Pose measuring device for accuracy improvements in kinematics

Abstract

In summary, this involves controlling a kinematic system to assume a (previously obtained / predefined) target pose. For this purpose, a pose measurement system is integrated into the kinematic system. The pose measurement system comprises a first and a second solid body. The pose measurement system is configured to measure a solid body pose. The solid body pose is a pose that the first and second solid bodies assume relative to each other. The pose measurement system is integrated into the kinematic system by: (i) attaching the first solid body to the end effector and attaching the second solid body to the base, or (ii) attaching the second solid body to the end effector and attaching the first solid body to the base. One or more solid body poses are measured by the attached pose measurement system when the kinematic system is controlled with the respective configuration vector(s).The target configuration vector is determined based on the target pose and the measured solid pose(s), and the kinematics are controlled with the determined target configuration vector.

Description

[0001] The present invention relates to devices and methods for increased accuracy (in particular motion accuracy and / or pose determination accuracy) in kinematics, and in particular in hexapods. BACKGROUND

[0002] For many years, increasing accuracy has been a key technical challenge in kinematics, especially parallel kinematics. Inaccuracies can often be traced back to an insufficiently precise realization of the kinematic geometry parameters (i.e., manufacturing inaccuracies in the kinematics). Consequently, so-called parameter identification, also known as kinematic calibration, has been proposed.

[0003] The challenge remains to find simple, cost-effective, and practical methods that, in addition to increasing accuracy, ideally offer further advantages. One of the objectives underlying the present invention is therefore to improve the accuracy of kinematics. SUMMARY

[0004] The problem is solved according to the invention by the features of the independent claims. Some advantageous embodiments are the subject of the dependent claims.

[0005] The invention is based on the idea of ​​increasing the accuracy of a hexapod by incorporating a pose measuring system, referred to here as a pose transmitter, into the hexapod, the accuracy of which can thereby be transferred to the hexapod.

[0006] According to a first aspect of the present invention, a method for controlling a kinematic system to assume a target pose is provided. The method comprises (i) a step of incorporating a pose measurement system into the kinematic system, (ii) a step of obtaining the target pose, (iii) a step of determining a target configuration vector based on the target pose and one or more measured solid poses, and (iv) a step of controlling the kinematic system with the target configuration vector. The pose measurement system comprises a first and a second solid, is configured to measure a solid pose, and the solid pose is the pose that the first and second solid assume relative to each other.Furthermore, the pose measurement system is integrated into the kinematics by: (i) attaching the first solid body to the end effector and attaching the second solid body to the base, or (ii) attaching the second solid body to the end effector and attaching the first solid body to the base. The one or more measured solid body poses are measured by the attached pose measurement system when the kinematics are actuated with the respective configuration vector(s).

[0007] According to a second aspect, which is provided in addition to the first aspect, determining the target configuration vector comprises the following steps: (i) calculating the target configuration vector based on the target pose and an inverse kinematics of the kinematics; (ii) controlling the kinematics with the target configuration vector; (iii) measuring, by means of the attached pose measurement system, a solid pose that the first and second solids assume relative to each other when controlling the kinematics with the calculated target configuration vector; and (iv) adjusting the target configuration vector based on a deviation of the measured solid pose from the target pose.

[0008] According to a third aspect, which is provided in addition to the second aspect, adjusting the target configuration vector includes: (i) determining a correction of the target configuration vector based on the solid pose measured when the kinematics are driven with the target configuration vector; and (ii) adjusting the target configuration vector based on the correction.

[0009] According to a fourth aspect, provided in addition to the third, the correction corresponds to a difference between: (a) the configuration vector assigned to the solid pose measured when the kinematics are controlled with the target configuration vector according to the inverse kinematics, and (b) the configuration vector assigned to the target pose according to the inverse kinematics. Furthermore, when adjusting the target configuration vector with the correction, the correction is added to the target configuration vector.

[0010] According to a fifth aspect, which is provided in addition to the second, third, or fourth aspect, the following steps will be repeated when determining the target configuration vector until a predetermined condition is met: (i) controlling the kinematics with the target configuration vector; (ii) measuring, by the attached pose measurement system, the solid pose that the first and second solids assume relative to each other when controlling the kinematics with the target configuration vector; and (iii) adjusting the target configuration vector based on the deviation of the measured solid pose from the target pose.

[0011] According to a sixth aspect, which is provided in addition to the fifth aspect, the predetermined condition is a condition on: (a) being a magnitude of a last determined correction; and / or (b) indicating a magnitude of a deviation of a last measured solid pose from the target pose.

[0012] According to a seventh aspect, which is provided in addition to one of the previous aspects, when calculating the target configuration vector based on the target pose and the inverse kinematics, a configuration vector that is assigned to the target pose according to the inverse kinematics is calculated as the target configuration vector.

[0013] According to an eighth aspect, provided in addition to the first, second, third, fourth, fifth, or sixth aspects, the procedure involves measuring a zero-state solid pose using the mounted pose measurement system. The zero-state solid pose is the pose that the first and second solid bodies assume relative to each other when the kinematics are actuated with a zero-configuration vector. This zero-configuration vector is assigned to a zero-state kinematic position according to the inverse kinematics. Furthermore, when calculating the target configuration vector based on the target pose and the inverse kinematics, a configuration vector is calculated as the target configuration vector, which is assigned to a modified target pose according to the inverse kinematics. The modified target pose is derived from the target pose by applying a pose transformation that is the inverse of a pose transformation corresponding to the measured zero-state solid pose.

[0014] According to a ninth aspect, a pose measurement system is provided. The pose measurement system is designed for integration into a kinematic system and comprises a first and a second solid body. The pose measurement system is configured to be integrated into the kinematic system by: (i) attaching the first solid body to the end effector and attaching the second solid body to the base, or (ii) attaching the second solid body to the end effector and attaching the first solid body to the base. Furthermore, the pose measurement system includes means for measuring the pose assumed by the first and second solid bodies relative to each other.

[0015] According to a tenth aspect, a computer program is provided. The computer program includes instructions which, when executed by a computer, cause it to: (i) determine the target configuration vector and control the kinematics according to a procedure provided according to one of the first to eighth aspects.

[0016] According to an eleventh aspect, a computer-readable data carrier is defined as one on which a computer program is stored according to the tenth aspect.

[0017] According to a twelfth aspect, a method for calibrating a pose measurement system is provided. The pose measurement system comprises a first and a second solid and is configured to measure a solid pose that the first and second solid assume relative to each other. Furthermore, the pose measurement system is configured to assign a calibrated configuration vector to a measured solid pose, which indicates a calibrated solid pose corresponding to the measured solid pose.In this process, (i) the calibrated configuration vector is obtained by adding a configuration vector correction to an uncalibrated configuration vector, (ii) the uncalibrated configuration vector is assigned to the measured solid pose according to an inverse kinematics, (iii) the configuration vector correction is assigned to the uncalibrated configuration vector according to a predetermined assignment rule, and (iv) the predetermined assignment rule is based on an interpolation which is based on a plurality of support points, each containing a predetermined configuration vector and a predetermined configuration vector difference.The procedure for calibration according to the twelfth aspect comprises the following steps for each of a plurality of solid poses: (i) measuring the solid pose with the pose measurement system; (ii) measuring the solid pose with an external pose measurement system; (iii) determining a configuration vector difference, which is a difference between a first and a second configuration vector, wherein (a) the first configuration vector is assigned to the solid pose measured with the pose measurement system according to an inverse kinematics, and (b) the second configuration vector is assigned to the solid pose measured with the external pose measurement system according to the inverse kinematics; and (iv) storing the first configuration vector and the configuration vector difference as one of the support points. BRIEF DESCRIPTION OF THE FIGURES

[0018] Further details, advantages and features of the invention will become apparent from the following description and the drawings, to which express reference is made with regard to all details not described in the text. The drawings show: Fig. 1 a schematic representation of an exemplary serial kinematics; Fig. 2 a schematic representation of an exemplary parallel kinematic system; Fig. 3 a schematic representation of the relationship between configuration space, workspace, direct kinematics and indirect kinematics; Fig. 4-6 a schematic representation of a hexapod with built-in pose indicator; Fig. 7 a schematic representation of a kinematic interface, a so-called three-groove kinematic coupling; Fig. 8 a schematic representation for the exemplary calibration of a pose transmitter using a hexapod; Fig. 9 a schematic representation to illustrate how an exemplary hexapod and an exemplary virtual hexapod relate to each other; Fig. 10 a flowchart with exemplary steps for calibrating a pose transmitter; Fig. 11 a flowchart with exemplary steps for the calibrated positioning of a hexapod; Fig. 12. A flowchart with exemplary steps for the calibrated positioning of a hexapod using an iterative method and avoiding the calculation of indirect kinematics; and Fig. 13 a block diagram of an exemplary control device of the pose generator.

[0019] Identical reference numbers and symbols in the various illustrations denote identical or corresponding elements, according to certain embodiments. DETAILED DESCRIPTION

[0020] The present invention relates to methods for controlling a kinematic system (i.e., a kinematic machine, in particular, for example, a robot) using a pose measurement system, as well as control devices for controlling kinematic systems, which are configured to execute such methods using a pose measurement system. The invention further relates to computer programs corresponding to the methods for controlling a kinematic system presented herein, which include instructions that, when executed by a computer, cause the computer to carry out the respective method (i.e., in particular, to determine the target configuration vector and to control the kinematic system, for example, by generating corresponding control instructions for the kinematic system). Finally, the present invention also includes computer-readable data carriers on which such a computer program is stored.The invention further relates to pose measuring systems suitable for use in the methods according to the invention, as well as methods for calibrating said pose measuring systems. Kinematics / Kinematic machines

[0021] In robotics, a fundamental distinction is made between the main classes of serial and parallel kinematics, which are also referred to here as "parallel kinematics." While serial kinematics consist of a series of links (e.g., linear and / or rotary axes) forming an open kinematic chain, parallel kinematics consist of a number of closed kinematic chains. This will be explained in more detail below. Furthermore, there are also so-called hybrid kinematics, which represent a combination of the aforementioned parallel and serial kinematics.

[0022] The term "hexapod" is frequently used in this disclosure, referring in particular to a Stewart platform. This term is used only because the explicit embodiments relate to a hexapod. However, the invention is not limited to a specific type of kinematics / kinematic machine and encompasses parallel kinematics as well as serial, hybrid kinematics and, in particular, stacked kinematics, e.g., with rotary adjusters. Unless otherwise apparent from the context, the term "hexapod" in this disclosure can therefore be replaced by the term "kinematic machine" or "kinematics" or, where appropriate, "parallel kinematic machine" or "parallel kinematics".

[0023] In particular, the present invention is also applicable to parallel kinematics that are not hexapods. In this sense, the term hexapod can be understood as a representative of an equivalence class of parallel kinematics with 6 degrees of freedom (DOF) of the end effector or pose, where the representative here stands for the entire class. Parallel kinematics with a DOF of less than 6 DOF are also included, as are hybrid kinematics and other kinematics.

[0024] It should be noted that in this disclosure, the term "kinematics" is frequently used as a shorthand for "kinematic machine," and the term "parallel kinematics" as a shorthand for "parallel kinematic machine." However, the term "kinematics" is also used to denote the direct / indirect kinematics of the kinematic machine, i.e., the mappings between the configuration space and the working space of the kinematic machine. Thus, direct / indirect kinematics does not refer to machines / devices / robots, but rather to mappings between the configuration space and the working space of the kinematic machine. Serial kinematics

[0025] Serial kinematics refers to the classic design of an open kinematic chain, in which the individual axes of motion are arranged sequentially, building upon one another in series. A kinematic chain is thus understood as a series of multiple bodies (the links of the chain) that are movably connected to each other by joints. The individual links of the chain can be rigid bodies or, for example, length-adjustable elements. Depending on the robot, they are also referred to as arm / leg components.

[0026] A joint connects two links and can have different degrees of freedom. The arrangement and type of joints and links influence the trajectories that the individual links can describe. Kinematic chains play an important role in the planning and calculation of the possible movements of industrial and other robots.

[0027] A classic example of serial kinematics is a SCARA robot (short for "Selective Compliance Assembly Robot Arm"), in which a specific pose is usually realized with two different configuration vectors.

[0028] Fig. Figure 1 schematically shows an example of serial kinematics with multiple links and joints. As indicated, these links can move linearly in one or more directions, can perform a rotational movement in a plane, have a joint structure, and / or are length-adjustable. Parallel kinematics

[0029] Parallel kinematics refers to a kinematic system consisting of a number of closed kinematic chains. In practice, parallel rod kinematics are frequently used for the parallel axes of motion, coupling two planes that move relative to each other. Thus, each drive is directly connected to the (end) effector (e.g., a tool carrier). As a result, unlike in serial kinematics, the drives are not subjected to the masses of all subsequent links and drives. Since all drives move simultaneously, i.e., parallel to each other, the loads are distributed more evenly across all guide elements. The resulting low moving masses enable extreme dynamics with high speeds and accelerations, while maintaining high mechanical precision.Another difference from serial mechanisms is that in parallel kinematics, the drives, especially the motors and gearboxes, remain stationary. This optimizes not only the dynamics and performance of such robots, but also their energy efficiency. Parallel kinematics are therefore often used when simple motion sequences with high repeatability and speed are required. Classic examples of parallel kinematics are hexapods and delta robots.

[0030] Fig. Figure 2 schematically shows an exemplary parallel kinematic system with 6 rod kinematics 210, i.e., legs 210 of variable length, and 12 passive joints 220. The kinematic system shown is therefore a Stewart platform, which is also referred to as a hexapod in this document. Furthermore, the bed 200 supporting the legs is also referred to as the base 200, and the movable platform 230 is also referred to as the gondola 230. Pose space (pose, pose parameters)

[0031] A pose is generally understood to be the position and orientation of an object. Pose space here refers to the space of theoretically conceivable poses, that is, the set of possible positions of a rigid body in space. A pose thus corresponds to an element of pose space. Pose space can be identified with the special Euclidean group SE(3), which consists of all rotations and translations in Euclidean space. A pose can therefore be identified with an element of the special Euclidean group SE(3). Each element of SE(3) corresponds to (exactly) one pose (and vice versa). More precisely, an element of SE(3) is identified with the pose that results when that element is applied to a given reference pose.

