WO2019096479A1

WO2019096479A1 - Method and means for operating a robot assembly

Info

Publication number: WO2019096479A1
Application number: PCT/EP2018/076216
Authority: WO
Inventors: Kirill SAFRONOV; Christian Scheurer
Original assignee: Kuka Deutschland Gmbh
Priority date: 2017-11-17
Filing date: 2018-09-27
Publication date: 2019-05-23
Also published as: DE102017010718A1

Abstract

A method according to the invention for operating a robot assembly having at least one robot arm (10) comprises the steps: − capturing (S10) a three-dimensional contour of the surroundings (110) of the actual surroundings (100) of the robot assembly; − determining (S30) a surroundings model and/or a free space (200) for the robot assembly on the basis of the captured contour of the surroundings; and − operating (S50) the robot assembly on the basis of the determined surroundings model and/or free space.

Description

description

Method and means for operating a robot arrangement

The present invention relates to a method and means for operating a robot assembly having at least one robot arm and a robot assembly having the means and a computer program product for carrying out the method.

From within the company, it is known to model robot cells using CAD models of the objects in the robot cell and to plan collision-free robot trajectories based on such cell or environmental models.

Unfortunately, such models require great expertise and effort, especially when they need to be updated as the robot cell is changed using CAD models of added, repositioned, and / or removed objects.

The object of the present invention is to improve the operation of a robot assembly. This object is achieved by a method having the features of claim 1. Claims 9 to 11 provide a means for operating a robot assembly according to a method described herein, a robot assembly having means described herein, and a computer program product for performing a method described herein. The subclaims relate to advantageous developments.

According to one embodiment of the present invention, a robot arrangement has one or more robot arms, which in one embodiment (one) one or more, in particular at least three, in one embodiment at least six, in particular at least seven, joints or (motion) axes , in particular rotary joints or axes, and / or drives for moving the robot arm, in particular adjusting its joints, has / have. The present

Invention may with particular advantage in such multi-, in particular

six-axis or redundant, robot arms are used. According to an embodiment of the present invention, a method of operating the robot assembly comprises the steps of:

- Especially three-dimensional, detecting a three-dimensional

Environmental contour of a real environment of the robot assembly;

Determining an environment model on the basis of or in dependence on this detected environmental contour and / or determining a free space for the or

one or more of the robot arm (s) based on, or in response to, said detected environmental contour, in an embodiment based on or in dependence on said environment model; and

- Operating the robot assembly based on or in dependence thereon

determined environment model and / or free space.

By determining an environment model or free space for a robot arm on the basis of a detected environmental contour of a real environment of the

Robot arrangement can advantageously operate the robot arrangement in one embodiment, in particular the robot arm based on the environment model or

within the free space (ambient) without collision or with reduced

Collision risk procedure, be moved autonomously in one execution.

In particular, in one embodiment, the effort and / or the necessary

Expertise compared to (purely) CAD based environment models can be reduced.

The robot assembly is stationary in one embodiment or relative to its

Environment fixed. Likewise, in particular, one or more of the robotic arm (s) may be mobile in one embodiment (relative to the environment of the robotic assembly), in particular (in each case or together) on a mobile, in particular mobile, platform. In this case, the real environment of the robot assembly may, in particular, be an environment (in) of or approached by or with this mobile robot assembly or platform

be to be approached position. The environment model or the free space is determined in this case in a version for this position. In this case, in one embodiment, the mobile robot arrangement or platform can in the meantime also move away from this position, the robot arrangement being based on or in

Depending on this determined environment model and / or free space is operated when they or their mobile platform is (still or again) in this position In particular, the platform is resting in this position or resting. In one embodiment, the method described herein may also be applied to several such

Positions of a mobile robot assembly or platform are performed.

A free space for a robot arm in the sense of the present invention may in one embodiment be a one-way or more, in particular all-sided, limited Cartesian or joint coordinate space within which the robot arm, in particular when the platform is stationary, can or must proceed.

