Method and apparatus for determining an optimized sequence of movements of a robotic device
阅读说明：本技术 用于确定机器人装置的优化移动序列的方法和设备 (Method and apparatus for determining an optimized sequence of movements of a robotic device ) 是由 P·S·施密特 F·维恩肖费尔 于 2019-02-04 设计创作，主要内容包括：本发明涉及一种方法,所述方法用于确定机器人装置的优化移动序列,以用于移动第一对象,使得第一对象被带到目标位置中而不管关于第二对象的第一对象的位置的不确定性且/或不管机器人装置的位置的不确定性。所述方法包括：模拟机器人装置的移动部段,同时考虑第一对象的位置的不确定性和/或机器人装置的位置的不确定性；以及确定机器人装置的优化移动序列,同时考虑指定第一对象的至少一个起始位置和目标位置的模拟移动部段和边界条件。确定优化移动序列,通过所述优化移动序列,机器人装置能够将所述第一对象可靠地引导至其目标位置中。(The invention relates to a method for determining an optimized movement sequence of a robotic device for moving a first object such that the first object is brought into a target position regardless of an uncertainty of the position of the first object with respect to a second object and/or regardless of an uncertainty of the position of the robotic device. The method comprises the following steps: simulating a moving section of the robotic device while taking into account an uncertainty of the position of the first object and/or an uncertainty of the position of the robotic device; and determining an optimized sequence of movements of the robotic device while taking into account simulated movement segments and boundary conditions specifying at least a start position and a target position of the first object. An optimized movement sequence is determined by which the robotic device can reliably guide the first object into its target position.)
1. A method for determining an optimized movement sequence (8) of a robotic device (3) for moving a first item (1, 1a-1 d) such that the first item (1, 1a-1 d) is brought to a target posture (ZP) regardless of an uncertainty of a posture of the first item (1, 1a-1 d) with reference to a second item (2) and/or regardless of an uncertainty of a posture of the robotic device (3), wherein the method comprises:
simulating (S1) a moving section of the robotic device (3) while taking into account the uncertainty of the pose of the first item (1, 1a-1 d) and/or the uncertainty of the pose of the robotic device (3); and
determining (S2) the optimized movement sequence (8) of the robotic device (3) while taking into account simulated movement sections and constraints indicative of at least a Start Posture (SP) and the target posture (ZP) of the first item (1, 1a-1 d), wherein
The robotic device (3) having a compliance for partially compensating the robotic device (3) for the uncertainty of the pose of the first item (1, 1a-1 d) and/or the uncertainty of the pose of the robotic device (3); and wherein
Further simulating the moving section while taking into account a predetermined compliance of the robotic device (3).
2. The method of claim 1, wherein the first and second light sources are selected from the group consisting of,
it is characterized in that the preparation method is characterized in that,
the determination of the optimized movement sequence (8) comprises determining an optimized compliance of the robotic device (3).
3. The method according to claim 1 or 2,
it is characterized in that the preparation method is characterized in that,
said uncertainty of said first item (1, 1a-1 d) and/or said pose of said robotic device (3) is a selected value of a statistical uncertainty distribution, in particular a value selected by random sampling.
4. Method according to one of claims 1 to 3,
it is characterized in that the preparation method is characterized in that,
-simulating the plurality of moving sections while taking into account different values of the uncertainty of the pose of the first item (1, 1a-1 d) and/or the robotic device (3).
5. Method according to one of claims 1 to 4,
it is characterized in that the preparation method is characterized in that,
-the simulation (S1) of the movement section involves randomly determining (S11) a plurality of movement sections, all starting from the same movement section starting pose of the robotic device (3); wherein the method further comprises:
randomly selecting (S12) a moving segment target pose for one of the plurality of simulated moving segments, and defining (S13) this selected moving segment target pose as a new moving segment start pose; and
repeating the simulating of the moving segment (S11), the random selection of the moving segment target pose (S12), and the defining of the selected moving segment target pose as a new moving segment starting pose (S13) until the moving segment target pose of one of the plurality of simulated moving segments is within a predetermined target area.
