Method and assembly system for automated machine based product assembling

The invention refers to a method for automated machine based product assembling according to the preamble claim 1 and an assembly system for automated machine based product assembling according to the preamble claim 6 .

In discrete manufacturing, products are often assembled from parts . For example , parts are inserted into other parts , press- fitted or otherwise j oined to one another to create the finished product . Application domains are for example industry manufacturing, automotive manufacturing and electronics manufacturing .

Automating these j oining processes can be complex since high accuracies , so for instance less than one millimeter, may be required when inserting parts . Assembly processes are usually automated by constructing special-purpose assembly machines . Often, each step in the assembly process is automated by a special-purpose machine and these machines are arranged in an assembly line where the work piece is transported from one machine to next until the product is completely assembled .

This traditional approach of constructing a special-purpose assembly line is only applied i f high volumes of products are produced on the assembly line . Only then will the high investment in the assembly line by j usti fied . For this reason, a low-volume or possibly also high-mix assembly is rarely automated .

Furthermore , these machines include fixtures and other devices to minimi ze inaccuracies . Such machines cannot be easily re-tasked to assemble other products . Flexible assembly machines exist according to the non-patent literature from "Sawik T. (1999) Flexible Assembly Systems — Hardware Components and Features. In: Production Planning and Scheduling in Flexible Assembly Systems. Springer, Berlin, Heidelberg, https : //doi . org/ 10.1007/ 978-3- 642-58614-9 1" .

But these machines are either highly specialized, e.g. according to post-published European Patent Application with the Application No. 20203916.0-1202 for pick & place of components in "Printed Circuit Board <PCB>"- production or they require a high manual effort in programming, e.g. in body in white lines in car production.

To minimize inaccuracies, traditionally machines include fixtures and other devices to precisely orient parts for assembly. These specialized fixtures and devices add to the complexity of assembly machines and underline their quality as special-purpose machines.

It is an objective of the invention to propose a method and an assembly system for automated machine based product assembling, by which efforts of (i) setting up an automated product assembling and (ii) repurposing the automated product assembling to assemble different products, in particular products of distinguishing application domains such as automotive and electronics, are reduced.

This objective is solved with regard to a method defined in the preamble of claim 1 by the features in the characterizing part of claim 1.

The objective is further solved with regard to an assembly system defined in the preamble of claim 6 by the features in the characterizing part of claim 6.

The main idea of the invention according to the claims 1 and 6 is to assemble a product on a modular basis from product parts to be assembled, wherein the product assembling is split into two separate operations , by i ) automated machine fetching and automated machine placing the product parts part-by-part from a delivery area on optically locali zed distribution fixtures of a modular semantic fixture system at a hand-over area in the course of a logistic distribution operation, ii ) automated machine fetching and automated machine placing a product part from optically locali zed distribution fixtures at the hand-over area on optically locali zed assembly fixtures of the modular semantic fixture system at an assembly workspace in the course of an assembly operation, iii ) computing and executing by automated machine motion generation

- primary kinematic machine-motion-sequences to enable the automated machine fetching and the automated machine placing in the course of the distribution operation and

- secondary kinematic machine-motion-sequences to enable the automated machine fetching and the automated machine placing in the course of the assembly operation, iv) providing a hybrid, at least reactive and deliberative automated machine architecture to enable or ensure , based on a "world model for automated machines" , a three-dimensional model of an assembly environment , information - in particular geometric information - of the product and the product parts and an " automated machine workflow" , the execution of the primary kinematic machine-motion-sequences and the secondary kinematic machine-motion-sequences via the automated machine motion generation by

- initiali zing the "world model for automated machines" according to a configuration information configuring an automated machine and the assembly environment and

- instantiating the "machine workflow" and updating the "world model for automated machines" with the information of the product and the product parts .

A technical explanation, what " the hybrid, at least reactive and deliberative machine architecture" is about and what it meant by, is given in https : / Zen . wikipedia . or g/ wiki /Robo tic paradigm according to the version from August 29 , 2020 and in https : Z Zde . ikipedia . orgZ ikiZAutonomer mobiler Roboter according to the version from October 12 , 2020 .

The advantage of the idea solving the addressed problem ( obj ective of the invention) and summari zed above lies in the reduction of setup costs per assembly . This cost will be mainly driven by the cost of generating a digital model of the product to be assembled, e . g . , design . Since such models are already used today, for manual assembly, the cost of setting up an automated assembly unit will be close to zero using the idea .

One key aspect of the idea is an automated machine motion generation with a collision- free motion or path planning of the automated machine within a workspace of the automated machine executing - as two kinematic chains - primary kinematic machine-motion-sequences enabling a logistic distribution operation of the automated machine and secondary kinematic machine-motion-sequences enabling an assembly operation of the automated machine .

