input device can be composed of a rotational and a translatory change. The rotational change can be determined by means of three components, which can be determined unambiguously by the three rotational degrees of freedom. The translational change can also be determined by means of three components, which can be determined unambiguously by the three translatory degrees of freedom. Thus, the components of the commanded change correspond to the pose of the input device, for example in one Cartesian reference system, the translational axes x, y and z and the rotations about these same axes, which are described for example by the angle α, β, γ. If a different coordinate system or another representation is selected to describe the commanded change, the components can also be described in this coordinate system or in this representation.

The direction of the deviation may also be described according to the components, and accordingly includes both the direction of translation and the sense of rotation. A component runs in the direction of the deviation if and only if it runs in the positive or negative direction of the deviation, ie it is rectified or opposite. The deviation points in the positive direction from the input point to the reference point, or from an orientation of the input coordinate system to an orientation of the output coordinate system.

The determination of a correction term when at least one component of the commanded change in the pose of the input device is in the direction of the deviation is advantageous since only in this case a correction term according to the preferred method is needed. Since the control of the manipulator by means of the input device typically has a real-time requirement, the calculation of the correction term is time-critical. If the correction term is calculated only when it is needed, then calculation time can be saved. The manipulator system can thus operate efficiently and quickly.

Since the correction term runs in the direction of said at least one component of the change in the commanded pose of the input device, ie in its positive or negative direction, the correction term influences the new pose of the manipulator, which is determined in method step d), so that the deviation is at least reduced.

In an illustrative, one-dimensional example, the position of the reference point (corresponding to the pose of the manipulator) is shifted by 5 units in the positive direction relative to the position of the input point (corresponds to the pose of the input device). The deviation is therefore +5. If a change of the position of the input point is now commanded, which for example amounts to 3 units in the positive direction (+3), then a correction term is determined. In the example mentioned, the scaling of the manipulator system is 1: 1, so that when converting the commanded change of the position of the input point into a commanded change in the position of the reference point, a commanded change of the position of the reference point would be determined by +3 in the positive direction. Without a correction term, this commanded change in the position of the reference point would be implemented so that after moving the reference point of the manipulator, the deviation would still be +5. If a correction term is then determined which at least reduces the deviation, this could be, for example, 2 units in the negative direction (-2). In determining a new position of the reference point of the manipulator through the combination of the commanded change of the position of the input point and the previously determined correction term, for example, a new position would be the unit in the positive direction away from the original pose of the reference point. However, since the input point would be moved by 3 units in the positive direction, the reference point would only move by 1 unit in the positive direction, would result in a new, reduced deviation of +3 (+5 - 2 = +3).

As shown in the simplified example, the deviation can be reduced with the described method during the manipulation of the manipulator. This is advantageous, as it allows deviations to be compensated successively until the pose of the input device and the pose of the manipulator are synchronized again. This synchronization takes place without deliberately correcting the operator and can therefore take place in the background and virtually unnoticed. The control of the manipulator can be done without major adverse effects, even if the pose of the input device and the pose of the manipulator are initially out of sync.

Preferably, a commanded change in the pose of the input device is converted to a commanded change in pose of the manipulator by means of a scaling factor "f" and the scaling factor is preferably 0 <f ≤ 1. Scaling the conversion of the pose of the input device to a change in pose of the manipulator, in a first case allows to convert small movements of the input / input coordinate system of the input device into large movements of the reference point / output coordinate system of the manipulator. Conversely, in a second case, large movements of the input point / input coordinate system of the input device may be converted into small movements of the reference point / output coordinate system of the manipulator.

In the first case, for example, large manipulators can be controlled by the input device. This allows moving large objects or tools by means of the manipulator over long distances. According to the second case, the scaling factor f is preferably selected in the range 0 <f ≦ 1, so that, for example, movements on the input device are converted into small movements of the manipulator. Such an application is found, for example, in minimally invasive surgery. The movements of the input point of the input device can thus be converted into very small, precise movements of the manipulator.

Preferably, the correction term is determined such that when at least one component of the commanded change in the pose of the input device is in the positive direction of the deviation, the manipulator in the corresponding component is moved less by a translational and / or rotational amount than commanded Change the pose of the input device, which is converted into a commanded change in the pose of the manipulator commanded.

If the correction term is selected, the input device can catch up with the manipulator during the movement of the manipulator. This leads at least to a reduction of the deviation and preferably in the case of repeated commanded change of the pose of the input device in the positive direction of the deviation to a complete elimination of the deviation and thus to synchronization. In particular, the deviation may already be completely synchronized in the positive direction of the deviation during the first commanded change in the pose of the input device. Preferably, the manipulator is moved in the direction of the commanded change by a translatory and / or rotational amount less than commanded until the input device has overtaken the manipulator. Thus, while the full synchronization is delayed, the operator experiences immediate feedback, in the form of movement of the manipulator, to his input on the input device. This simplifies the control of the manipulator system. In particular, an override is avoided by the operator, as this experiences an immediate feedback.