6. The method of any one of claims 1 to 5,
it is characterized in that the preparation method is characterized in that,
the optimized movement sequence (8) of the robotic device (3) is determined by putting together at least two of the simulated movement sections (S2).
7. Method according to one of claims 1 to 6,
it is characterized in that the preparation method is characterized in that,
-the simulation (S1) of the moving section takes into account at least one robotic device performance, a first article performance of the first article (1, 1a-1 d) and/or a second article performance of the second article (2); wherein the robotic device performance includes at least:
a maximum force that can be exerted by the robotic device (3);
a model representation of the robotic device (3); and/or
Possible movements of the robotic device (3); and/or wherein the at least one of the first,
the first article property and/or the second article property comprises at least:
a coefficient of friction of the first article (1, 1a-1 d) and/or the second article (2);
the size and/or shape of the first article (1, 1a-1 d) and/or the second article (2);
a model representation of the first item (1, 1a-1 d) and/or the second item (2);
the weight of the first article (1, 1a-1 d) and/or the second article (2); and/or
Material properties of the first article (1, 1a-1 d) and/or the second article (2).
8. Method according to one of claims 1 to 7,
it is characterized in that the preparation method is characterized in that,
the determination (S2) of the optimized movement sequence (8) comprises:
calculating (S3) a cost function for each simulated moving segment; and
determining (S3) the optimized movement sequence (8) based on the calculated cost function.
9. The method of claim 8, wherein the first and second light sources are selected from the group consisting of,
it is characterized in that the preparation method is characterized in that,
determining a plurality of movement sequences based on the simulated movement segments;
calculating (S21) a total cost function for each movement sequence based on the cost functions of the movement segments of the respective movement sequence; and
selecting a movement sequence having a smallest total cost function from the plurality of movement sequences as the optimized movement sequence (8).
10. The method of claim 8, wherein the first and second light sources are selected from the group consisting of,
it is characterized in that the preparation method is characterized in that,
determining a first movement sequence based on the simulated movement segment;
determining a moving sequence portion comprising at least two moving segments;
calculating the cost function of the moving sequence portion; and
further tracking the portion of the sequence of movements if the cost function of the portion of the sequence of movements is lower than the cost function of the first sequence of movements.
11. Method according to one of claims 8 to 10,
it is characterized in that the preparation method is characterized in that,
the cost function includes indications of:
the mass of the simulated moving section;
a length of a path corresponding to the simulated movement section;
a duration of execution of the simulated movement segment;
energy consumption of the execution of the simulated moving segments; and/or
Forces occurring on the first item (1, 1a-1 d) and/or the second item (2) during execution of the simulated movement section.
12. Method according to one of claims 1 to 11,
it is characterized in that the preparation method is characterized in that,
the uncertainty of the pose of the first item (1, 1a-1 d) represents an uncertainty of a model for the first item (1, 1a-1 d) and/or the second item (2) and/or a production tolerance of the first item (1, 1a-1 d) and/or the second item (2).
13. A computer program product causing a method according to one of claims 1 to 12 to be performed on a program-controlled apparatus.
14. An apparatus (4) for determining an optimized movement sequence (8) of a robotic device (3) for moving a first item (1, 1a-1 d) such that the first item (1, 1a-1 d) is brought to a target posture (ZP) regardless of an uncertainty of a posture of the first item (1, 1a-1 d) with reference to a second item (2) and/or regardless of an uncertainty of a posture of the robotic device (3), wherein the apparatus (4) comprises:
a simulation unit (5) for simulating a moving section of the robotic device (3) while taking into account the uncertainty of the pose of the first item (1, 1a-1 d) and/or the uncertainty of the pose of the robotic device (3); and
a determination unit (6) for determining the optimized movement sequence (8) of the robotic device (3) while taking into account the simulated movement sections and constraints indicative of at least a Starting Posture (SP) and the target posture (ZP) of the first item (1, 1a-1 d), wherein,
the robotic device (3) having a compliance for partially compensating the robotic device (3) for the uncertainty of the pose of the first item (1, 1a-1 d) and/or the uncertainty of the pose of the robotic device (3); and wherein the one or more of the one,
the moving section is further simulated by the simulation unit (5) while taking into account a predetermined compliance of the robotic device (3).
