DE102023210548B3 - controlling a robot - Google Patents

Abstract

The invention relates to a method for controlling a robot, comprising the steps: providing (S10) a first data processing based at least partially on machine learning for determining a target movement of a first robot-fixed reference; providing (S10) a second data processing based at least partially on machine learning for determining a target movement of a second robot-fixed reference; determining (S20) a target axis movement of the robot based on a target movement of the first reference determined by means of the first data processing and a target movement of the second reference determined by means of the second data processing, wherein in this determination of the target axis movement the determined target movement of the first reference is prioritized over the determined target movement of the second reference; and controlling (S30) the robot on the basis of the determined target axis movement. The invention also relates to a system or computer program (product).

Description

The present invention relates to a method and a system for controlling a robot as well as a computer program or computer program product for carrying out the method.

The DE 10 2022 203 406 A1 relates to a method for creating a program for carrying out a robot application that includes multiple actions of a first robot, the method comprising the steps of: capturing programming inputs for programming these actions of the first robot; and creating the program on the basis of these captured programming inputs; wherein, in order to create the program for different robots, in particular by means of a context, input names are uniformly specified and the programming inputs are interpreted on the basis of the input name used by the respective programming input and the input name used for this content determined by the respective programming input, in particular by means of JSON-LD; and/or at least one of the programming inputs comprises at least one indication of a coordinate space, one indication of a control method of the first robot, one indication of a path of the first robot, one indication of a pose of the first robot, one indication of a tool action of the first robot, and/or one indication of a sensor type and/or action of the first robot.

The DE 10 2018 208 088 A1 relates to a method for controlling a robot, wherein a follow-up movement is determined on the basis of a guide force externally applied to the robot, wherein a deviation between a direction of a desired movement and a direction of the determined follow-up movement is increased when approaching a predetermined axis position, axis speed and/or axis acceleration limit.

The DE 10 2019 220 619 B3 relates to a method for moving a robot-fixed reference, in particular an end effector, of a robot, in particular a multi-axis robot arm, on the basis of a number, in particular a sequence, of predetermined discrete permissible poses of the reference, the method comprising the step, in particular repeated several times: exerting drive forces on the robot by its drives for moving the reference to follow a manual guide force exerted on the robot on the basis of a difference between a current pose of the reference and a first pose of the permissible poses closest to it, such that the drive forces, in the event of a deviation between the current pose and the first pose, cause a return force on the reference that is dependent on the difference, the return force corresponding to a return force dependent on the difference from the current pose to the first pose, the component of which in the direction of a connection between the first pose and a second pose of the permissible poses adjacent to it is masked out, so that the component of a total force caused by the drive forces on the reference in the direction of the connection is independent of the difference; and/or at least a portion of the drive forces is dependent on speeds of the drives and/or a speed of the reference, in particular linearly.

The DE 10 2018 116 053 A1 relates to a robot system comprising a robot including at least one sensor that detects an applied external force and a controller that controls the robot, wherein the controller changes an operation mode of the robot when performing lead-through learning according to a position and a pattern of the external force detected by the sensor.

The DE 10 2019 205 651 B3 relates to a method for executing an application using a robot, in particular an application with planned environmental contact of the robot with its environment, wherein an agent trained using reinforcement learning specifies a target value to a controller of the robot and the controller commands drives of the robot based on this target value and a kinematic, in particular dynamic, model of the robot. Additionally or alternatively, several similar controllers are trained using reinforcement learning on the basis of the same learning application with different boundary conditions, an artificial neural network is trained using the trained controllers using supervised learning, and the robot is controlled using the trained neural network in order to execute the specified application. Additionally or alternatively, an agent is trained using reinforcement learning on the basis of the same learning application with different boundary conditions and the robot is controlled using the trained neural network in order to execute the specified application.

The DE 10 2016 212 958 A1 relates to a method for controlling a redundant manipulator, wherein the manipulator has several axes of movement and the method comprises the following steps: determining a primary task for controlling the manipulator, wherein the primary task comprises traversing a movement path of the manipulator; determining at least one secondary task task for controlling the manipulator, wherein the secondary task comprises a reduction of the effective mass and/or the effective inertia; calculating n alternatives of possible axis angles of the axes of motion of the manipulator to fulfill the primary task, and calculating the axis angular velocities for the n alternatives; and controlling the manipulator taking into account the n alternatives and the secondary task.

The object of the present invention is to improve the control of a robot.

This object is achieved by a method having the features of claim 1. Claims 8, 9 protect a system or computer program or computer program product for carrying out a method described here. The subclaims relate to advantageous further developments.

According to one embodiment of the present invention, a robot has at least one robot arm and/or two or more robot legs and/or a movable base; it can in particular consist of these and/or be a mobile and/or humanoid robot. In one embodiment, the robot, in a further development its robot arm and/or its movable base, in particular each or together, has at least three, preferably at least six, in a further development at least seven, axes of movement or degrees of freedom of movement or joints, which are referred to here as axes of the robot. In one embodiment, one or more of these axes are rotational axes of movement or degrees of freedom or rotary joints and/or one or more of these axes are translational axes of movement or degrees of freedom or linear joints. The movable base can in particular be guided by a guide, in particular one or more rails, and thus be movable along one or more axes, which then represent corresponding (translational movement) axes of the robot or its base. Likewise, the base can be freely or unboundly movable, in which case horizontal movements and a rotation about a vertical yaw axis represent corresponding (movement) axes of the robot or its base. In one embodiment, one or more of the axes of the robot are moved or adjusted by drives, preferably electric motor drives, of the robot. A movement of the robot in one or more of its axes or an adjustment of one or more of the axes of the robot or a movement of the robot through or as a result of an adjustment (in) one or more of its axes is referred to here as the axis movement of the robot and is usually referred to as dq/dt (q ∈ ℜ ^f with the number f of axes of the robot). Accordingly, an axis movement of the robot in the sense of the present invention is in particular a change in its axis or joint coordinates.

