METHOD FOR HAPTIC INSPECTION OF AN OBJECT

Description

[0001] The application relates to a method for the haptic testing of a physical object according to the preamble of claim 1.

[0002] The manufacturing industry is undergoing a transformation. The variety of product variants is increasing. At the same time, series production and development times are decreasing, with the goal of meeting customer needs flexibly and as individually as possible. At the same time, the demands on the quality of components, assemblies, and systems are increasing.

[0003] In addition to pure functionality, quality criteria also include subjective impressions regarding appearance and value, especially for high-priced capital goods. In addition to visual characteristics, tactile features are becoming increasingly important.

[0004] Robot-based measurement methods are being used on an increasingly large scale. These enable the reproducible assessment of products based on a wide range of objective physical quality criteria.

[0005] According to DE 10 2005 006 575 A1, there is an innovative combination of measurement robotics and methods from psychophysics and experimental psychology. Objective and subjective criteria are combined to conduct a holistic evaluation of product quality.

[0006] Robot-based measuring methods can be found, for example, in DE 10 2016 106 488 B3 or DE 10 2015 102 238 A1.

[0007] A method of the type mentioned above can be found in EP 2 216 144 A2.

[0008] The present invention addresses the problem of managing the increasing complexity of products and the development, production, and evaluation steps required for their manufacture. Objective and subjective quality criteria should be measurable.

[0009] To achieve this, the invention provides a method having the features of claim 1.

[0010] According to the proposed solution, a test bench is fully virtualized, featuring the measuring robot required for quality testing. The virtual test bench is coupled with the physical test bench to perform virtual simulations based on relevant condition information. Quality testing is performed on virtual and physical prototypes, with a coupling between the virtual and physical test benches. The result is a hybrid quality inspection.

[0011] In a virtual test bench, the test equipment, particularly in the form of a robot, is virtualized and coupled with the functional, virtual object. This creates a functional, virtual test system. A method for virtual quality testing is created in which the complete functional behavior of the physical object, such as a door, and the physical test equipment, such as a robot, is simulated in order to test possible test scenarios before the actual setup. For example, the accessibility of test points, measurement speeds, space requirements, elimination of disturbances, and measurements of actuation forces and torques can be determined.

[0012] If the simulated measurements or tests are successful, the simulated processes are transferred to a real system so that the real test process can be programmed.

[0013] This method allows real systems to be programmed at global locations, for example, via the cloud. This results in enormous time savings for the user and reliable proof of the accuracy of the measurements, e.g., defined measuring points, etc.

[0014] The hybrid quality control provides a testing method that describes the planning of the quality control test with the help of the virtual test bench, describes the implementation of a hybrid quality test with the help of the coupled test bench to optimize the position accuracy and movement pattern of the physical measuring robot, describes the coupling of the subjective measurement data determined with the hybrid prototype with the objective measurement data determined in the coupled test bench to a perception quality of the prototype.

[0015] A virtualization of a particularly robot-based quality inspection is carried out with the aim of developing future products through early virtual evaluation of objective and subjective quality criteria. to be able to produce faster and more cost-effectively in fewer development steps by avoiding physical prototypes, to be able to adapt better to the needs of the customer by incorporating subjective evaluation criteria at an earlier stage and to be able to control product quality through targeted, early design to the needs of the customer in such a way that only the actually required, perception-oriented quality is implemented, thus enabling more cost-effective production.

[0016] The invention is defined by claim 1. Further advantageous embodiments are defined by the dependent claims.

[0017] They show: Fig. 1 is a schematic diagram of a first test system not belonging to the invention, and Fig. 2 is a schematic diagram of a second test system.

[0018] The first test system is explained below using an industry-relevant application in the form of a door test simulation. In this application, a measuring robot is used to test a vehicle door for various qualitative and quantitative quality characteristics. These can include, for example, the actuation force, opening speed, or damping properties of the vehicle door system.

[0019] In other words, in the exemplary embodiment, the vehicle door is an object that is physically provided and virtually recreated.

