Motion Capture Suit Using Fourier Transform for Movement Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Classifying different body positions and movements using only data from body sensors without visual data is challenging, particularly in distinguishing between static positions and dynamic movements.
Innovation Solution
A motion capture glove with sensors connected to the back and fingers, transmitting data to a computing system for real-time processing using a support vector machine trained with Fourier-transformed axis and angle data to predict body positions and movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If only sensor data from the body is used for movement classification, then the system portability and wearability are improved, but the classification accuracy between static positions and dynamic movements deteriorates
Solution Approach 1:
The patent transforms joint angle data from the time domain to the frequency domain using Fast Fourier Transformation. This parameter transformation allows the system to extract frequency-based features that are independent of time dependency, enabling the classifier to distinguish between static positions (low frequency or zero frequency) and dynamic movements (higher frequency components) using only wearable sensor data, thus resolving the contradiction between wearability and classification accuracy
2Measurement precision
If Fast Fourier Transformation is applied to joint angle data, then time dependency is removed and classification capability is improved, but computational complexity increases
Solution Approach 1:
The patent applies Fast Fourier Transformation as a preliminary processing step to convert joint angle data into the frequency domain before classification. By performing this transformation in advance, the system extracts essential frequency features that capture the nature of movement versus static positions, enabling the subsequent Support Vector Machine classifier to make accurate distinctions without having to process raw time-dependent data, thus balancing computational complexity with classification accuracy
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Achieves high accuracy in classifying hand movements, with linear and RBF support vector machines demonstrating 99.93% and 99.97% accuracy respectively for static positions, and 99.8% and 84.52% for dynamic movements, enabling effective real-time movement classification.
Implementation Method 1
transforming each joints-angle data to a corresponding frequency domain using a fast Fourier transformation to remove any time dependency value
Data Source
AI summary
In one aspect, a computerized process useful for movement classification using a motion capture suit includes the step of providing the motion capture suit worn by a user. The motion capture suit comprises a set of position sensors and a Wi-Fi system configured to communicate a set of position sensor data to a computing system. The process includes the step of providing the computing system to: receive a set of position data from the motion capture suit for a specified time window of data comprising X, Y and Z axis positions and a joints-angle data for each position sensor of the set of position sensors, transforming each joints-angle data to a corresponding frequency domain using a fast Fourier transformation to remove any time dependency value, after the fast Fourier data transformation, train a support vector machine using the X, Y and Z axis positions data and the frequency domain data as input, using the support vector machine to predict a set of body positions and movements.