Adaptive Sensor Sampling for Low-Power Biometric Signal Approximation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wearable devices face power consumption challenges due to the need for frequent sensor sampling, which reduces battery life and is exacerbated by motion artifacts, necessitating a more efficient method to maintain signal quality while minimizing power usage.
Innovation Solution
Implementing custom distributions of sampling intervals and transmit power levels for sensors, utilizing methods such as random subsampling, phase-locked loops, and uncertainty estimation to adjust sampling rates and power levels dynamically based on signal quality and biometric prediction needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequent sensor sampling is performed to maintain signal quality and accurate biometric predictions, then measurement precision is improved, but power consumption increases
Solution Approach 1:
The patent implements dynamic sampling rate adjustment where the sensor sampling frequency is continuously adapted based on detected motion levels. During periods of low motion, sampling rates are reduced to conserve power, while during high motion periods, sampling rates increase to maintain signal quality. This dynamic adaptation resolves the contradiction by making the sampling strategy context-dependent rather than static.
Solution Approach 2:
The system changes the sampling interval parameter based on motion detection results. When motion artifacts are detected to be below a threshold level, the sampling interval is extended (lower sampling rate), and when motion artifacts exceed the threshold, the sampling interval is reduced (higher sampling rate). This parameter modification allows the system to maintain measurement precision only when necessary, reducing overall power consumption.
2Measurement precision
If high sampling rates are used to capture motion artifacts and maintain signal accuracy, then measurement precision is improved, but battery life is reduced
Solution Approach 1:
The patent employs periodic motion detection and sampling rate adjustment cycles. Instead of continuous high-rate sampling, the system periodically assesses motion levels and adjusts sampling rates accordingly. This periodic action allows the system to maintain biometric prediction accuracy during relevant periods while extending battery life through reduced sampling during irrelevant periods when motion artifacts are minimal.
Solution Approach 2:
The system applies partial sampling by collecting only the necessary amount of data required for accurate biometric predictions during low-motion periods, rather than continuously collecting excessive data. This partial action approach maintains sufficient measurement precision for health monitoring while significantly reducing the total number of samples taken, thereby extending battery life.
3Measurement precision
If sensor power levels are increased to maintain signal quality during motion, then measurement precision is improved, but power consumption increases
Solution Approach 1:
The patent applies different power levels locally based on motion detection results. Instead of using high transmit power continuously, the system adjusts power levels locally in response to detected motion conditions. During low-motion periods, lower power levels are sufficient, while during high-motion periods, power is increased only when needed to maintain signal quality, thus resolving the power-precision contradiction.
Data Source
AI summary
A method of creating a low-powered approximation of one or more completed sets of data is provided. A lower resolution is sampled, at a predetermined interval distribution, of one or more of sensors: accelerometer, electrodermal activity sensor, photoplethysmographic (PPG) sensor; impedance sensor, gyroscopic sensor, and/or a radio sensor. A combined uncertainty from the sensors is determined. Error change in a predicting a signal is estimated. The predetermined interval distribution is modified based upon the combined uncertainty and using phase-locked loops at one or more targeted frequencies, which are adjusted in real-time. The sampling of the sensors is modified, in real time based on the estimated error change, to be one of random, sparse, and/or high resolution.