Sensor Signal Quality Filtering for Biological Event Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Sensor devices face challenges in accurately detecting and predicting events due to inadequate and low-quality signal data, particularly in early stages of deployment and repetitive or cyclic processes, leading to unreliable performance in health care, industrial, and environmental monitoring applications.
Innovation Solution
The development of systems and methods that include wearable devices with sensors connected to computing devices for real-time data analysis, quality detection algorithms, and machine learning techniques to filter out unusable samples and improve signal processing, enabling the detection of breathing events and trends in respiratory monitoring and other repetitive processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor devices are deployed in early stages or repetitive processes, then monitoring coverage is improved, but signal quality and data reliability deteriorate
Solution Approach 1:
The system performs preliminary actions by collecting and storing large amounts of sensor data during deployment phases before adequate signal patterns are established. This preparatory data accumulation enables future algorithm training and event detection even when initial signal quality is insufficient for reliable monitoring.
Solution Approach 2:
The system changes parameters by adjusting algorithm sensitivity thresholds and detection criteria based on the phase of deployment. During early stages with inadequate signal data, the system modifies detection parameters to accommodate lower signal quality, then gradually tightens criteria as adequate signal patterns accumulate over time.
2Measurement precision
If quality detection algorithms and machine learning techniques are implemented, then event detection accuracy is improved, but device complexity increases
Solution Approach 1:
The system segments the complex monitoring task into distinct phases: data collection phase, algorithm training phase, and event detection phase. Each phase uses simplified processing appropriate to its purpose, with complex machine learning algorithms applied only during training rather than real-time operation, thereby reducing overall system complexity while maintaining detection accuracy.
Solution Approach 2:
The system introduces an intermediary training phase that acts as a mediator between raw sensor data and final event detection. During this intermediate phase, algorithms learn from accumulated data and establish detection criteria, which then enables simpler real-time detection operations without requiring full algorithmic complexity during active monitoring.
Data Source
AI summary
Systems and methods for detecting the quality of signals captured by a sensor device monitoring a biological function. Sensor data associated with the sensor device is received, the sensor data representing time-series measurement samples of one or more parameters associated with the biological function, the sensor data including usable and unusable samples of the time-series measurements. Data representing two or more features of samples of the time-series measurements is extracted and filtered to reduce outliers in the extracted data based on an expected outlier ratio. A machine learning algorithm is then trained to identify events based on the filtered extracted data.