[0032] Poses are specified by so-called pose parameters. Specific values ​​for the pose parameters define a pose and thus correspond to a pose specification. The choice of parameterization (i.e., the pose parameters) is not unique. There are many parameterizations of the workspace, or rather of SE(3). One possible parameterization is the specification of a displacement plus a cardan angle. For example, three Cartesian coordinates (displacement) and three orientation angles (orientation) can be used as pose parameters. The displacement can be specified, for example, by the distance to a reference point in Cartesian coordinates. Cardan angles can be used to specify the orientation. Instead of cardan angles, actual Euler angles or other angles can, of course, be used as orientation angles.

[0033] However, it is also possible to specify a pose (uniquely) using other coordinates or parameters. Therefore, in the following, the term "pose parameter" generally refers to those coordinates or parameters used to specify a pose. Each of the pose parameters corresponds to a specific value. These values ​​can be combined into a vector of pose parameters that uniquely identifies a pose. Conversely, a pose is defined or specified by providing a vector of pose parameters, i.e., by specifying a value for each of the pose parameters. The pose parameters can also be "normal" world coordinates (for example, three spatial coordinates to specify the position and three angles to specify the orientation).

[0034] In general, pose specifications / pose parameters can refer to the position and orientation of the object within the workspace. The pose is then described spatially, meaning that the position and orientation are described in "real" three-dimensional space, the so-called workspace. The coordinate system belonging to the workspace, the world coordinate system, lies "unchanging" in space and is independent of the robot's movement. The pose is then also said to be given in world coordinates.

[0035] In general, pose specifications / pose parameters can refer to other, even changing, reference poses or reference bodies, or be understood as a specification of the object's position and orientation in the workspace. A pose specification in the workspace can also be understood as a specification with respect to a reference body; namely, a specification with respect to a reference body that is stationary in the workspace (in particular, the basis of the kinematics can be understood as such a reference body). Thus, each pose uniquely corresponds to the position and orientation of two solid bodies relative to each other, where one can be understood as the reference body for the other. Hexapodope

[0036] The pose of a hexapod, also referred to as the hexapod pose in this disclosure, describes the state of its respective kinematics. The state of the kinematics refers to the pose of a specific link in the kinematic chain, which is particularly relevant for a given application. Specifically, the state of the kinematics can be the pose of an end effector of the respective kinematics. The end effector, for example, refers to the last link in the kinematic chain. It is generally the component or assembly responsible for performing the actual handling task. In other words, the effector brings about the actual interaction of the robot (i.e., the kinematics) with its environment. An end effector can be, in particular, a tool, a tool carrier, a gripper, or a platform to be moved (e.g., in the case of hexapods).

[0037] Hexapod pose specifies, for example, the pose of the end effector in the workspace, or the pose of the end effector relative to the basis of the kinematics. An example of how to parameterize the pose of a hexapod (hexapod pose) is using the pose parameters (X, Y, Z, U, V, W). Here, (X, Y, Z) specify the position in Cartesian axes / coordinates and can, for example, have the unit "millimeters". (U, V, W) are angles, such as Cardan angles (360 / 360), a subtype of Euler angles, and specify the orientation or rotation.

[0038] It should also be noted that a robot is generally not capable of actually realizing every arbitrary pose, e.g., of the end effector, within pose space. Firstly, the poses that can actually be assumed are limited by the geometry of the kinematics, particularly the length of the limbs. Furthermore, a robot does not necessarily have six degrees of freedom at its joints. In this case, the space of actually achievable poses will generally not be six-dimensional, and a pose can be specified by fewer than six parameters. For example, it is conceivable that the robot's freedom of movement (e.g., the position of the effector) is restricted to a single plane. Configuration space (joint coordinates)

[0039] The configuration space refers to the space of possible configurations of the individual machine components (joints, arms, etc.). It thus has the dimension of the independent degrees of freedom of the kinematics. These degrees of freedom can be joint angles and / or, for example, lengths of length-adjustable elements (arms / legs). The individual joint coordinates (angles, lengths) can be combined into a configuration vector, i.e., a vector in the configuration space. This corresponds to a representation of the configuration space as the Cartesian product of the individual value ranges corresponding to the joint coordinates (angle ranges and / or length ranges).

[0040] In a Stewart platform (also called a hexapod), for example, the configuration space is given by the six variable leg lengths, and a configuration vector specifies a corresponding length for each of the six legs. The configuration space K HexThe length of a hexapod can therefore be understood as the Cartesian product of the individual intervals corresponding to the possible lengths of the legs. KHex=∏i=16[Lmin(i),Lmax(i)]=[Lmin(1),Lmax(1)]×….×[Lmin(6),Lmax(6)] where L min (i) and L max (i) denote the smallest or the largest possible (permissible or used) length of the i-th leg. Direct kinematics

[0041] As in Fig. As illustrated in Figure 3, direct kinematics, forward kinematics, or forward transformation deals with the question of how the pose (position and orientation) of the end effector can be determined from given joint angles and / or given lengths of length-adjustable limbs of a robot. It is the logical counterpart to indirect kinematics. Direct kinematics thus allows the calculation of the pose from given leg lengths and / or leg angles, i.e., from a given configuration vector, and corresponds to a mapping from configuration space to workspace or pose space. Indirect kinematics

[0042] Indirect kinematics, also known as inverse kinematics or reverse transformation, converts the position and orientation of the effector, given by world coordinates or pose parameters, into individual joint coordinates. As in Fig. As illustrated in Figure 3, indirect kinematics is the logical counterpart to direct kinematics and corresponds to a mapping from the workspace or pose space to the configuration space. Indirect kinematics thus allows the calculation of joint angles and / or limb lengths from a given pose. It should be noted that indirect kinematics need not be unique. In other words, a given pose can be realized by different configuration vectors. This is frequently the case, for example, with SCARA robots.

[0043] Calculating indirect kinematics is typically necessary when a specific pose (target pose) is to be assumed, and therefore a joint configuration corresponding to the target pose is required. The kinematics can then be controlled using the determined configuration vector, for example, to assume the target pose. This control may involve calculating a trajectory in configuration space based on the current configuration vector and the configuration vector corresponding to the target pose as the start and end points of said trajectory. The kinematics can thus be made to follow a trajectory whose endpoint corresponds to the target pose. Invention in the context of the prior art

[0044] To determine the pose of a hexapod during operation, the leg deflections can be measured and the direct kinematics calculated. Such pose determinations are standard practice in kinematics. Alternatively, poses can be measured directly, for example, for calibration purposes. Direct pose measurements are based on pose measurements on rigid bodies, which are performed, for example, with coordinate measuring machines, laser trackers, or laser tracers. Special kinematic systems are also suitable for direct pose measurements on rigid bodies. These systems grip the rigid body at its tool center point, determine its own pose, and assign it to the rigid body. Direct pose measurements can be used within the framework of kinematic calibration—also known as parameter identification. Alternatively, calibration based on error compensation can be performed using such data; this is then referred to as mapping procedures. Mapping can take place in the workspace or in the configuration space (joint space).

[0045] All measuring systems used to perform pose measurements for the purpose of increasing the accuracy of a hexapod during calibration, or even simply to obtain redundant pose measurements with high accuracy, should possess a reasonable level of accuracy. Generally, this accuracy is achieved either through high-precision manufacturing of the measuring systems, through calibration of the measuring system, or both. Coordinate measuring machines combine high-precision manufacturing, calibration, and the use of high-precision sensors.

[0046] Direct pose measurements can also be used in the regular operation of hexapods by integrating pose sensors into the kinematics. Pose control in the form of "visual servoing" based on these pose measurements is possible, as is the acquisition of redundant pose data for safety-critical applications such as surgical procedures, but above all for increasing accuracy.

[0047] The particular challenge with direct position measuring systems lies in the accuracy requirements. Modern hexapods, for example, exhibit a repeatability of up to 0.1 micrometers in the X-direction. The corresponding relative repeatability, i.e., based on a travel range of -50 mm to 50 mm, thus results in a relative repeatability value of 1E-6 (i.e., 10 -6This corresponds approximately to the error caused by length changes of legs / joints due to thermal effects when considering the coefficient of thermal expansion of iron or aluminum as hexapod materials within a very narrow temperature range, such as that found in a temperature-controlled laboratory or a thermal chamber. This is therefore also the threshold of accuracy, the exceeding of which suggests the use of materials with particularly low coefficients of expansion, such as Zerodur or Invar.

[0048] Progress in increasing the accuracy of hexapods is measured by the challenge of shifting the absolute accuracy towards the currently achievable repeatability. Considering the aforementioned relative repeatability, it becomes clear that common industrial measuring systems on the market, such as those consisting of cameras and markers, exhibit insufficient accuracy because too many measurement uncertainties are introduced and comprehensive calibration is not possible. Only very high-quality measuring systems, such as those using a laser tracer or based on probing by a coordinate measuring machine, offer the required accuracy, but these are unsuitable for integration as position sensors in hexapods.

[0049] The method of increasing accuracy described in this disclosure utilizes a pose sensor capable of detecting the poses of a kinematic system / hexapod. More precisely, this disclosure relates to direct pose measurement systems that can be used as assemblies within the hexapod, operate autonomously, are calibratable, and are referred to as pose sensors. This primarily addresses tasks related to increasing accuracy based on direct pose measurements using, in principle, autonomously operating, calibratable assemblies for direct pose measurement, i.e., the pose sensors presented here. Direct pose measurements by pose sensors can, for example, be used / performed outside the contour of hexapods for the calibration of the pose sensor. In this state, where the pose sensor is not installed, it is not considered an assembly of the hexapod.For (direct) pose measurement by a pose transmitter, laser tracers, visual locating devices based on cameras and optical markers, etc., are particularly suitable.

[0050] As mentioned above, direct measurement systems are inherently more accurate than kinematic systems that implement poses. High-precision measurements of hexapod accuracy can currently only be performed using coordinate measuring machines or laser tracers. Therefore, an integrated pose sensor can increase the accuracy of the hexapod by allowing it to inherit the accuracies of a direct measurement system. Pose giver

[0051] As briefly mentioned earlier, the present invention relates to arrangements, methods for using, methods for employing, and methods for calibrating pose-measuring assemblies. Such pose-measuring assemblies are also referred to in the present disclosure as "pose transmitters," "pose measuring devices," or, for example, "pose measuring systems," and it should be noted that these terms are used interchangeably.

[0052] In this context, a pose generator is defined as a system for pose measurement comprising two rigid bodies (or solids) and providing information to determine or provide the pose of one rigid body relative to the other. This pose of one solid body of the pose generator relative to the pose of the other solid body of the pose generator is referred to in this disclosure as the solid body pose. The solid body pose is thus the relative pose / position of one solid body relative to the other solid body, or the pose that the first and second solid bodies assume relative to each other. More precisely, the solid body pose corresponds to the "pose" (i.e., position and orientation) of a coordinate system attached to one solid body within a coordinate system attached to the other solid body. The pose generator is therefore configured to measure the solid body pose.

[0053] The accuracy of a hexapod is limited by its thermal stability. Using and calibrating a position sensor largely solves this problem. The position sensor can be calibrated before installation (or recalibrated after removal), simplifying calibration. In particular, the thermal stability of a hexapod can be improved by using a position sensor when performing optical measurements, as optical measurements are much less affected by temperature fluctuations than tactile measurements, which are influenced by material properties such as the coefficient of thermal expansion. Repairs to the hexapod, such as replacing legs, are also possible without loss of accuracy or recalibration.

[0054] In the embodiments presented here, the installation in the hexapod is reversible, which can offer further advantages (such as use in a different hexapod, separate recalibration in a separate state). A simple replacement of one position sensor with another of the same design is therefore also possible. This configuration, as shown in the figures, also allows for the easy replacement of the position sensor in the hexapod, whereby the hexapod can be immediately operated in a calibrated state again after the replacement.

[0055] A pose sensor like this can therefore be used to increase the accuracy of a hexapod. In particular, the thermal stability of hexapods can be significantly improved by using non-contact pose sensors, especially those that operate with optical measurements. Thus, a pose sensor can be used to increase the movement accuracy of the hexapod and / or to approach poses more precisely. Furthermore, pose control based on measured poses from a pose sensor is possible, similar to how "visual servoing" works. A pose sensor can also be used to calibrate a hexapod (e.g., to determine geometric parameters more accurately).

[0056] Furthermore, in safety-critical applications, pose transmitters can provide redundant pose data to monitor the accuracy of positioning (e.g., more precisely and / or reliably). Pose transmitters can also eliminate the need for measuring devices for the deflection of the hexapod legs, thus saving costs. Installation of the pose sensor

[0057] As mentioned previously, a float switch comprises two solid bodies, also referred to as the "first" and "second" solid / rigid body. These two solid bodies can move relative to each other and are suitable for installation in a hexapod. The installation of a float switch in a hexapod will now be explained in more detail. Generally, the float switch is installed in the hexapod by: • Attaching the first solid to the end effector and attaching the second solid to the base, or • Attaching the second solid to the end effector and attaching the first solid to the base.

[0058] The integration of the pose sensor into the hexapod / kinematics refers to the installation / attachment of one of the two solid bodies to the base and end effector of the kinematics. One of the solid bodies is attached to the base of the hexapod, and the other to the nacelle. In other words, one of the solid bodies is installed so that it moves with the end effector of the kinematics when the robot moves—there is a fixed relationship between the pose of this installed solid body and the pose of the end effector. The other solid body, however, is installed so that it does not move with the end effector of the kinematics when the robot moves—there is a fixed relationship between the pose of this installed solid body and the pose of the base of the kinematics. Installation in and removal from a hexapod should be possible without a (significant) loss of the pose sensor's accuracy.Such a fastening, in particular a reversible fastening, of a solid body to an end effector or nacelle can be achieved, for example, using screws. It should be noted, however, that the use of screws is only one example and the present invention is not limited to this.