15. The apparatus of claim 14, for performing the method of one of claims 2 to 12.
Technical Field
The present invention relates to a method for determining an optimized sequence of movements (sequence) of a robotic device and an apparatus for performing such a method.
Background
Robotic devices can be used in industrial installations, for example to move articles (in particular workpieces) and/or bring them together. The first item can be generally referenced to a target pose referenced to the second item by a sequence of movements of the robotic device. The sequence of movements may need to be optimized such that the first item is correctly brought to the target pose even in case of disturbances and/or errors in the model representation of the item.
The document "object-based frame for motion planning under sensing and control uncertainty" by Lavalle et al (International journal of robotics research, 1/1998) relates to the optimization of sequences of robot movements. This is achieved by simulating and evaluating future states while taking into account uncertainty in the pose of the item to be moved.
The document "Parts assembly planning with uncertainty with simulation-aided physical reasoning" by Kim et al (about the 2017 IEEE conference on robotics and automation, 29
The document "Collision-probability constrained PRM for a manipulator with basic pose uncertainty" (Intelligent robots and Systems, 2009) relates to the planning of the movement of a robot arranged on a moving platform.
On this background, it is an object of the present invention to provide an improved determination of an optimized movement sequence of a robotic device.
Disclosure of Invention
According to a first aspect, a method is presented for determining an optimized movement sequence of a robotic device for moving a first item such that the first item is brought to a target pose, regardless of an uncertainty of a pose of the first item with reference to a second item and/or regardless of an uncertainty of a pose of the robotic device. The method comprises the following steps:
simulating a moving section of the robotic device while taking into account uncertainty of the pose of the first article and/or uncertainty of the pose of the robotic device; and
an optimized sequence of movements of the robotic device is determined while considering simulated movement segments and constraints indicative of at least a starting pose and a target pose of the first item.
For example, a robotic device is a robot, a robotic arm, or a device capable of performing a particular sequence of movements and/or movement segments. The robotic device can be used in automated manufacturing processes, particularly in industrial installations or in packaging processes.
The robotic device is particularly suitable for bringing a first item to a target pose with reference to a second item. For example, in this case, the first article and the second article (also referred to as "articles" hereinafter together) can be brought together. The first item can be inserted into or placed onto the second item by the robotic device in order to subsequently attach the items to one another. The final pose of the first item with reference to the pose of the second item can be referred to as the target pose.
A "gesture" of an object is understood to mean in particular a position and/or an orientation of this object. The terms "target gesture" and "starting gesture" can be similarly defined.
A sequence of movements can be understood to mean a series of individual movements of the robotic device. The movement sequence can include a variety of movement segments. In particular, a plurality of moving sections can be put together such that they form a moving sequence. The movement sequence and/or the movement section can indicate, for example, a direction in which the robotic device should move the first article along the coordinate system, a speed at which the robotic device should move the first article, and/or a force that the robot should apply in this case, particularly in the context of impedance control.
The simulation of the mobile section particularly relates to physical simulation. The moving section can be simulated by using a simulator. For example, the simulator is a solid state simulator that reproduces the dynamic system (first article and robotic device). In particular, each moving segment can be described by a moving segment start pose and a moving segment target pose. During the simulation, the randomly moving segments can be determined starting from the initial pose. Multiple instances of the analog movement segment can be correlated to form a movement sequence.