• - Bereitstellen einer ersten wenigstens teilweise auf maschinellem Lernen basierenden Datenverarbeitung zur Ermittlung einer Soll-Bewegung einer ersten roboterfesten Referenz, in einer Ausführung dx₁/dt (mit x₁ ∈ ℜⁿ¹; n1 ∈ {1, 2, ...,6});
• - Bereitstellen einer zweiten wenigstens teilweise auf maschinellem Lernen basierenden Datenverarbeitung zur Ermittlung einer Soll-Bewegung einer zweiten roboterfesten Referenz, in einer Ausführung dx₂/dt (mit x₂ ∈ ℜⁿ²; n2 ∈ {1, 2, ...,6});
• - Ermitteln einer Soll-Achs-Bewegung des Roboters basierend auf einer mittels der ersten Datenverarbeitung ermittelten Soll-Bewegung der ersten Referenz und einer mittels der zweiten Datenverarbeitung ermittelten Soll-Bewegung der zweiten Referenz, in einer Ausführung dq_s/dt = dq_s/dt(dx₁/dt, dx₂/dt); wobei bei diesem Ermitteln der Soll-Achs-Bewegung die ermittelte Soll-Bewegung der ersten Referenz gegenüber der ermittelten Soll-Bewegung der zweiten Referenz (höher) priorisiert wird; und
• - Steuern des Roboters auf Basis der (solcherart) ermittelten Soll-Achs-Bewegung.

According to an embodiment of the present invention, a method for controlling the robot comprises the steps:

• - Providing a first data processing based at least partially on machine learning for determining a target movement of a first robot-fixed reference, in an embodiment dx ₁ /dt (with x ₁ ∈ ℜ ⁿ¹ ; n1 ∈ {1, 2, ...,6});
• - Providing a second data processing based at least partially on machine learning for determining a target movement of a second robot-fixed reference, in an embodiment dx ₂ /dt (with x ₂ ∈ ℜ ⁿ² ; n2 ∈ {1, 2, ...,6});
• - Determining a target axis movement of the robot based on a target movement of the first reference determined by means of the first data processing and a target movement of the second reference determined by means of the second data processing, in one embodiment dq _s /dt = dq _s /dt(dx ₁ /dt, dx ₂ /dt); wherein in this determination of the target axis movement, the determined target movement of the first reference is given (higher) priority over the determined target movement of the second reference; and
• - Controlling the robot based on the (thus) determined target axis movement.

By determining target movements of robot-fixed references using data processing based at least partially on machine learning, desired target movements can be specified very easily, quickly, precisely and/or reliably, especially by less experienced users. By dividing into or specifying individual target movements of at least two robot-fixed references in combination with a (higher) prioritization of one of these target movements compared to the other target movements, redundancies can be advantageously used or resolved and/or machine learning can be used particularly advantageously, in particular the target axis movement can be determined particularly efficiently and/or reliably.

• - Bereitstellen (je) einer weiteren wenigstens teilweise auf maschinellem Lernen basierenden Datenverarbeitung zur Ermittlung einer Soll-Bewegung (je) einer weiteren roboterfesten Referenz, in einer Ausführung dx_i/dt (mit x_i ∈ ℜⁿⁱ; ni ∈ {1, 2, ...,6});

wobei die Soll-Achs-Bewegung basierend auch auf einer mittels dieser weiteren Datenverarbeitung(en) ermittelten Soll-Bewegung(en) dieser weiteren Referenz(en) ermittelt wird, in einer Ausführung dq_s/dt = dq_s/dt(dx₁/dt, dx₂/dt, dx₃/dt (, ...));
wobei bei diesem Ermitteln der Soll-Achs-Bewegung die ermittelten Soll-Bewegungen der ersten und zweiten Referenz gegenüber dieser ermittelten Soll-Bewegung dieser wenigstens einen weiteren Referenz (höher) priorisiert werden, in einer Ausführung bei ermittelten Soll-Bewegungen von zwei oder mehr weiteren Referenzen die Soll-Bewegungen der ersten und zweiten Referenz gegenüber diesen ermittelten Soll-Bewegungen dieser weiteren Referenzen (höher) priorisiert werden und zusätzlich für eine vorgegebene, in einer Weiterbildung feste oder variierende, Reihen- bzw. Rangfolge der weiteren Referenzen die ermittelte Soll-Bewegung der jeweils ranghöchsten bzw. in der Reihenfolge vordersten bzw. obersten weiteren Referenz gegenüber den ermittelten Soll-Bewegung der restlichen weiteren Referenzen (höher) priorisiert wird.In one embodiment, the method comprises the step(s) or two or more of:

• - Providing (each) a further data processing based at least partially on machine learning for determining a target movement (each) of a further robot-fixed reference, in an embodiment dx _i /dt (with x _i ∈ ℜ ⁿⁱ ; ni ∈ {1, 2, ...,6});

wherein the target axis movement is also determined based on a target movement(s) of these further reference(s) determined by means of this further data processing(s), in one embodiment dq _s /dt = dq _s /dt(dx ₁ /dt, dx ₂ /dt, dx ₃ /dt (, ...));
wherein, in this determination of the target axis movement, the determined target movements of the first and second reference are given (higher) priority over this determined target movement of this at least one further reference, in one embodiment, when target movements of two or more further references are determined, the target movements of the first and second reference are given (higher) priority over these determined target movements of these further references, and additionally, for a predetermined, in a further development fixed or varying, sequence or ranking of the further references, the determined target movement of the respective highest-ranking or foremost or uppermost further reference is given (higher) priority over the determined target movements of the remaining further references.