The feedback is preferably experienced by the operator by direct visual contact with the manipulator. If the scaling chosen so small that a direct visual feedback is not possible, or a direct visual feedback, for example due to the spatial separation of input device and manipulator is not possible, a display device is preferably used to output the feedback to the operator. Such a display device may comprise a display on which at least a part of the manipulator, in particular the reference point and / or the output coordinate system of the manipulator is imaged. The image of the reference point and / or the output coordinate system of the manipulator may comprise a schematic representation or be based on real image data. Other feedback systems, such as haptic feedback, are also possible.

Preferably, the correction term is determined such that, if at least one component of the commanded change in the pose of the input device is in the negative direction of the deviation, the manipulator in the corresponding component is further moved by a translational and / or rotational amount than the commanded change the pose of the input device, which is converted into a commanded change in the pose of the manipulator, is commanded. A corresponding determination of the correction term makes it possible to reduce the deviation even if the change in the pose of the input device is in the negative direction of the deviation, since the manipulator can at least partially catch up with the input device. This is advantageous, since such a very fast synchronization can be achieved. However, it should be noted that the manipulator performs another movement than commanded by the operator so that potential hazards may arise. These may be, for example, a collision with surrounding structures or persons. In this case, preferably, the movement of the manipulator is additionally monitored in order to ensure a safe movement. The monitoring includes, for example, a collision monitoring, a workspace monitoring a singularity monitoring and / or other monitoring mechanisms.

The correction term preferably comprises a translatory and / or rotational correction term. The provision of a translatory and / or rotational correction term is advantageous since the correction term can be calculated specifically. If, for example, only one rotational correction term is required, the determination of the translatory correction term can be omitted, thus saving computing time. This also applies vice versa. According to method step d), the new pose of the manipulator is determined from the combination of the correction term and the commanded change.

Preferably, the input device is associated with an input point, wherein the input point is the origin of an input coordinate system, and wherein the position of the input point and the orientation of the input coordinate system determine the pose of the input device, and wherein the manipulator is assigned a reference point, the reference point being the origin of an output coordinate system and the position of the reference point and the orientation of the output coordinate system determine the pose of the manipulator. Thus, changing the pose of the input device or manipulator includes changing the position of the input point or datum point and changing the orientation of the input coordinate system or the output coordinate system.

zu:

Preferably, the new position of the reference point of the manipulator, according to the new pose of the manipulator, is the vector sum of the original position of the reference point at time t _k , according to the original pose of the manipulator, the translational correction term and the commanded change of the position of the reference point of the manipulator. This results in the new position of the reference point

to:

ursprüngliche Position des Bezugspunktes ist,

der Korrekturterm ist und

die kommandierte Änderung der Position des Bezugspunktes ist.Where the

original position of the reference point,

the correction term is and

the commanded change is the position of the reference point.

The consideration of the commanded change of the position of the reference point of the manipulator allows the inclusion of scaling factors in the synchronization. This is particularly advantageous because the scaling factor "f" may change during operation of the manipulator system. It is thus possible, for example, for the operator, in order to be able to reach a specific position quickly, to select a larger scale and then to reduce it for precise work.

Preferably, the respective components of the translational correction term are equal to the corresponding components of the commanded change in the position of the reference point of opposite sign when the corresponding component of the commanded change in the position of the reference point is in the positive direction of the corresponding component of the deviation, and preferably the respective components of the translational correction term equals zero when the respective component of the commanded change in the position of the reference point is in the negative direction of the corresponding component of the deviation.

beispielsweise in seinen einzelnen Komponenten i ∊ {x, y, z} bezogen auf die kommandierte Änderung der Position des Eingabepunktes

wie folgt bestimmt werden:

In a Cartesian coordinate system, the translatory correction term

For example, in its individual components i ε {x, y, z} with respect to the commanded change in the position of the input point

be determined as follows:

According to the exemplary formula, a commanded movement of the reference point of the manipulator is compensated by the correction term until the deviation is synchronized when the commanded movement of the input point of the input device is in the positive direction of the deviation. If the commanded movement of the input point runs in the negative direction of the deviation, then the correction term in this component is equal to zero and the deviation is not reduced. Thus, a simple, robust and efficient synchronization method for the translation can be created. Since the reference point of the manipulator is not moved in this case or according to the commanded movement, this determination of the correction term and the resulting movement of the manipulator are safe.