An environment model in the sense of the present invention describes, in particular approximates, in one embodiment the real environment of the robot arrangement in mathematical, in particular numerical, form, in particular in the form of

Surfaces and / or three-dimensional object (model) s or corresponding data.

In one embodiment, the environmental contour is contactless, in particular optically, and / or with the aid of at least one sensor, in particular a 2D or 3D sensor, in particular a 2D or 3D camera, in an embodiment using exactly or only one such sensor, in particular three-dimensional, captured. A 2D sensor is moved in one embodiment to, or when detecting the environmental contour. In one embodiment, the environmental contour detected by means of at least one, in particular moving, 2D sensor is detected by calculating the three-dimensional surrounding contour from the environment detected with the aid of the sensor, in particular using a structure-by-motion method known per se.

By a non-contact, in particular optical, detection, this can be carried out in one embodiment quickly and / or without obstruction of the robot assembly by a 3D sensor, in particular a 3D camera, in a particularly fast and / or compact execution, by a 2D Sensor, in particular a 2D camera, particularly simple and / or inexpensive in one embodiment.

In a further development, the 3D sensor may have a predetermined pattern, the one

For this purpose, the light source is imaged onto the real environment, or its distortion detected, via travel-time measurements of light distances to the real environment, in one embodiment a so-called ToF camera, in particular a PMD sensor or sensors the like, in particular, detect interferences between measuring and object beams and / or with the aid of (micro) lens arrays or the like in addition to a brightness of pixels and a light direction of rays that lead to a pixel capture or be set up for this purpose or . be used. As a result, the detection in one embodiment can be carried out particularly quickly and / or compactly.

In one embodiment, the sensor, which is freely movable in space for this purpose in an embodiment, for or during detection (by a user or a

Operator) guided manually, in particular pivoted.

Additionally or alternatively, in one embodiment, it communicates wirelessly or via at least one line with a processing means which determines the environmental model and / or the free space on the basis of data detected or transmitted by the sensor.

In one embodiment, the sensor is arranged on a mobile phone, in particular a smartphone, tablet computer or another handheld device such as a personal digital assistant ("PDA") or the like, in particular integrated or detachably attached.

In one embodiment, by manually guiding the, in particular freely movable, sensor and / or wireless communication, the surrounding contour can be detected simply, quickly and / or without obstruction of the robot arrangement, by pivoting a larger environment with a small sensor detection range can be scanned. By using a Handheids, in particular

Smartphones or the like, devices that are otherwise used can advantageously also be used for detecting the surrounding contour. Communication via a line can reduce the risk of interference in one embodiment.

In one embodiment, one or more robotic arm-resistant elements,

In particular, a robot-guided tool and / or the (entire) robot arm, identified in the detected environmental contour, in an embodiment using image processing, in particular recognition, are located in one embodiment or a pose of the element (s) in the detected environmental contour and / or to each other. In a further development, one or the following explained (s)

robot-guided sensor for detecting one or the three-dimensional fine contour of the real environment of the robot arrangement in the detected surrounding contour, in an embodiment using image processing, in particular recognition,

identified, in particular localized.

Additionally or alternatively, in one embodiment, a position of the robot arm, in particular a position of one or more, in an embodiment of all, joints or axes of the robot arm, in which the surrounding contour is detected or detected, detected in a development using sensors at the joints and / or drives. In one embodiment, the detected position of the robot arm is wirelessly transmitted to one or the processing means for determining the environment model and / or free space.

The environment model and / or the free space is in an embodiment based on or in dependence on this identification or localization of or

determined roboterarmfesten elements and / or this detected position of the robot arm.

In a further development, the identified robot arm is eliminated from the detected environmental contour, in particular for determining the environmental model

filtered out or adjusted the detected environmental contour accordingly.