The simulator is particularly suitable for simulating complex moving sections. In fact, the movement sequence can include movements in up to six degrees of freedom (three rotations and three translations). Furthermore, contact of the robotic device with the first item and/or contact between two items can produce a certain amount of non-linearity that is difficult to model. Furthermore, the process in which the first item is brought to the target posture is a dynamic process, which is why at least the speed can also be included in the simulation.
The uncertainty of the posture of the first item refers in particular to the uncertainty as to the exact position present for the first item. The uncertainty of the pose of the first item is defined with particular reference to the second item, which is why the uncertainty of the pose of the second item can also be covered by the uncertainty of the pose of the first item. The uncertainty of the pose of the first article can be due to the first article and/or the second article having an amount of elasticity and/or a specific coefficient of friction that is not considered by the model representation of the first article and/or the second article. The uncertainty of the pose of the first article is not particularly always constant and can vary from article to article.
The uncertainty of the robotic device describes in particular the uncertainty as to the exact position where the robotic device is located. The uncertainty of the robotic device can also be a reflection of an inaccurate model representation of the robotic device. Further, the uncertainty of the robotic device can represent an interruption in the movement of the robotic device.
The described method can involve a plurality of mobile sections being simulated, in particular twenty mobile sections or more. Each of these simulations can account for uncertainty in the pose of the first item and/or uncertainty in the pose of the robotic device. Hereinafter, the uncertainty of the pose of the first article and the uncertainty of the pose of the robotic device are collectively referred to as "uncertainty of the pose".
In particular, each simulation of a moving segment takes into account a different value for the uncertainty of the pose. Furthermore, the performance of the robotic device (also referred to as robotic device performance) and/or the article performance of the first article and/or the second article may also be considered, as will be described in more detail below.
The simulation of the moving section particularly results in a variety of simulated moving sections being obtained. For example, the individual simulated moving segments are then evaluated with respect to their quality. They can also be combined to form an optimized movement sequence.
The evaluation of the simulated movement sections leads in particular to an optimized simulation sequence of the robot device being determined. To determine the optimal movement sequence, the starting pose and the target pose of the first item can be considered as constraints. The starting pose is in particular the pose in which the robotic device moves the first item according to the movement sequence and guides it into the first item preceding in the target pose.
The optimized movement sequence is a movement sequence that specifically allows the robotic device to guide the first item into the target pose even if there is uncertainty in the pose of the first item and/or the robotic device. That is, two different first articles moved by the robotic device along the same optimized sequence of movements can each be directed into a target pose even though they are affected by different uncertainties.
The determination of the optimized movement sequence allows in particular a reliable movement of the first item to the target pose. In particular, an optimized sequence of movements is provided that guides the first item into the desired target pose despite model errors and other variable disturbances. For example, it is not necessary to use complex methods to try to reduce model errors and/or to exclude disturbances as much as possible.
The optimal movement sequence can be flexibly determined since this determination is performed automatically. In particular, the optimal movement sequence can be determined even for movements that should be performed only once or a few times with the robotic device, since the optimal movement sequence can be determined with little effort. The described method can be used in a particularly versatile manner, since the optimized movement sequence is determined while taking into account the uncertainty of the pose of the first item and/or the uncertainty of the pose of the robotic device. When determining the optimized movement sequence, reference may also be made to a plan for optimizing the movement sequence, in particular a kinodynamic plan.
In an embodiment, the method further comprises: an optimized sequence of movements is performed with the robotic device such that the first item is directed into a target pose with reference to the second item.
According to one embodiment, the robotic device has a compensation capability for partially compensating the robotic device for uncertainty in the pose of the first article and/or uncertainty in the pose of the robotic device; and wherein
The determination of the optimal movement sequence comprises determining an optimal compensation performance of the robotic device or further simulating the movement section while taking into account a predetermined compensation performance of the robotic device.
The compensation performance of the robotic device (which is used to partially compensate the robotic device for pose uncertainties) is also referred to below as "compliance". This is, for example, the spring action or damping behavior of the robotic device. Furthermore, the robot apparatus can also be compensated for uncertainty in the posture based on compliance control of the robot apparatus, particularly based on impedance monitoring of the robot apparatus.
