Robot End Effector Marking for NC Program Training Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The training process for robots is time-consuming and expensive due to errors in movement caused by forces applied during actuation, leading to the need for multiple parts to be used and modified to adjust Numerical Control (NC) programs, especially in industries with expensive components like aerospace.
Innovation Solution
A system and method using a non-destructive end effector that applies visible marks to parts without altering them, allowing robots to train multiple times on a single part to tune NC programs, utilizing an extendable punch and reflective adhesive tape to create photogrammetric targets for accurate alignment and error detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If robots are trained by performing work on multiple parts to detect and correct location errors, then manufacturing precision is improved, but loss of time and loss of substance increase due to consuming multiple parts
Solution Approach 1:
The patent applies the copying principle by creating virtual copies of physical parts through 3D scanning and digital modeling. These digital twins serve as training objects for robot NC program development, allowing multiple iterations of training and testing without consuming physical parts. The virtual environment replicates the geometric and spatial characteristics of actual parts, enabling accurate robot path planning and parameter optimization before real-world implementation.
Solution Approach 2:
The patent employs disposable virtual training objects that can be freely created, modified, and discarded in the digital environment. These virtual parts serve as cheap substitutes for expensive physical components, allowing extensive training iterations without material cost. The virtual objects can be rapidly generated from 3D scans and modified without manufacturing constraints, enabling efficient robot training.
2Manufacturing precision
If robots are trained by performing work on multiple parts to detect and correct location errors, then manufacturing precision is improved, but loss of time and loss of substance increase due to consuming multiple parts
Solution Approach 1:
The patent applies the copying principle by creating virtual copies of physical parts through 3D scanning and digital modeling. These digital twins serve as training objects for robot NC program development, allowing multiple iterations of training and testing without consuming physical parts. The virtual environment replicates the geometric and spatial characteristics of actual parts, enabling accurate robot path planning and parameter optimization before real-world implementation.
Solution Approach 2:
The patent employs disposable virtual training objects that can be freely created, modified, and discarded in the digital environment. These virtual parts serve as cheap substitutes for expensive physical components, allowing extensive training iterations without material cost. The virtual objects can be rapidly generated from 3D scans and modified without manufacturing constraints, enabling efficient robot training.
3Manufacturing precision
If NC programs are adjusted based on errors detected from working on multiple parts, then manufacturing precision is improved, but device complexity increases due to the training process
Solution Approach 1:
The patent replaces the mechanical training process with an information-based digital system. Instead of physically moving robots through multiple training iterations on actual parts, the system uses 3D scanning, digital twin creation, and virtual simulation to pre-calculate and optimize NC programs. This substitution of mechanical training with digital modeling simplifies the overall process while maintaining or improving precision.
Solution Approach 2:
The patent applies preliminary action by performing all necessary training, testing, and optimization in the virtual environment before actual manufacturing begins. The digital twin allows pre-validation of NC programs, pre-identification of potential errors, and pre-optimization of robot paths. This preliminary work in the virtual domain eliminates the need for complex iterative training on physical equipment.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods are provided for placing non-destructive marks onto a part via an end effector of a robot. One embodiment is a system comprising an end effector of a robot. The end effector includes an extendable punch that places targets onto a part, and supports that hold a strip of reflective adhesive tape between the punch and the part. Extending the punch cuts out a target from the strip and applies an adhesive side of the target to the part, and retracting the punch leaves a reflective side of the target visible on the part.