By dividing into or specifying individual target movements of more than two robot-fixed references in combination with a corresponding, in particular hierarchical or weighted, prioritization of these target movements, redundancies can be used or resolved particularly advantageously and/or machine learning can be used particularly advantageously, in particular the target axis movement can be determined particularly efficiently and/or reliably.

In a preferred embodiment, the first robot-fixed reference is a reference fixed to the robot end flange, preferably a tool center point (TCP) of the robot. This allows the robot to carry out desired TCP movements particularly advantageously. The second robot-fixed reference can, for example, be an elbow-fixed reference of a robot arm, a base-fixed reference of a movable base of the robot or the like. A determined target movement of one or more of the references, preferably the first and/or the second and/or the one or more of the further reference(s), describes or specifies in one embodiment (each) a one-, two- or three-dimensional translational movement or, in particular, a temporal change in a one-, two- or three-dimensional position and/or a one-, two- or three-dimensional rotational movement or, in particular, a temporal change in a one-, two- or three-dimensional orientation of the (respective) reference, preferably in the Cartesian or visual or work space of the robot. As a result, redundancies can be used or resolved particularly advantageously and/or machine learning can be used particularly advantageously, in particular the target axis movement can be determined particularly efficiently and/or reliably. If a target movement of a reference k is referred to as dx _k /dt in the usual way, then according to an embodiment of the present invention a target axis movement dq _s /dt of the robot is determined based on the target movements of the references dx _k /dt, k = 1, 2(, 3, ...) determined by means of the data processing (dq _s /dt = dq _s /dt(dx ₁ /dt, dx ₂ /dt(, dx ₃ /dt(, ...))), whereby these target movements of the references are each determined by means of one of the data processing operations and are given (higher) priority in a predetermined sequence or ranking compared to the lower-ranking or subsequent reference(s).

In one embodiment, a determined target movement of a reference, compared to which the determined target movement(s) of the other reference(s) are (higher) prioritized, describes or indicates a change, in particular a temporal change, in the axis or joint coordinates of the robot (dq _last /dt). This makes it possible to approach desired target configurations as far as possible. In a further development, the corresponding data processing then determines this target movement on the basis of the current axis or joint coordinates and/or a current value of at least one path parameter and/or a current value of at least one environmental parameter and/or a current value of at least one process parameter. For a more compact representation of this embodiment, in which even a most subordinate target movement is defined directly in the axis or joint coordinate space or determined by the corresponding data processing, the robot as a whole or the entirety of its axes is generally referred to as a robot-fixed reference in the sense of the present invention, wherein preferably one or more of the robot-fixed references mentioned here, in one embodiment at least the first and in a further development also the second robot-fixed reference and/or the or one or more of the further robot-fixed reference(s), (each) comprises a reference that is fixed in place with respect to a specific member of the robot, in particular a (reference) point, a coordinate system or the like, is preferably.

