Robot Posture Planning for Variable Grasp and Workpiece States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot posture determination methods struggle to efficiently account for variations in grasping states of workpieces, leading to prolonged trial and error in determining optimal robot postures for tasks like grasping and moving objects.
Innovation Solution
A robot posture determination device and method that acquires work, grasping, and workpiece posture information, and uses a searching unit to associate and search for sequences of robot postures that minimize operation time by considering relative positional relationships, kinematics, and potential interference with the environment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a robot posture determination method based on skilled worker intuition and trial and error is used, then the robot can perform predetermined work, but the time required to determine the optimal posture becomes excessively long
Solution Approach 1:
The patent replaces the mechanical trial-and-error approach with a computer-based automated calculation system. The determination unit computes optimal postures by calculating movement times for multiple candidate postures and selecting the best combination, substituting human intuition and physical experimentation with algorithmic optimization.
Solution Approach 2:
The patent performs preliminary calculation of movement times for all candidate postures before final determination. By pre-calculating and storing movement time data for various posture combinations, the system prepares optimization information in advance, enabling rapid selection of the optimal posture without iterative trial and error during actual operation.
2Device complexity
If the robot hand grasping state is specified in advance, then the posture determination can be simplified, but the system cannot account for variations in grasping states of the workpiece
Solution Approach 1:
The patent makes the grasping state dynamic rather than fixed. The determination unit calculates optimal postures for multiple different grasping states and selects the best combination, allowing the system to adapt to variations in workpiece grasping requirements. This dynamic approach enables the robot to optimize posture based on the specific grasping state needed for each task.
Solution Approach 2:
The patent changes the grasping state parameters by considering multiple candidate grasping positions and attitudes. By varying the grasping state parameters and calculating optimal postures for each variation, the system can handle different workpiece types and grasping requirements without increasing overall system complexity.
3Measurement precision
If all combinations of workpiece posture and hand part posture are evaluated, then the optimal posture can be determined accurately, but the calculation complexity and time required increase significantly
Solution Approach 1:
The patent segments the posture determination problem into separate evaluation stages. The determination unit first evaluates workpiece postures, then hand part postures for each workpiece posture, and finally robot body postures. This segmented approach breaks down the complex optimization problem into manageable sub-problems, reducing calculation complexity while maintaining accuracy.
Solution Approach 2:
The patent evaluates multiple candidate postures beyond the minimum single optimal posture. By calculating and comparing a set of candidate postures with different movement times, the system ensures accurate determination of the optimal posture while managing calculation complexity through controlled enumeration of candidates rather than exhaustive search.
Data Source
AI summary
An acquisition unit acquires work information related to work to be performed by a robot having a hand part, grasping information indicating a plurality of candidates for the relative positional relationship between the hand part and a workpiece, workpiece posture information indicating a plurality of candidates for a posture adoptable by the workpiece, and kinematics information regarding the robot. A generation unit generates, for a plurality of target points on a path of the robot for performing the work, a graph which includes a node corresponding to each combination of the workpiece posture information, the posture of the hand part grasping the workpiece, and the posture of the robot, and an edge connecting nodes corresponding to combinations that are transitionable between target points, and in which an estimated operation time corresponding to the transition between the nodes connected by the edge is given to the edge as a weight.