A 259.6 μW HRV-EEG Processor With Nonlinear Chaotic Analysis During Mental Tasks.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Wearable Technology

Background:

• Mental task monitoring is crucial for understanding cognitive states.
• Existing methods for analyzing physiological signals like heart rate variability (HRV) and electroencephalography (EEG) can be computationally intensive and error-prone.
• The need for efficient, low-power hardware solutions for real-time physiological data analysis is growing.

Purpose of the Study:

• To present a novel system-on-chip (SoC) designed for mental task monitoring.
• To integrate nonlinear chaotic analysis (NCA) with HRV and EEG signal processing.
• To achieve high accuracy and low power consumption in a compact hardware implementation.

Main Methods:

• Development of a system-on-chip (SoC) incorporating nonlinear chaotic analysis (NCA).
• Utilizing an independent component analysis (ICA) accelerator to improve HRV extraction accuracy.
• Implementing NCA acceleration for calculating Largest Lyapunov Exponents (LLE) and linear features (mean, standard deviation, sub-band power).

Main Results:

• Achieved a significant reduction in HRV extraction error from 5.94% to 1.84% using an ICA accelerator.
• Demonstrated high confidence levels (95%) in mental task monitoring measurements.
• Fabricated a chaos-processor in 0.13 μm CMOS technology with remarkably low power consumption (259.6 μW).

Conclusions:

• The proposed SoC with NCA provides an accurate and efficient solution for mental task monitoring.
• Hardware acceleration significantly enhances the performance and reduces the error in physiological signal analysis.
• The low power consumption makes this technology suitable for wearable and portable applications.