Motion Capture Sensor Error Compensation via Time-Filtered Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion capture techniques face challenges in maintaining high accuracy over time due to error accumulation and require extensive learning data, making them labor-intensive and costly.
Innovation Solution
An information processing apparatus and method that calculates learning labels from sensor values, including angular velocity and acceleration, to learn inference parameters for posture, speed, and position, focusing on accurate data within a predetermined time frame to reduce error and simplify the learning process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If measurement results are used after a predetermined time has elapsed, then more measurement data is available, but measurement accuracy decreases due to error accumulation
Solution Approach 1:
The system pre-calculates and stores correction values during periods when high-accuracy reference measurements are available (when error accumulation is minimal). These correction values are then applied to subsequent measurements to compensate for error accumulation over time, enabling accurate measurements even after extended periods without reference updates.
2Measurement precision
If learning is performed using enormous amounts of learning data to achieve high accuracy for all people, then detection accuracy is improved, but labor and cost increase
Solution Approach 1:
The system enables each user to perform self-learning by simply wearing the sensor unit and performing their own movements. The learning unit automatically generates personalized correction values based on the user's unique movement patterns without requiring external intervention or large datasets from multiple users. This self-service approach achieves high detection accuracy for each individual while minimizing labor and cost.
3Measurement precision
If only measurement results within a predetermined time frame are used, then measurement accuracy is maintained, but the measurement duration is limited
Solution Approach 1:
The system continuously monitors measurement accuracy and automatically triggers reference measurements when error accumulation reaches a threshold. The correction values derived from these reference measurements are fed back to compensate for accumulated errors, enabling the system to maintain high measurement accuracy over extended periods indefinitely, rather than being limited to fixed time frames.
Data Source
AI summary
The present disclosure relates to an information processing apparatus, an information processing method, and a program capable of implementing highly accurate motion capture.The posture, speed, and position of the sensor unit are calculated by integrating the angular velocity and the acceleration detected by the sensor unit. Since at least one of the calculated posture, speed, or position of the sensor unit is obtained by integration, the accuracy decreases with the lapse of time, and thus an inference parameter for inferring at least one of the posture, speed, or position of the sensor unit is learned using only a value having a small error with an elapsed time shorter than a predetermined time. The present invention can be applied to a motion capture apparatus.