Novel Tunable Pseudoresistor-Based Chopper-Stabilized Capacitively Coupled Amplifier and Its Machine Learning-Based...

Area of Science:

• Electronics Engineering
• Biomedical Engineering
• Signal Processing

Background:

• Bio-potential amplifiers are crucial for acquiring physiological signals.
• High common-mode rejection ratio (CMRR) is essential for accurate bio-signal measurement.
• Existing amplifier designs often face challenges with linearity and common-mode signal suppression.

Purpose of the Study:

• To develop a high-gain, high-CMRR FinFET-based bio-potential amplifier.
• To introduce a novel technique for reducing common-mode gain and enhancing CMRR.
• To investigate the performance of tunable pseudoresistors for improved linearity and adjustable low cut-off frequency.

Main Methods:

• A capacitively coupled chopper-stabilized circuit with a feedback buffer was employed.
• Conventional pseudoresistors were replaced with a tunable parallel-cell configuration.
• A chopper spike filter was integrated to mitigate switching spikes.
• A machine learning model (CNN+LSTM) was developed for arrhythmia diagnosis.

Main Results:

• The amplifier achieved a mid-band gain of 42.6 dB and a bandwidth of 6.96 Hz to 621 Hz.
• The noise efficiency factor (NEF) was 6.1 with a power dissipation of 0.92 µW.
• The CMRR reached 106.9 dB with the feedback buffer, and the CNN+LSTM model achieved 99.12% accuracy for arrhythmia diagnosis.

Conclusions:

• The proposed FinFET bio-potential amplifier offers superior CMRR and linearity for bio-signal acquisition.
• The novel CMRR reduction technique and tunable pseudoresistors provide significant performance improvements.
• The integrated machine learning model demonstrates high accuracy in diagnosing arrhythmias from bio-signals.