Robot Arm Force Control Tuning Under Workpiece Variation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for setting force control parameters in robots with force control systems are inefficient, as they often result in overfitting and are not suitable for real-world operations with variations in manufacturing and gripping conditions, making it difficult to balance required force and torque while maintaining productivity.
Innovation Solution
A method that adjusts force control parameters by iteratively updating candidate values based on working time and applied force, with noise addition to simulate variations in work conditions, until convergence is achieved, allowing for the determination of optimal force control parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If force control parameter is set by trial-and-error method with limited conditions, then the parameter can be determined for specific conditions, but the parameter becomes overfitted and not suitable for real operation with variations
Solution Approach 1:
The patent transforms the static force control parameter setting into a dynamic optimization process. By introducing iterative optimization with noise addition, the system continuously adapts parameters based on feedback from working time and force data, enabling the parameters to dynamically adjust to variations in manufacturing and gripping conditions rather than being fixed for specific conditions only
Solution Approach 2:
The patent systematically changes force control parameters through iterative optimization. The optimization process modifies parameters like stiffness and damping coefficients based on performance feedback, and the added noise introduces controlled variations to explore the parameter space, preventing overfitting while finding optimal settings for real-world variations
2Reliability
If trial-and-error method is used to set force control parameter, then the parameter can be optimized for specific work conditions, but the process requires repeated work trials and is time-consuming
Solution Approach 1:
The patent implements a feedback mechanism where the optimization process uses measured working time and force data from actual operations to update force control parameters. This closed-loop feedback allows the system to learn from real performance data and automatically adjust parameters, eliminating the need for repeated manual trial-and-error testing while improving reliability
Solution Approach 2:
The optimization system performs self-adjustment of force control parameters using automatically collected operational data. The system serves itself by autonomously optimizing parameters based on its own performance metrics without requiring external intervention or repeated manual testing, significantly reducing the time investment required
3Productivity
If force control parameter is optimized for specific conditions without variations, then the parameter achieves required performance under those conditions, but it cannot handle variations in manufacturing and gripping
Solution Approach 1:
The patent performs preliminary optimization with noise addition before actual production. By pre-optimizing parameters with introduced variations and then fine-tuning with actual operational data, the system prepares robust parameters in advance that can handle manufacturing and gripping variations, preventing productivity losses from parameter mismatch during production
Data Source
AI summary
A method of adjusting a force control parameter includes a first step of moving a robot arm based on first information on work start position and orientation of the robot arm and a candidate value of a force control parameter, and acquiring second information on a working time taken for the work and third information on a force applied to the robot arm during the work, and a second step of acquiring an updated value obtained by updating of the candidate value of the force control parameter based on the acquired second information and third information, wherein the first step and the second step are repeatedly performed with noise added to the first information until the acquired working time or force applied to the robot arm converges, and a final value of the force control parameter is obtained.