Robot Hand Grasp Planning and Arm Motion Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional grasp planning methods for humanoid robots integrate motion and grasp planning, leading to inefficient, unstable, and unnatural object grasping, as they do not separate the planning processes, resulting in prolonged grasping times and potential failure to achieve the desired grasp for the working purpose.
Innovation Solution
A robot system that separates grasp planning for its hands from motion planning for its arms, using a database to store grasp motions, a grasp motion generation unit to calculate scores based on hand positions and obstacles, and an arm path generation unit to create paths based on high-scoring grasp motions, allowing for more efficient and stable grasping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If motion planning and grasp planning are integrated and executed simultaneously, then the system can operate with a unified planning framework, but the calculation time increases and the system may fall into infinite loops
Solution Approach 1:
The patent divides the integrated planning system into two separate modules: motion planning and grasp planning. These modules operate independently with distinct calculation processes, allowing parallel execution and reducing total computation time while avoiding infinite loops that occur when both are coupled in a single unified framework.
2Reliability
If grasp planning is performed through multiple grasp motions, then the robot can achieve proper grasping of target points, but the grasping process becomes slow and unstable
Solution Approach 1:
The system pre-calculates and stores multiple grasp motions in a database before actual grasping occurs. During execution, the robot quickly retrieves pre-planned grasp motions that are suitable for different objects, avoiding real-time calculation delays and ensuring stable grasping through pre-validated motion sequences.
Solution Approach 2:
The patent introduces a grasp score parameter that evaluates different grasp motions based on multiple criteria (stability, naturalness, task suitability). By changing the evaluation parameters and selecting motions with optimal scores, the system achieves both speed and stability in grasping operations.
3Measurement precision
If the final desired pose is judged after all motion planning and grasping planning are completed, then comprehensive evaluation is achieved, but the robot hand motion may fall into an infinite loop
Solution Approach 1:
The system implements intermediate feedback mechanisms where grasp scores are calculated and evaluated during the grasp planning phase, before motion execution begins. This allows the system to validate grasp feasibility early, provide feedback for motion planning adjustments, and avoid infinite loops by detecting potential issues before full execution.
Data Source
AI summary
A robot and a control method thereof which execute grasp planning of hands of the robot separately from motion planning of arms of the robot so as to apply a result of the grasp planning of the hands of the robot to the motion planning of the arms of the robot, and thus more rapidly, naturally and stably grasp an object in a grasp manner suited to a desired working purpose and judge whether or not grasping is executable prior to the motion planning of the arms of the robot, thereby more stably grasping the object.