des Ausgabekoordinatensystems des Manipulators, gemäß der neuen Pose des Manipulators, das Matrixprodukt der ursprünglichen Orientierung

des Ausgabekoordinatensystems des Manipulators zum Zeitpunkt t_k, gemäß der ursprünglichen Pose des Manipulators, des rotatorischen Korrekturterms

und der kommandierte Änderung der Orientierung

des Ausgabekoordinatensystems des Manipulators. Somit ergibt sich die neue Orientierung des Ausgabekoordinatensystems

beispielsweise zu:

Preferably, the new orientation

the output coordinate system of the manipulator, according to the new pose of the manipulator, the matrix product of the original orientation

the output coordinate system of the manipulator at time t _k , according to the original pose of the manipulator, the rotational correction term

and the commanded change of orientation

the output coordinate system of the manipulator. This results in the new orientation of the output coordinate system

for example:

The consideration of the commanded change of the orientation of the output coordinate system allows the inclusion of scaling factors in the synchronization. This is particularly advantageous if the scaling factor "f" can change during operation of the manipulator system as described above.

um die Koordinatenachsen eines Bezugskoordinatensystems, gleich den jeweiligen Drehwinkeln der kommandierte Änderung

der Orientierung des Ausgabekoordinatensystems des Manipulators mit umgekehrtem Drehsinn, wenn die entsprechenden Drehwinkel der kommandierten Änderung

der Orientierung des Ausgabekoordinatensystems den gleichen Drehsinn des entsprechenden Drehwinkels der Abweichung R_r aufweisen, und vorzugsweise sind die jeweiligen Drehwinkel des rotatorischen Korrekturterms

gleich null, wenn die entsprechenden Drehwinkel der kommandierten Änderung

der Orientierung des Ausgabekoordinatensystems den entgegengesetzten Drehsinn des entsprechenden Drehwinkels der Abweichung R_r aufweisen.Preferably, the respective rotational angles of the rotational correction term

around the coordinate axes of a reference coordinate system, equal to the respective rotation angles of the commanded change

the orientation of the output coordinate system of the manipulator with reverse rotation when the corresponding rotation angle of the commanded change

the orientation of the output coordinate system have the same direction of rotation of the corresponding rotation angle of the deviation R _r , and preferably, the respective rotation angle of the rotational correction term

equal to zero, if the corresponding rotation angle of the commanded change

the orientation of the output coordinate system have the opposite direction of rotation of the corresponding rotation angle of the deviation R _r .

Analogously to the exemplary determination of the translatory correction term, in the determination of the rotational correction term, a commanded movement of the output coordinate system of the manipulator is compensated by the correction term until the deviation is eliminated when the commanded movement of the input coordinate system of the input point is in the direction of rotation of the deviation. If the commanded movement of the input coordinate system of the input point runs contrary to the direction of rotation of the deviation, the correction term in this component is equal to zero and the deviation is not reduced. Thus, a simple, robust and efficient synchronization method for the rotation can be created. Since the output coordinate system of the manipulator is not moved in this case or according to the commanded movement, these determinations of the correction term and the resulting movement of the manipulator are safe.

des rotatorischen Korrekturterms

derart skaliert, dass der Drehwinkel φ_Δ der Bewegung des Ausgabekoordinatensystems des Manipulators in die zuvor bestimmte, neue Orientierung

kleiner oder gleich dem Drehwinkel

der kommandierten Änderung

der Orientierung des Ausgabekoordinatensystems ist, d. h.

Somit kann sichergestellt werden, dass die tatsächlich ausgeführte Drehung bei der Bewegung des Ausgabekoordinatensystems des Manipulators in die neue Orientierung nicht größer ist, als die kommandierte Änderung der Orientierung, d. h. als der kommandierte Winkel des Ausgabekoordinatensystems. So kann beispielsweise vermieden werden, dass der Drehwinkel der Bewegung des Ausgabekoordinatensystems umschlägt, und so die Abweichung entgegen des ursprünglichen kommandierten Drehsinns synchronisiert. Folglich kann mit der Skalierung des Drehwinkels der Bewegung des Ausgabekoordinatensystems ein sicheres Synchronisationsverfahren bereitgestellt werden. Dies ist insbesondere in sicherheitskritischen Anwendungen, wie beispielsweise der minimal invasiven Chirurgie, vorteilhaft.Preferably, the resulting angle of rotation becomes

of the rotational correction term

scaled such that the rotation angle φ _{Δ of} the movement of the output coordinate system of the manipulator in the previously determined, new orientation

less than or equal to the angle of rotation

the commanded change

the orientation of the output coordinate system, ie

Thus, it can be ensured that the actual rotation performed during the movement of the output coordinate system of the manipulator into the new orientation is not greater than the commanded change of the orientation, ie as the commanded angle of the output coordinate system. For example, it can be avoided that the rotation angle of the movement of the output coordinate system turns over, and thus synchronized the deviation against the original commanded sense of rotation. Consequently, with the scaling of the rotation angle of the movement of the output coordinate system, a secure synchronization method can be provided. This is particularly advantageous in safety-critical applications, such as minimally invasive surgery.