[0059] Fig. Figure 4 shows an example of how to install / attach a float switch to a hexapod. Fig. 5 shows the one in Fig. Four hexapods shown from a different perspective, in which the fastening of the upper disc can be seen. Fig. 6 shows the one already in the Fig. 4 and Fig. The 5 hexapods shown have their components pulled apart, i.e., they are shown as exploded views.

[0060] As can be seen, the float switch comprises the two rigid bodies 402 and 405, which are already inserted / installed in the hexapod. The two rigid bodies 402 and 405 of the float switch are shown here in a disc shape and are also referred to as discs.

[0061] As can be seen, the rigid body 405 is attached to the base 403. The outer diameter of the lower disc 405 corresponds to the inner diameter of the base 403, which is designed as a perforated disc 403 (a circular disc with a circular hole). The base 403 and the lower disc 405 are detachably fitted together and screwed to each other with a round disc / perforated disc 404. The screws with heads 409, 410, and 411 are screwed into the base 403, and the screws 406, 407, and 408 are screwed into the lower disc 405. The perforated disc 403 has six round holes for the screws. The lower disc 405, attached to the base, is therefore stationary after installation (it does not move when the end effector 401 moves).

[0062] The gondola 401 of the exemplary hexapod is (also) designed as a perforated disk to which the rigid body 402 is attached. The upper disk 402 is flush with the top surface of the gondola 401. The two parts 401 and 402 are attached to each other with a circular disk, analogous to the attachment of the lower disk 405 to the base 403. This is shown in Fig. 4 is not explicitly shown, which is why we refer to the above. Fig. 5 is referred to.

[0063] As in Fig. As shown in Figure 5, the gondola 401 and the upper disc 402 of the position sensor are detachably inserted into one another and screwed together with a round / perforated disc 504. The outer screws with heads 509, 510, and 511 are screwed into the gondola 401, and the inner screws 506, 507, and 508 are screwed into the upper disc 402. After installation, the disc 402 is moved along with the gondola 401 (thus, the movement of the end effector 401 is also affected).

[0064] In the Fig. 4, Fig. 5 to Fig. In the explicit example presented, the underside of the mounted lower disk 405 is flush with the underside of the base 403. This facilitates the precise and reproducible anchoring of the lower disk 405 to a surface on which the hexapod is placed. This allows the pose of the lower disk 405, and thus also the pose of the upper disk 402, to be referenced to a world coordinate system, provided that the poses of the two disks relative to each other were accurately measured during the calibration of the pose generator.

[0065] The flush finish of the lower disc with the underside of the base also offers the advantage that the lower disc can be anchored to the base surface via a kinematic interface, and the platform's pose can be determined relative to a world coordinate system. This flush finish also proves advantageous in increasing thermal stability. To further enhance thermal stability, one or two discs can be manufactured from a material with an exceptionally low coefficient of thermal expansion. This ensures, in particular, that the relative positions of the LEDs or light sensors, as in the exemplary embodiment, are thermally stabilized. Alternatively, depending on the application, it may be structurally advantageous to raise the lower disc above the level of the base, for example, with a plinth, in order to reduce the measurement volume and adapt it to the working space.Questions regarding the depth of field of optical sensors can play a role here. Kinematic interface

[0066] A kinematic interface is a device for the rigid and detachable connection of two rigid bodies, whereby the two rigid bodies can be fixed relative to each other in the same pose in a precisely reproducible and deterministic manner by means of this interface. A kinematic interface generally comprises at least two coordinated parts, which are referred to as interface parts in the present disclosure. Each of the two rigid bodies to be connected has one of the interface parts of the kinematic interface, and the precisely reproducible connection is effected by the kinematic interface. More precisely, one of the interface parts is attached to each of the two rigid bodies. In this respect, reference is made to application DE 102020124136, the contents of which are hereby incorporated into this application.

[0067] The interface components are designed to fit together, following the "plug and socket" concept in electrical engineering. Crucially, the plug and socket can be coupled in precisely one way, thus eliminating symmetries like those found in a Europlug. Ideally, the interface components should be easily detachable and reattached. Self-locking mechanisms, snap-in connections, magnetic retention, and similar features are advantageous.

[0068] Interface components are functional features of rigid bodies. Kinematic interfaces serve to create a rigid connection between two rigid bodies, thereby forming a new rigid body. For this to occur, each of the two rigid bodies must have an interface component. Kinematic interfaces thus provide a connection option for rigid bodies. They are designed so that the connection is detachable, as deterministic and reproducible as possible, and is rigid and stiff. The kinematic interfaces should be dimensionally stable under forces and moments. Kinematic coupling - triple-groove kinematic coupling

[0069] A kinematic coupling is an example of a kinematic interface. A kinematic interface is, for example, a large-scale groove portion of a three-groove kinematic coupling. Such a three-groove kinematic coupling is found in Fig. Figure 7 is shown schematically. As can be seen, it consists of two parts, namely a grooved part 701, which has three grooves 701a, 701b, 701c, and a ball part 702, which has three balls 702a, 702b, 702c.

[0070] Each of the three spheres of the spherical part has point contact at two points in a groove of the grooved part when positively engaged, resulting in contact at six points. This creates static determinacy. Therefore, there is exactly one way to insert the spherical part (the "plug") into the grooved part (the "socket"). It should also be noted that the representation of the grooved part 415, as well as the kinematic interface in Fig. 7 is only schematic, and the three grooves do not meet at 120 degrees, for example, to break the symmetry and thus achieve a uniqueness of the attachment with the kinematic interface.

[0071] Without loss of generality, the interface parts of the kinematic interface in the present disclosure are intuitively referred to as the sphere part and groove part, respectively, corresponding to the interface parts in the three-groove kinematic coupling. Making the reference coordinate system tangible

[0072] If, for example, no reference artifact was used when acquiring the calibration data of the pose sensor, the pose sensor's reference coordinate system may not be readily available. Generally, the pose sensor's reference coordinate system can be made available by attaching a reference artifact to one of the pose sensor's solid components (especially the moving body) using a kinematic interface. One interface part of the kinematic interface is located on the pose sensor's solid component, and the other interface part is located on the reference artifact. The reference artifact can thus be uniquely attached to the pose sensor's solid component. The pose sensor's reference coordinate system can then be made available using the reference artifact attached to the pose sensor's solid component and, for example, a coordinate measuring machine.Only a few measurements in suitable positions are required for this.

[0073] The reference coordinate system is defined based on the collected calibration data. The accuracy achieved regarding the position and location of the defined reference coordinate system, as well as the accuracy of a calibration, depends on precise pose measurements during calibration. Precise pose measurements require highly accurate pose markers on artifacts, such as the cuboids mentioned below, whose pose can be reliably determined metrologically. Therefore, it is recommended to use reference artifacts during calibration that are specifically designed to enable highly accurate metrological determination of coordinate systems.

[0074] A coordinate system is considered tangible here if its position and orientation with respect to a spatial registration are defined by a coordinate transformation. For example, the reference coordinate system of a hexapod is tangible if its origin coordinates and its orientation relative to the coordinate system of a calibration artifact can be specified, if the latter is connected to the gondola via a kinematic interface, and if the hexapod has assumed a designated, distinguished pose, which is typically its initialization pose.

[0075] In other words, a tangible reference coordinate system of a robot is a coordinate system that defines the robot's directions of movement and pivot point, and whose origin and axe orientation can be determined by measuring the pose of a rigid body, such as one attached to the robot's end effector. Accordingly, the reference coordinate system of the pose sensor is considered tangible or intangible if it can be determined / captured metrologically by measuring the pose of a suitable rigid body attached to the pose sensor, or if it cannot be determined / captured metrologically.

[0076] For example, if a grooved section is located on the second / moving solid body of the pose sensor, and a reference artifact is placed on this grooved section, which has a spherical component belonging to the grooved section, then the coordinate system of the reference artifact is rigidly connected to the second / moving solid body. The position of the pose sensor's reference coordinate system (i.e., the second / moving solid body) can be specified with high precision in the pose sensor's initialization pose using this coordinate system. The required coordinate transformation consists of the combination of two coordinate transformations. The first coordinate transformation is between the coordinate system of the calibration artifact and the coordinate system of the reference artifact. The second coordinate transformation is between the reference coordinate system of the pose sensor and the coordinate system of the calibration artifact.Both coordinate transformations must be known to the robot's controller. The coordinate transformation between the coordinate system of the calibration artifact and the coordinate system of the reference artifact is independent of the pose transmitter.

[0077] The term "reference artifact" here refers to an interface-bearing artifact that has a pose marking to which a spatial registration is attached. An interface-bearing artifact is a rigid body equipped with an interface component. The reference artifact thus has a defined, attached coordinate system, which can be referred to as the coordinate system of the reference artifact. Reference artifacts can be cuboids whose outer surfaces represent a pose marking. Optionally, the cuboids can be suitable for moving a mass placed upon them. In particular, plate-shaped cuboids can have markings on their upper surface relating to their spatial marking, in the direction of the outer surfaces. In a preferred embodiment, the pose marking consists of stop surfaces on the plate-shaped cuboids.

[0078] The calibration artifact is, in principle, an arbitrarily selected reference artifact used to relate the first coordinate systems of all reference artifacts. The position of the first coordinate system of a reference artifact relative to the position of the first coordinate system of the calibration artifact, represented as a datum consisting of six real numbers describing a coordinate transformation, is a datum unique to each reference artifact. Therefore, if such a registered reference artifact is attached to the grooved portion of a three-groove kinematic coupling, a coordinate system can be defined based on its position marking, the position of which is known relative to the calibration artifact. Thus, the position of the reference coordinate system can also be determined if its position relative to the calibration artifact is known.

[0079] Furthermore, the term "pose marker" refers to a marking on a rigid body that allows a pose in space to be measured relative to a reference pose of that rigid body. Pose markers make it possible to attach a coordinate system to a rigid body, the origin and orientation of which can be determined metrologically using the pose marker. For example, three non-collinearly arranged spheres, rigidly attached to the rigid body, are suitable as pose markers. The position of an attached coordinate system can be determined by the coordinates of their centers. Another pose marker can be implemented using a fixed cube, since a coordinate system can be established on just three pairs of non-parallel planes.

[0080] A spatial registration is a rule that assigns a (attached) coordinate system to a rigid body based on a pose marker attached to that rigid body. This rule thus uniquely defines, for each type of pose marker used, how the attached coordinate system is positioned relative to the pose marker. Using a spatial registration, a pose can be assigned to a rigid body that has a pose marker, provided a reference coordinate system exists in which its pose can be defined.

[0081] For example, this is in Fig. The grooved part 415 shown is such a kinematic interface part, with the help of which the coordinate system of the pose transmitter can be made tangible. For this purpose, a reference artifact with a corresponding spherical part is attached to the grooved part 415.

[0082] The attached reference artifact serves as a surface feature to assign a metrologically measurable reference coordinate system to the pose transmitter. By equipping the upper disk 402 (more generally, the solid body that moves along with it after installation in the hexapod) with a kinematic interface component, the position of the pose transmitter's reference coordinate system can be defined and made tangible. This tangible reference coordinate system is achieved by attaching a reference artifact to the kinematic interface component of the upper disk. The position and orientation of the pose transmitter's reference coordinate system (relative to an external coordinate system) can then be metrologically determined at any time by measuring the pose of the reference artifact.

[0083] In the Fig. In the example shown, only the upper disc 402 is equipped with a kinematic interface; the lower disc 405, however, is not equipped with any elements of a kinematic interface. If the lower disc 405 is to be used to connect to a world coordinate system, the lower disc can be attached to the base with a kinematic interface, and the hexapod can also be attached to the base (e.g., also with a kinematic interface). The lower disc of the pose generator and the base of the hexapod do not necessarily have to be attached to each other (e.g., with a kinematic interface), although this is also possible. If no direct attachment is possible, the calibrated pose generator, in its calibrated state, enforces a pose that is oriented to the base. It is still initially an "absolute pose."From this absolute pose, a pose relative to the substrate can be specified if the rototranslation between the lower and upper disks in the calibrated zero pose has been determined during the calibration measurements.

[0084] The base of the hexapod can then be attached and, after a rotating alignment that aligns the hexapod's movement directions as closely as possible with the reference coordinate system of the float transmitter, secured to the substrate. The idea is that the float transmitter is an autonomously operating system that is only moved by the hexapod. High-precision alignment according to 901 to 906 is compromised if the base and lower disc are not attached to each other as intended. However, the hexapod is rotated before being attached to the substrate so that 901 to 906 are adhered to as closely as possible. The function of assuming calibrated poses still works. If the lower disc is oriented to the substrate via the kinematic interface, the base and lower disc can still be screwed together, and the base can then be additionally secured to the substrate in some way. Measurement of the solid state pose

[0085] In general, the pose sensor comprises means for measuring the solid pose that the first and second solids assume relative to each other. These "means for measuring a solid pose" are attached to the two solids and will be described in more detail later. The pose sensor can thus provide data through measurements that make it possible to determine the solid pose. In particular, a pose sensor is capable of measuring poses in the installed and / or removed state. As already mentioned, the accuracy in the installed state is (practically) identical to that in the removed state. An increase in the accuracy of the hexapod by installing the pose sensor is based on the fact that the pose of the hexapod is either considered to be the pose of the (calibrated, high-precision) pose sensor or the pose of the hexapod is determined based on the pose of the (calibrated, high-precision) pose sensor.

[0086] The pose generator can further include devices / means for generating and evaluating the signals, means for controlling the signals, means for outputting the determined poses, means / devices for signal transmission, and power supply. It is important that the pose generator comprises the upper and lower disks (i.e., the two solid bodies) on which means for measuring the solid body pose are mounted.