Additionally or alternatively, in a further development, the detected environmental contour, the environmental model and / or the free space are aligned or positioned and / or oriented relative to a robot-arm-fixed, in particular roboterarm base or flange-resistant reference, in particular a reference (coordinate) system a deviation between the roboterarmfesten identified in the detected environmental contour and their pose, which results from the detected position of the robot arm, is minimal.

As a result, in one embodiment, the detected environmental contour, the

Environmental model or the free space with a roboterarmfesten reference matched or matched or aligned to this and thereby advantageously operated the robot assembly, in particular controlled.

In an embodiment, the method comprises the step of: detecting a

Three-dimensional fine contour of the real environment of the robot assembly after detecting the surrounding contour, in an embodiment based on the determined environment model and / or free space. In one embodiment, the

Fine contour recorded with higher accuracy and / or slower than the surrounding contour and / or with the help of at least one robot-guided sensor and the

Robot arrangement operated on the basis of the determined fine contour, in particular an environmental model and / or, in an embodiment based on this environmental model, a collision space for the robot arm determined.

In one embodiment, the environmental contour, in particular with the aid of the manually guided sensor, is thus first roughly recorded and the environmental model and / or free space determined on the basis of this environmental contour (coarse), and subsequently the fine contour recorded with greater accuracy using the robot-guided sensor, wherein the Robot arrangement for this purpose based on the roughly determined

Environment model and / or free space is controlled.

In one embodiment, based on the determined fine contour, the roughly determined environment model for the environmental model and / or the roughly determined free space for the collision space are modified, in particular specified. Likewise, that can

Environmental model and / or the collision space to be redetermined, with a

Collision space for a robot arm according to the present invention in one embodiment may also be a single or multi, especially all sides, limited Cartesian or joint coordinate space, within which the

Robot arm, in particular when the mobile platform is stationary, and which in one embodiment has a higher accuracy than the (roughly determined) free space, an environmental model according to the present invention corresponding to the real environment of the robot arrangement in mathematical, in particular numerical, shape, in particular in the form of surfaces and / or three-dimensional object models or corresponding data, describes, in particular approximated, and in one embodiment has a higher accuracy than the (roughly determined) environment model. As a result, in one embodiment, an advantageous at least two-stage

Method with an initial coarse detection of the environmental contour and the resulting determination of the environmental model and / or free space and a subsequent more precise detection of the fine contour and based thereon

Determination of the environmental model and / or collision space proposed for the robot arm.

In one embodiment, the environment model and / or the free space, in one embodiment additionally or alternatively (also) the environmental model and / or the

Collision space, (each)

- Based on geometric objects, in particular primitives, in one

Executing primitives from a group containing one or more cylinders,

Balls and / or polyhedra, in particular cuboid, wherein these

geometric objects, in one embodiment are independent of a priori known objects in the vicinity of the robot arm, are given or are and / or, in particular by means of pattern recognition, in the detected

Environmental contour to be determined;

on the basis of, in particular, a maximum (possible or achievable working space of the robot arm, and / or

based on a connection and / or approximation of detected points in the

Environment, in particular meshing, interpolation or the like, determined.

By using given geometrical objects, the detected environmental contour can advantageously be supplemented and / or corrected in one embodiment, whereby the omission of the use of a priori known objects, in particular CAD models of such objects, can simplify and / or accelerate the method.

In one embodiment, the free space is at least partially limited by the working space of the robot arm. As a result, in one embodiment, areas of the surrounding contour which the robot arm, in particular when the platform is stationary, can not be excluded in the determination of the free space due to kinematic or structural restrictions. Nevertheless, such areas can be visualized in one embodiment in the detected environmental contour or contained in the determined environment model. In one embodiment, the working space of the robot arm is determined on the basis of its identification, in particular from a database or base determined the robot arm associated specific working space and this used in the determination of the environment model, free space, environmental model and / or collision space, in particular the free - and / or collision space through this specific

Workspace at least partially limited, which has been determined on the basis of the identification of the robot arm. In mobile robot arrangements, the

Working space of the robot arm in particular be his workspace at immovable Roboterarmbasis or platform.