• - auf Basis einer aktuellen Pose dieser Referenz, insbesondere also die erste Datenverarbeitung die Soll-Bewegung der ersten Referenz auf Basis einer aktuellen Pose der ersten Referenz und/oder die zweite Datenverarbeitung die Soll-Bewegung der zweiten Referenz auf Basis einer aktuellen Pose der zweiten Referenz und/oder die (jeweilige) weitere Datenverarbeitung die Soll-Bewegung der (jeweiligen) weiteren Referenz auf Basis einer aktuellen Pose der (jeweiligen) weiteren Referenz; und/oder
• - auf Basis eines aktuellen Wertes wenigstens eines Bahnparameters; und/oder
• - auf Basis eines aktuellen Wertes wenigstens eines Umweltparameters; und/oder
• - auf Basis eines aktuellen Wertes wenigstens eines Prozessparameters, wobei in einer Weiterbildung dieser bzw. wenigstens einer dieser Bahnparameter (jeweils) eine Prozesszeit, insbesondere einen (aktuellen) Zeitstempel, und/oder abgefahrene Strecke umfasst und/oder dieser bzw. wenigstens einer dieser Umweltparameter (jeweils) von einer, insbesondere ein-, zwei- oder mehrdimensionalen und/oder variablen, Pose eines oder mehrerer Hindernisse(s), insbesondere relativ zum Roboter, abhängt, diese Pose insbesondere angeben kann und/oder dieser bzw. wenigstens einer dieser Prozessparameter (jeweils) von einem Zustand eines mithilfe des Roboters durchgeführten Prozesses abhängt, diesen Zustand insbesondere angeben kann. In einer Ausführung wird der Wert des bzw. wenigstens eines der Umweltparameter(s) und/oder der Wert des bzw. wenigstens eines der Prozessparameter(s) sensorisch erfasst, vorzugsweise mithilfe wenigstens einer Kamera und/oder wenigstens eines Ultraschallsensors und/oder wenigstens eines Radarsensors und/oder wenigstens eines Lidarsensors und/oder wenigstens eines Kraft- und/oder Momentensensors oder dergleichen. Eine Pose im Sinne der vorliegenden Erfindung umfasst in einer Ausführung eine ein-, zwei- oder dreidimensionalen Position und/oder eine ein-, zwei- oder dreidimensionale Orientierung, vorzugsweise im kartesischen bzw. Anschauungs- bzw. Arbeitsraum des Roboters. Bezeichnet τ einen (aktuellen Wert eines) Bahnparameter(s), insbesondere eine Prozesszeit und/oder abgefahrene Strecke, und y einen (aktuellen Wert eines) Umweltparameter(s) und/oder Prozessparameter(s), insbesondere eine Pose eines oder mehrerer Hindernisse(s), so wird in einer Ausführung mittels der wenigstens teilweise auf maschinellem Lernen basierenden Datenverarbeitung i die Soll-Bewegung der jeweiligen Referenz i entsprechend gemäß dx_i/dt = f_i(x_i, τ, y) oder dx_i/dt = f_i(x_i, τ) oder dx_i/dt = f_i(x_i, y) oder dx_i/dt = f_i(x_i) oder dx_i/dt = f_i(τ, y) oder dx_i/dt = f_i(τ) oder dx_i/dt = f_i(y) ermittelt. Natürlich kann die Soll-Bewegung einer oder mehrerer der Referenzen auch noch von weiteren Parametern abhängen. Dabei können zwei oder mehr der Datenverarbeitungen jeweils alle auf Basis des (aktuellen Werts des)selben Bahnparameters und/oder des (aktuellen Werts des)selben Umwelt- bzw. Prozessparameters ermitteln und/oder verschiedene Datenverarbeitungen verschiedene Arten von Eingangsparametern verwenden. So kann es, rein exemplarisch, vorteilhaft sein, eine Pose eines Hindernisses bei der Ermittlung der Soll-Bewegung(en) der zweiten und gegebenenfalls weiteren Referenz(en) zu berücksichtigen, nicht aber bei der Ermittlung der Soll-Bewegung der ersten Referenz, so dass ein Ausweichen gegenüber diesem Hindernis nur durch Soll-Bewegung der entsprechenden Referenz(en) realisiert wird, während die Ermittlung der Soll-Bewegung der ersten Referenz rasch(er) und/oder einfach(er) durchgeführt werden kann. In einer Ausführung umfasst das Verfahren ein Ermitteln einer aktuellen Pose der jeweiligen Referenz und/oder eines aktuellen Wertes des (jeweiligen) Bahnparameters und/oder eines aktuellen Wertes des (jeweiligen) Umwelt- bzw. Prozessparameters, insbesondere mittels eines oder mehrerer roboterseitigen und/oder eines oder mehrerer vom Roboter beabstandeten bzw. umgebungsseitigen Sensoren.

In one embodiment, at least one of the data processing operations determines the target movement of the corresponding reference, in particular the first data processing operation determines the target movement of the first reference and/or the second data processing operation determines the target movement of the second reference and/or the (respective) further data processing operation determines the target movement of the (respective) further reference, (respectively)

• - based on a current pose of this reference, in particular the first data processing the target movement of the first reference based on a current pose of the first reference and/or the second data processing the target movement of the second reference based on a current pose of the second reference and/or the (respective) further data processing the target movement of the (respective) further reference based on a current pose of the (respective) further reference; and/or
• - based on a current value of at least one orbit parameter; and/or
• - based on a current value of at least one environmental parameter; and/or
• - based on a current value of at least one process parameter, wherein in a further development this or at least one of these path parameters (each) comprises a process time, in particular a (current) time stamp, and/or distance traveled and/or this or at least one of these environmental parameters (each) depends on a pose of one or more obstacles, in particular relative to the robot, in particular one-, two- or multi-dimensional and/or variable, and/or can indicate this pose and/or this or at least one of these process parameters (each) depends on a state of a process carried out with the aid of the robot, in particular can indicate this state. In one embodiment, the value of the or at least one of the environmental parameters and/or the value of the or at least one of the process parameters is recorded by sensors, preferably with the aid of at least one camera and/or at least one ultrasonic sensor and/or at least one radar sensor and/or at least one lidar sensor and/or at least one force and/or torque sensor or the like. A pose in the sense of the present invention comprises, in one embodiment, a one-, two- or three-dimensional position and/or a one-, two- or three-dimensional orientation, preferably in the Cartesian or visual or working space of the robot. If τ denotes a (current value of a) path parameter(s), in particular a process time and/or distance traveled, and y denotes a (current value of an) environmental parameter(s) and/or process parameter(s), in particular a pose of one or more obstacles, then in one embodiment the target movement of the respective reference i is determined by means of the data processing i based at least partially on machine learning according to dx _i /dt = f _i (x _i , τ, y) or dx _i /dt = f _i (x _i , τ) or dx _i /dt = f _i (x _i , y) or dx _i /dt = f _i (x _i ) or dx _i /dt = f _i (τ, y) or dx _i /dt = f _i (τ) or dx _i /dt = f _i (y). Of course, the target movement of one or more of the references can also depend on further parameters. In this case, two or more of the data processing operations can all be determined on the basis of the (current value of) the same path parameter and/or the (current value of) the same environmental or process parameter and/or different data processing operations can use different types of input parameters. For example, it can be advantageous to take a pose of an obstacle into account when determining the target movement(s) of the second and possibly further reference(s), but not when determining the target movement of the first reference, so that avoiding this obstacle is only achieved by the target movement of the corresponding reference(s), while determining the target movement of the first reference can be carried out more quickly and/or more easily. In one embodiment, the method comprises determining a current pose of the respective reference and/or a current value of the (respective) path parameter and/or a current value of the (respective) environmental or process parameter, in particular by means of one or more robot-side sensors and/or one or more sensors spaced apart from the robot or on the environment.