When determining an optimized movement sequence, compliance can be determined (in particular calculated). Alternatively, the compliance may already be an input value for a simulation of the moving section, such that the moving section is simulated while taking into account a predetermined compliance of the robotic device.
The compliance of the robotic device is particularly useful for preventing the robotic device from damaging itself, the first item and/or the second item, as it is held beside it/them due to an uncertainty.
According to another embodiment, the uncertainty of the pose of the first item and/or the robotic device is a selected value of a statistical uncertainty distribution.
In particular, an uncertainty of the pose of the first item is modeled based on a first uncertainty distribution, and an uncertainty of the pose of the robotic device is modeled based on another uncertainty distribution. The uncertainty can in particular assume different values. The distribution of these values may be expressed as a statistical uncertainty distribution. From this uncertainty distribution, the uncertainty of the pose of the first article and/or the robotic device can be selected, the values of which are used in simulating the moving section.
According to another embodiment, the uncertainty of the pose of the first item and/or the robotic device is selected from a statistical uncertainty distribution by random sampling.
In particular, that is, the uncertainty of the pose of the first item and/or the robotic device is selected from an uncertainty distribution on a random basis. The selection by random sampling is particularly useful for reproducing the uncertainty density function.
According to another embodiment, a plurality of moving sections are simulated while taking into account different values of uncertainty of the pose of the first article and/or the robotic device.
For example, different uncertainty values are for multiple simulations of the same mobile segment. These different uncertainty values can be derived from the statistical uncertainty distribution.
According to another embodiment, the simulation of the movement sections involves randomly determining a plurality of movement sections, which all originate from the same movement section starting pose of the robot device. The method further comprises the following steps:
randomly selecting a moving segment target pose for one of a plurality of simulated moving segments and defining the selected moving segment target pose as a new moving segment starting pose;
the simulation of the movement segment, the random selection of the movement segment target pose, and the specification of the selected movement segment target pose as a new movement segment starting pose are repeated until the movement segment target pose of one of the plurality of simulated movement segments is within the predetermined target area.
The starting point for the simulated moving segment is initially, in particular, the moving segment starting pose. The moving section starting posture can be a starting posture of the first article. A plurality of moving segments can initiate a gesture simulation from the moving segments. These movement sections are in particular randomly movement sections, all of which have a movement section starting posture as initial posture. However, a single simulated moving segment can have different moving segment target poses.
During the simulation, a single moving segment target pose is selected, in particular randomly, from these multiple moving segment target poses. The selected moving segment target pose is defined as the new starting point, i.e., the new moving segment starting pose. From this new moving segment start pose, multiple moving segments can again be simulated randomly, which in turn can have a different (new) moving segment target pose. One of the (new) moving segment target poses can again be randomly selected in order to use it as a further moving segment starting pose.
The steps of simulating a movement segment, randomly selecting a movement segment target pose, and specifying the selected movement segment target pose as a new movement segment start pose can be iteratively repeated. In particular, they are repeated until a moving segment target pose simulating a moving segment is detected in the target area. The target area comprises in particular the target gesture.
When the moving segment target pose is in the target region, the simulation can be stopped and the individual moving segments that result in the target region can be put together in order to obtain an optimized movement sequence.
According to another embodiment, the optimized movement sequence of the robotic device is determined by putting at least two of the simulated movement sections together.
According to another embodiment, the simulation of the moving section takes into account at least one robotic device performance, a first item performance of the first item, and/or a second item performance of the second item.
According to another embodiment, the robotic device performance includes at least:
a maximum force that can be applied by the robotic device;
a model representation of the robotic device; and/or
Possible movements of the robotic device.
The maximum force that can be applied by the robotic device is in particular the force limit of the robotic device in a particular direction. The model representation of the robotic device can be a CAD (computer aided design) model and for example indicate how large the robotic device is. Possible movements of the robotic device also particularly represent those movements that the robotic device is capable of performing.