By mesh formation, interpolation or the like, the surrounding contour can be detected coarser in one embodiment, in particular the number of detected points in the environment can be reduced. In one embodiment, the environmental model and / or the free space, in one embodiment additionally or alternatively (also) the environmental model and / or the collision space, are determined (respectively) based on a 3D occupancy grid, in particular on a voxel basis ("3D occupancy grid") , which represents a particularly advantageous approximation.

In one embodiment, when determining the environment model and / or free space and / or environmental model and / or collision space (respectively), a safety distance to the detected ambient or fine contour is provided or maintained or

considered, in particular added. As a result, the security can be increased.

In one embodiment, on the basis of the determined environment model and / or free space and / or the determined fine contour, in particular of the environmental model based thereon and / or collision space, (in each case) a movement, in particular trajectory or trajectory, of the robot arrangement, in particular of the robot arm, in particular a work movement or track for performing a

Work process, a retraction movement or orbit from an error,

in particular Kollisionspose or situation, and / or a search movement

or track for detecting the fine contour using the robot-guided sensor, in particular autonomously or by the robot arrangement itself and / or in advance or offline or during operation of the robot arrangement or online, given, in particular planned, controlled, in particular regulated, and / or monitored.

This represents in each case particularly advantageous modes for operating the robot arrangement.

In one embodiment, the real environment of the robotic assembly may include a

Workplace, in particular a robot cell, have, in particular be, wherein the robot arm is arranged in one embodiment when detecting the surrounding contour at the workplace, in another embodiment detects the surrounding contour without the robot arm and this is then arranged at the workplace. Accordingly, a future environment (still to be detected when detecting) is also referred to as the real environment of the robot arrangement in the sense of the present invention.

In particular, when the robotic arm is located at the workplace when detecting the environmental contour, determining a pose of the robotic arm-fixed reference and the detected environmental contour relative to each other can be simplified and / or their precision and / or reliability improved. Conversely, detecting the surrounding contour without robotic arm can simplify detection.

According to one embodiment, an (operating) means for operating the robot arrangement, in particular hardware and / or software, in particular program technology, for implementing a method described here is set up and / or has:

- Detection means for, in particular three-dimensional, detecting a

three - dimensional environmental contour of a real environment of

Robotic assembly;

- Processing means for determining an environment model based on or in response to this detected environmental contour and / or determining a free space for the or one or more of the robotic arm (s) based on or in dependence on this detected environmental contour, in a design based on or depending on this environment model; and

- Means for operating the robot assembly based on or in dependence on this determined environment model and / or free space. In one embodiment, the (operating) means and / or his (e) means comprises: means for non-contact, in particular optical detection of the surrounding contour. Additionally or alternatively, in one embodiment, the detection means on at least one sensor, in particular 2D or 3D sensor, in particular a 2D or 3D camera on.

In a development, the sensor for detecting manual guidance, in particular pivoting, and / or for communication via at least one line or for wireless communication with the processing means is arranged and / or arranged on a handheld device, in particular a smartphone.

In one embodiment, the (operating) means and / or its agent comprises:

Means for identifying at least one roboterarm element, in particular a robot-guided tool and / or the robot arm, in the detected environmental contour, and / or means for detecting a position of the robot arm, and / or means for determining the environment model and / or the free space based on Identification of the roboterarmfesten element and / or the detected position; and or

Means for detecting a three-dimensional fine contour of the real environment of the robot assembly after detecting the surrounding contour with higher

Accuracy and / or slower than the surrounding contour and / or using at least one robot-guided sensor and operating the robot assembly, in particular determining an environmental model and / or a collision space for the robot arm, based on the determined fine contour; and or