gemäß der nachstehenden Formel bestimmt:

wobei die Matrizen

Rotationsmatrizen sind, die einer um den Faktor m skalierten Rotation um die Rotationsachen

entsprechen, wobei die Matrizen

einen Drehwinkel

aufweisen und wobei der Drehwinkel

sich zu

berechnet, wobei

der Drehwinkel der kommandierten Änderung

der Orientierung des Eingabekoordinatensystems der Eingabevorrichtung ist, und der Drehwinkel φ ~_r sich zu φ ~_r = m·φ ~_r, berechnet, wobei der Drehwinkel φ ~_r das Produkt des kleineren der beiden Winkel von

und dem Betrag des Skalarprodukts der Rotationsachen

ist:

und wobei φ_r der Drehwinkel der Abweichung R_r ist. Vorzugsweise berechnet sich der Faktor m dabei zu:

Preferably, the correction term

determined according to the formula below:

where the matrices

Rotation matrices are those of a rotation scaled by the factor m rotation

match, with the matrices

a rotation angle

and wherein the angle of rotation

to

calculated, where

the angle of rotation of the commanded change

the orientation of the input coordinate system of the input device is, and the rotation angle φ ~ _r to φ ~ _r = m · φ ~ _r , calculated, wherein the rotation angle φ ~ _r, the product of the smaller of the two angles of

and the magnitude of the dot product of the rotation matters

is:

and where φ _{r is} the rotation angle of the deviation R _r . Preferably, the factor m is calculated to be:

The described calculation ensures that the actual rotation performed when the output coordinate system moves to the new orientation is not greater than the commanded change in orientation. The calculation of the rotational correction term in the rotation matrix representation is advantageous because this calculation can be carried out quickly, whereby computing time and computing capacity can be saved. This is particularly advantageous for real-time requirements.

Preferably, the reference point is the tool center point (TCP) of the manipulator. The choice of the TCP as the reference point of the manipulator allows a simple conversion of the commanded changes of the input point, since the pose of the end effector of the manipulator is described via the TCP. The TCP is an imaginary reference point that is located at a suitable location on the end effector of the manipulator and typically defines the tool or tool coordinate system. To describe which pose the end effector should occupy, it is sufficient to define the position of the TCP and the orientation of a coordinate system assigned to the TCP in space. If the reference point and TCP are identical, further conversions of the pose to other coordinate systems are unnecessary in order to control the manipulator or the end effector by means of the input device.

In particular, the abovementioned objects are also achieved by a manipulator system comprising at least one manipulator and an input device connected to the manipulator, wherein the manipulator system is configured to carry out a method according to the claims 1 to 13. Preferably, the manipulator system comprises a display device, wherein the display device is adapted to display a change in the pose of the manipulator.

The display device allows a feedback of the pose of the manipulator to the operator. If the scaling chosen so small that a direct visual feedback is not possible, the display device can scale the feedback accordingly. Likewise, the display device allows visual feedback even when the input device and manipulator are spatially separated. Such a display device may comprise a display on which at least a part of the manipulator, in particular the reference point and / or the output coordinate system of the manipulator is imaged. The image of the reference point of the manipulator may comprise a schematic representation or be based on real image data. Other feedback systems, such as haptic feedback, are also conceivable.

4. Description of the figures

In the following, preferred embodiments of the invention will be described in more detail with reference to the figures. Showing:

1 a schematic representation of a manipulator system;

2 a flowchart of the method for synchronization;

3A C by way of example the determination of a new position of the reference point of the manipulator;

4A C by way of example a further determination of a new position of the reference point of the manipulator;

5A -C by way of example the determination of a new orientation of the output coordinate system of the manipulator; and

6A C by way of example a further determination of a new orientation of the output coordinate system of the manipulator.

