Related Experiment Video
Updated: Jul 3, 2025
Construction of a Wireless-Enabled Endoscopically Implantable Sensor for pH Monitoring with Zero-Bias Schottky Diode-based Receiver
Published on: August 27, 2021
Towards Flexible and Low-Power Wireless Smart Sensors: Reconfigurable Analog-to-Feature Conversion for Healthcare
Mikhail Manokhin1, Paul Chollet1, Patricia Desgreys1
1C2S Team, ComElec Department, Laboratoire de Traitement et Communication de l'Information (LTCI), Télécom Paris, Institut Polytechnique de Paris, 19 Place Marguerite Perey, 91120 Palaiseau, France.
Analog-to-feature (A2F) conversion using non-uniform wavelet sampling (NUWS) significantly cuts energy use in wireless sensors for tasks like ECG anomaly detection and human activity recognition (HAR). This energy-efficient approach offers flexible sensor systems.
Area of Science:
- Signal Processing
- Embedded Systems
- Biomedical Engineering
Background:
- Wireless sensor systems require energy-efficient methods for data acquisition and processing.
- Analog-to-feature (A2F) conversion with non-uniform wavelet sampling (NUWS) shows promise in reducing energy consumption for specific tasks.
- Existing methods like Nyquist sampling can be energy-intensive for continuous monitoring applications.
Purpose of the Study:
- To extend the application of the generic A2F converter using NUWS to human activity recognition (HAR).
- To evaluate the energy efficiency of the A2F converter compared to Nyquist-based approaches for both ECG anomaly detection and HAR.
- To define valuable features and hardware specifications for a complete A2F converter circuit design.
Main Methods:
- Simulations were performed to train and evaluate neural network (NN) classifiers for ECG anomaly detection and HAR.
- A key component, the integrator within the A2F converter, was modified and simulated at the circuit level.
- Energy consumption of the communication chain with the A2F converter was estimated and compared to Nyquist-based systems.
Main Results:
- The A2F converter with NUWS demonstrated significant energy savings, consuming at least 20 times less energy than Nyquist for arrhythmia detection.
- For HAR tasks, the A2F converter consumed at least 5 times less energy compared to the Nyquist approach.
- Valuable features and hardware specifications for the A2F converter circuit were defined based on simulation results.
Conclusions:
- A2F conversion utilizing NUWS is a highly effective strategy for developing energy-efficient wireless sensor systems.
- The proposed A2F converter architecture is adaptable for diverse applications, including biomedical monitoring and HAR.
- This technology holds significant potential for flexible, low-power sensor deployments in various fields.