$$\text{d}{q}_s/\text{d}t=J_1{}^{\#}\cdot \text{d}{x}_1/\text{d}t+\left(1-J_1{}^{\#}\cdot J_1\right)\cdot J_2{}^{\#}\cdot \left[\text{d}{x}_2/\text{d}t+\left(1-J_2{}^{\#}\cdot J_2\right)\cdot J_3{}^{\#}\cdot \text{d}{x}_3/\text{d}t\right]$$

mit der jeweiligen Pseudoinversen Ji^# der jeweiligen Jacobimatrix J_i = ∂(dx_i/dt)/ ∂(dq/dt) und dem jeweiligen Nullraumoperator N_i = (1 - J_i ^#·J_i). Sofern die nachrangigste Soll-Bewegung eine Soll-Bewegung im Achs- bzw.In one embodiment, one or more of the (higher) prioritized determined target movements of robot-fixed references are prioritized over one or more other of the determined target movements of robot-fixed references (each) by means of a projection of the (respective) other target movement(s) into a null space of the (respective) one, (higher) prioritized target movement, in one embodiment in the form $$\text{d}{q}_{s}l\text{d}t=J_1{}^{\#}\cdot \text{d}{x}_1/\text{d}t+\left(1-J_1{}^{\#}\cdot J_1\right)\cdot J_2{}^{\#}\cdot \text{d}{x}_2/\text{d}t$$

$$\text{d}{q}_s/\text{d}t=J_1{}^{\#}\cdot \text{d}{x}_1/\text{d}t+\left(1-J_1{}^{\#}\cdot J_1\right)\cdot J_2{}^{\#}\cdot \left[\text{d}{x}_2/\text{d}t+\left(1-J_2{}^{\#}\cdot J_2\right)\cdot J_3{}^{\#}\cdot \text{d}{x}_3/\text{d}t\right]$$

with the respective pseudoinverse Ji ^# of the respective Jacobian matrix J _i = ∂(dx _i /dt)/ ∂(dq/dt) and the respective null space operator N _i = (1 - J _i ^# ·J _i ). If the most subordinate desired movement is a desired movement in the axis or

joint coordinate space, the corresponding Jacobian matrix or pseudoinverse can be the identity matrix.

In one embodiment, the prioritization or one or more of the prioritizations is done via a weighting, for example in the form: $$\text{d}{q}_s/\text{d}t=w_1\cdot J_1{}^{\#}\cdot \text{d}{x}_1/\text{d}t+w_2\cdot J_2{}^{\#}\cdot \text{d}{x}_2/\text{d}t\left(+w_3\cdot J_3{}^{\#}\cdot \text{d}{x}_3/\text{d}t\left(+\dots \right)\right)$$

The prioritization weights w ₁ , w ₂ , ... can be constant or can also be varied over time. If the most subordinate target movement is a target movement in the axis or joint coordinate space, the corresponding Jacobian matrix or pseudoinverse can be the identity matrix (...+ w _ast· dq _last /dt).

In one embodiment, one or more of the (higher) prioritized determined target movements of robot-fixed references are prioritized over one or more other of the determined target movements of robot-fixed references (each) by means of a cascade of quadratic programs, in a further development taking into account movement restrictions, in particular hardware limits and/or virtual limits, in particular for safety spaces or the like.

In one embodiment, one or more of the (higher) prioritized determined target movements of robot-fixed references are prioritized over one or more other of the determined target movements of robot-fixed references (each) by means of a saturation-in-the-nullspace algorithm (“SNS algorithm”), in a further development taking into account movement restrictions, in particular hardware limits and/or virtual limits, in particular for safety spaces or the like.

By means of the aforementioned prioritizations, the target axis movement can be determined in a particularly advantageous manner, in particular simply, quickly and/or robustly, without the present invention being limited thereto.

In one embodiment, the provision of the data processing based at least partially on machine learning comprises training the data processing, in a further development a demonstration of training movements of the robot by a human user and machine learning based on these training movements of the robot, in particular machine learning of determining a target movement of the respective reference and/or training the respective data processing based on these training movements of the robot.

In a further training, the demonstration of training movements of the robot by a human user includes the user exerting forces on the compliantly controlled robot ("hand guiding") and/or the user moving a tracked demonstrator and/or the user remotely controlling the robot. Additionally or alternatively, in a further training, the demonstration is carried out using a virtual reality application or augmented reality application.

This allows the robot's training movements to be specified very easily, quickly, precisely and/or reliably, especially by inexperienced users.

• - aktuelle Werte von einem oder mehreren Eingangsparametern der (jeweiligen) Datenverarbeitung(en), beispielsweise aktuellen Posen einer oder mehrerer der Referenzen und/oder aktuelle Werte wenigstens eines Bahnparameters und/oder aktuelle Werte wenigstens eines Umweltparameters und/oder aktuelle Werte wenigstens eines Prozessparameters; und
• - aktuelle Bewegungen der (jeweiligen) roboterfesten Referenz(en) ermittelt und ein Ermitteln der (jeweiligen) Soll-Bewegung(en) der (jeweiligen) roboterfesten Referenz(en) mittels der (jeweiligen) Datenverarbeitung(en) derart maschinell gelernt, dass diese Eingangsparameter möglichst gut auf diese Bewegungen der (jeweiligen) roboterfesten Referenz(en) abgebildet werden. Beim Demonstrieren von Trainings-Bewegungen werden in einer Ausführung zunächst Daten im Achs- bzw. Gelenkkoordinatenraum erfasst und mittels Vorwärtskinematik in den Arbeitsraum bzw. einen oder mehrere Task-Space(s) projiziert.

