AI-Based Physiological Parameter Estimation in Wearable Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic devices designed to measure physiological parameters require complex, expensive, and energy-consuming sensors, leading to increased device costs, reduced battery efficiency, and wasted memory resources when multiple sensors are included.
Innovation Solution
An electronic device uses an artificial intelligence model, such as a deep neural network, to estimate interdependent physiological parameters from directly obtainable bio-signals, determining if additional sensors are needed based on the estimated values, thereby reducing the necessity for multiple sensors and optimizing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensors are included to directly measure physiological parameters, then measurement precision is improved, but device complexity increases and manufacturing cost increases
Solution Approach 1:
The patent uses machine learning models to create virtual copies of physiological parameter measurements. Instead of using multiple physical sensors to directly measure parameters like blood pressure and blood sugar, the system trains ML models using data from wearable sensors to predict these parameters. This virtual copying approach achieves comparable measurement precision without the complexity of multiple sensors.
Solution Approach 2:
The patent replaces the mechanical sensor-based measurement system with an information-processing system using machine learning. Rather than physically measuring blood pressure with a cuff or blood sugar with a needle, the system substitutes these mechanical measurement methods with algorithms that process data from accelerometers, gyroscopes, and other wearable sensors to estimate the physiological parameters.
2Measurement precision
If multiple sensors are included to directly measure physiological parameters, then measurement precision is improved, but manufacturing cost increases
Solution Approach 1:
The patent replaces expensive physical measurement systems with virtual copies generated by machine learning models. By training models to predict blood pressure, blood sugar, and other parameters from data collected by inexpensive wearable sensors, the system eliminates the need for costly medical-grade sensors while maintaining measurement precision.
Solution Approach 2:
The patent uses low-cost wearable sensors (accelerometers, gyroscopes, photodetectors) that can be mass-produced and disposed of, replacing expensive, durable medical sensors. The ML models process data from these inexpensive sensors to achieve accurate physiological parameter estimation, significantly reducing manufacturing costs.
3Measurement precision
If multiple sensors are included to directly measure physiological parameters, then measurement precision is improved, but energy consumption increases
Solution Approach 1:
The patent creates virtual measurements of physiological parameters using ML models trained on data from low-power wearable sensors. Instead of activating multiple high-power sensors continuously, the system uses machine learning to predict parameters like blood pressure and oxygen saturation from minimal sensor data, dramatically reducing energy consumption while maintaining measurement precision.
Solution Approach 2:
The patent uses a minimal subset of sensors (accelerometer, gyroscope, photodetector) to gather sufficient data for ML models to estimate multiple physiological parameters. Rather than deploying a full array of sensors for direct measurement, the system uses partial sensing combined with computational inference to achieve accurate results with lower energy expenditure.
4Measurement precision
If multiple sensors are included to directly measure physiological parameters, then measurement precision is improved, but memory resource usage increases
Solution Approach 1:
The patent uses ML models to generate virtual copies of physiological parameter data rather than storing data from multiple physical sensors. The models process and compress information from wearable sensors to estimate multiple parameters, reducing the memory resources needed to store and manage sensor data while maintaining measurement precision.
5Reliability
If multiple sensors are included to directly measure physiological parameters, then reliability is improved, but ease of operation worsens due to system complexity
Solution Approach 1:
The patent replaces complex multi-sensor systems with ML-based virtual measurement systems. The machine learning models automatically process data from simple wearable sensors to reliably estimate physiological parameters, eliminating the complexity of coordinating multiple sensors while maintaining or improving measurement reliability through advanced algorithms.
Data Source
AI summary
Disclosed herein is an electronic device and a control method thereof. The control method of an electronic device includes: obtaining a bio-signal from at least one sensor, determining a first physiological parameter based on the bio-signal, estimating a second physiological parameter including a specified correlation with the first physiological parameter, and providing information about the estimated second physiological parameter.