Means for determining the environment model and / or free space on the basis of, in particular independent of a priori known objects in the vicinity of the robot arm, predetermined geometric objects, in particular primitives, based on a working space of the robot arm, and / or based on a compound and / or Approximation of detected points in the environment; and or

Means for prescribing, controlling and / or monitoring a movement of the

Robot arrangement, in particular a working movement for performing a working process, a retraction movement from a fault pose and / or a search movement for detecting the fine contour using the robot-guided sensor, based on the determined environment model and / or free space and / or the determined fine contour. A means in the sense of the present invention may be designed in terms of hardware and / or software, in particular a data or signal-connected, preferably digital, processing, in particular microprocessor unit (CPU) and / or a memory and / or bus system or multiple programs or program modules. The CPU may be configured to execute instructions implemented as a program stored in a memory system, to capture input signals from a data bus, and / or

Output signals to a data bus. A storage system may comprise one or more, in particular different, storage media, in particular optical, magnetic, solid state and / or other non-volatile media. The program may be such that it is capable of embodying or executing the methods described herein, so that the CPU may perform the steps of such methods and thus, in particular, operate the robot assembly. In one embodiment, a computer program product may include, in particular, a non-volatile storage medium for storing a program or a program stored thereon, wherein execution of this program causes a system or a controller, in particular a computer, to do so method described herein or one or more of its steps.

In one embodiment, one or more, in particular all, steps of

Method completely or partially automated, in particular by the (operating) means or its (e) means.

In one embodiment, the detection of a three-dimensional environmental contour comprises the detection of a plurality of staggered three-dimensional partial environmental contours of the real environment and their joining to the captured (overall) environmental contour. In this case, in an embodiment in the detected partial environmental contours, in each case the robot arrangement or a part of the

Localized robot arrangement, identified in particular as described here, and the detected partial environmental contours then assembled on the basis of these localizations to the overall environment contour, in particular (in each case) transformed into one or the common (s) reference (coordinate) system, wherein the (respective) transformation results from the localization of the robot arrangement in the (respective) partial environmental contour. In one embodiment, the partial environmental contours become put together so that the robot arrangements located in them

to match.

Additionally or alternatively, in one embodiment, the detection of a three-dimensional fine contour can also include the detection of a plurality of three-dimensional partial fine contours of the real environment and their joining

Include (total) fine contour. In this case, in one embodiment, the robot arrangement or a part of the robot arrangement is located, in particular identified in the manner described here, and the partial fine contours are then combined on the basis of these localizations to form the overall fine contour, in particular (in each case) into one or the common (s)

Reference (coordinate) system transformed, wherein the (respective) transformation results from the localization of the robot assembly in the (respective) partial fine contour. In one embodiment, the partial fine contours are assembled so that the robot arrangements located within them coincide. Additionally or alternatively, in one embodiment, determining the environmental and / or the

Environmental model, the determination of several sub-environmental or environmental models and their joining to the (overall) environmental or environmental model include. In one embodiment, in the determined part-environment or

Environmental models each located the robot assembly or a part of the robot assembly, in particular identified in the manner described here, and the partial environmental or environmental models then based on these localizations for

(Overall) environmental or environmental model combined, in particular (in each case) transformed into one or the common reference (coordinate) system, whereby the (respective) transformation from the localization of the robot arrangement in the (respective) Environmental or environmental model results. In one embodiment, the sub-environmental models are assembled such that the robot arrays located within them match.

Further advantages and features emerge from the subclaims and the exemplary embodiments. This shows, partially schematized: 1 shows a robot arrangement with a robot arm and a means for operating the robot arrangement according to an embodiment of the present invention; and

FIG. 2 shows a method for operating the robot arrangement according to FIG

Embodiment of the present invention.

1 shows a robot arrangement with a six-axis robot arm 10, which is arranged in a robot cell 100 that is delimited by cell walls, for example fences 21.

In the cell 100, two objects 22, for example tables, shelves, machine tools, conveyors or the like are arranged by way of example.