According to another embodiment, the first article property and/or the second article property (also referred to hereafter as "article properties" collectively) comprises at least:
a coefficient of friction of the first article and/or the second article;
the size and/or shape of the first article and/or the second article;
a model representation of the first item and/or the second item;
the weight of the first and/or second article; and/or
Material properties of the first article and/or the second article.
The model representation of the first and/or second item can be a CAD model. Since the optimal movement sequence is determined while taking into account the robot device performance and/or the article performance, the optimal movement sequence can be determined in a dedicated manner.
According to another embodiment, the determination of the optimized movement sequence comprises:
calculating a cost function for each simulated moving segment; and
an optimized movement sequence is determined based on the calculated cost function.
The cost function is in particular an indication of the suitability of the analog mobile section.
According to one embodiment, the method further comprises:
determining a plurality of movement sequences based on the simulated movement segments;
calculating a total cost function for each movement sequence based on the cost functions of the movement segments of the respective movement sequence; and
selecting the movement sequence having the smallest total cost function from the plurality of movement sequences as the optimized movement sequence.
A mobile sequence compiled from a plurality of mobile segments can have a total cost function that is the sum of the cost functions of the individual mobile segments of the mobile sequence. The optimized move sequence can be the move sequence whose total cost function is minimal over multiple passes.
According to one embodiment, the method further comprises:
determining a first movement sequence based on the simulated movement segment;
determining a moving sequence portion comprising at least two moving segments;
calculating a cost function for the portion of the moving sequence; and
the portion of the movement sequence is further tracked if the cost function of the portion of the movement sequence is lower than the cost function of the first movement sequence.
The movement sequence is in particular determined in steps. With each step, a mobile segment can be appended to an existing mobile sequence portion, and a check can be performed to determine if the cost function of the resulting mobile sequence portion exceeds a predetermined value. The predetermined value can be a cost function of the previously determined movement sequence, i.e. the first movement sequence. In particular, if the cost function of the part of the moving sequence is already higher than the cost function of the first moving sequence, the part of the moving sequence is not tracked further. That is, for example, no further movement sections are appended in order to obtain a complete movement sequence. In particular, it can be ensured that only parts of the movement sequence that are good in terms of the cost function are further tracked. This allows a significant reduction in computational effort. In particular, a solution for optimizing a movement sequence can be found with little computational effort.
According to a further embodiment, the cost function comprises an indication of:
simulating the mass of the moving section;
a length of a path corresponding to the simulated movement section;
simulating a duration of execution of the mobile segment;
simulating energy consumption for execution of the mobile section; and/or
Forces occurring on the first and/or second article during execution of the simulated movement section.
The path (also referred to as the path of movement) is in particular the path along which the first article moves when the robotic device executes the simulated movement section.
According to another embodiment, the uncertainty of the pose of the first item represents an uncertainty of a model for the first item and/or the second item and/or a production tolerance of the first item and/or the second item.
For example, the uncertainty of the model used can be an error in the model representation of the first item and/or the second item. For example, such an error can be that the model does not take into account the elasticity of the first and/or second item.
According to a second aspect, a computer program product is presented, which causes a program-controlled apparatus to perform the method according to the first aspect or according to an embodiment of the first aspect.
For example, a computer program product, such as a computer program means (means), can be provided or supplied as a storage medium, such as for example a memory card, a USB memory stick, a CD-ROM, a DVD or other form of file downloadable from a server in a network. In a wireless communication network, this can be achieved by transmitting an appropriate file containing a computer program product or a computer program appliance, for example.
According to a third aspect, an apparatus is presented for determining an optimized movement sequence of a robotic device for moving a first item such that the first item is brought to a target pose despite an uncertainty of a pose of the first item with reference to a second item and/or despite an uncertainty of a pose of the robotic device. The apparatus comprises:
a simulation unit for simulating a moving section of the robotic device while taking into account an uncertainty of the pose of the first article and/or an uncertainty of the pose of the robotic device; and
a determination unit for determining an optimized sequence of movements of the robotic device while taking into account simulated movement segments and constraints indicative of at least a starting pose and a target pose of the first item.