[0020] A virtual testing robot system for simulation and subject-based product validation in immersive development environments is described.

[0021] Based on the Fig. 1 The aim is to explain how a person 10 perceives haptic features on a vehicle door 12 without having to perceive the real, i.e. physical, vehicle door 14 itself.

[0022] For this purpose, Fig. 1a It is intended that an initial analysis of the physical object 14 or prototype "vehicle door" is performed with regard to its function and underlying physical mechanisms. Based on this data, the vehicle door 14 is virtualized. In this process, the virtual vehicle door 12 can also be optimized with regard to its function using data stored in libraries.

[0023] In particular, the physical object, also referred to as the physical prototype, is to be virtualized, based in particular on the Functional Digital Mock-Up (FDMU or FMU) approach. This involves coupling a virtual CAD-based geometric model with a behavioral simulation. A software combination of Dymola for behavioral modeling and Unity3D for geometric representation can be used to construct the specific virtual prototype. The virtual prototype is then hybridized by coupling it with a haptic input device, based in particular on the Smart Hybrid Prototyping (SHP) approach. A multi-axis force and torque sensor system was integrated into the measuring robot.

[0024] According to Fig. 1b The door 12 is represented virtually, for example, in a head-mounted display used by a person. Additionally, the measuring robot 16 is used, which is positioned in the spatial position of the virtual vehicle door 12. The robot 16 is equipped with a movable arm in which force or torque sensors are integrated in order to detect the forces or torques acting on the robot 16, which in turn can be used to draw conclusions about the actuation forces, opening speeds, or damping properties of the virtual door 12.

[0025] If the person 10, i.e. the real person, detects, for example, the virtual handle of the door 12 - the movement of the person or the hand is detected by markings on the hand using, for example, cameras, i.e., for example, an optical tracking system - a section of the robot arm or an element emanating from it is in this position, so that when the door 12 is virtually closed or opened, the arm or element moves with it, with the result that the acting forces / moments can be detected. The person 10 thus gains a haptic impression of the closing or opening movement of the door 12, with the result that parameters relevant to the closing or opening process of a vehicle door, taking into account the subjective impressions of the person 10, can be detected based on the data determined by the robot 16.

[0026] A virtual prototyping of the physical object is performed, with properties of the physical object being mapped onto the virtual object. Using a real measuring device, particularly a robot, virtual movements of the virtual object are then recorded to obtain data that can be considered in the design of the physical object underlying the virtual object.

[0027] In the example, if the Fig. 1 a real person is required to determine quality parameters, the embodiment of the Fig. 2 , which reflects the invention, is aimed at the possibility of virtually determining characteristics or parameters that are mapped to physical objects without the need for a real person.

[0028] According to the invention, a physical test bench is virtualized. The physical test bench comprises a measuring robot, test bench technology such as measurement technology, and a prototype environment such as a partial vehicle. This approach builds on the approaches of robot simulation and couples them with the approaches of the Functional Digital Mock-Up (FDMU or FMU). The virtual test bench and the virtual prototype are then coupled, as previously explained, to form the virtual test system. A quality inspection of various quality characteristics is then carried out using the virtual test system.

[0029] In particular, the invention is characterized by the fact that, in addition to the use of established approaches to robot modeling and simulation, models are used to integrate robot control, particularly the multi-axis force and torque sensors. This enables the virtualization of the measuring robot and the test bench technology. The previously explained virtualization of the physical prototype is extended into a virtual test system to provide a tool for the simple virtualization of the physical test bench elements. A method for virtual quality testing is enabled by the virtual test system created with this tool.

[0030] In the left illustration of the Fig. 2 A test bench 100 is shown purely schematically, in which the function of a vehicle door 102 is to be recorded as a physical object. The test bench 100 has a testing robot 104, which also has a movable arm 106 with force and torque sensors, which interacts with the motor vehicle door 102 as a physical object or physical prototype to close or open it. Additional equipment (test bench technology) may be located in the test bench 100, without this being discussed in detail.