In one version, when demonstrating training movements of the robot

• - current values of one or more input parameters of the (respective) data processing(s), for example current poses of one or more of the references and/or current values of at least one trajectory parameter and/or current values of at least one environmental parameter and/or current values of at least one process parameter; and
• - current movements of the (respective) robot-fixed reference(s) are determined and the (respective) target movement(s) of the (respective) robot-fixed reference(s) are determined by machine learning using the (respective) data processing(s) in such a way that these input parameters are mapped as well as possible to these movements of the (respective) robot-fixed reference(s). When demonstrating training movements, in one embodiment, data is first collected in the axis or joint coordinates space and projected into the workspace or one or more task spaces using forward kinematics.

In one embodiment, an algorithm automatically or a user, preferably by means of input, decides how many and/or which target movements or robot-fixed references or task spaces with which prioritization are relevant or are to be used.

Additionally or alternatively, in one embodiment, the provision of the data processing based at least partially on machine learning comprises in particular training the data processing, specifying training movements of the robot by an automatic motion planner, in a further development path planner, and machine learning based on these training movements of the robot.

This makes it possible to specify particularly advantageous training movements for the robot.

• - Werte von einem oder mehreren Eingangsparametern der (jeweiligen) Datenverarbeitung(en), beispielsweise Posen einer oder mehrerer der Referenzen und/oder Werte wenigstens eines Bahnparameters und/oder Werte wenigstens eines Umweltparameters und/oder Werte wenigstens eines Prozessparameters; und
• - Bewegungen der (jeweiligen) roboterfesten Referenz(en) ermittelt und ein Ermitteln der (jeweiligen) Soll-Bewegung(en) der (jeweiligen) roboterfesten Referenz(en) mittels der (jeweiligen) Datenverarbeitung(en) derart maschinell gelernt, dass diese Eingangsparameter möglichst gut auf diese Bewegungen der (jeweiligen) roboterfesten Referenz(en) abgebildet werden.

In one version, the robot’s training movements are specified for this

• - values of one or more input parameters of the (respective) data processing(s), for example poses of one or more of the references and/or values of at least one trajectory parameter and/or values of at least one environmental parameter and/or values of at least one process parameter; and
• - Movements of the (respective) robot-fixed reference(s) are determined and a determination of the (respective) target movement(s) of the (respective) robot-fixed reference(s) is machine-learned by means of the (respective) data processing(s) in such a way that these input parameters are mapped as well as possible to these movements of the (respective) robot-fixed reference(s).

Thus, in one embodiment, one or more of the mappings or models dx _i /dt = f _i (x _i , τ, y) or dx _i /dt = f _i (x _i , τ) or dx _i /dt = f _i (x _i , y) or dx _i /dt = f _i (x _i ) or dx _i /dt = f _i (τ, y) or dx _i /dt = f _i (τ) or dx _i /dt = f _i (y) are machine learned by demonstrating training movements of the robot by a human user and/or specifying training movements of the robot by an automatic motion planner.

In one embodiment, the robot is redundant with respect to the target movement of the first reference determined by means of the first data processing, or different axis movements of the robot can cause the same target movement of the first reference. In addition or alternatively, in one embodiment, the robot is redundant with respect to the target movement of the second reference determined by means of the second data processing, or different axis movements of the robot can cause the same target movement of the second reference. In addition or alternatively, in one embodiment, the robot is (respectively) redundant with respect to the target movement of the (respective) further reference determined by means of the or one of the further data processing, or different axis movements of the robot can cause the same target movement of the (respective) further reference. Such redundancies can be used or resolved particularly advantageously by the inventive prioritization when determining the target axis movement.

• - Variieren der bzw. einer oder mehrerer der Priorisierungen (über die Zeit (hinweg) bzw. während des Steuerns des Roboters); und/oder
• - Vorgeben der bzw. einer oder mehrerer der Priorisierungen, in einer Weiterbildung der zeit-abhängigen Prioritäten;

auf. Dadurch kann vorteilhaft auf sich ändernde Randbedingungen reagiert werden.In one embodiment, one or more of the prioritizations mentioned here are specified by a user or automatically by an algorithm and/or varied over time, in particular time-dependent priorities are specified by the user or automatically by an algorithm. Accordingly, in one embodiment, the method comprises the step

• - Varying the prioritization or one or more of the prioritizations (over time or during control of the robot); and/or
• - Specifying one or more of the prioritizations in a further development of the time-dependent priorities;

This allows for a favorable response to changing boundary conditions.

In one embodiment, the machine learning of one or more of the data processing operations is limited in such a way that the resulting, preferably non-linear, dynamic system in the task space dx _u /dt = f(X) has a, preferably global or local, attractor X*, in particular in such a way that starting from an arbitrary state X (with global stability or attractor) or starting from a State in the local vicinity of X* (with local stability or attractor) the learned system is always guaranteed to converge to the attractor state X*. Such restrictions for learning a stable dynamic system dx _u /dt = f(X) can be based in particular on the Lyapunov stability theory.