The respective units, for example the simulation unit or the determination unit, can be implemented in hardware and/or also in software. In the case of a hardware embodiment, the respective unit can be in the form of a device or a part of a device, for example in the form of a computer or a microprocessor or in the form of a control computer of the vehicle. In the case of a software implementation, the respective units can be in the form of a computer program product, a function, a routine, a part of program code, or an executable object.
According to an embodiment, the device is adapted for performing the method according to the first aspect or according to an embodiment of the first aspect.
The described embodiments and features for the proposed method apply correspondingly for the proposed device.
Further possible implementations of the invention also include combinations of features or embodiments not explicitly mentioned above or below in connection with the exemplary embodiments. Those skilled in the art will also add separate aspects as an improvement or supplement to the corresponding basic form of the invention.
Drawings
Further advantageous configurations and aspects of the invention are the subject of the dependent claims and also the subject of the following exemplary embodiments of the invention. The invention is explained in more detail below on the basis of preferred embodiments with reference to the drawings.
FIG. 1 shows an example of a manufacturing apparatus in which a first article is in a starting position;
FIG. 2 shows an example of a manufacturing apparatus in which a first item is in a target pose;
fig. 3 shows a method for determining an optimized movement sequence of a robotic device according to a first embodiment;
fig. 4 shows a method for determining an optimized movement sequence of a robotic device according to a second embodiment;
FIG. 5 shows an example of a first item being moved along an optimized sequence of movements; and
fig. 6 shows an apparatus for determining an optimized movement sequence of a robotic device according to an embodiment.
In the figures, elements that are identical or have the same function have been provided with the same reference numerals, unless otherwise indicated.
Detailed Description
Fig. 1 shows an example of a
The first article 1 is a cylindrical piece made of plastic. The size of the first article 1 is selected such that the first article 1 can be inserted into the
The robot apparatus 3 is a robot for assembling components. The robotic device 3 is programmable to perform different sequences of movements. For example, it can guide a first article 1 into a
Fig. 1 depicts how the
The posture of the first article 1 represents a combination of the position and orientation of the first article, and is defined relative to the posture (position and orientation) of the
The pose of the first item 1 with reference to the
Further, the posture of the robot apparatus 3 also has a certain degree of uncertainty.
The movement sequence specified for the robotic device 3 reliably guides the first article 1 into its target posture ZP, regardless of the uncertainty of the posture of the first article 1 and regardless of the uncertainty of the posture of the robotic device 3. Such a movement sequence is particularly referred to as an "optimized movement sequence".
The robotic device 3 has a compensation capability (hereinafter also referred to as "compliance") for at least partially compensating for the uncertainty in the pose of the first article 1 and/or the uncertainty in the pose of the robotic device 3. For this purpose, the robot device 3 has a spring system, not depicted.
The optimized movement sequence can be determined using a method for determining an optimized movement sequence of the robotic device 3. Fig. 3 shows an example of such a method according to the first embodiment.
In an (optional) preparation step S0, a robotic device 3 and two
In step S1, the movement section of the robot device 3 is simulated. The movement section corresponds to a movement sequence of the robot apparatus 3. The mobile section is simulated using a solid state simulator. At least thirty mobile sections are simulated in total. The simulation of all moving sections is achieved while taking into account the same predetermined robot performance and the same first and second article performance of the first and
The simulation of the different movement sequences takes into account different values of the uncertainty of the attitude of the first article 1 and the uncertainty of the attitude of the robotic device 3. The uncertainty values are selected from a statistical uncertainty distribution on a random basis.