[0031] The measuring robot 104 and the test bench equipment are analyzed with regard to their functions and underlying physical mechanisms. Based on this data, the measuring robot 104 and the test bench equipment are virtualized. A virtual object 202—in this exemplary embodiment, a vehicle door—is then implemented in a virtual test bench 200. The virtual test bench 200 includes—modeled on the physical test bench 100—a virtual measuring robot 204 with an arm 206 having movable limbs. Virtual force and torque sensors are also integrated.

[0032] When virtualizing the test bench, the approaches of Smart Hybrid Prototyping (SHP) and Functional Digital Mock-Up (FDMU or FMU) are used in particular.

[0033] In particular, a software combination of Dymola (for behavioral modeling) and Unity3D (for geometric representation) is used to build the virtual prototype. The virtual prototype can then be compared with reality, refined, and verified. According to the previously explained explanations, the planning of the physical test bench test can thus be carried out, i.e., the results from the virtual test system can be used to plan and optimize the physical test system.

[0034] Virtual quality control planning is then carried out to carry out test bench tests using the virtual test bench 200.

[0035] The tests on the physical object 102 are then transmitted based on the data obtained in the virtual test bench 200. The object to be tested, in the exemplary embodiment the vehicle door 102, is also located in the physical test bench 100.

[0036] An analysis then follows of the deviations and problems occurring in the physical test bench 100 regarding the positioning accuracy and movement patterns of the measuring robot 104 during interaction with the physical prototype 102. The deviations and problems are implemented in the virtual test bench 200 using an optimization and error correction algorithm in order to resolve the deviations and problems. The sensors of the physical test bench 100 are coupled to the virtual test bench 200 so that status information can be imported into the virtual test bench 200. Conversely, a coupling exists between the virtual test bench 200 and the physical test bench 100 in order to import the developed optimizations and error corrections into the physical test bench 100.

[0037] Additional tests can be carried out to specifically trigger position inaccuracies or faulty movement patterns in the physical test bench 100.

[0038] Feedback is established between the physical and virtual test benches 100 and 200 to achieve optimization. A virtual quality inspection is implemented within a real one.

[0039] In further development, the physical test bench is linked to the virtual test bench in order to enable the virtual test bench to detect deviations in the physical test bench using status information, to derive corrections through simulation and to feed these back into the physical test bench. This solution approach is in Fig. 2 marked with the arrow P (planning of the physical test bench test).

[0040] The teaching of the invention provides opportunities to reduce development times and manage the increasing complexity in the development of customized products with partially abstract quality criteria. The ability to virtualize prototypes and make them haptic experiences using a physical measuring robot will significantly reduce the construction of real prototype structures or even eliminate them altogether. Due to the virtualization of prototypes, the measuring robot is able to haptically simulate any prototype system. The possibility of a completely virtual testing system provides a new software tool with which the haptic impressions of human test subjects can be predicted entirely in virtual space, without necessarily having to resort to hardware components.This represents a tool, particularly for providers of complex systems with long supply chains and a high diversity of components, as is prevalent in the automotive industry, with which development times and costs can be reduced and the entire development and production process can be made more flexible.

[0041] Highlights include: a) The physical device can be provided in the spatial position of the virtual object. b) A robot arm with movable limbs can be used as the physical device with integrated force and / or moment sensors. c) During interaction with the virtual object, the forces and / or moments acting on the physical device can be determined and compared with forces and / or moments characteristic of the movement of the physical object. In the event of deviations, feedback can be provided to influence or change properties assigned to the physical object, such as opening speed or damping. d) At least one function and / or mechanism of action of the physical object can be determined and the virtual object can be designed accordingly.

[0042] According to the invention, a physical test bench is virtualized. The physical test bench comprises a measuring robot, test bench technology such as measurement technology, and a prototype environment such as a partial vehicle. This approach builds on the approaches of robot simulation and couples them with the approaches of the Functional Digital Mock-Up (FDMU or FMU). The virtual test bench and the virtual prototype are then coupled, as previously explained, to form the virtual test system. A quality inspection of various quality characteristics is then carried out using the virtual test system.