[0087] Signal processing, for example, can optionally be performed in the hexapod's controller or in a separate controller for the float switch, which is irrelevant to the float switch's function. The float switch's controller, the separate hexapod controller (if present), and / or the sensors for measuring the solid body float can, for example, be capable of exchanging information via wireless communication.

[0088] The means of performing the calculations can therefore include, in particular, circuits, calculators, computers, and the like. If the pose generator includes means for determining the solid pose based on the solid pose measurements, it can, for example, automatically measure and output the pose of the moving disk relative to its zero position. If the pose generator is calibratable, it can output (e.g., optionally) the uncalibrated pose, the calibrated pose, the uncalibrated virtual configuration vector, and / or the calibrated virtual configuration vector (virtual configuration vectors are explained further below).

[0089] Such a 1300 controller of the pose generator is in Fig. Figure 13 illustrates this. The controller or control device 1300 can be implemented in any hardware. For example, it can run as software on a programmable processor 1310. Alternatively, a specialized hardware unit can represent the control device 1300. A combination of specialized hardware and programmable hardware is possible. In an advantageous embodiment, the control device 1300 is distributed, and its functions can be performed on one or more hardware units 1310, 1320 (e.g., processor(s), circuit(s), in particular integrated circuit(s)). In particular, the functions of controlling the serial or parallel kinematics can be performed by a local controller, and the calculation of the interpolation and / or the division of space into spatial units can be performed in an external device, such as a computer. Other configurations are possible.The controller can also include a memory 1350 on which the result of a calibration measurement of the pose sensor can be stored (e.g. the configuration vector correction F(K). M )) or the measured support points). The Controller 1300 does not need to be permanently attached to one of the two solid bodies of the float transmitter; it is also possible for it to exchange data wirelessly with the sensors / markers attached to the solid bodies.

[0090] It should be noted at this point that in the present disclosure, the term "measured / uncalibrated solid pose" refers to a pose that is determined / calculated (directly) from the results of the distance / direction measurements of the sensors attached to the pose transmitter. This measured / uncalibrated solid pose can be inaccurate because, for example, (i) the "transmitters" and / or "receivers" are not attached exactly at the intended positions and / or orientations, (ii) the cameras / sensors may have optical distortions that can / must be compensated for by corrective parameters, and (iii) the cameras are not aligned exactly in the respective intended direction (i.e., 2 angles / degrees of freedom).Since the calculation of the measured / uncalibrated solid pose based on the results of the distance / direction measurements is also based on the positions on the solids of the pose transmitter where the transmitters / receivers are attached, this attachment error is transferred to the measured / uncalibrated solid pose P calculated in this way. M Furthermore, the term calibrated solid float refers to a solid float that results from adding a float correction to an uncalibrated solid float, which is explained further below in connection with the calibration of the float transmitter. Means for measuring solid state pose

[0091] As already mentioned, the term poser encompasses two rigid bodies, e.g., those relating to Fig. 5 to Fig. Six previously described disk-shaped solid bodies, designated as upper disk 402 and lower disk 405. These two rigid bodies, as assemblies, carry transmitters and receivers of signals, which are also referred to here as means for measuring the solid-state pose.

[0092] In other words, the means for measuring a solid-state pose consist of one or more "transmitters" and one or more "receivers." These transmitters and receivers are attached to the two solids. The term "transmitter" is used here in a general sense, as it can refer to both active transmitters (e.g., LEDs) and passive transmitters, such as markers (especially fiducials / April tags). The "receivers" are the sensors / detectors capable of receiving the signal from the transmitter or recognizing the markers.

[0093] In particular, each receiver may be able to: (i) determine the distance of a particular transmitter / marker from that receiver; and / or (ii) determine the direction in which a particular transmitter / marker is located from that receiver. In general, it may be possible for a receiver to determine the distance and / or direction of only one transmitter, of several transmitters, or of all transmitters.

[0094] Since the transmitters and receivers are attached to the solids at known / predetermined positions, the solid's pose can be determined from a sufficient amount of distance and / or direction information. More generally, the receivers are measuring devices capable of determining the distance / direction between a predetermined / known point on the first solid and a predetermined / known point on the second solid. Furthermore, the distance and / or direction information provided by all receivers should be sufficient to uniquely determine the solid's pose. Means for measuring a solid pose

[0095] The pose measurement of the solid body can be performed, for example, using light sensors / LEDs. More precisely, one of the two solid bodies can have a light sensor and the other a LED. These light sensors and LEDs are arranged on the solid bodies in such a way that the light sensors can locate the LEDs in their installed state, allowing the relative position of the two solid bodies to be determined (i.e., the solid body pose can be determined).

[0096] Such an arrangement of light sensors and light-emitting diodes is also used in the exemplary embodiment of the Fig. 4, Fig. 5 to Fig. Figure 6 illustrates this. As can be seen, three light sensors 450, which can also be implemented as small cameras 450, are located on the lower solid 405, and three light-emitting diodes 560 are located on the upper solid 402. More precisely, when installed, the light sensors 450 are located on the top surface of the lower disk 405 (i.e., on the side of the lower disk facing the upper disk). The light-emitting diodes 560 are located on the underside of the upper disk 402 (i.e., on the side of the upper disk facing the lower disk). The sensors 450 can thus locate the light-emitting diodes 560. Therefore, the sensors 450 can locate the upper disk 402 and determine its Cartesian coordinates. More precisely, the positions of the light-emitting diodes relative to the light sensors can be determined from the measurements of the sensors 450. From the positions of the light-emitting diodes, the pose of the solid can then be determined.

[0097] For example, the sensors could be pinhole cameras, with each camera measuring the direction in which each LED is located relative to the camera. Thus, for instance, two angles are measured when a specific LED is targeted. The three cameras can then work together to determine the relative Cartesian coordinates (distances and angles / directions) of the LEDs to the cameras. Theoretically, two cameras are sufficient, but with three cameras, the redundancy of the measurement results can be used to determine the accuracy of the LED position determination. Theoretically, distance measurement sensors, such as laser interferometry, could also be used, but this is too expensive for many practical applications.

[0098] Instead of using LEDs, optical markers (e.g., 3D scan markers), April tags, or similar devices can also be used, and the cameras can be replaced by sensors operating on the pinhole camera principle. Advantageously, only one of the three LEDs is illuminated at any given time, so that the LEDs are targeted sequentially. In the exemplary embodiment, the pose sensor operates optically. Other methods of pose measurement are also conceivable, such as those based on time-of-flight measurements of, for example, ultrasonic pulses, or interferometric measurements. Pose sensors that do not operate without contact are not excluded according to the invention. Likewise, the number of light sensors or LEDs is not limited to three.

[0099] In this example, the upper disc acts as a signal transmitter, and the lower disc as a signal receiver; the reverse assignment is possible. Designs are also conceivable in which one or more discs function both as signal transmitters and as signal sensors, which also includes identical designs for the upper and lower discs. Determination of the solid state pose

[0100] In principle, the pose generator measures (or outputs) the pose of the first solid relative to the second solid, or the pose of a rigid body relative to a defined initial pose of that rigid body. More precisely, based on the measurement results of the means used to measure the solid's pose, the relative pose and / or the absolute pose can be determined. If the pose generator includes a corresponding circuit or circuits (e.g., the [insert circuit here]), these circuits can determine the relative pose and / or absolute pose based on the measurement results, and the pose generator can output the relative / absolute pose, for example, via a suitable interface. The output can be sent to a user and / or the hexapod's controller. The output relative / absolute pose may (but does not have to) be calibrated. The terms relative pose and absolute pose are explained below. Relative pose

[0101] If the pose of the first solid is denoted by P1 and the pose of the second solid by P2, both with respect to the same freely selectable coordinate system, then the relative pose P 12 of the second solid in relation to the first solid given by P12=P2∘P1−1, where P1−1 The inverse element of SE(3) is P1 (poses are thus understood as elements of SE(3), whereby in the present disclosure elements of SE(3) are also referred to as rototranslations). It should be noted that the poses P1 and P2 are not determined by the pose-giver (and generally cannot be determined by the pose-giver alone). The relative pose P 12 is independent of any joint movement of the two solid bodies.

[0102] More precisely, pose P1 is the pose that a coordinate system attached to the first solid body assumes in an external coordinate system. The rototranslation corresponding to pose P1 transforms this external coordinate system into the coordinate system attached to the first solid body. Pose P2 is the pose that a coordinate system attached to the second solid body assumes in an external coordinate system. The rototranslation corresponding to pose P2 transforms the external coordinate system into the coordinate system attached to the second solid body. In principle, various external coordinate systems can be used, but this generally only complicates the calculations / representation unnecessarily and is therefore not discussed in this disclosure. The relative pose P 12 The transformation is what rotates / moves P1 to P2. Absolute pose

[0103] To simplify notation, the term "absolute pose" can be introduced / used here. For this purpose, the relative poses P can be selected. 12 Between the first and second solid, one can be distinguished, i.e., selected. This establishes a canonical configuration, here as the relative null pose P. null12 designated.

[0104] If the second solid body is the solid body that moves along with the hexapod after installation, then the relative zero position P can be determined. null12 e.g. written as Pnull12=Pnull2∘P1−1, where the pose P null2 is an excellent pose of the second solid. The poses P null2 , P2 and P1 of the relative pose P 12 and the relative zero pose P null12 are both referenced to the same coordinate system. The relative zero pose P null12It can also be determined by a predetermined configuration vector after installation in the hexapod. The relative zero pose P null12 This is the solid-state pose that the pose generator measures when the predetermined configuration vector is commanded to the hexapod (e.g., to assume the zero position).

[0105] Using the relative zero pose P null12 Can a pose be assigned to the solid P2, which is also referred to here as absolute pose P? abs is called. More precisely, the absolute position P abs based on the relative zero pose P null12 defined by Pabs=P12∘Pnull12−1= (P2∘P1−1)∘(Pnull2∘P1−1)−1= P2∘P1−1∘P1∘Pnull2−1= P2∘Pnull2−1

[0106] In particular, pose P null2 of the second solid, since for P2 = P null2 The identical mapping (here also referred to as the zero pose) for the absolute pose yields: P abs= id. Determining the absolute pose of one of the two rigid bodies (usually the moving rigid body) thus corresponds to determining the pose of that rigid body relative to a distinguished initial pose of that rigid body. It should also be noted that the transformations into equations to P abs This also demonstrates, in particular, that pose P1 actually plays no role for P abs plays, which is one reason for the introduction / use of P abs as well as their designation as an "absolute" pose.

[0107] Hexapods often use "absolute poses" themselves. The coordinate system of the gondola in its zero state then serves as the reference coordinate system for specifying the hexapod's pose. Such "absolute" poses are frequently used, particularly in application engineering. The pose of a hexapod—that is, its platform—is referenced to a pose of the platform in its zero position. This eliminates the need to include two different poses, P1 and P2, in the descriptions. The absolute pose is aligned to the pivot point and the directions of movement during calibration measurements, i.e., to a reference coordinate system. Therefore, the absolute pose is referenced to a coordinate system defined during calibration, which might be, for example, a reference coordinate system.In the case of a hexapod, this is determined by surface features of the end effector or, in the explicit example, by surface features of the upper disk (in particular by a reference artifact attached to the end effector / moving disk with a kinematic interface), and determines both the directions of movement and the pivot point.

[0108] Regarding the use of the float switch in the hexapod, P1 and P2 no longer need to be explicitly carried. However, floats P1 and P2 can come into play if the float switch is sold as a separate product or if it is calibrated. P1 and P2 then appear (albeit somewhat "hidden") during the calibration of the float switch. Afterwards, the float switch's absolute pose, as described here, can be used.

[0109] The term "absolute pose" therefore does not refer to different coordinate systems or poses of other rigid bodies, but to an initial pose of the rigid body. Only a reference coordinate system needs to be specified, which can be done during the calibration of the pose generator. Context is provided by the fixed installation of the hexapod; the movable platform of the hexapod is inextricably linked to its base, and the two disks of the pose generator are fixed in a defined position. This simplifies the notation. When the pose generator outputs a pose, this pose does not need to refer to the lower disk (the stationary rigid body), since its pose (or the rotational translation between the lower and upper disks) does not need to be recorded during calibration.A measuring system (observer) mounted on the lower disk can measure defined changes in the position of the upper disk relative to its defined initial position, even without knowing its own position. Only a reference coordinate system is required.

[0110] However, when calibrating the pose transmitter, a defined rototranslation between the two disks in the zero position of the pose transmitter is required; it is the same as after installation in the hexapod, but this rototranslation does not need to be known quantitatively (in numbers) or particularly precisely.

[0111] This pose of the second / upper / moving solid (here P2) is, however, reduced to a zero pose (P null2 ) related, i.e., measured and provided, is the rototranslation (P abs ) between the zero position (P null2) of the moving / moving disk to the current position (P2) of the moving / moving disk. A pose of one body relative to the pose of the other body can be calculated from this if, during the calibration of the pose transmitter, the pose of one body relative to the other body in the zero position is measured using external measuring instruments.

[0112] A pose of the upper movable solid is determined (e.g., by a dedicated circuit in the pose sensor) with respect to a coordinate system / reference pose that is not necessarily numerically precise. This determination is repeatable, however, because the pose is (uniquely) defined by the relative position of the two disks to each other, and the pose can therefore always be obtained, especially since pose markers and sensors are permanently attached to the disks. The measured pose is compared with the initial pose obtained during calibration based on the same coordinate system / reference pose. Using "absolute poses" thus allows for various simplifications.

