Robot Object Grasping With Confidence-Guided Action Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robots struggle to accurately execute untrained tasks and face reduced disambiguation ability when object attributes are similar, leading to low task execution success rates and efficiency.
Innovation Solution
An object grasping method that determines a second confidence of candidate objects based on state description information and first confidence, performs preset actions like grasping and questioning, and uses reward information to determine optimal actions, improving accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If robots use pre-trained tasks for object grasping, then task execution reliability is improved, but adaptability to untrained tasks deteriorates
Solution Approach 1:
The system performs preliminary confidence assessment and object identification before executing the grasping action. By evaluating candidate objects against the target object description in advance and calculating confidence scores, the system prepares the optimal grasping plan without requiring pre-training for every specific task scenario, thus maintaining reliability while improving adaptability.
2Measurement precision
If robots ask more questions to disambiguate objects, then object identification accuracy is improved, but interaction time increases
Solution Approach 1:
The system performs partial questioning by only asking questions when the confidence score of the identified object falls below a threshold. When confidence is sufficient, no further questions are asked, thus reducing unnecessary interactions. This partial action approach maintains high identification accuracy while minimizing interaction time.
Solution Approach 2:
The system uses confidence scores as feedback to determine whether additional questioning is needed. The confidence score reflects the current identification accuracy, and based on this feedback, the system decides whether to ask clarifying questions or proceed directly to grasping, thereby optimizing the balance between accuracy and time.
3Manufacturing precision
If robots perform comprehensive object analysis, then grasping accuracy is improved, but computational complexity increases
Solution Approach 1:
The system performs preliminary analysis by comparing candidate objects against the target object description and calculating confidence scores before making the final grasping decision. This preliminary action structure breaks down the comprehensive analysis into manageable steps, improving grasping accuracy while controlling computational complexity through structured processing.
Solution Approach 2:
The system replaces complex mechanical analysis with a confidence-based decision-making mechanism. Instead of performing exhaustive physical and mechanical analysis of each candidate object, the system uses a simplified confidence score calculation that substitutes comprehensive mechanical evaluation with a more efficient computational approach, maintaining accuracy while reducing complexity.
Data Source
AI summary
The present disclosure provides an object grasping method, an apparatus, a computer device and a storage medium, the method includes: in response to receiving a target object grasping instruction, determining a second confidence of each of candidate objects being a target object based on state description information of the target object and a first confidence of each of the candidate objects in a current scene having each preset state; determining a target candidate object and a reward information after performing each preset action for the target candidate object based on the first confidence and the second confidence; and determining a target preset action to be performed based on the reward information, and performing the target preset action.