1 shows a manipulator system 1 which is an input device 200 with an entry point 210 and an input coordinate system 210k and a manipulator 100 with a reference point 110 and an output coordinate system 110k includes. The input device 200 is with the manipulator 100 connected and set up the manipulator 100 to control. Preferably, the input device is 200 a sensitive manipulator, which changes the pose of the input device 200 can capture. By changing the pose of the input device 200 can the manipulator 100 For example, be controlled such that the reference point 110 on the trajectory 111 which is indicated by the operator via the input device 200 is given, is moved. A display device 300 gives to the operator Feedback regarding the pose of the manipulator 100 , For example, take a camera 310 real image data of the manipulator 100 and sends them to the display device 300 on which the image data is displayed.

der Pose der Eingabevorrichtung 200 erfasst, und in eine kommandierte Änderung

der Pose des Manipulators 100 konvertiert. In Verfahrensschritt 14 wird überprüft, ob zumindest eine Komponente der kommandierten Änderung

der Pose der Eingabevorrichtung 200 in der Richtung der Abweichung p →_r, R_r verläuft. Ist dies der Fall, so wird in einem Verfahrensschritt 15 zumindest ein Korrekturterms

bestimmt, wobei der Korrekturterm

so bestimmt wird, dass der Korrekturterm

in Richtung der genannten zumindest einen Komponente verläuft. In Verfahrensschritt 16 wird eine neuen Pose

des Manipulators 100 bestimmt, wobei die neue Pose

des Manipulators 100 sich aus der Kombination der kommandierten Änderung

der Pose der Eingabevorrichtung 200 und des zuvor bestimmten Korrekturterms

ergibt, wobei sich die Abweichung p →_r, R_r zumindest verringert. In Verfahrensschritt 17 wird der Manipulator 100 in die zuvor bestimmte, neue Pose

des Manipulators 100 bewegt, sodass die neue Pose

bis zu einem zweiten Zeitpunkt t_k+1 erreicht wird. 2 shows a flowchart of a method 10 for continuous synchronization. After the start of the process, first in the process step 11 Check if a deviation p → _r , R _r in the pose of the manipulator 100 and the input device 200 is present at a first time t _k . This deviation may be in step 12 in amount and direction (translatory direction and direction of rotation) are determined. In one process step 13 becomes a commanded change

the pose of the input device 200 captured, and in a commanded change

the pose of the manipulator 100 converted. In process step 14 it checks whether at least one component of the commanded change

the pose of the input device 200 in the direction of deviation p → _r , R _r . If this is the case, then in one process step 15 at least one correction term

determined, the correction term

is determined so that the correction term

extends in the direction of said at least one component. In process step 16 will be a new pose

of the manipulator 100 determined, taking the new pose

of the manipulator 100 arising from the combination of the commanded change

the pose of the input device 200 and the previously determined correction term

results, with the deviation p → _r , R _r at least reduced. In process step 17 becomes the manipulator 100 into the previously determined, new pose

of the manipulator 100 moves, leaving the new pose

until a second time t _{k + 1 is} reached.

The 3A C show by way of example the determination of a new position of the reference point 110 , as part of a pose of the manipulator 100 , It shows 3A the deviation 211 between input point 210 and the reference point 110 , The deviation 211 at time t _{k is} determined in the x direction to p _{r, x} = 2 and in the y direction to p _{r, y} = 1. The deviation 211 is shown as an arrow with closed tip. As in a first case in 3B initially, a change p _{M, Δ of} the input point is shown 210 , which is represented by an X, commands two (2) units in the x-direction. The reference point 110 is represented by a small circle, and is moved to a new position p _{s, k + 1} at time t _{k + 1} due to the resulting movement p _Δ from a position p _{s, k} at time t _k . The resulting motion p _Δ results from the commanded change of the position p _{S, Δ} , of the reference point 110 and the correction term p _{r, Δ} . As it turns out, the commanded change of the position of the input point p _{M, Δ} in the positive direction of the deviation p _{r, x} . Consequently, the correction term p _{r, Δ} is determined so that the reference point is moved by a translational amount less than commanded by the commanded change in the position of the reference point p _{S, Δ} = 2. In the case shown, the correction term is determined to be p _{r, Δ} = -1. The combination of the correction term with the commanded change in the position of the reference point p _{S, Δ} = 2 results in a resulting movement p _Δ = 1.

As in 3C is shown, in a second case, a change of the input point p ' _{M, Δ} , which is represented by an X, a unit in the negative x-direction (-1) commanded. The reference point is represented by a small circle, and is moved to a new position p ' _{s, k + 1} at time t' _{k + 1} due to the resulting movement p ' _Δ from a position p' _{s, k} at time t _k . The resulting motion p ' _Δ results from the commanded change of the position p' _{S, Δ} , of the reference point 110 and the correction term p ' _{r, Δ} . As it turns out, the commanded change of the position of the input point p ' _{M, Δ} in the negative direction of the deviation p' _{r, x} . Consequently, the correction term p ' _{r, Δ} is determined such that the reference point is moved further by a translatory amount than commanded by the commanded change in the position of the reference point p' _{S, Δ} = -1. In the case shown, the correction term is determined to be p _{r, Δ} = -1. The combination of the correction term with the commanded change of the position of the reference point p _{S, Δ} = -1 results in a resulting motion p _Δ = -2.

