Automatic Robot Program Generation via State Space Discretization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for programming industrial robots are inadequate in handling additional degrees of freedom and variations in work piece movements or process parameters, leading to unreliable and non-robust robot programs, especially when dealing with moving work pieces and unknown parameter fluctuations.
Innovation Solution
A system and method for automatically generating optimized robot programs that involve discretizing state spaces to simulate and evaluate different parameter combinations and motion patterns, iteratively modifying tool paths to minimize cost functions and ensure reliable operation despite variations, using virtual robot controllers for precise simulation and collision avoidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If conventional offline programming methods are used, then programming time is reduced compared to online teaching, but the generated robot programs are not robust against work piece movements and parameter variations
Solution Approach 1:
The system performs preliminary simulation and optimization of robot programs against various disturbances (work piece movements, parameter variations) before actual execution. By pre-evaluating and optimizing programs for multiple possible states, the system ensures robustness without requiring time-consuming online adjustments during production.
Solution Approach 2:
The system uses virtual robot controllers and simulation environments to provide feedback on program performance under various conditions. This feedback loop allows iterative optimization of robot programs to handle work piece movements and parameter variations, improving robustness while maintaining efficient offline programming.
2Measurement precision
If virtual robot controllers with accurate models are implemented, then simulation precision is improved, but system complexity and computational requirements increase
Solution Approach 1:
The system creates virtual copies (models) of the robot controller and work cell environment to perform simulations. These virtual models replicate the essential behaviors and characteristics of the physical system, enabling accurate prediction of program performance without requiring the physical robot to be present or modified.
Solution Approach 2:
The virtual robot controller serves multiple functions: it simulates robot behavior, evaluates program robustness, optimizes trajectories, and predicts performance under various disturbances. This multi-functional approach consolidates multiple tools into a single system, managing complexity while providing comprehensive offline programming capabilities.
3Manufacturing precision
If state space discretization is performed to evaluate multiple parameter combinations, then program optimization is improved, but computational time and resources increase
Solution Approach 1:
The system divides the continuous state space into discrete states, allowing systematic evaluation of different work piece positions, orientations, and parameter variations. This segmentation enables comprehensive optimization without requiring infinite computational resources, as the discretized space can be efficiently explored and optimized.
Data Source
AI summary
Workflow charts for processing (e.g., treating, machining) a workpiece with a tool of an industrial robot are automatically generated. An initial chart has a plurality of tool paths for a workpiece in a defined target position and for defined process parameters. The tool path determines the desired movement of the tool along the workpiece. A state space describing variable parameter values that impact the workpiece processing are defined. Each point in the space represents a concrete combination of possible parameter values. The space is discretized into individual states. The processing of the workpiece is simulated using the initial chart for one or several discrete states and the simulated process results are evaluated according to a pre-definable criterion. The initial chart is iteratively modified, subsequently workpiece processing is simulated using the modified chart for at least one discrete state, and the simulated processing results are evaluated with a pre-definable cost function.