The automated machine motion generation approach on the one hand prevents unintended collision of the automated machine , e . g . , the robot or the two robots . Furthermore , the automated machine motion generation is robust to expected disturbances and tolerances .

For this reason, the solution - the proposed method and the proposed assembly system - solved the addressed problem enables at least one automated machine connected with or assigned to an automated-machine-control-unit which are forming a technical , functional subsystem of the assembly system and designed such that inter alia an automated machine architecture is provided . The proposed assembly system with the at least one automated machine is modular and flexible by design .

The modular and flexible assembly system with the at least one automated machine , consisting of one or more robots forming a robot system, includes besides the automated-machine- control-unit for instance further a product database that holds assembly instructions and geometric information about the product , motion generation software , a three-dimensional model of the environment , a kinematic description of the at least one automated machine respectively the robot system, a system of modular semantic fixtures and a vision system based on at least one camera ( cf . claim 9 ) .

So , the automated machine can be extended to produce new products or variants by adding or removing fixtures and other devices . The machine it is based on one or more robots ( cf . claims 5 and 11 ) . The robots can handle various parts during assembly and automated-machine-control-unit automates the programming of the machine by computing all required motions .

To enable the logistic distribution operation and the automated machine of the automated machine the assembly system consists - for example , but preferably - one robot that retrieves parts for the product assembly, called as a " logistics robot" , and a further robot that assembles the product parts to assemble the product , called as " assembly robot" .

The automated-machine-control-unit computes motions for both robots based on the three-dimensional model of the environment , the kinematic description of the robot system, the layout of the modular semantic fixtures , and the geometric information of the product and its product parts . To compute motions , the automated-machine-control-unit uses the motion generation software and implements the motions , e . g . , by applying a motion planning algorithm . Flexibility in the assembly process is achieved by relying on the modular system of fixtures at the handover area and assembly workspace . The fixtures contain a machine-readable marker that identi fies the fixture and allows it to be precisely located in the workspace ( cf . claims 3 and 8 ) . The camera of the vision system is used to locate the fixtures whenever fixtures are added, moved, or removed . The fixtures ensure that product parts are presented and grasped in a known, precisely defined location and orientation . These fixtures can be speci fic to the product parts being assembled but they can be added and removed as required by the assembly process . For this reason, the assembly workspace remains modular and flexible even though product-specific fixtures are used .

The automated-machine-control-unit computes motions for the logistics robot so that the logistics robot retrieves product parts from a designated delivery area in the assembly environment . The logistics robot retrieves parts to be assembled from the delivery area and places them in fixtures at a handover location that is also reachable by the assembly robot .

The automated-machine-control-unit then computes motions for the assembly robot to grasp the parts from the fixtures in the handover area . The assembly robot grasps the parts and places them at fixtures in an assembly workspace , automated- machine-control-unit computes motions for the assembly robot so that the assembly robot assembles parts to other product parts as instructed in the assembly instructions . After assembly is finished the product is placed in a fixture in the handover area . The logistics robot then takes the product from this fixture and places it at a defined output area ( cf . claims 2 and 7 ) .

The delivery, handover, and output areas as well as the location of the fixtures are speci fied in the three-dimensional model of the assembly environment . To account for inaccuracies in parts delivery at the delivery area, the logistics robot can use the camera of the vision system to locate the product parts .

An extension of this system is a way of speci fying or encoding logistics and assembly instructions in machine-readable identi fiers of the fixtures ( cf . claims 4 and 10 ) . In this way, the process described above would be encoded using the modular semantic fixtures . For example , the logistics robot would first read the identi fier of all fixtures in the handover area . It would then decode the identi fiers or do a corresponding database lookup to determine which product parts to put on those fixtures . It could then retrieve the product parts from the delivery area and place these parts on the fixtures . Similarly, the assembly robot could read the identi fiers of all fixtures to determine all required assembly steps and delivery locations in the handover area .

For clari fication, the assembly system with exactly two robots was explained above , but it is also possible to use within the assembly system with one robot or more than two robots . I f only one robot is used, this robot is used to implement the function of the logistics and assembly robots . I f more than two robots are used then the function of one or both robots is duplicated, e . g . , for performance reasons . Furthermore , it is also possible to use within the assembly system other kinematics instead of robots , such as gantries or delta robots ( cf . claims 5 and 11 ) .

With the invention being presented above , a flexible assembly system or cell can be constructed . This system or cell allows to automate low-volume assembly tasks since it can be adapted to new products with moderate ef fort . All that needs to be changed are the product database , the fixtures , i f they are product-speci fic, and the assembly sequence . I f the assembly sequence is encoded in the fixtures , only the fixtures and product database need to be changed . The system hardware and automated-machine-control-unit , the description of the envi- ronment , the kinematic description of the system, and the motion generation software can be reused .