The 4A C show by way of example the determination of a further new position of the reference point 110 according to the exemplary formula described above:

It shows 4A the deviation between input point 210 and the reference point 110 , The deviation 211 at time t _{k is} determined in the x direction to p _{r, x} = 2 and in the y direction to p _{r, y} = 1. As in 4B In a first case, a change in the input point p _{M, Δ} , which is represented by an X, is commanded to two units in the positive x direction (+ 2). The reference point 110 becomes represented by a small circle. As it turns out, the commanded change points the position of the entry point 210 p _{M, Δ} in the positive direction of the deviation p _{r, x} . Consequently, the correction term p _{r, Δ} is determined so that the reference point 110 is moved only when the deviation in the x-axis is synchronized. Since the commanded change in position of the input point P _{M, Δ} = 2 just corresponding to the deviation of the input point is 110 not moved. The deviation p _{r, x} in the x direction is synchronized at time t _{k + 1} .

As in 4C 1, a change of the input point p ' _{M, Δ} , which is represented by an X, is first commanded into a unit in the negative x-direction (-1). The reference point 110 is represented by a small circle. As it turns out, the commanded change of the position of the input point p ' _{M, Δ} in the negative direction of the deviation p _{r, x} . Accordingly, the deviation can, according to this example, not be compensated, so that the correction term to p _{r, Δ} is determined = 0th The reference point 110 is moved according to the commanded change in the position of the reference point p _{S, Δ} = -1.

The 5A C show by way of example the determination of a new orientation of the output coordinate system 110k of the manipulator 100 , as part of a pose of the manipulator 100 , It shows 5A the deviation R _r between the input coordinate system 210k the input device 200 and the output coordinate system 110k , For clarity, only one axis of the input coordinate system will be used 210k and the output coordinate system 110k in the 5A -C shown. The deviation at time t _{k is} determined to be R _r = 60 °. As in 5B In a first exemplary case, a change in the orientation of the input coordinate system is first shown 210k R _{M, Δ} , which is represented by an axis marked with an X, is commanded by 45 ° clockwise. The output coordinate system 110k is represented by an axis marked with a small circle, and due to the resulting movement R _{Δ is moved} from an orientation R _{s, k} at time t _k to a new orientation R _{s, k + 1} at time t _{k + 1} in the clockwise direction. The resulting movement R _Δ results from the commanded change of the orientation R _{S, Δ of} the output coordinate system 110k and the correction term R _{r, Δ} . As it turns out, the commanded change of the input coordinate system indicates 210k R _{M, Δ} in the direction of rotation of the deviation R _r . Consequently, the correction term R _{r, Δ} is determined so that the output coordinate system 110k is less far moved by a rotational amount than commanded by the commanded change in the orientation of the output coordinate system R _{S, Δ} = 45 °. In the case shown, the correction term is determined, for example, as R _{r, Δ} = -30 °. The combination of the correction term with the commanded change in the orientation of the output coordinate system R _{S, Δ} = 60 ° gives a resulting movement R _Δ = 15 °. The deviation at time t _{k + 1} is therefore 15 °.

As in 3C In a second exemplary case, a change in the orientation of the input coordinate system R ' _{M, Δ} , which is represented by an axis designated by an X, is first commanded by 60 ° in the counterclockwise direction. The output coordinate system represented by an axis indicated by a small circle becomes a new orientation R ' _{s, k + 1} at time t' due to the resulting movement R ' _Δ from an orientation R' _{s, k} at time t _{k. k + 1} moves. The resulting movement R ' _Δ results from the commanded change of the orientation R' _{S, Δ} , of the output coordinate system 110k and the correction term R ' _{r, Δ} . As it turns out, the commanded change of the orientation of the input coordinate system R ' _{M, Δ points} in the negative direction of the deviation R' _r . Consequently, the correction term R ' _{r, Δ} is determined so that the output coordinate system 110k is further moved by a rotational amount than commanded by the commanded change in the orientation of the output coordinate system R ' _{S, Δ} = -60 °. In the case shown, the correction term is determined, for example, as R _{r, Δ} = -45 °. The combination of the correction term with the commanded change in the orientation of the output coordinate system R _{S, Δ} = -60 ° gives a resulting motion R _Δ = -105 °. The deviation at time t ' _{k + 1} is therefore 15 °.