• - Mittel zum Bereitstellen einer ersten wenigstens teilweise auf maschinellem Lernen basierenden Datenverarbeitung zur Ermittlung einer Soll-Bewegung einer ersten roboterfesten Referenz, wobei das Mittel in einer Weiterbildung die erste Datenverarbeitung umfassen, insbesondere sein, kann;
• - Mittel zum Bereitstellen einer zweiten wenigstens teilweise auf maschinellem Lernen basierenden Datenverarbeitung zur Ermittlung einer Soll-Bewegung einer zweiten roboterfesten Referenz, wobei das Mittel in einer Weiterbildung die zweite Datenverarbeitung umfassen, insbesondere sein, kann;
• - Mittel zum Ermitteln einer Soll-Achs-Bewegung des Roboters basierend auf einer mittels der ersten Datenverarbeitung ermittelten Soll-Bewegung der ersten Referenz und einer mittels der zweiten Datenverarbeitung ermittelten Soll-Bewegung der zweiten Referenz; wobei bei diesem Ermitteln der Soll-Achs-Bewegung die ermittelte Soll-Bewegung der ersten Referenz gegenüber der ermittelten Soll-Bewegung der zweiten Referenz priorisiert wird; und
• - Mittel zum Steuern des Roboters auf Basis der ermittelten Soll-Achs-Bewegung.

According to one embodiment of the present invention, a system for controlling a robot, in particular hardware and/or software, in particular program technology, is set up to carry out a method described here and/or has:

• - means for providing a first data processing based at least partially on machine learning for determining a target movement of a first robot-fixed reference, wherein the means in a further development can comprise, in particular be, the first data processing;
• - means for providing a second data processing based at least partially on machine learning for determining a target movement of a second robot-fixed reference, wherein the means in a further development can comprise, in particular be, the second data processing;
• - means for determining a target axis movement of the robot based on a target movement of the first reference determined by means of the first data processing and a target movement of the second reference determined by means of the second data processing; wherein, in this determination of the target axis movement, the determined target movement of the first reference is prioritized over the determined target movement of the second reference; and
• - Means for controlling the robot based on the determined target axis movement.

In one embodiment, the system has means for providing at least one further data processing based at least partially on machine learning for determining a target movement of a further robot-fixed reference, which in a further development can comprise, in particular be, the further data processing, wherein the means for determining a target axis movement is set up to determine the target axis movement based also on a target movement of this further reference determined by means of this further data processing, wherein in this determination of the target axis movement the determined target movements of the first and second reference are prioritized over this determined target movement of this further reference.

A means in the sense of the present invention can be designed in terms of hardware and/or software, in particular at least one, preferably data- or signal-connected, particularly digital, processing unit, particularly microprocessor unit (CPU), graphics card (GPU) or the like, and/or one or more programs or program modules. The processing unit can be designed to process commands that are implemented as a program stored in a storage system, to detect input signals from a data bus and/or to output output signals to a data bus. A storage system can have one or more, particularly different, storage media, particularly optical, magnetic, solid-state and/or other non-volatile media. The program can be designed in such a way that it embodies or is capable of carrying out the methods described here, so that the processing unit can carry out the steps of such methods and thus in particular can control the robot. In one embodiment, a computer program product can have, in particular be, a storage medium, in particular a computer-readable and/or non-volatile one, for storing a program or instructions or with a program or instructions stored thereon. In one embodiment, executing this program or these instructions by a system or a controller, in particular a computer or an arrangement of several computers, causes the system or the controller, in particular the computer(s), to carry out a method described here or one or more of its steps, or the program or the instructions are set up for this purpose.

In one embodiment, one or more, in particular all, steps of the method are fully or partially computer-implemented or one or more, in particular all, steps of the method are fully or partially automated, in particular by the system or its means.

In one embodiment, the system comprises the robot.

• 1: ein Verfahren nach einer Ausführung der vorliegenden Erfindung; und
• 2: ein System nach einer Ausführung der vorliegenden Erfindung.
• 2 zeigt ein System nach einer Ausführung der vorliegenden Erfindung mit einer Steuerung 20 zum Steuern eines Roboters 10.
• 1 zeigt ein Verfahren nach einer Ausführung der vorliegenden Erfindung zum Steuern des Roboters 10 mittels der Steuerung 20.

Further advantages and features emerge from the subclaims and the embodiments. This shows, partly schematically:

• 1 : a method according to an embodiment of the present invention; and
• 2 : a system according to an embodiment of the present invention.
• 2 shows a system according to an embodiment of the present invention with a controller 20 for controlling a robot 10.
• 1 shows a method according to an embodiment of the present invention for controlling the robot 10 by means of the controller 20.

In a step S10, several machine learning-based data processing 21, 22 of the controller 20 are trained. To do this, a human user (not shown) demonstrates target movements of the robot 10, for example by exerting forces on the compliantly controlled robot and thereby moving it. Likewise, he can also remotely control the robot 10 using a remote control, for example a joystick, a 3D space mouse, a tablet or the like (not shown) or move a tracked demonstrator (not shown) instead of the robot 10.

During these target movements, current poses of a first robot-fixed reference in the form of the TCP 11 and a second robot-fixed reference in the form of the elbow 12 are determined and saved. Additionally or alternatively, an assigned value of a path parameter, for example a process time or distance traveled, and/or an assigned value of an environmental parameter, for example a current pose of one or more obstacles 30, can be determined and saved.

$$22:\text{ d}{x}_{12}/\text{d}t=f_2\left(x_{\mathrm{12,}}\mathrm{\tau },y_{30}\right).$$

The data processing units 21, 22 learn to determine a target movement of the TCPs 11 or elbow 12 on the basis of the stored poses of the TCPs 11 or elbow 12, the stored values of the path parameters and/or the stored poses of the obstacles 30, for example in the form: $$21:\text{ d}{x}_{11}/\text{d}t=f_1\left(x_{\mathrm{11,}}\mathrm{\tau },y_{30}\right);$$

$$22:\text{ d}{x}_{12}/\text{d}t=f_2\left(x_{\mathrm{12,}}\mathrm{\tau },y_{30}\right).$$

Preferably, several target movements that are (only) slightly different from one another or similar are demonstrated one after the other, whereby it is particularly advantageous for the obstacles 30 to have different poses or execute different movements when demonstrating similar or identical target movements of the robot.