[0113] If details regarding the later installation of the float switch are known, the initial position of the absolute float can be selected during calibration so that it corresponds (almost) exactly to the installation position of the float switch (i.e., is as identical as possible). The "initial position" of the absolute float is determined by the choice of P null2 defines and designates the solid body pose in which the pose giver P abs = outputs id. The installation position refers to the solid-state pose that the two solids assume relative to each other when the pose sensor is installed and the hexapod is in its zero pose. The pose sensor can therefore be calibrated accordingly (e.g., by appropriately selecting P). null2 ) that it is relatively / “as” as possible exactly P abs = outputs id when the solid pose P 12This corresponds to the relative zero position of the nacelle to the base of the hexapod into which the pose generator will later be installed. The details for the subsequent installation include, in particular, the approximate installation positions and orientations of the solids in the base and nacelle of the hexapod. If the zero position is commanded after installation in this hexapod, the pose generator will already output a pose that is close to the zero position (which is convenient). If the pose generator is to be sold separately, it can be helpful to measure the rotational translation of the two solids (discs) relative to each other during calibration and provide this measurement with the pose generator. Adjusting the pose sensor during installation in the hexapod

[0114] If the absolute pose is determined (e.g., by a circuit in the pose sensor), then—as described above—the pose of the upper disk is determined relative to its pose in a defined zero position. This property of absolute measurement means that the installation of the upper and / or lower disk in the hexapod does not require the highest precision. Inaccuracies during installation primarily lead only to larger pose deviations between a calibrated and an uncalibrated hexapod (i.e., between the output of the pose sensor and a pose calculated based on the direct kinematics of the hexapod). Therefore, the installed pose sensor generally does not provide an exact P. abs = id off when the null pose is commanded, even if P null2e.g., the calibration was chosen accordingly, which does not necessarily pose a problem for accuracy but only means a (in the "practical" case described above, small) redefinition of the zero pose of the hexapod.

[0115] Tasks such as those encountered in the kinematic calibration of kinematic systems, namely the determination of unknown parameters, are generally no longer necessary. Determining the position of mounting holes on the discs, the precise position of sensors or cameras on the discs, and their orientation are already implicitly solved during the calibration process. The identification of the poses of the lower and upper discs with respect to a reference coordinate system is eliminated. This makes the use of the pose sensor flexible and easy to apply. Hexapodope

[0116] In general, the accuracy of the hexapod can be increased by adjusting the measurement results (or the (possibly calibrated) output pose P). AUS ) of the pose generator to determine the pose actually assumed by the hexapod at a given time and / or a target configuration vector at which the hexapod assumes a given target pose.

[0117] According to the present disclosure, for example, the hexapodpose can be defined as pose P AUS The built-in (possibly calibrated) pose transmitter is used. If the hexapod is then to be operated in a calibrated mode, the pose P output by the calibrated pose transmitter (possibly calibrated) applies. AUSas the pose of the hexapod. The precision of the pose-giver is thus transferred directly and (practically) without distortion to the hexapod, which is thereby also "calibrated". In addition to the accuracy of the pose-giver itself, the concept described in the present revelation, according to which the pose of the hexapod is ultimately equated with the pose of the pose-giver, also contributes to an increase in accuracy. Calibration of the pose sensor

[0118] If a direct float measuring system, i.e., a float transmitter, is installed in a hexapod, and the goal is to meet specific accuracy requirements for the hexapod, then this direct float measuring system should inherently possess a corresponding level of accuracy.

[0119] A first approach to implementing direct pose measurement in a hexapod could consist of individual cameras and optical markers mounted separately on the gondola and / or base. Such an arrangement must be calibrated, as the precise coordinates of the markers, the camera(s), and their optical parameters (distortions) relative to each other and to a reference coordinate system must be known. This information is expressed as manufacturing and optical imaging "parameters," which must either be implemented with high precision or determined. The problem of high-precision manufacturing, or alternatively, the precise determination of manufacturing parameters, arises analogously in connection with the so-called kinematic calibration (parameter identification) in hexapod calibration. High-precision manufacturing is expensive, and so-called parameter identification typically yields unusable results.The problem of parameter identification for kinematic parameters in hexapods thus reappears in this case with direct pose measurement systems, if these are to be calibrated in this way. While precise knowledge of the parameters of the hexapod itself no longer plays a role, the parameters of the setup for direct pose measurement—i.e., the position vectors of the markers, sensors, etc.—become crucial. Therefore, nothing would be gained here; only a shifting of the parameter identification problem from one location to another would be achieved.

[0120] Such high-precision manufacturing of a pose sensor to achieve high accuracy is expensive, but not excluded within the scope of the invention. A more cost-effective realization of the required accuracy can be achieved by calibrating the pose sensor if the expense of high-precision manufacturing is to be avoided.

[0121] If a float sensor were permanently installed in the hexapod, and could only be calibrated in this installed state, then the hexapod would be unavailable for use during recalibration. Furthermore, hexapods could not be retrofitted with a float sensor to achieve a calibrated state, and the methods for moving the float sensor into the required calibration poses during a calibration process, as well as the methods for measuring deviations in these poses, would be limited.

[0122] After calibration, the (calibrated) pose measurement system (e.g., a corresponding circuit of the pose measurement system) is set up to determine the position of a (or each) assumed solid body pose: • to assign (and output) a calibrated configuration vector, and / or • to assign (and output) a calibrated solid body pose.

[0123] The associated calibrated configuration vector or the associated calibrated solid pose corresponds to the assumed solid pose. The present disclosure provides a corresponding calibration method for this purpose. Here, the pose transmitter is calibrated to increase its accuracy, i.e., the accuracy of the measurement of the solid pose or the output relative and / or absolute pose.

[0124] As explained in more detail below, the following steps, for example, can be performed for calibration: a) The pose generator measures the assumed solid pose, whereby it displays the measured / uncalibrated solid pose P M receives. P M This could be, for example, an uncalibrated absolute float, as described above. It should be noted that the following steps b) and c) can be omitted if the float provider only uses the measured / uncalibrated solid float P. Moutputs as the assigned pose. b) The measured / uncalibrated configuration vector K M is achieved using a virtual indirect kinematics IK v () based on the uncalibrated solid pose P M calculated, i.e., K M =IK v (P M ). K M This corresponds to the uncalibrated solid pose P M It should be noted that the following step c) can be omitted if the pose generator only provides the measured / uncalibrated configuration vector K. M outputs as an assigned configuration vector (or possibly also the measured / uncalibrated solid pose P) M outputs). For example, the indirect / direct kinematics of the hexapod can be used as indirect / direct virtual kinematics. c) A configuration vector correction F(K) is applied. M ) to the measured / uncalibrated configuration vector K Madded, from which the assigned / output configuration vector correction K is derived. AUS results in K AUS =K M +F(K M The configuration vector correction F(x) is a mapping that assigns a correction to each configuration vector and is determined during the calibration of the pose transmitter and then stored (for example, in a non-volatile memory of the pose transmitter). d) The assigned / output / calibrated solid pose P AUS This can then be done using virtual direct kinematics DK. v () based on the calibrated configuration vector correction K M calculate, i.e., P AUS = DK v (P AUS The virtual direct kinematics DK v () is the inverse mapping of the virtual indirect kinematics IK. v ().

[0125] Furthermore, it should be noted that in step c) a mapping procedure in the configuration space was used to calibrate the pose transmitter. In other words, the configuration vector correction F(K) M ) a function in the configuration space of the virtual hexapod.

[0126] If a mapping method is used in the workspace, a mapping F*(x) is determined (and stored) during calibration, which is assigned to each measured / uncalibrated P M Solid body pose, a pose correction F*(P M ) assigns. The calibrated solid pose P AUS which corresponds to the assumed solid body pose and is possibly output, then results according to P AUS = P M +F*(P M ). The calibrated configuration vector P AUS which corresponds to the assumed solid pose and is output if necessary, then results, for example, from the calibrated solid pose P determined in this way. AUSand the indirect virtual one according to K AUS = IK v (P Aus ). The indirect / direct kinematics of the hexapod can also be used here, for example, as indirect / direct virtual kinematics. External measurement of the solid pose to determine the calibrated solid pose

[0127] External measurements of the solid float are performed to calibrate the float sensor. "External" here refers to a measurement by a different external measuring system than the float sensor itself. This external system should be able to measure the solid float with higher accuracy than the float sensor.

[0128] During the calibration of a pose generator, it is (sequentially) brought into a plurality of poses. This plurality of solid poses is explained in more detail below and is referred to in this disclosure as the pose measurement set {P SOLL (i)} is denoted.

[0129] In other words, during calibration, the first and second solid bodies are moved into a plurality of solid body poses. The (plurality of) solid body poses into which the second solid body is moved relative to the first solid body during calibration is denoted here by {P SOLL (i)}, denoted. Here, “i” is an index used to identify a single solid pose among the plurality of solid poses in the pose measurement set {P SOLL (i)} to identify (e.g.: 1 ≤ i ≤ N, where N is the number of solid poses in the pose measurement set). Should the pose measurement set {P} be applied to individual poses within the pose measurement set? SOLL (i)} When referenced, the term P is usually used SOLL (i) is used (i.e., without curly brackets).

[0130] These solid-state poses can be defined as the poses measured by the pose generator or as the externally measured poses. In other words, the calibration for each pose P SOLL(i) the Posen measurement quantity {P SOLL (i)}, such that the first solid and the second solid are moved relative to each other, such that: • the pose giver an uncalibrated pose P M outputs, which the pose P SoLL (i) corresponds (e.g. such that P M = P SoLL (i) is output); or • the external measuring system a calibrated pose P EXTERN measures the pose P SoLL (i) corresponds (e.g. such that P EXTERN = P SoLL (i) is output).

[0131] The solid body poses P measured by the pose transmitter during calibration M (i) are the respective externally measured poses P EXTERN (i) assigned, which here e.g. are measured by a coordinate measuring machine and are designated as P EXTERN (i) are designated. More precisely, each of a plurality of solid body poses P measured by the pose transmitter is designated. M(i) assigned the corresponding externally measured solid pose P EXTERN (i). Each of the solid poses P M (i) is thus an externally measured solid pose P EXTERN (i) assigned. These externally measured solid poses P EXTERN (i) are also referred to in the present disclosure as a calibrated solid pose.

[0132] The poses P measured by the coordinate measuring machine (or coordinate measuring machine (CMM)) EXTERN(i) can in turn be poses that refer to the zero pose of the pose generator. In particular, the measured poses can refer to the initial pose of the artifact in the zero pose of the pose generator. These measured poses must then be referenced to the datum coordinate system of the pose generator. This is possible, firstly, by having the CMM take measurements of surface features of the pose generator; measurements on the grooved portion of a kinematic interface are suitable. A second possibility is to reference the poses defined by the pose generator to the datum coordinate system of the pose generator by means of a least squares calculation with those measured by the coordinate measuring machine.

[0133] It is noted that calibrating pose sensors may be less complex than calibrating hexapods, since the upper disc of a non-contact, direct optical pose measuring system can be rotated and moved without resistance or friction, for example, using a CNC-controlled positioning coordinate measuring machine and a series-mounted, high-precision, also CNC-controlled 3-DOF kinematic system to orient the disc. The direct pose measuring system can also be calibrated according to the invention even if it is not installed in a hexapod. Movement by hexapods

[0134] In Fig. Figure 8 shows an example of the calibration of a pose sensor using a hexapod. The pose sensor, consisting of the two disks 804 and 805, is installed in a hexapod 803 for calibration. An artifact 802 is attached to the pose sensor. This artifact is a rigid body designed to enable precise pose measurements. It can be a highly precisely manufactured cuboid. Cuboids can be easily probed by a coordinate measuring machine, and the orientation of the cuboid and its position vector can thus be determined in the coordinate system of the coordinate measuring machine. Figure 801 shows the probe head of the coordinate measuring machine. The small black circles represent small spheres with which objects can be probed. Probing means measuring the relevant sphere coordinates at the moment of its contact with a solid body.

[0135] In Fig. 8. A hexapod acts as the mover for the pose transmitter, guiding it sequentially through the majority of poses. The procedure presented here, which uses a hexapod to move the pose transmitter and a coordinate measuring machine for external pose measurement, exemplifies the calibration process. In practice, a faster method would be sought, for example, using a motorized coordinate measuring machine with a 3DOF kinematic system capable of performing rotary movements and offering high accuracy to move and simultaneously measure the pose transmitter, thus eliminating the need for the hexapod. Movement achieved through highly precise serial kinematics consisting of a pivoting mechanism and a coordinate measuring machine

[0136] An elegant method for obtaining calibration data involves using a high-precision kinematic system to move the upper disk into defined poses. During calibration, the pose generator (e.g., exactly one of the two solid bodies, while the other remains stationary) is guided by this high-precision kinematic system. The poses defined by the high-precision kinematic system can then be used as the calibrated solid body poses.

[0137] For high-precision kinematics, special kinematic systems can be used, for example, a hybrid kinematic system consisting of a 3D Cartesian serial, high-precision kinematic system mounted on a 3DOF rotary axis. High-precision Cartesian kinematic systems are available as motorized coordinate measuring machines. High-precision 3DOF rotary kinematic systems must be designed as a modification of high-precision 2DOF kinematic systems. High-precision 2DOF rotary kinematic systems are already known from the design of laser tracers. Calibration using mapping methods

[0138] The calibration of the pose sensor presented here is based on a mapping method. Mapping methods work with representations that assign a correction vector to each pose in the workspace or a correction vector to each vector in the configuration space. Mapping in the workspace

[0139] For the workspace, such mappings are obtained by determining correction vectors for a finite number of measured poses, which allow conclusions to be drawn about correction vectors in the entire workspace or configuration space. The poses P measured by the pose transmitter during calibration M (i) can, according to such a mapping procedure, in the workspace, for example, the correction vectors Δ(i)=P EXTERN (i)-P M (i) are assigned.

[0140] If, after calibration, the pose sensor measures a solid pose for which no explicit measurement was performed during calibration, an interpolation value is used as the correction vector for this solid pose, which is based on the available pairs (P M (i), Δ(i)) are based on solid-state poses and corresponding correction vectors. In the mapping procedure in the workspace, calibrated poses are assigned to uncalibrated poses of the pose generator.

