ML-Based Validation for Simplified Gait Analysis Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional gait analysis in clinical settings requires complex devices and trained professionals, limiting accessibility and convenience for patients, while simplified devices like wearable insoles need validation to ensure their accuracy in detecting medical conditions.
Innovation Solution
A machine learning model trained on data from production devices is used to validate test devices by categorizing individuals as having or not having a target condition, based on sensed data from both control and target groups, with the model's output compared to known categorizations to determine a match value that exceeds a predetermined threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional complex devices are used for gait analysis, then measurement precision is improved, but device complexity increases and accessibility decreases
Solution Approach 1:
The patent creates a digital copy or virtual model of the complex production device's functionality through machine learning models. The ML model learns from data collected by the production device and replicates its analytical capabilities, enabling the test device to achieve similar measurement precision without requiring the physical complexity of the original device.
Solution Approach 2:
The patent replaces the mechanical and physical complexity of traditional gait analysis devices with a computational system. Instead of requiring force plates, motion capture cameras, and complex mechanical sensors, the system uses machine learning algorithms processed from simpler sensor data to achieve equivalent or superior measurement precision.
2Ease of operation
If simplified test devices are used, then ease of operation is improved, but reliability decreases due to lack of validation
Solution Approach 1:
The patent implements a feedback mechanism where the machine learning model is trained and validated using data from both the simplified test device and the complex production device. The model continuously learns from comparisons between the two devices, adjusting its parameters to ensure that the test device's measurements reliably correlate with the production device's gold standard measurements.
Solution Approach 2:
The patent performs preliminary validation and training actions before deploying the test device for actual use. The machine learning model undergoes extensive training phases where it learns from labeled data, and validation phases where its predictions are compared against known outcomes. This preliminary work ensures reliability is established before the device is used in practice.
3Measurement precision
If machine learning models are trained on production device data, then measurement precision is maintained, but loss of time increases during training and validation
Solution Approach 1:
The patent applies partial training strategies where the machine learning model is trained on a carefully selected subset of the most informative data rather than exhaustively training on all available data. The training process focuses on the most critical features and cases that provide maximum learning value, reducing training time while maintaining or improving measurement precision.
Solution Approach 2:
The patent performs data preprocessing, feature extraction, and dataset preparation in advance before the actual model training begins. By organizing and preparing the training data beforehand, the actual training process is accelerated significantly, as the model receives pre-processed, ready-to-use features rather than raw data requiring extensive processing during training.
Data Source
AI summary
Embodiments disclosed herein are directed to systems and methods for validating a test device using machine learning models generated based on a production device. The test device may be a simpler or more updated device in reference to a production device. Aspects of validating a model based on subsets of clinical data are also disclosed. Aspects of identifying features to determine individual signatures are also disclosed. Aspects of an example gait analysis device are also disclosed.