A robot controller 1 1 performs a method of operating the robot assembly according to an embodiment of the present invention, which will be explained below with reference to FIG. 2.

In a step S10, a three-dimensional environmental contour 110 of the real environment in the form of the robot cell 100 is detected by means of a manually guided by a user 3D camera of a smartphone 30, which this pivots this and transmits the corresponding data wirelessly to the robot controller 1 1. In a modification, a 2D camera can also be used and the three-dimensional contour of the surroundings can be calculated, for example, by means of a so-called "structure from motion" method.

The detected environmental contour 1 10 is indicated in phantom in Fig. 1, wherein a deviation of this roughly detected environmental contour 1 10 of the real environment 100 is schematized and exaggerated thereby indicated that the detected

Surrounding contour 1 10 partially within or outside the real environment 100 is shown.

In a step S20, the robot controller 1 1 identifies the robot arm 10 in the detected surrounding contour 1 10, which is indicated cross-hatched in FIG. In addition, in step S20, it detects its position based on its detected joint or axis angle.

In a step S30, it determines based on the detected environmental contour 1 10, the robot arm 10 identified therein and its detected position

Environment Model.

For this purpose, it removes the identified robot arm 10 from the detected environmental contour 1 10.

In addition, it identifies 1 10 geometric primitives in the detected environmental contour. For this purpose, a cuboid 120 is indicated hatched by way of example, with the in

Environment model, the right in Fig. 1 object 22 is approximated.

In addition, the robot controller 1 1 directs the environmental model or the detected environmental contour 1 10 relative to a robot arm base fixed coordinate system, with respect to which the position of the robot arm 10 has been detected on the basis of its detected joint or Achswinkel, such that the position of the in the detected ambient contour 1 10 identified robot arm 10 as little as possible deviates from this detected on the basis of its detected joint or axis angle position.

For example, the robot controller 1 1 can rotate the environment model or the detected environmental contour 1 10 around an axis perpendicular to the plane of the drawing of FIG. 1 and / or move it in the plane of the drawing until the orientation or position of the rocker and the hand of the identified

Robotic arm as well as possible with the position of rocker and hand according to the detected joint or axis angle matches.

In a step S40, the robot controller 11 determines 1 based on this roughly determined environment model, which in turn based on the coarse detected

Environmental contour 1 10 has been determined, roughly a free space 200, the limit 210 is indicated by dashed lines in Fig. 1. As can be seen from this, this free space 200 is also limited by the constructively maximum possible working space 300 of the robot arm, which is indicated in dotted lines in FIG.

In a step S50, the robot controller 1 1 then moves the robot arm 10 within this roughly determined free space 200 in order to precisely record a fine contour 130 of the surroundings 100 with a robot-guided 3D camera 13, which in the context of the drawing accuracy of FIG. 1 with the surroundings 100 coincides and therefore is not separately recognizable in Fig. 1.

In a step S60, the robot controller 11 analogous to step S30, S40 can now determine a (more accurate) environmental model and / or a (more accurate) collision space for the robot arm 10 on the basis of the acquired fine contour 130.

As an alternative to scanning the environment 100 with the robot-guided 3D camera 13, step S60 can also be omitted and the robot controller 11 in step S50, for example, a retraction movement of the robot arm 10 from a

Planning and controlling the collision situation or another autonomous movement functionality on the basis of the roughly determined free space 200.

Although exemplary embodiments have been explained in the foregoing description, it should be understood that a variety of modifications are possible.

Thus, in particular, a partial environmental contour can in each case be detected several times and the robot arm can be located in it, wherein these mutually offset partial environmental contours are then joined to the or an overall environmental contour in accordance with the respective robot arm located therein

be transformed according to each other or in a common reference (coordinate) system. It should also be noted that it is the exemplary

The explanations are merely examples of the scope of protection

Applications and the structure should in no way limit. Rather, the expert is by the preceding description a guide for the

Implementation of at least one exemplary embodiment given, wherein various changes, in particular with regard to the function and arrangement of the components described, can be made without departing from the scope, as it is apparent from the claims and these equivalent

Feature combinations results.