[0141] However, model calculations for parallel kinematics show that mapping in the workspace is usually considerably worse than mapping in the joint space. Such error compensation in the workspace is typically implemented by additively correcting the vector of pose parameters to be commanded, given, for example, by Cartesian coordinates and cardan angles, using a correction vector. A correction vector is determined in each calibration pose and then obtained by interpolation at every point in the workspace. The vector of pose parameters has dimension 6, and its components are usually denoted by X, Y, Z, U, V, and W. The correction vector in a pose is determined as a function of that pose.Model calculations show that the error compensation described above in the workspace yields significantly worse results than the compensation in the configuration space (joint space) that follows and is not suitable for high-precision kinematics, but can only be used in small subsets of the workspace. Mapping in the configuration space

[0142] Mapping functions for the configuration space are obtained for hexapods by determining correction vectors in the configuration space for a finite number of configuration vectors, which allow a conclusion to be drawn about correction vectors in the configuration space.

[0143] In order to apply a mapping procedure in the configuration space to the pose generator (which has no joints / configuration space), the pose generator is assigned a "virtual hexapod" in the present disclosure. This provides the pose generator with a conceptual / virtual configuration space.

[0144] In such a calibration procedure, a plurality of solid poses (here also {P}) are used for each. SOLL (i)} denotes) the solid pose measured with the pose measuring system (therefore for each solid pose P SOLL (i) ∈ {P SOLL (i)} a corresponding measured pose P M (i) is obtained).

[0145] For each measured pose P M (i) can then be a configuration vector K, here referred to as the measured configuration vector M (i) determine which of the solid pose P measured by the pose measuring system M (i) is assigned according to an inverse kinematics. This inverse kinematics is also referred to here as virtual inverse / indirect kinematics IK. v denoted. The measured configuration vector K M (i) therefore results according to K M (i)=IK v (P M (i)).

[0146] Similarly, P is assigned to each pose SOLL(i) the majority of solid poses {P SOLL (i)}, the solid pose measured with an external pose measurement system (thereby determining for each solid pose P SOLL (i) ∈ {P SOLL (i)} a corresponding externally measured pose P EXTERN (i) is obtained). Measurement with the external pose measuring system generally has a higher accuracy than measurement with the pose transmitter. For each externally measured pose P EXTERN (i) can then be the configuration vector K EXTERN (i) determine which of the solid pose P measured by the external pose measurement system EXTERN (i) according to the inverse kinematics IK v is assigned, i.e., K EXTERN (i) = IK v (P EXTERN (i)).

[0147] For each pose P SOLL (i) the majority of solid poses {P SOLL(i)} a configuration vector difference ΔK(i) can then be determined, which is a difference between the first and the second configuration vector, as ΔK(i) = IK v (P EXTERN (i))-IK v (P M (i)). For each pose P SOLL (i) the majority of solid poses {P SOLL (i)} can then be used to calculate the configuration vector difference ΔK(i), for example, together with the pose P measured by the pose transmitter. M (i) are stored as a support point. From these support points (K M (i), ΔK(i)) can then be interpolated to obtain a configuration vector correction function F(x), for which, for example, F(P M determine whether (i))= ΔK(i) can hold.

[0148] Essentially, this means that indirect kinematics are used to assign configuration vectors to solid-state poses. Since the pose generator has no joints, these configuration vectors are also referred to as virtual. In principle, this indirect kinematics can be chosen freely. However, for the sake of simplicity, it is advisable to use the indirect kinematics of the hexapod into which the pose generator will later be integrated. Therefore, for the sake of clarity, it is generally assumed in this disclosure that the indirect / direct kinematics used to define the virtual hexapod and the indirect / direct kinematics of the hexapod to be calibrated are identical.

[0149] In Fig. Figure 9 shows an example of the relationship between a hexapod and a virtual hexapod. Figure 901 shows the uncalibrated hexapod in its zero pose, with the corresponding virtual hexapod, 902, to its right. Both hexapods are superimposed, practically identical. For clarity, the virtual hexapod is shown shifted to the right. The uncalibrated zero pose of the hexapod can be defined at the factory by specifying a predetermined configuration vector K. Figure 903 shows a hexapod in its downward position, compressed, and to its right, in Figure 904, is the corresponding virtual hexapod. Figure 905 shows a tilted hexapod, with the corresponding virtual hexapod to its right in Figure 906. The right-hand images are identical to the left-hand images to demonstrate the equality of the poses. These illustrations are included here to visually demonstrate a crucial relationship.The virtual hexapod, much like a puppet manipulated by the hexapod, always assumes its poses relatively precisely. This is because the pose generator measures its own (solid-state) pose, which changes in the same direction as the hexapod's pose, and can model a virtual hexapod—with a predefined kinematic model—in this pose. Therefore, the aim is to integrate the pose generator into the hexapod as precisely as possible to ensure minimal deviation between the two poses.

[0150] In 907, the hexapod is shown in its uncalibrated null position. To the right of it is the virtual hexapod 908. Fig. The virtual hexapod exaggerates its actually very small pose deviation compared to the pose of the uncalibrated hexapod to the point of being clearly visible. The virtual hexapod 908 is tilted to the left. Figure P AUS(NULL) indicates the read pose of the pose sensor when the uncalibrated hexapod is in the zero pose. Thus, one can P AUS (NULL) can be interpreted as a rototranslation between the pose of the uncalibrated hexapod in 907 and the virtual hexapod in 908. This rototranslation between the pose of the uncalibrated hexapod and the pose of the pose giver theoretically always holds, regardless of the pose of the uncalibrated hexapod. The determination of P AUS (NULL) serves only as an optional pre-positioning parameter when positioning the pose generator. If the virtual hexapod of the pose generator is to assume a pose according to 910 that consists of reducing the height of its platform, the tilt in the other direction is applied in the poses of the uncalibrated hexapod 909, according to P. AUS (ZERO) -1, and is tilted to the right here. Pose 910 of the pose generator is considered the pose of the calibrated hexapod. Pose 909 of the uncalibrated hexapod serves only to align the virtual hexapod. In order to determine the connection of the virtual hexapod's coordinate system to external coordinate systems, the upper plate of the pose generator has characteristic surface features, as shown in Fig. 6 is shown as part of a kinematic interface.

[0151] For this purpose, poses are approached throughout the entire work area and the poses P measured by the uncalibrated pose transmitter are recorded. M (i) with the respective actual poses P EXTERN(i) compared. If necessary, adjustment procedures must be applied after the calibration measurements to relate the poses to each other. This is because an initially defined reference coordinate system of the pose transmitter is not yet available during the measurement process. The pose transmitter is iteratively / sequentially moved into the calibration poses P M (i) moves. And the respective realized pose P EXTERN (i) measured directly / externally. Since the described calibration procedure for pose transmitters is based on a large number of pose measurements regularly distributed in the configuration space, it also encompasses the entire working space. More precisely, the configuration vector correction F(K) M ) based on the support points distributed in the configuration space, can be extended to a function over the entire configuration space by interpolation (e.g., by linear interpolation).

[0152] The pose generator can be configured for the hexapod in which it is to be used. This means that in every pose of the hexapod, the pose generator measures and / or outputs a pose that corresponds as closely as possible to the hexapod's pose. This can be achieved in particular by using an absolute pose P. abs of the pose giver, which is done by a corresponding choice of the distinguished zero pose P null2The second solid is aligned with the hexapod. The pose generator and the hexapod refer to essentially the same reference coordinate system when specifying poses. The hexapod pose is typically defined as the rototranslation of the upper platform into a distinguished zero pose relative to the current pose. Rototranslation denotes an element of SE(3). This zero pose is defined by a predefined configuration vector, i.e., predefined leg lengths of the hexapod. The imaginary kinematic model of the pose generator can correspond to the kinematic model of the hexapod. Therefore, the direct / indirect kinematics of the hexapod can be used as the virtual direct / indirect kinematics. Thus, the leg lengths of the hexapod correspond to the leg lengths of the virtual hexapod.

[0153] The configuration vectors {K SOLL(i)} for which measurements are to be taken during calibration, are conveniently located, for example, in an orthogonal grid in the joint space. Each of these configuration vectors K SOLL (i) ∈ {K SOLL(i)} is assigned an additive correction vector ΔK(i). Correction vectors ΔK(i) for the entire joint space are preferably obtained using multilinear interpolation. Such a mapping method can also be applied to posers, which only have a "working space" but no joint space. Parameters of the posers, such as the target coordinates of three targeted LEDs assigned to a pose, as in the exemplary embodiment, can be considered metaphorically as being moved by an actuator and thus define a joint space. However, it proves particularly advantageous to think of a virtual hexapod as the joint space in order to arrive at a joint space. Initially, modeling via a computationally easy-to-control hexapod, which is a classic Stewart platform and has simple ball joints at the ends of its legs, seems sensible.

[0154] Calibration of the pose transmitter is therefore possible using a virtual joint space based on a mapping procedure within the joint space. For this purpose, the pose transmitter can be assigned a multitude of poses. M (i) are moved. In these poses, a comparison of the current pose P is made. M (i) with the desired position P EXTERN(i) are taken, which are then further evaluated as a vector of virtual leg length differences. These measurements can be performed by moving the upper disk with a pose transmitter and determining the resulting poses with a coordinate measuring machine, for example, by probing three non-collinear spheres that are moved along with the gondola of the hexapod. If the spheres are pose markers of a reference artifact attached to the kinematic coupling of the upper disk, calibration data and a tangible reference coordinate system are obtained immediately. Alternatively, poses can be measured with a laser tracer by locating three reflectors sequentially in each pose.

[0155] Calibration using mapping methods in the configuration space is therefore a calculus based on direct kinematics. For simplicity, the direct kinematics of the hexapod (or the virtual direct / indirect kinematics of the hexapod type) into which the pose transmitter is later integrated can be used. Within this concept, pose deviations from calibration measurements can thus be expressed as a vector of virtual leg length errors using indirect kinematics. This is because every pose in the workspace can be expressed by virtual leg lengths using indirect kinematics. The virtual leg length errors obtained, for example, through multilinear interpolation in a pose are then added to the virtual leg lengths obtained through measurement. Applying direct kinematics then results in a corrected pose. Based on this calculus, a calibrated pose of the pose transmitter can be determined for every measured pose.However, the virtual kinematics does not have to correspond to the kinematics in which the pose generator is to be used, and one can choose a virtual kinematics whose indirect and direct kinematics are as easy / particularly easy to calculate as possible.

[0156] This approach to determining a calibrated pose, as just described, where the indirect kinematics of the pose generator do not correspond to the indirect kinematics of the hexapod, is computationally intensive due to the necessary additional calculations of direct kinematics required to implement a calibrated pose for the pose generator. Calculating direct kinematics is usually extremely computationally intensive, especially for hexapods. The calculus that avoids this significant computational effort and does not require calculations of direct kinematics is described in Fig. 10 and the associated figure description. Computationally intensive calculations are avoided, since only indirect kinematics need to be calculated at any given time, and pose corrections can be made indirectly within the joint space. This requires that the indirect kinematics of the virtual hexapod is identical to the indirect kinematics of the hexapod into which the pose transmitter is integrated. Joint measurement quantity & Pose measurement quantity

[0157] In general, for calibration, a set of poses to be measured is first configured using a plurality of configuration vectors {K SOLL (i)} given. This set of configuration vectors {K SOLL (i)} is also referred to in the present disclosure as joint measurement set {K SOLL (i)}. The set of poses to be measured is also referred to in the present disclosure as the set of poses to be measured {P SOLL (i)} called. The poses of the pose measurement set {P SOLL(i)} are the configuration vectors of the joint measurement set {K SOLL (i)} each according to the virtual direct kinematics DK used v () assigned, i.e., P SOLL (i) = DK v (K SOLL (i)), for each value of the index “i”. The pose P SOLL (i) of number i is thus defined by the configuration vector K SOLL (i) specified. Is the joint measurement set {K SOLL Given (i)}, the pose measurement set {P} can therefore be derived from this using virtual direct kinematics. SOLL (i)} calculate.

[0158] Conversely, the configuration vectors of the joint measurement set {K SOLL (i)} each of the poses of the pose measurement set {P SOLL (i)} according to the virtual indirect kinematics IK used v () assigned, i.e., K SOLL (I) = IK v (P SOLL (i)), for each i. It should therefore be noted that the pose measurement quantity {P SOLL(i)} can generally also be directly defined by a set of poses in pose space. From this, the joint measurement set {K} can be derived using virtual indirect kinematics. SOLL (i)} calculate. Determination of calibrated configuration vectors or calibrated poses

[0159] Generally, P is used for each pose SOLL (i) the Posen measurement quantity {P SOLL (i)}, determines a configuration vector correction ΔK(i) and / or a pose correction ΔP(i).

[0160] The pose correction ΔP(i) can be defined as the difference between a pose P measured by the external system. EXTERN (i) and a pose P measured by the pose transmitter M (i) can be calculated (e.g. ΔP(i) = P EXTERN (i)-P M (i)). For this purpose, the first solid and the second solid are moved relative to each other into a solid pose, which is called pose P SoLL(i) corresponds, and in this solid pose the measurements P will then be EXTERN (i) and P M (i) performed. When measuring P EXTERN (i) and P M (i) the first and second solid bodies are therefore in the same solid body pose. One of the poses P SoLL (i) The corresponding pose can be, for example, a pose in which either PM(i) = P SOLL (i) or P EXTERN (i) = P SoLL (i) applies (In practice |P M (i) - P SOLL (i)|<ε or |P EXTERN (i) - P SOLL (i)|<ε, chosen for a given “small” ε). However, it can also be the solid pose that results when the P SOLL (i) is commanded by a hexapod to which the two solid bodies are attached as in the example of the Fig. 8 are attached for calibration purposes.

[0161] A configuration vector correction ΔK(i) can be performed using virtually indirect kinematics based on the calibrated solid pose P EXTERN(i) and the uncalibrated solid pose P M (i) can be calculated. For example, for the pose P SOLL (i) the configuration vector correction ΔK(i) = IK v (P EXTERN (i))-IK v (P M (i)) be calculated.