In step S2, an optimized movement sequence is determined using the movement segments simulated in step S1. This is achieved while taking into account the starting posture SP and the target posture ZP of the first article 1. In step S2, the multiple instances of the simulated movement sections are put together such that they form a movement sequence of the robotic device 3 that allows the robotic device 3 to guide the first article 1 from the starting posture SP into the target posture ZP. Such a movement sequence, which is composed of a plurality of simulated movement segments, corresponds to an optimized movement sequence. Based on the optimized movement sequence, the robot apparatus 3 is able to reliably guide the first article 1 into the target posture ZP, regardless of the uncertainty of the posture of the first article 1 and regardless of the uncertainty of the posture of the robot apparatus. The elasticity applied by the robotic device 3 is determined as part of an optimized movement sequence.
Fig. 4 shows a method for determining an optimized movement sequence of a robotic device according to a second embodiment. Step S0 of the method according to the second embodiment corresponds to step S0 already described in relation to the method according to the first embodiment (fig. 3) and will therefore not be described again.
In the method according to the second embodiment, the simulation step S1 includes steps S10 to S13. In step S10, a plurality of moving segments all starting from the same moving segment start posture of the robot apparatus are randomly determined. The moving section starting posture for the first pass is the starting posture of the first article 1. The simulated moving segments have different moving segment target poses.
In step S11, it is determined whether one of the moving segment target postures is in the target area. The target area can comprise a target pose of the first item 1.
If there is no moving segment target gesture in the target region, steps S12 and S13 are performed. In step S12, a moving segment target pose of one of the plurality of simulated moving segments is randomly selected. In step S13, the moving segment target posture selected in step S12 is defined as a new moving segment start posture.
After steps S12 and S13, steps S10 and S11 are performed again. That is, a plurality of moving segments are simulated again, this time all of the moving segments have the same new moving segment start posture as the moving segment start posture.
Steps S10, S11, and possibly steps S12 and S13 are repeated until it is determined in step S11 that one of the moving section target postures obtained based on the simulation is in the target area.
In this case, step S2 is executed. Step S2 includes steps S20, S21, S22, S23, and S24. In step S20, simulated moving segments that have caused the moving segment target posture to exist in the target area are combined. These combined movement segments form a movement sequence.
In step S21, a total cost function of the movement sequence formed in step S20 is calculated. This is achieved by calculating and aggregating cost functions of the individual mobile segments forming the mobile sequence. In this case, the cost function is an indication of the duration of execution of the simulated moving segment. It is desirable to minimize the duration of execution of the entire movement sequence as this allows for an increase in the amount of manufacturing in a given time. Thus, the calculation of the total cost function serves to simplify the determination of the optimized movement sequence.
In step S22, it is determined whether the total cost function is lower than a previously calculated total cost function of the previously determined (first) movement sequence. If this is the case, the movement sequence determined in step S20 is maintained (step S23). If the total cost function is not lower than the previously calculated total cost function of the previously determined (first) movement sequence, the movement sequence determined in step S20 is rejected (step S24). If the movement sequence determined in step S20 is the first movement sequence, step S23 is performed.
Steps S1 and S2 can be repeated as many times as necessary to minimize the movement sequence determined in step S20. The repetition is depicted in fig. 4 using arrow W. After a predetermined number of repetitions (i.e., passes), the last movement sequence held in step S23 is determined as the optimized movement sequence in step S3.
In step S4, the robotic device 3 moves the first article 1 along the optimized movement sequence specified in step 3 and in doing so guides the first article 1 into the target posture ZP. This movement of the first article 1 is shown in fig. 5.
Fig. 5 shows an example of a
Fig. 5 depicts postures I to V of the
In each of the postures I to V depicted in fig. 5, the
In each of the poses II, III and IV,
The
Fig. 6 shows an
The
The
Although the present invention has been described based on exemplary embodiments, it can be modified in various ways. For example, moving sequence portions may be determined and a cost function calculated for those moving sequence portions while the moving segments are actually being simulated. In particular, only the part of the movement sequence having a cost function lower than the cost function of the previously determined movement sequence is further tracked. It is also conceivable that the robot device 3 places or presses the first article 1 on the