Motion Data Screening with Key Point Validation for Robot Gestures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in securing and processing motion data for robot gesture generation, leading to inefficient and noisy learning processes due to the inclusion of erroneous data, which degrades the quality of robot interactions.
Innovation Solution
A system and method for screening motion data by extracting key point information, detecting and correcting errors, and analyzing motion variations to selectively construct high-quality data for learning, using a reference motion model and weight functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large amount of motion data is collected to improve robot gesture generation, then the diversity and quality of gesture expressions improve, but the data processing complexity and computational resources required increase significantly
Solution Approach 1:
The patent extracts only the essential motion key point information from large-scale motion data sets rather than processing all raw data. The motion data extraction unit identifies and extracts key points that are most relevant for robot gesture generation, filtering out redundant information and reducing computational complexity while maintaining gesture diversity.
Solution Approach 2:
The patent performs preliminary screening and validation of motion data before the main learning process. By pre-processing data to identify valid motion patterns and remove erroneous data in advance, the system reduces the computational burden during training while ensuring high-quality input data for gesture generation.
2Quantity of substance
If all collected motion data is used for learning, then more motion patterns are covered, but erroneous data and noise degrade the learning quality and generation effectiveness
Solution Approach 1:
The patent implements preliminary validation and screening of motion data using predetermined criteria and reference motion models before data is used for learning. This pre-processing step identifies and filters out erroneous data, ensuring that only high-quality valid motion data enters the learning process, thereby maintaining both data volume and learning quality.
Solution Approach 2:
The patent employs feedback mechanisms where the system continuously evaluates motion data quality during the screening process. Validity determination units assess whether extracted motion key points meet predetermined criteria, and this feedback loop ensures that erroneous data is identified and excluded, improving overall learning quality while preserving useful motion patterns.
3Reliability
If motion data screening is performed to improve data quality, then learning effectiveness improves, but the processing time and computational overhead increase
Solution Approach 1:
The patent divides the motion data screening process into multiple independent modules: motion key point extraction, validity determination, and motion variation analysis. Each module handles a specific aspect of data quality assessment independently, allowing for efficient parallel processing and reducing overall processing time while maintaining comprehensive data quality checks.
Solution Approach 2:
The patent applies different screening criteria and validation methods to different types of motion data and key points based on their specific characteristics. Rather than applying a uniform screening process to all data, the system tailors validation approaches to local data characteristics, improving efficiency by focusing computational resources where they are most needed.
Data Source
AI summary
Provided is a method of screening motion data. The method includes extracting motion key point information from a motion data set, determining whether motion data corresponding to the extracted motion key point information are valid data based on at least one of predetermined auxiliary information and a pre-stored reference motion model, and screening motion data for learning from the motion data set based on a motion variation of the motion data when the motion data are valid data as a result of the determination.