[0043] In particular, the invention is characterized by the fact that, in addition to the use of established approaches to robot modeling and simulation, models are used to integrate robot control, particularly the multi-axis force and torque sensors. This enables the virtualization of the measuring robot and the test bench technology. The previously explained virtualization of the physical prototype is extended into a virtual test system to provide a tool for the simple virtualization of the physical test bench elements. A method for virtual quality testing is enabled by the virtual test system created with this tool.

[0044] In the left illustration of the Fig. 2 A test bench 100 is shown purely schematically, in which the function of a vehicle door 102 is to be recorded as a physical object. The test bench 100 has a testing robot 104, which also has a movable arm 106 with force and torque sensors, which interacts with the motor vehicle door 102 as a physical object or physical prototype to close or open it. Additional equipment (test bench technology) may be located in the test bench 100, although this will not be discussed in detail.

[0045] The measuring robot 100 and the test bench equipment are analyzed with regard to their functions and underlying physical mechanisms. Based on this data, the measuring robot 100 and the test bench equipment are virtualized. A virtual object 202—in this exemplary embodiment, a vehicle door—is then implemented in a virtual test bench 200. The virtual test bench 200 includes—modeled on the physical test bench 100—a virtual measuring robot 204 with an arm 206 having movable limbs. Virtual force and torque sensors are also integrated.

[0046] When virtualizing the test bench, the approaches of Smart Hybrid Prototyping (SHP) and Functional Digital Mock-Up (FDMU or FMU) are used in particular.

[0047] In particular, a software combination of Dymola (for behavioral modeling) and Unity3D (for geometric representation) is used to build the virtual prototype. The virtual prototype can then be compared with reality, refined, and verified. This allows planning of the physical test bench test to proceed according to the previously explained principles, i.e., the results from the virtual test system can be used to plan and optimize the physical test system.

[0048] Virtual quality control planning is then carried out to carry out test bench tests using the virtual test bench 200.

[0049] The tests on the physical object 102 are then transmitted based on the data obtained in the virtual test bench 200. The object to be tested, in the exemplary embodiment the vehicle door 102, is also located in the physical test bench 100.

[0050] An analysis then follows of the deviations and problems occurring in the physical test bench 100 regarding the positioning accuracy and movement patterns of the measuring robot 104 during interaction with the physical prototype 102. The deviations and problems are implemented in the virtual test bench 200 using an optimization and error correction algorithm in order to resolve the deviations and problems. The sensors of the physical test bench 100 are coupled to the virtual test bench 200 so that status information can be imported into the virtual test bench 200. Conversely, a coupling exists between the virtual test bench 200 and the physical test bench 100 in order to import the developed optimizations and error corrections into the physical test bench 100.

[0051] Additional tests can be carried out to specifically trigger position inaccuracies or faulty movement patterns in the physical test bench 100.

[0052] Feedback is established between the physical and virtual test benches 100 and 200 to achieve optimization. A virtual quality inspection is implemented within a real one.

[0053] In further development, the physical test bench is linked to the virtual test bench in order to enable the virtual test bench to detect deviations in the physical test bench using status information, to derive corrections through simulation and to feed these back into the physical test bench. This solution approach is in Fig. 2 marked with the arrow P (planning of the physical test bench test).

[0054] The teaching of the invention provides opportunities to reduce development times and manage the increasing complexity in the development of customized products with partially abstract quality criteria. The ability to virtualize prototypes and make them haptic experiences using a physical measuring robot will significantly reduce the construction of real prototype structures or even eliminate them altogether. Due to the virtualization of prototypes, the measuring robot is able to haptically simulate any prototype system. The possibility of a completely virtual testing system provides a new software tool with which the haptic impressions of human test subjects can be predicted entirely in virtual space, without necessarily having to resort to hardware components.This represents a tool, particularly for providers of complex systems with long supply chains and a high diversity of components, as is prevalent in the automotive industry, with which development times and costs can be reduced and the entire development and production process can be made more flexible.