In a step S20, a target axis movement of the robot 10 is determined based on a target movement dx ₁₁ /dt of the first reference determined by means of the first data processing 21 and a target movement dx ₁₂ /dt of the second reference determined by means of the second data processing 22, for example in the form: $$d{q}_s/\text{d}t={\mathrm{<figure-callout\; id="11"\; label="J"\; filenames="DE102023210548B3\_0007.png"\; state="\{\{state\}\}">J</figure-callout>}}_{11}{}^{\#}\cdot \text{d}{x}_{11}/\text{d}t+\left(1-{\mathrm{<figure-callout\; id="11"\; label="J"\; filenames="DE102023210548B3\_0007.png"\; state="\{\{state\}\}">J</figure-callout>}}_{11}{}^{\#}\cdot J_{11}\right)\cdot {\mathrm{<figure-callout\; id="12"\; label="J"\; filenames="DE102023210548B3\_0007.png"\; state="\{\{state\}\}">J</figure-callout>}}_{12}{}^{\#}\cdot \text{d}{x}_{12}/\text{d}t.$$

Through this projection of the target movement dx ₁₂ /dt into the zero space of the target movement dx ₁₁ /dt when determining the target axis movement dq _s /dt, the determined target movement of the first reference is prioritized over the determined target movement of the second reference.

In a step S30, the controller 20 controls the robot to execute this determined target axis movement dq _s /dt. This results in new poses of the TCPs 11 or elbow 12, a new value of the path parameter and, if necessary, new poses of the obstacles 30, on the basis of which, as described above, the data processing 21, 22 again determine a target movement dx ₁₁ /dt or dx ₁₂ /dt of the respective robot-fixed reference 11 or 12 and the controller 20 again determines a target axis movement dq _s /dt based on this (see step S20) and controls the robot 10 to execute this new determined target axis movement (see step S30).

In the present disclosure, "has an X" generally does not imply an exhaustive list, but is a shortened form of "has at least one X" and also includes "has two or more Xs" and "has Y in addition to X." Although exemplary embodiments have been explained in the foregoing description, it should be noted that a variety of variations are possible.

In particular, in addition to or as an alternative to a demonstration of target movements of the robot by a human user, target movements of the robot can be specified by an automatic motion planner of the controller and used to train the data processing units 21, 22.

list of reference symbols

10

robot (arm)

11

first robot-proof reference

12

second robot-proof reference

20

steering

21

first data processing

22

second data processing

30

obstacle

q1,..., q7

joint angle

Claims (9)

Method for controlling a robot, comprising the steps: - providing (S10) a first data processing based at least partially on machine learning to determine a target movement of a first robot-fixed reference; - providing (S10) a second data processing based at least partially on machine learning to determine a target movement of a second robot-fixed reference; - determining (S20) a target axis movement of the robot based on a target movement of the first reference determined by means of the first data processing and a target movement of the second reference determined by means of the second data processing; wherein, in this determination of the target axis movement, the determined target movement of the first reference is prioritized over the determined target movement of the second reference; and - controlling (S30) the robot on the basis of the determined target axis movement; wherein the provision of the data processing based at least partially on machine learning comprises - demonstrating training movements of the robot by a human user and/or - specifying training movements of the robot by an automatic motion planner and comprises machine learning based on the training movements of the robot. procedure according to claim 1 , characterized by the step: - providing at least one further data processing based at least partially on machine learning for determining a target movement of a further robot-fixed reference; wherein the target axis movement is also determined based on a target movement of this further reference determined by means of this further data processing; wherein, in this determination of the target axis movement, the determined target movements of the first and second reference are prioritized over this determined target movement of this further reference. Method according to one of the preceding claims, characterized in that at least one of the data processing operations determines the desired movement of the corresponding reference on the basis of a current pose of this reference and/or a current value of at least one path parameter and/or a current value of at least one environmental parameter and/or a current value of at least one process parameter. Method according to the preceding claim, characterized in that at least one path parameter comprises a process time and/or distance traveled and/or at least one environmental parameter depends on a pose of at least one obstacle and/or a value of at least one environmental parameter and/or at least one process parameter is detected by sensors. Method according to one of the preceding claims, characterized in that at least one of the determined target movements of a robot-fixed reference is compared with at least one other of the determined target movements of a robot-fixed reference by means of a projection of the other target Movement into a null space that is prioritized by a target movement or a weighting or a cascade of quadratic programs or a saturation-in-the-nullspace algorithm. Method according to one of the preceding claims, characterized in that a demonstration of desired movements of the robot by a human user comprises an exertion of forces on the compliantly controlled robot by the user and/or a movement of a tracked demonstrator by the user and/or a remote control of the robot by the user and/or is carried out with the aid of a virtual reality application or augmented reality application. Method according to one of the preceding claims, characterized in that the robot is redundant with respect to the desired movement of the first reference determined by means of the first data processing and/or with respect to the desired movement of the second reference determined by means of the second data processing. System for controlling a robot, which is arranged to carry out a method according to one of the preceding claims. Computer program or computer program product, the computer program or computer program product containing instructions which, when executed by one or more computers or a system, claim 8 cause the computer or systems to perform a process according to any of the Claims 1 until 7 to carry out.

Images