[0162] Further details regarding calibration are also provided below using the example in Fig. The example shown in point 10 explains, as a flowchart illustrating the measurement of a pose by the pose giver. More precisely, it is in Fig. 10 an example explained in which (i) the pose giver to move into the desired poses P SOLL (i) in a hexapod as with respect to Fig. 8 explained is installed, and (ii) the two solid bodies of the pose generator are moved such that the pose generator assumes a pose P M (i) measures the position P that is close to the pose selected (in this iteration). SOLL (i) lies, i.e. for the |P M (i) - P SOLL(i)|<ε, for a given "small" value ε. Furthermore, in the example the Fig. 10 a joint measurement set {K SOLL (i)} is given. As already explained, according to this “map” of configuration vectors {K SOLL (i)} thus also a set of poses to be measured {P SOLL (i)} given. Configuration vectors {K SOLL (i)} can, for example, be located as a 6-dimensional grid in the configuration space of the virtual hexapod of the pose giver.

[0163] In step S1010, one of the configuration vectors K is therefore SOLL (i) the joint measurement set {K SOLL (i)} selected which is here considered K SOLL (i) (also without curly braces) is denoted to better distinguish it from the set {K SOLL (i)} to be able to distinguish. The pose giver (i.e., its solid body) should therefore be moved so that it assumes pose P SOLL (i)) measures. The following steps of Fig. Figure 10 therefore only shows the determination of a single calibrated configuration vector P EXTERN (i), namely the calibrated configuration vector P EXTERN (i) the selected configuration vectors K SOLL (i) corresponds. During calibration of the pose transmitter, the in Fig. 10 steps S1020, S1030, S1040, S1050, S1060, S1070, S1080, and S1090 described below, i.e., for each of the configuration vector of the joint measurement set {K SOLL (i)} executed.

[0164] The selected configuration vector K SOLL (i) is used as the starting value of the hexapodpose K HEXAPOD specified (step S1020). More precisely, the initial value of the configuration vector K is given. HEXAPOD equal to the selected configuration vector K SOLL (i) set.

[0165] In step S1030, the hexapod, in which the pose transmitter is installed for calibration, is then assigned this selected pose P. SOLL(i) is commanded. More precisely, the hexapod is given the current K at each iteration of the loop (steps S1030 to S1070). HEXAPOD targeted.

[0166] The hexapod then assumes a pose (this is generally not P). SOLL (i)) and the float generator performs a measurement of the solid float in step S1040. The float generator thus measures its uncalibrated measured float P in step S1040. M The uncalibrated configuration vector K M is derived from its uncalibrated measured pose P M calculated and queried, using, for example, a virtual indirect kinematics according to K M =IK v (P M ) is used. This configuration vector K M It is therefore calculated from the pose P calculated by the pose giver. M , which is determined from its sensor signals. The sensor signals detect the markings of the pose transmitter.

[0167] The configuration vector K MIn the present disclosure, this is also referred to as the measured configuration vector or uncalibrated configuration vector, since it results directly from the measurements of the sensors attached to the solid bodies of the position sensor. More precisely, the uncalibrated configuration vector K is M the configuration vector of the measured / uncalibrated solid pose P M according to the virtual indirect kinematics IK used, it is assigned, i.e., K M = IK v (P M ).

[0168] In step S1050, it is checked whether the uncalibrated pose P M the selected pose P SOLL (i) is implemented with sufficient accuracy. In the present example, this check is performed in the configuration space. It is therefore checked whether the magnitude of the difference |IK v (P M )-K SOLL (i)| less than (or less than or equal to) the given value ε.

[0169] If this is not the case (No in S1050; ||IKv (P M )-K SOLL (i)||≥ ε), then a correction ∇ is determined in step S1060. The correction ∇ can, for example, be the difference between the chosen configuration vector K. SOLL (i) and measured configuration vector K M be, for example ∇= K SOLL (i)-IK v (P M ), as in Fig. 10 is shown. This correction in step S1070 to the current value of K. HEXAPOD added, and the hexapod is updated in step S1030 with the new value of the configuration vector K. HEXAPOD targeted.

[0170] The loop defined by steps S1030, S1040, S1050, S1060, and S1070 is executed repeatedly until the check in step S1050 returns "Yes". If step S1050 returns "Yes", then |IK v (P M )-K SOLL (i)|< ε, in the exemplary embodiment a coordinate measuring machine measures the pose of the artifact of Fig. 8 (in step S1080). This pose, here as P* EXTERN (i) is saved. The poses P* are then stored. EXTERN (1) to P* EXTERN (MaxIndex) converted to the reference coordinate system of the pose generator, which is identical to the reference coordinate system of the hexapod, and then further down with P EXTERN These converted poses P EXTERN (i) correspond to the calibrated poses of the pose generator which correspond to the uncalibrated poses K SOLL (i) of the pose giver P M =DK v (K SOLL (i)) are assigned as a function. In simplified terms, the external measurement is performed in step S1080, and the external pose P is determined based on the external measurement. EXTERN (i) determines the selected pose P SOLL (i) corresponds. Thus, measurement for the configuration vector K SOLL (i) completed (step S1090). Use of calibration data

[0171] The calibration data can be used to determine calibrated poses and / or calibrated configuration vectors of the pose sensor from the marker measurement results of the pose sensor, i.e., from a pose measured by the pose sensor / uncalibrated pose within a pose. The calibration data can therefore be used to obtain corrections even for solid poses that are not included in the pose measurement set {P SOLL (i)} are included.

[0172] The poses in which these calibration measurements are performed can be defined by grid points of a six-dimensional grid in configuration space, where the grid fills the configuration space. Each of these grid points corresponds to a configuration vector of the joint measurement set {K}. SOLL (i)}.

[0173] Be K M such a configuration vector of the joint measurement set {K SOLL(i)}. The pose transmitter is moved to the pose for which the internal pose determination, based on the sensor data, determines a P M delivers. It applies to the virtually inverse kinematics IK. v the equation K M = IK v (P M As before, the index "M" refers to data directly related to the evaluation of the sensor data from the pose transmitter. An external pose measurement, for example performed with a coordinate measuring machine, yields the corresponding measured pose P*. EXTERN , which is related to the coordinate system of the virtual hexapod and is subsequently denoted by P EXTERN is designated as the virtual inverse kinematics IK. v K delivers EXTERN = IK v (P EXTERN The difference vector Δ=K EXTERN -K M represents the result of one of the calibration measurements; it will be K M Assigned and saved. The virtual hexapod is now in pose P. M, which results directly from evaluating the sensor data, one obtains K M = IK v (P M ). The corrected configuration vector in this pose K AUS = K M + Δ = K M + (K EXTERN - K M ) = K EXTERN and corresponds to the externally measured correct pose P, which can be calculated using direct kinematics EXTERN =DK v (K EXTERN ).

[0174] If the pose transmitter measures an uncalibrated pose P after calibration, M , which of a pose of the pose measurement set {P SOLL (i)} corresponds (i)} to (i) M = P SOLL (i), for a particular i), then • the corrected / calibrated solid P AUS (i) and • the corrected / calibrated configuration vector K AUS (i) =IK v (P EXTERN (i)).

[0175] If pose correction ΔP(i) is stored, the calibrated solid P can be used.EXTERN (i) from the uncalibrated pose P M Calculate by adding the correction using P EXTERN (i) = P M +ΔP(i). If configuration vector correction ΔK(i) is stored, the calibrated configuration vector K can be calculated. EXTERN (i) from the uncalibrated pose P M Calculate by adding the correction ΔK(i) using K EXTERN (i) = IK v (P M ) +ΔK(i).

[0176] The grid points {K SOLL (i)} measured difference vectors ΔK(i) can be used as support points of a function F that corresponds to each point in the configuration space, as K M The term denotes the assignment of such a difference vector to the virtual hexapod. The preferred function is a multilinear interpolation. F(K M ) corresponds to K M belonging configuration vector correction.

[0177] To each K MA calibrated K can now be selected in the configuration space. AUS calculate with K AUS = K M + F(K M ). K AUS is the calibrated configuration vector of the virtual hexapod. DK v (K AUS ) is the calibrated pose of the pose generator. The index "AUS" indicates that the data indexed in this way is output by the pose generator as calibrated configuration vectors or calibrated poses. Due to the one-to-one correspondence of direct and indirect kinematics, one can write P AUS (P M ) = DK(IK v (P M ) + F(IK v (P M )). This allows one to go from the uncalibrated pose to the calibrated pose of the pose generator, and as shown above, from the uncalibrated configuration vector of the pose generator to the calibrated configuration vector of the pose generator.

[0178] In summary, the calibrated configuration vector K is obtained. AUSthat is, by adding a configuration vector correction F(K) M ) to the respective uncalibrated configuration vector K M The uncalibrated configuration vector K M is the measured solid pose P M according to a (virtual) inverse kinematics IK v assigned (i.e., K) M = IK v (P M )). The configuration vector correction F(K M ) is the uncalibrated configuration vector K M assigned according to a predetermined assignment rule. The assignment rule therefore specifies K M the configuration vector correction F(K M )) to, you are through the figure K M - F(K M ) given.

[0179] This predetermined allocation rule Figure K M → F(K M) is based on an interpolation which is based on a plurality of support points, each containing a predetermined configuration vector and a predetermined configuration vector difference. The support points are therefore the pairs obtained through the calibration measurements (K ). SOLL (i), ΔK(i)). Control procedures

[0180] The present disclosure provides, in particular, methods for controlling a kinematic system. The kinematic system is to be controlled in such a way that it assumes a (predefined) target pose (as precisely as possible). To realize this target pose as precisely as possible, the control method employs a pose generator.

[0181] As in Fig. As shown in 11, the procedure includes • one step S1100 of the integration of a pose sensor / pose measuring system into the kinematics; • a step S1120, in which the target pose (here also referred to as desired pose P) SOLL is obtained (designated); • a step S1140 of determining a target configuration vector, which here is K HEXAPOD is referred to as, • a step S1160 of controlling the kinematics with the specified target configuration vector K HEXAPOD . Determining the target configuration vector K HEXAPOD - Step S1140

[0182] In general, the target configuration vector is determined based on the target pose and one or more measured solid poses, which are measured by the attached pose measurement system when the kinematics are controlled with the respective configuration vector(s).

[0183] The hexapod is controlled with a configuration vector and moves into the corresponding pose. The integrated pose sensor measures the relative position of the two solid bodies of the pose sensor when the hexapod is in that pose. This process can be repeated for other configuration vectors. The measured solid pose(s) for these configuration vector(s) are used to determine a target configuration vector that will result in an (accurate) realization of the target pose when the hexapod is controlled with this target configuration vector.

[0184] Further details of an exemplary realization of a target pose will also be provided below using the following examples. Fig. 12 explains, which shows an example of calibrated positioning of a hexapod using a pose generator in a flowchart.

[0185] As in Fig. As shown in 12, determining S1140 of the target configuration vector K is possible. HEXAPOD The following steps include: • Calculate S1210, S1220, S1230, S1240 of the target configuration vector based on the target pose and the inverse kinematics of the kinematics. • Controlling S1250 of the kinematics with the (current) target configuration vector K HEXAPOD In step S1250, the hexapod is commanded into a pose defined by the (current) configuration vector K. HEXAPOD is defined. In other words, the hexapod is defined with the current configuration vector K. HEXAPOD targeted. • Measure S1260 through the fixed pose measuring system, a solid body pose, which the first and second solid bodies are when the kinematics are controlled with the (current) target configuration vector K. HEXAPOD take place relative to each other. In step S1260, for example, the configuration vector K is taken. AUS of the pose giver and / or the solid body pose P AUSThe poser's information was queried. • Adjusting S1290 of the (current) target configuration vector K HEXAPOD based on a deviation of the measured solid pose from the target pose.

[0186] The goal of positioning in Fig. It is 12 that the pose giver sets the calibrated pose P SOLL within its predetermined error limit. This is equivalent to the fact that the pose giver has a P AUS corresponding calibrated configuration vector K AUS outputs in accordance with this condition. That is, the target configuration vector K should be displayed. HEXAPOD can be found where, if the hexapod has this target configuration vector K HEXAPOD The pose generator is controlled by a configuration vector K. AUS = K SOLL = IK(P SOLL outputs ), where IK() is the inverse kinematics of the hexapod. Positioning into the desired pose P SOLL This therefore usually involves the calculation of a configuration vector K.SOLL with the indirect kinematics IK() of the hexapod according to K SOLL = IK(P SOLL ), which in Fig. 12 is shown in step S1210. The resulting target configuration vector K SOLL The target configuration vector K should then ultimately be occupied by the pose generator, i.e., by the pose generator's virtual hexapod. SOLL This is therefore the target configuration vector of the pose generator or of a virtual hexapod associated with the pose generator.

[0187] In other words, the built-in pose sensor, which is calibrated and / or has a higher accuracy than the hexapod, should be set to pose P. SOLL The poser is considered to be in pose P. SOLL when he outputs / measures these (P AUS = P SOLL ), or one of the poses P SOLL outputs / measures the corresponding configuration vector K AUS = K SOLL = IK(P SOLL ).

[0188] One advantage of the in Fig. The advantage of the method described in section 12 is that the positioning process can completely avoid the need to calculate direct kinematics, which can be very complex for hexapods compared to indirect kinematics. Positioning in the workspace (where the pose generator of the assumed pose is a solid pose P) AUS (assigns and outputs), instead of, as presented here, a positioning in the configuration space (where the pose generator of the assumed pose is a configuration vector K). AUS (assigns and outputs) is also feasible.