[0055] Highlights include: a) The physical device can be provided in the spatial position of the virtual object. b) A robot arm with movable limbs can be used as the physical device with integrated force and / or moment sensors. c) During interaction with the virtual object, the forces and / or moments acting on the physical device can be determined and compared with forces and / or moments characteristic of the movement of the physical object. In the event of deviations, feedback can be provided to influence or change properties assigned to the physical object, such as opening speed or damping. d) At least one function and / or mechanism of action of the physical object can be determined and the virtual object can be designed accordingly.

[0056] The invention is particularly characterized by e) a method for the haptic testing of a physical object by means of a physical testing device in the form of a robot with an arm having movable limbs through interaction between the physical testing device and the physical object, comprising the method steps of analyzing the physical testing device and generating a virtual testing device from the physical testing device, providing a virtual object that corresponds to the physical object, providing first data that are determined during an interaction between the virtual testing device and the virtual object, using the first data for the interaction between the physical testing device and the physical object, wherein the interaction between the virtual testing device and the virtual object is influenced depending on second data determined during the interaction between the physical testing device and the physical object,and / or vice versa. f) The interaction between the virtual test device and the virtual object can be influenced depending on second data determined during the interaction between the physical test device and the physical object. g) A robot with a movable arm that interacts with the physical object can be used as the physical test device. h) A virtual robot with virtually integrated sensors for measuring force and / or torque can be used as the virtual test device. i) The virtual test device can be coupled to the physical test device, whereby data generated from simulations performed in the virtual test device, for example, relating to corrections in the interaction between the object and the test device or to properties of the object, is fed back to the physical test device.

Claims

1. A method for haptic testing of a physical object (102) by means of a physical testing device (104) in the form of a measurement robot, with an arm having movable members and with a sensor system for measuring force and / or torque by interaction between the physical testing device and the physical object (102), comprising the method step - provision of a virtual object (202), corresponding to the physical object (102), in the form of a virtual CAD-based geometry model, by means of a computer, characterized in that the physical measurement robot (104) used in the method is one in which the sensor system for measuring force and / or torque is integrated in the arm having movable members as a force and torque sensor system, that the physical measurement robot (104) and test rig equipment form a physical test rig (100), which is analyzed in respect of its functions and its underlying physical effect mechanisms by means of a computer, that the physical measurement robot (104) and the test equipment are virtualized on the basis of the analysis by means of the computer, wherein virtual force and torque sensors are integrated in a virtual measurement robot (204) generated during virtualization, that the virtual object (202) is implemented in a virtual test rig (200), wherein the virtual test rig (200) comprises the virtual measurement robot (204), that first data determined during an interaction between the virtual measurement robot (204) and the virtual object (202) is provided, that the first data is used for the interaction between the physical measurement robot (104) and the physical object (102), that an analysis of divergences and problems occurring in the physical test rig (100) and relating to positional accuracy and movement sequences of the physical measurement robot (104) is performed during the interaction with the physical object (102), wherein the divergences and problems are implemented in the virtual test rig (200) by means of an optimization and troubleshooting algorithm in the form of second data, wherein the sensor system of the physical test rig (100) is coupled to the virtual test rig (200), wherein status information in the form of divergences and problems is input as second data into the virtual test rig (200), wherein the interaction between the virtual measurement robot (204) and the virtual object (202) is influenced depending on the second data determined during the interaction between the physical measurement robot (104) and the physical object (102), in order to enable the virtual test rig (200) to detect the divergences in the physical test rig (100) using the status information of the physical test rig (100), and to deduce corrections by simulation and feed said corrections back into the physical test rig in the form of first data.

2. The method according to claim 1, characterized in that the first data generated on the basis of simulations performed with the virtual measurement robot (204) relates to corrections in the interaction between the virtual object (202) and the virtual measurement robot (204), or to properties of the virtual object (202).