Power-aware feature selection for optimized Analog-to-Feature converter

Back, Antoine; Chollet, Paul; Fercoq, Olivier; Desgreys, Patricia

doi:10.1016/j.mejo.2022.105386

Cited by 4 publications

(13 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thanks to more advanced software used, the reproduced simulation results of binary arrhythmia detection in ECG signals are slightly better than those shown in [5]. The optimized SFS algorithm limited to three parallel feature extractors achieves a 98.17% classification accuracy with eight extracted features and 2.6 µJ energy consumption, whereas 3.35 µJ were previously required to extract 10 features for a 98% accuracy.…”

Section: A Arrhythmia Detectionmentioning

confidence: 91%

“…However, the extracted analog features are digitized by an ADC to reduce the amount of data used for further classification either at the sensor level or after transmission to an aggregator. In the case of ECG arrhythmia detection, 6-bit precision appeared to be enough to maintain the same classification accuracy while performing the feature selection by the SFS on digitized data [5]. Similar studies should be performed for HAR to determine the ADC specifications.…”

Section: Specifications For a Hardware Implementationmentioning

confidence: 95%

“…Following the adopted methodology [5], a combination of two feature selection algorithms is applied during simulations. Firstly, only the 100 best features are chosen according to the Information Gain (IG) criterion without building an ML model.…”

Section: Analog-to-feature Conversionmentioning

confidence: 99%

“…This work aims to design a generic, reconfigurable A2F converter for various low-frequency signals. A2F conversion based on Non-Uniform Wavelet Sampling (NUWS) [4] has already been proposed for binary arrhythmia detection in ECG signals in [5], showing benefits over the A2I approach and conventional sampling in terms of energy consumption. Herein we adopt arrhythmia detection as the first application and use the same methodology to build a classification model.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Flexible Analog-to-Feature Converter for Wireless Smart Healthcare Sensors

Manokhin,

Chollet,

Desgreys

2023

2023 21st IEEE Interregional NEWCAS Conference (NEWCAS)

Self Cite

View full text Add to dashboard Cite

Analog-to-Feature (A2F) conversion based on Non-Uniform Wavelet Sampling (NUWS) has demonstrated the ability to drastically reduce the energy consumption in wireless sensors while employed for electrocardiogram (ECG) anomaly detection. The underlying idea is to extract relevant features from the analog signal and perform the classification in the digital domain. We adopt the same approach for a human activity recognition (HAR) task, considered as a second application for a proposed generic A2F converter. By extracting only 16 features from the inertial signals of the UCI-HAR data set and using these features as inputs for a simple Neural Network, we achieved an 87.7% accuracy in multiclass classification. From the simulation results, we defined the relevant features and the hardware specifications required for a complete circuit design and chip fabrication.

show abstract

Section: A Arrhythmia Detectionmentioning

confidence: 91%

Section: Specifications For a Hardware Implementationmentioning

confidence: 95%

Section: Analog-to-feature Conversionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Flexible Analog-to-Feature Converter for Wireless Smart Healthcare Sensors

Manokhin,

Chollet,

Desgreys

2023

2023 21st IEEE Interregional NEWCAS Conference (NEWCAS)

Self Cite

View full text Add to dashboard Cite

show abstract

“…In this paper, extending the findings presented in [15], our objective is to design a generic, reconfigurable A2F converter suitable for processing several types of signals with low sampling frequencies (below hundreds of kilohertz). Our work follows the one carried out in [16], where the A2F conversion approach has been applied for binary arrhythmia detection in electrocardiogram (ECG) signals, outperforming alternative acquisition techniques (conventional Nyquist rate sampling and CS) in terms of power efficiency and hardware simplicity. Effectively, ECG wearable sensors allowing for cardiac activity measurement represent an example of sensors constituting body area sensor networks (BASNs), which are strongly constrained in the energy available for sensor consumption.…”

Section: Introductionmentioning

confidence: 99%

Towards Flexible and Low-Power Wireless Smart Sensors: Reconfigurable Analog-to-Feature Conversion for Healthcare Applications

Manokhin,

Chollet,

Desgreys

2024

Sensors

Self Cite

View full text Add to dashboard Cite

Analog-to-feature (A2F) conversion based on non-uniform wavelet sampling (NUWS) has demonstrated the ability to reduce energy consumption in wireless sensors while employed for electrocardiogram (ECG) anomaly detection. The technique involves extracting only relevant features for a given task directly from analog signals and conducting classification in the digital domain. Building on this approach, we extended the application of the proposed generic A2F converter to address a human activity recognition (HAR) task. The performed simulations include the training and evaluation of neural network (NN) classifiers built for each application. The corresponding results enabled the definition of valuable features and the hardware specifications for the ongoing complete circuit design. One of the principal elements constituting the developed converter, the integrator brought from the state-of-the-art design, was modified and simulated at the circuit level to meet our requirements. The revised value of its power consumption served to estimate the energy spent by the communication chain with the A2F converter. It consumes at least 20 and 5 times less than the chain employing the Nyquist approach in arrhythmia detection and HAR tasks, respectively. This fact highlights the potential of A2F conversion with NUWS in achieving flexible and energy-efficient sensor systems for diverse applications.

show abstract