The 6A C show by way of example the determination of a further new orientation of the output coordinate system 110k , analogous to the method, which in connection with the 4A -C for the position has been described. For clarity, only one axis of the input coordinate system will be used 210k and the output coordinate system 110k in the 6A -C shown. It shows 6A the deviation between input coordinate system 210k and the output coordinate system 110k , The deviation at time t _{k is} determined to be R _r = 60 °. As in 6B In a first exemplary case, a change in the orientation of the input coordinate system R _{M, Δ} , which is represented by an axis marked by an X, is first commanded by 45 ° in the clockwise direction. The output coordinate system 100k is represented by an axis marked with a small circle. As it turns out, the commanded change of the orientation of the input coordinate system R _{M, Δ} in the positive direction of the deviation R _r . Consequently, the correction term R _{r, Δ} is determined so that the output coordinate system 110k is only moved when the deviation is synchronized. Since the commanded change of the orientation of the input coordinate system R _{M, Δ} = 45 ° is smaller than the deviation, the output coordinate system becomes 110k not moved. The deviation R _r is partially synchronized at time t _{k + 1} . There remains a deviation of 15 °.

As in 6C 1, in a second exemplary case, a change of the input coordinate system R ' _{M, Δ} , represented by an axis marked X, is first commanded by 45 ° in the counterclockwise direction. The output coordinate system 110k is represented by an axis marked with a small circle. As it turns out, the commanded change of the orientation of the input coordinate system R ' _{M, Δ} counter to the direction of rotation of the deviation R _r . Consequently, according to this embodiment, the deviation can not be compensated, so that the correction term is determined to be R _{r, Δ} = 0. Thus, the output coordinate system becomes 110k according to the commanded change in the orientation of the output coordinate system R _{S, Δ} = 45 ° counterclockwise moves.

LIST OF REFERENCE NUMBERS

1

manipulator system

10

method

11

Check for deviation

12

Determine the deviation

13

Capture a commanded change

14

check the direction

15

Determining a correction term

16

Determine a new pose of the reference point

17

Moving the reference point

100

manipulator

110

Reference point of the manipulator

110k

Output coordinate system of the reference point

111

Trajectory of the manipulator

200

input device

210

Input point of the input device

210k

Input coordinate system of the input point

211

Arrow, deviation

300

display device

310

camera

p _r , R _r

Deviation (position, orientation)

p _MΔ , R _MΔ

Commanded change of the pose of the entry point

p _SΔ , R _SΔ

Commanded change of the pose of the reference point

p _rΔ , R _rΔ

Correction term (position, orientation)

p _{s, k} , R _{s, k}

Original pose of the reference point

p _{s, k + 1} , R _{s, k + 1}

New pose of the reference point

t _k

First time

_{tk + 1} ; _{t'k + 1}

Second time

f

scaling factor

φ _rΔ

Rotation angle of the correction term

φ _Δ

Rotation angle of movement

_φsΔ

Angle of rotation of the commanded change of orientation

e _r , e _MΔ

rotational axes

m

factor

Claims (15)

Procedure ( 10 ) for continuously synchronizing the pose of a manipulator ( 100 ) and the pose of an input device ( 200 ), wherein the input device ( 200 ) for controlling the manipulator ( 100 ), and the method comprises the following method steps: a) determining ( 12 ) of the deviation (p → _r , R _r ) in the pose of the manipulator ( 100 ) and the pose of the input device ( 200 ) at a first time t _k ; b) Capture ( 13 ) a commanded change

the pose of the input device ( 200 ), and convert the commanded change

the pose of the input device ( 210 ) into a commanded change

the pose of the manipulator ( 100 ); c) determining ( 15 ) at least one correction term

if at least one component of the commanded change

the pose of the input device ( 100 ) in the direction of the deviation (p → _r , R _r ), the correction term

is determined so that the correction term

extends in the direction of said at least one component; d) determining ( 16 ) a new pose

of the manipulator ( 100 ), being the new pose

of the manipulator ( 100 ) from the combination of the commanded change

the pose of the input device ( 200 ) and the previously determined correction term

yields, wherein the deviation (p → _r , R _r ) at least decreases; and e) moving ( 17 ) of the manipulator ( 100 ) in the previously determined, new pose

of the manipulator ( 100 ), so the new pose

until a second time t _{k + 1 is} reached. The method of claim 1, wherein a commanded change

the pose of the input device ( 200 ) into a commanded change

the pose of the manipulator ( 100 ) is converted by means of a scaling factor (f) and the scaling factor is preferably 0 <f ≦ 1. Method according to one of the preceding claims, wherein the correction term

is determined so that, if at least one component of the commanded change

the pose of the input device ( 200 ) in the positive direction of the deviation (p → _r , R _r ), the manipulator ( 100 ) is moved less by a translatory and / or rotational amount in the corresponding component than by the commanded change

the pose of the input device ( 200 ), which in a commanded change

the pose of the manipulator ( 100 ) is commanded. Method according to one of the preceding claims, wherein the correction term