The flexible modular assembly system or cell will allow for a smaller footprint and a smaller capital investment compared to a traditional production line . While it is unlikely that it will have the same performance as an optimi zed specialpurpose line , several modular assembly systems or cells can be used in parallel to scale performance as required . And while the components of special-purpose production lines rarely are reused in future production lines , the modular assembly system or cell can be re-tasked to many future applications , leading to a reduction in waste .

Moreover, advantageous further developments of the invention arise out of the following description of a preferred embodiment of the invention according to FIGURES 1 and 2 . They show :

FIGURE 1 - as a demonstration setup - an assembly system for automated machine based product assembling within a workspace ,

FIGURE 2 structure and design of the assembly system according to the FIGURE 1 .

FIGURE 1 shows - as a demonstration setup - an assembly system ASY for automated machine based product assembling within a workspace WS as part of an assembly environment AEV to assemble a product PD on a modular basis from product parts PDP to be assembled, wherein the product assembling is split into two separate operations OP . For doing so the assembly system ASY includes two automated machines AMA and an automated- machine-control-unit AMACU connected with the two automated machines AMA executing the two operations OP .

As a modi fication of the assembly system ASY depicted in the FIGURE 1 it is also possible that only one automated machine AMA or more than two automated machines AMA are included in the assembly system ASY, wherein the two operations are executed either both through the one automated machine AMA or they are spread over the more than two automated machines AMA.

The two operations OP are a logistic distribution operation DOP and an assembly operation AOP and the automated machine AMA is preferably a robot , a gantry or a delta robot executing the distribution operation DOP and the assembly operation AOP .

In the course of the logistic distribution operation DOP one of the two automated machines AMA - in the FIGURE 1 the left one - ( i ) fetches in a first step from a delivery area DA of the workspace WS , where the product parts PDP are located, product parts PDPDA part-by-part for the distribution to assemble the product PD in an automated manner and ( ii ) places in a second step the fetched product parts PDPDA part-by-part each on a distribution fixture DF of a modular semantic fixture system at a hand-over area HOA of the workspace WS for product parts to handover . This modular semantic fixture system with the distribution fixtures DF is product part speci fic and assembly process flexible .

In order to carry the automated machine fetching and the automated machine placing the said automated machine AMA includes a camera CAM for locali zing the distribution fixture DF in the hand-over area HOA optically . To locali ze the distribution fixtures DF, they are marked by automated machine- readable markers MAR, which precisely locate the distribution fixtures DF at the hand-over area HOA. In addition, according to a preferred embodiment the automated machine-readable markers MAR include identi fiers ID identi fying the distribution fixtures DF to facilitate the fetching and placing process . In the course of the assembly operation AOP the other of the two automated machines AMA - in the FIGURE 1 the right one - ( i ) fetches in a first step from the distribution fixture DF a product part PDPDF of the product parts PDPDF and ( ii ) places each in a second step the fetched product part PDPDF on an assembly fixture AF of the modular semantic fixture system at an assembly workspace AWS of the workspace WS for product parts to assemble . The modular semantic fixture system with the fixtures DF, AF is also product part speci fic and assembly process flexible .

In order to carry the automated machine fetching and the automated machine placing the said automated machine AMA also includes a camera CAM for locali zing the distribution fixtures DF at the hand-over area HOA and the assembly fixtures AF of the modular semantic fixture system at the assembly workspace AWS optically . To also locali ze the assembly fixtures AF, they are marked also by automated machine-readable markers MAR, which precisely locate the assembly fixtures AF at the assembly workspace AWS . Again, in addition according to a preferred embodiment the automated machine-readable markers MAR include identi fiers ID identi fying the assembly fixtures AF to facilitate the fetching and placing process .

As soon as the product PD is assembled according to the assembly operation AOP it is placed by the other automated machine AMA - in the FIGURE 1 the right one - in or on a product fixture PF of the modular semantic fixture system at the hand-over area HOA and afterwards the product PD being placed in or on the product fixture PF is handed over to an output area OPA of the workspace WS by the one automated machine AMA - in the FIGURE 1 the left one .

FIGURE 2 shows structure and design of the assembly system ASY including - according to the FIGURE 1 - the at least one automated machine AMA and the automated-machine-control-unit AMACU, which form a technical , functional subsystem SSY . Regardless from the cited design of the automated machine AMA it is essential for the assembly purpose that the automated machine AMA is able to carry out the two operations OP, so

( i ) the logistic distribution operation DOP and ( ii ) the assembly operation AOP .