[0189] It should also be noted that it can be advantageous or practical if the float indicator is installed in the hexapod in a certain way, so that the float P measured by it AUS The position should correspond as closely as possible to the pose the hexapod assumes in its uncalibrated state. After calibration, the hexapod's position is considered to be pose P. AUSof the poser, which in the example of the Fig. 12 from the K he issued AUSThis can be calculated using direct kinematics. In particular, this can prevent distortions, as the calibrated pose of the pose generator "guarantees" the highest accuracy. However, precise installation is not necessary. In particular, as described below, an adapted inverse kinematics (AIK) can be used to compensate for, for example, installation inaccuracies. The pose of the uncalibrated hexapod is referenced to the pose that is factory-defined as the initial pose, referred to as the zero pose, and corresponds to the neutral element of SE(3). In this pose, the hexapod is vertically in the middle of its stroke range, it is not horizontally displaced, and its platform is not twisted. This distinguished pose is its zero pose, and the mapping of this pose is the identical mapping ID. This zero pose is factory-defined by a predetermined configuration vector of the hexapod. Iterative process

[0190] As in Fig. As shown in Figure 12, the actuation of S1250, the measurement of S1260, and the adjustment of S1290 can be performed repeatedly (e.g., in this order) until a predetermined condition is met. Whether the predetermined condition is met can be determined, for example, as shown in Figure 12. Fig. 12. In one step, S1270 is checked after measuring S1260. If the condition is met, the procedure ends and the current target configuration vector K is displayed. HEXAPOD is the result of determining the target configuration vector, which corresponds to the desired pose P SOLL corresponds.

[0191] In Fig. In step S1270, it is tested whether K AUS sufficiently close to K SOLL comes close. In Fig. In section 12, the criterion / condition is assumed to be, for example, that a norm of the difference vector of the two configuration vectors falls below a freely chosen threshold. Maximum norms are a suitable choice. It should be noted that, especially when an adapted inverse kinematics is used, the pose generator is usually already in a pose P in the first iteration of the loop. SOLL comes close.

[0192] The predetermined condition can generally be a condition on the magnitude (e.g., the absolute value of the maximum norm) of a last determined correction (i.e., the size of the adjustment in step S1290). In particular, the condition can be that the absolute value of the last correction is less than (or less than or equal to) a predetermined threshold. Alternatively or additionally, the predetermined condition can, for example, be a condition on the magnitude (e.g., the absolute value of the maximum norm) of a deviation of a last measured solid pose from the target pose. For example, the condition can be that the deviation of the solid pose P output by the pose generator is AUS from the target position P SOLL is less than (or less than or equal to) a predetermined threshold.

[0193] It is noted that the progress within the loop (steps S1250 to S1290) in Fig. 12 largely follows the course of the loop in Fig. 10 (steps S1030 to S1070) is equivalent with the difference that instead of an uncalibrated configuration vector K M the (possibly calibrated) configuration vector K AUS is queried and processed. Using adapted kinematics - determining the initial position

[0194] The branch at S1220 of the Fig. 12 determines whether or not to use an “adapted” inverse kinematics (AlK) (or an “adapted” target pose) to determine the starting value of K. HEXAPOD to determine for the iterative process. In the optional step S1220, for example, it can be determined whether an AIK should be used or not. This can be based on user input and / or accuracy requirements. If it is determined that no AIK should be used, then step S1230 is executed. If it is determined that an AIK should be used, then step S1240 is executed.

[0195] It is noted that each of steps S1220, S1230, and S1240 is optional. Some embodiments include step S1230 but not steps S1220 and S1240. Other embodiments include step S1240 but not steps S1220 and S1230. Still other embodiments include all steps S1220, S1230, and S1240. A (first) determination of K HEXAPOD This is done either in S1230 or S1240. S1290 can also be used as a (further) determination by K. HEXAPOD It is viewed here, but is usually considered an adaptation of K HEXAPOD designated.

[0196] An AIK (Automatic Interlocking Control) can be used to command the hexapod into an advantageous pose, which places the pose generator in a favorable starting position for the following steps (i.e., so that the initial corrections are smaller). This advantageous starting position is determined here based on the known pose deviation P. AUS (ID) determined. P AUS(ID) can be retrieved by querying K AUS This can be obtained when the uncalibrated hexapod is in its zero position. P AUS (ID) is therefore measurably accessible. The reason for this position deviation P AUS (ID) is mainly due to inaccuracies in the mechanics of the hexapod and inaccuracies in the installation of the pose transmitter in the hexapod.

[0197] Positioning can also be performed without pre-positioning the hexapod using AIK. However, AIK is recommended for every positioning operation of the hexapod, even if the pose generator remains unused, so that even without the pose generator, the uncalibrated pose of the hexapod largely corresponds to its calibrated pose.

[0198] In step S1230, the target configuration vector K is HEXAPOD based on the target pose P SOLLand the inverse kinematics IK() of the hexapod. Thus, the configuration vector corresponding to the target pose P is calculated. SOLL The target configuration vector K is assigned according to the inverse kinematics IK(). HEXAPOD = K SOLL = IK(P SOLL ).

[0199] In step S1240, the target configuration vector is determined according to K. HEXAPOD = IK(P SOLL ∘(P AUS (ID) -1 ) is determined, where IK() is the inverse kinematics of the hexapod. It is therefore the configuration vector that corresponds to the adapted target pose P. SOLL ∘(P AUS (ID) -1 according to the inverse kinematics IK(), is assigned as the target configuration vector K HEXAPOD calculated. The adjusted target pose P SOLL ∘(P AUS (ID) -1 This results from the target pose P SOLL by applying “∘” of a pose transformation (P AUS (ID) -1, which is the inverse of a pose transformation P corresponding to one of the measured zero solid poses AUS (ID) is.

[0200] The measured zero solid pose P AUS (ID) can be obtained by combining the hexapod with the zero configuration vector K. NULL = IK(ID) is controlled. The zero configuration vector K NULL Therefore, according to inverse kinematics IK(), the zero pose ID of the kinematics is assigned. The hexapod is then assigned the zero configuration vector K. NULL When controlled, it moves to its zero position ID. It should be noted that in practice, it is usually assumed that in K NULLNo pose errors can occur because pose errors only manifest themselves in pose changes. The concept of the absolute pose has been internalized, meaning that the initial pose cannot be incorrect because all other poses are referenced to it, and there is no other reference point. This is also intuitive because the hexapod's reference coordinate system is depicted in the construction drawings / manuals as being "stuck" to the gondola in the zero pose. The built-in pose sensor then measures the solid pose while the hexapod is in this zero pose ID, and the pose sensor assigns the pose P to the assumed solid pose. AUS (ID) to. The zero solid pose P AUS (ID) is therefore the solid pose that the built-in pose generator assigns to the solids when the hexapod is controlled with the zero configuration vector. Adapting S1290 from K HEXAPOD - Difference configuration vector

[0201] In general, adjusting the target configuration vector K (S1290) HEXAPOD (i) determining S1280 of a correction Δ of the target configuration vector Δ and (ii) adjusting the target configuration vector K HEXAPOD based on the correction Δ.

[0202] Step S1280 of the Fig. Figure 12 shows the determination of such a difference configuration vector Δ. The correction Δ is based on the result obtained when controlling the kinematics with the (current) target configuration vector K. HEXAPOD measured solid-state pose. If one could command the virtual hexapod, one would K = K AUS Command + Δ, and the virtual hexapod would be in the desired position P. SOLL Because K AUS + Δ = K AUS + K SOLL - K AUS = K SOLL The correction Δ can therefore be, for example, a difference between the configuration vector K measured / output by the pose transmitter. AUS(which is based on the solid body float measured by the float setter K AUS =IK(P AUS )) and the target configuration vector K SOLL are equivalent to.

[0203] In step S1290, we now apply the specific Δ to the current target configuration vector K. HEXAPOD of the hexapod. As in Fig. As shown in 12, when adjusting the target configuration vector K HEXAPOD with the correction Δ, e.g. adding the correction Δ to the target configuration vector K HEXAPOD will be. The target configuration vector K HEXAPOD so it will be on K HEXAPOD + Δ, updated / adjusted.

[0204] The adjustment by adding the vector Δ moves the hexapod to the configuration vector K. HEXAPOD + Δ, the virtual hexapod, but to a good approximation according to K AUS + Δ = K SOLLThis is because the hexapod and the virtual hexapod essentially lie on top of each other (and here the virtual direct / indirect kinematics are assumed to be the same as the direct / indirect kinematics of the hexapod). Such iterative realization of poses based on corrective vectors is referred to as "alignment" in hexapod applications; these are typical uses of hexapods. Typically, objects are aligned using a hexapod based on repeated measurements from external sensors. In these iterative positioning processes, well-estimated correction vectors emerge, which cannot be precisely determined, primarily due to the lack of an exact reference between the hexapod coordinate system and external coordinate systems in the prior art. The in Fig. The 12 approximation methods shown essentially correspond to the approximation methods used in alignments.

[0205] In summary, this involves controlling a kinematic system to assume a (previously obtained / predefined) target pose. For this purpose, a pose measurement system is integrated into the kinematic system. The pose measurement system comprises a first and a second solid body. The pose measurement system is configured to measure a solid body pose. The solid body pose is a pose that the first and second solid bodies assume relative to each other. The pose measurement system is integrated into the kinematic system by: (i) attaching the first solid body to the end effector and attaching the second solid body to the base, or (ii) attaching the second solid body to the end effector and attaching the first solid body to the base. One or more solid body poses are measured by the attached pose measurement system when the kinematic system is controlled with the respective configuration vector(s).The target configuration vector is determined based on the target pose and the measured solid pose(s), and the kinematics are controlled with the determined target configuration vector. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] DE 102020124136

[0066]

Claims

[1] A method for controlling a kinematic system to assume a target pose, wherein the method comprises the following steps: Integrating a pose measurement system into the kinematics, whereby - the Posen measuring system: (i) comprises a first and a second solid body, and (ii) is designed to measure a solid pose which the first and second solids assume relative to each other, and - the pose measurement system is integrated into the kinematics by: (i) Attaching the first solid body to the end effector and attaching the second solid body to the base, or (ii) Attaching the second solid to the end effector and attaching the first solid to the base; Maintaining the target pose; Determining a target configuration vector based on the target pose and one or more measured solid poses, wherein the one or more measured solid poses are measured by the attached pose measuring system when the kinematics are controlled with the respective configuration vector(s); and Controlling the kinematics using the target configuration vector. [2] The method according to claim 1, wherein determining the target configuration vector comprises the following steps: Calculating the target configuration vector based on the target pose and an inverse kinematics of the kinematics; Controlling the kinematics using the target configuration vector; Measuring, by means of the fixed pose measuring system, a solid-body pose that the first and second solid bodies assume relative to each other when the kinematics are controlled with the calculated target configuration vector; and Adjusting the target configuration vector based on a deviation of the measured solid pose from the target pose. [3] The method according to claim 2, wherein the adjustment of the target configuration vector comprises: Determining a correction of the target configuration vector based on the solid pose measured when the kinematics are controlled with the target configuration vector; and Adjusting the target configuration vector based on the correction. [4] The method according to claim 3, wherein The correction of a difference corresponds to: • the configuration vector that is assigned to the solid body pose measured when the kinematics are controlled with the target configuration vector according to the inverse kinematics, and • the configuration vector assigned to the target pose according to inverse kinematics; and When adjusting the target configuration vector with the correction, the correction is added to the target configuration vector. [5] The method according to any one of claims 2 to 4, wherein the following steps are repeatedly performed when determining the target configuration vector until a predetermined condition is satisfied: - controlling the kinematics with the target configuration vector; - the measurement, by the fixed pose measuring system, of the solid body pose that the first and second solid bodies assume relative to each other when the kinematics are controlled with the target configuration vector; and - adjusting the target configuration vector based on the deviation of the measured solid pose from the target pose. [6] The method according to claim 5, wherein the predetermined condition is a condition: - a measure of a recently determined correction; and / or - indicates the magnitude of a deviation of a last measured solid body pose from the target pose. [7] The method according to any one of claims 1 to 6, wherein, when calculating the target configuration vector based on the target pose and the inverse kinematics, A configuration vector that is assigned to the target pose according to inverse kinematics is calculated as the target configuration vector. [8] The method according to any one of claims 1 to 6, comprising: Measuring, by means of the fixed pose measuring system, a zero solid pose which the first and the second solid assume relative to each other when the kinematics are controlled with a zero configuration vector, wherein the zero configuration vector is assigned to a zero pose of the kinematics according to the inverse kinematics; and wherein, when calculating the target configuration vector based on the target pose and the inverse kinematics, a configuration vector that is assigned to a fitted target pose according to inverse kinematics, is calculated as the target configuration vector, where The adapted target pose is obtained from the target pose by applying a pose transformation that is inverse to a pose transformation corresponding to the measured zero solid pose. [9] A pose measurement system for installation in a kinematic system, comprising: a first and a second solid body, wherein The pose measurement system is integrated into the kinematics by: - Attaching the first solid to the end effector and attaching the second solid to the base, or - Attaching the second solid to the end effector and attaching the first solid to the base; and Means for measuring a solid-state pose that the first and second solid bodies assume relative to each other. [10] A computer program comprising instructions which, when executed by a computer, cause it to: To determine the target configuration vector and to control the kinematics according to any one of claims 1 to 8. [11] A computer-readable data carrier on which the computer program according to claim 10 is stored. [12] A method for calibrating a pose measurement system, wherein the pose measurement system: - comprises a first and a second solid body, - is designed to measure a solid pose that the first and second solids assume relative to each other, and - is set up to assign a calibrated configuration vector to a measured solid pose, which indicates a calibrated solid pose corresponding to the measured solid pose, wherein The calibrated configuration vector is obtained by adding a configuration vector correction to an uncalibrated configuration vector. the uncalibrated configuration vector of the measured solid pose is assigned according to an inverse kinematics, the configuration vector correction is assigned to the uncalibrated configuration vector according to a predetermined assignment rule, and the predetermined assignment rule is based on an interpolation which is based on a plurality of support points, each containing a predetermined configuration vector and a predetermined configuration vector difference; and The procedure comprises the following steps for each of a plurality of solid poses: - Measuring the solid body float with the float measuring system; - Measuring the solid body float using an external float measurement system; - Determining a configuration vector difference, which is a difference between a first and a second configuration vector, where (i) the first configuration vector of the solid pose measured by the Posen measurement system is assigned according to an inverse kinematics, and (ii) the second configuration vector is assigned to the solid pose measured with the external pose measurement system according to the inverse kinematics; and - Storing the first configuration vector and the configuration vector difference as one of the support points.

Images