is determined so that, if at least one component of the commanded change

the pose of the input device ( 200 ) in the negative direction of the deviation (p → _r , R _r ), the manipulator ( 100 ) is moved further in the corresponding component by a translatory and / or rotational amount than by the commanded change

the pose of the input device ( 200 ), which in a commanded change

the pose of the manipulator ( 100 ) is commanded. Method according to one of the preceding claims, wherein the correction term is a translatory

and / or rotational

Correction term includes. Method according to one of the preceding claims, wherein the input device ( 200 ) an input point ( 210 ) and the input point ( 210 ) the origin of an input coordinate system ( 210k ), and wherein the position of the input point ( 210 ) and the orientation of the input coordinate system ( 210k ) the pose of the input device ( 200 ), and wherein the manipulator ( 100 ) a reference point ( 110 ), the reference point ( 110 ) the origin of an output coordinate system ( 110k ), and wherein the position of the reference point ( 110 ) and the orientation of the output coordinate system ( 110k ) the pose of the manipulator ( 100 ). Method according to claim 6, wherein the new position

the reference point ( 110 ) of the manipulator ( 100 ), according to the new pose of the manipulator ( 100 ), the vector sum of the original position

the reference point ( 100 ) at the time t _k , according to the original pose of the manipulator ( 100 ), the translatory correction term

and the commanded change of position

the reference point ( 110 ) of the manipulator ( 100 ). Method according to one of claims 6 or 7, wherein the respective components

of the translatory correction term

equal to the corresponding components of the commanded change

the position of the reference point ( 110 ) of the manipulator ( 100 ) with the opposite sign, if the corresponding component of the commanded change (

the position of the reference point ( 110 ) is in the positive direction of the corresponding component of the deviation (p → _r ), and preferably equal to zero, if the respective component of the commanded change

the position of the reference point ( 110 ) in the negative direction of the corresponding component of the deviation (p → _r ). Method according to one of claims 6 to 8, wherein the new orientation (

the output coordinate system ( 210k ) of the manipulator ( 100 ), according to the new pose of the manipulator ( 100 ), the matrix product of the original orientation

the output coordinate system ( 210k ) of the manipulator ( 100 ) at the time t _k , according to the original pose of the manipulator ( 100 ), the rotational correction term

and the commanded change of orientation

the output coordinate system ( 110k ) of the manipulator ( 100 ). Method according to one of claims 6 to 9, wherein the respective rotational angles of the rotational correction term

around the coordinate axes of a reference coordinate system, equal to the respective rotation angles of the commanded change

the orientation of the output coordinate system ( 110k ) of the manipulator ( 100 ) with reversed rotation are when the corresponding rotation angle of the commanded change

the orientation of the output coordinate system ( 110k ) have the same direction of rotation of the corresponding angle of rotation of the deviation (R _r ), and are preferably equal to zero, if the corresponding rotation angle of the commanded change

the orientation of the output coordinate system ( 110k ) Have the opposite sense of rotation of the corresponding rotating angle of deviation (R _r). Method according to one of the preceding claims, wherein the resulting angle of rotation

of the rotational correction term

is scaled such that the rotation angle (φ _Δ ) of the movement of the output coordinate system ( 110k ) of the manipulator ( 100 ) in the previously determined, new orientation

the output coordinate system ( 110k ) of the manipulator ( 100 ) less than or equal to the angle of rotation

the commanded change

the orientation of the output coordinate system ( 110k )

Method according to one of the preceding claims, wherein the rotational correction term

determined according to the formula below:

where the matrices

Rotation matrices are those of a rotation scaled by the factor (m)

to the rotation things

match, and where the matrices

the rotation angle

and wherein the angle of rotation

to

calculated, where

the angle of rotation of the commanded change

the orientation of the input coordinate system ( 210k ) of the input device ( 200 ), and the angle of rotation (φ ~ _r ) is calculated to be φ ~ _r = m · φ ~ _r , where the angle of rotation (φ ~ _r ) is the product of the smaller of the two angles of

and the magnitude of the dot product of the rotation matters

is:

and where φ _{r is} the rotation angle of the deviation (R _r ). Method according to one of the preceding claims, wherein the reference point ( 110 ) the tool center point (TCP) of the manipulator ( 100 ). Manipulator system ( 1 ) comprising at least one manipulator ( 100 ) and one with the manipulator ( 100 ) connected input device ( 200 ), wherein the manipulator system is adapted to carry out a method according to claims 1 to 13. Manipulator system ( 1 ) according to claim 14, wherein the manipulator system ( 100 ) a display device ( 300 ), which display device ( 300 ) is adapted to change the pose of the manipulator ( 100 ).

Images