This means that in the case of a single robot both operations , the logistic distribution operation DOP and the assembly operation AOP, the single robot must be able to carry out both operations , whereas in the case of two robots or two gantries or two delta robots as depicted in the FIGURE 2 one robot carries out the logistic distribution operation DOP and the other the assembly operation AOP .

The automated machine AMA and the automated-machine-control- unit AMACU of the assembly system ASY, which work together technically and functionally, form a hybrid, at least reactive and deliberative automated machine architecture AMAAT . Such a machine architecture is described for example https : / Zen . ikipedia . org/wiki/ Robo tic paradigm according to the version from August 29 , 2020 and in https : / /de . wikipedia . org/wiki/Autonomer mobiler Roboter according to the version from October 12 , 2020 .

According to this automated machine architecture AMAAT ( i ) a "world model for automated machines" -module AMAWM-M, being part of the automated-machine-control-unit AMACU, implements a "world model for automated machines " as a digital twin to formulate correct behavioral sets of the automated machine AMA, so-called automated machine-behavioral sets , being used during a run-time of the automated machine AMA, ( ii ) a three- dimensional model of the assembly environment AEV - being stored in a data repository DRP and including inter alia model data of the distribution fixtures DE, the assembly fixtures AF, the hand-over area HOA, the distribution area DA, the output area OPA and the assembly workspace AWS within which and information INF, preferably geometric information, of the product PD and the product parts PDP, PDPDA, PDPDF with which the distribution operation DOP and the assembly operation AOP are executed, and ( iii ) a " automated machine work- flow" -module AMAWF-M, being also part of the automated- machine-control-unit AMACU, implements a workflow of the automated machine AMA, a so-called " automated machine workflow" , by which a piece of automated machine work to carry out the distribution operation DOP and the assembly operation AOP is passed in series of automated machine stages from initiation to completion .

The automated-machine-control-unit AMACU includes further an " automated machine motion generation" -module AMAMG-M and an " automated machine motion execution" -module AMAME-M . The " automated machine motion generation" -module AMAMG-M thereby provides prv a collision- free motion or path planning of the automated machine AMA within the workspace WS of the assembly environment AEV, where according to the FIGURE 1 the product PD is assembled on the modular basis from the product parts PDP to be assembled, to the " automated machine motion execu- tion" -module AMAME-M . Based on this provision the " automated machine motion execution" -module AMAME-M executes exe on the automated machine AMA primary kinematic machine-motionsequences MAMSki to enable the automated machine fetching and the automated machine placing of the at least one automated machine AMA in the course of the distribution operation DOP and secondary kinematic machine-motion-sequences MAMSk2 to enable the automated machine fetching and the automated machine placing of the at least one automated machine AMA in the course of the assembly operation AOP .

Within automated-machine-control-unit AMACU of the assembly system ASY the "world model for automated machines" -module AMAWM-M, the " automated machine workf low" -module AMAWF-M, the " automated machine motion generation" -module AMAMG-M and the " automated machine motion execution" -module AMAME-M form a technical , functional subsystem SSY . Forming this technical , functional subsystem SSY the "world model for automated ma- chines" -module AMAWM-M is updated upd by the " automated ma- chine workf low" -module AMAWF-M and the " automated machine motion generation" -module AMAMG-M is updated upd and initiali zed il z on one hand by the "world model for automated ma- chines" -module AMAWM-M and is requested req on the other by the " automated machine workf low" -module AMAWF-M .

Furthermore , the technical , functional subsystem SSY enables or ensures the execution of the primary kinematic machinemotion-sequences MAMSki and the secondary kinematic machinemotion-sequences MAMSk2 via the " automated machine motion generation" -module AMAMG-M and the " automated machine motion generation" -module AMAMG-M in such a way that

(a) a configuration information CINF configuring the automated machine AMA and the assembly environment AEV is inputted ipt into in the "world model for automated machines" -module AMAWM-M and thereby initiali zing il z the "world model for automated machines" and

(b) a design of the product PD and the product parts PDP, PDPDA, PDPDF in form of the information INF, being stored also in the data repository DRP, is inputted ipt

(bl ) into the " automated machine workf low" -module AMAWF-M thereby instantiating itt the " automated machine workflow" and

(b2 ) into in the "world model for automated machines" -module AMAWM-M thereby updating upd the "world model for automated machines " .

The configuration information CINF configuring the automated machine AMA and assembly environment AEV with the workspace WS as well as the design of the product PD and the product parts PDP, PDPDA, PDPDF in form of the information INF can be taken from the data repository DRP or databases in general by PUSH- or PULL-based data trans fer, but also a file-trans fer or a trans fer via an electronic interface is possible .