LIST OF REFERENCES

10 robotic arm

1 1 robot control

13 robot-guided 3D sensor

21 robot cell fence

22 object in robot cell

30 hand-held 3D camera

100 environment

1 10 detected environmental contour

120 geometric primitive

130 detected fine contour

200 free space

210 free space limit

300 working space (border)

Claims

claims

A method of operating a robot assembly having at least one

Robot arm (10), with the steps:

- detecting (S10) a three-dimensional environmental contour (110) of a real environment (100) of the robot assembly;

- determining (S30) an environmental model and / or a free space (200) for the robot arm based on the detected environmental contour; and

- Operating (S50) of the robot assembly based on the determined

Environment model and / or free space.

2. The method according to claim 1, characterized in that the environmental contour without contact, in particular optically, and / or by means of at least one sensor, in particular 2D or 3D sensor, in particular a 2D or 3D camera (30) is detected.

3. The method according to the preceding claim, characterized in that the sensor (30) for detecting manually guided, in particular pivoted, and / or wirelessly or via at least one line with a

Processing means for determining the environment model and / or free space communicates and / or on a handheld, in particular a smartphone, is arranged.

4. The method according to any one of the preceding claims, characterized by

at least one of the steps:

Identifying (S20) at least one roboterarmfesten element,

in particular a robot-guided tool (13) and / or the

Robot arm (10), in the detected environmental contour (1 10); and or

Detecting (S20) a position of the robot arm (10),

wherein the environmental model and / or the free space is determined on the basis of the identification of the roboterarmfesten element and / or the detected position.

5. The method according to any one of the preceding claims, characterized by the step: - Detecting (S50) a three-dimensional fine contour (130) of the real

Environment (100) of the robot assembly after detecting the

Environmental contour (1 10);

the fine contour with higher accuracy and / or slower than the

Environmental contour and / or detected using at least one robot-guided sensor (13) and operated the robot assembly based on the determined fine contour, in particular an environmental model and / or a collision space for the robot arm is determined (S60).

6. The method according to any one of the preceding claims, characterized in that the environment model and / or the free space

- Based on, in particular independent of a priori known objects in the vicinity of the robot arm, predetermined geometric objects, in particular primitives (120);

on the basis of a working space (300) of the robot arm; and or

based on a connection and / or approximation of detected points in the

Surroundings

is determined.

7. The method according to any one of the preceding claims, characterized in that on the basis of the determined environment model and / or free space and / or the determined fine contour movement of the robot assembly, in particular a working movement for performing a work process, a

Retraction movement ung from a Fehlpose and / or a search movement for detecting the fine contour using the robot-guided sensor, predetermined, controlled and / or monitored.

8. The method according to any one of the preceding claims, characterized in that the real environment of the robot assembly has a workplace, in particular a robot cell (100), wherein the robot arm is arranged when detecting the surrounding contour at the workplace or the surrounding contour is detected without the robot arm ,

9. means (1 1, 30) for operating a robot arrangement having at least one robot arm (10) which is set up to carry out a method according to one of the preceding claims and / or comprises:

- Detection means (30) for detecting a three-dimensional

Environmental contour (1 10) of a real environment (100) of the robot assembly;

- Processing means (1 1) for determining an environment model and / or a free space (200) for the robot arm based on the detected

Surrounding contour; and

- Means (1 1) for operating the robot assembly based on the determined environment model and / or free space.

10. Robot arrangement with a robot arm (10) and a means (1 1, 30) for

Operating the robot assembly according to any one of the preceding claims.

1 1. A computer program product having a program code stored on a computer-readable medium for performing a method according to any one of the preceding claims.