Neural Network Motion Coordination Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for measuring body balance and coordination between body parts require multiple sensors, making them impractical and inefficient.
Innovation Solution
An artificial neural network model training method and apparatus that infers coordination between body parts using similarity in motion data from a subset of body parts, reducing the need for multiple sensors by training a model with motion data from one part to predict coordination with others.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensors are attached to multiple body parts to measure coordination and balance, then measurement precision is improved, but device complexity and the number of sensors required increase
Solution Approach 1:
The patent uses motion data from one body part as a copy or proxy to infer the motion characteristics of other body parts. Instead of directly measuring all body parts with multiple sensors, the system captures motion data from a single part and uses machine learning models to generate inferred motion data for other parts, thereby reducing sensor requirements while maintaining measurement capability
Solution Approach 2:
The patent introduces machine learning models as an intermediary between the single sensor and the coordination measurement. The model acts as a mediator that processes the motion data from one body part and generates inferred motion data for other parts, enabling coordination measurement without direct sensing of all parts
2Measurement precision
If multiple sensors are used to measure motion of multiple body parts, then coordination inference accuracy is improved, but ease of operation deteriorates due to complex sensor attachment
Solution Approach 1:
The patent extracts and removes the requirement for multiple sensors from the system. By taking out the assumption that each body part needs its own sensor, the system achieves coordination measurement with minimal sensors, greatly simplifying the attachment process and improving ease of operation
3Reliability
If motion data from all body parts is collected using multiple sensors, then training data completeness is improved, but loss of substance increases due to more sensors and data processing
Solution Approach 1:
The patent generates inferred motion data as copies of the actual motion data that would be obtained from additional sensors. These inferred copies are created through machine learning models trained on available data, allowing the system to work with incomplete sensor data while maintaining training effectiveness
Data Source
AI summary
An artificial neural network model training method includes acquiring a plurality of motion data items including each motion data item for a plurality of parts of a moving body; calculating coordination between parts of the moving body based on correlation between the plurality of motion data items; and training an artificial neural network model using a training dataset including at least one motion data item among the plurality of motion data items as an input data item, and the coordination between the plurality of parts as a target variable.