GSR Board

Overview

The GSR Monitor Board is designed as a validation platform for the HEEPidermis SoC, currently under development at the Embedded Systems Laboratory (ESL), EPFL. This board facilitates testing and evaluation of key functionalities, particularly focusing on high-precision Galvanic Skin Response (GSR) measurement with low noise operation.

Features

Power Regulation & Battery Management

• Optimized for low-noise operation, ensuring high-quality analog signal acquisition.
• Supports external power through a LiPo/Li-Ion battery or a standard power supply unit (PSU).
• Integrated battery charger with load-sharing capability, allowing simultaneous charging and operation.
• Power regulation via two high-PSRR, low-noise LDOs with additional filtering stages to minimize high-frequency noise.

Precision Current DAC (iDAC)

• Features a 16-bit current DAC (iDAC) capable of generating currents in the range 40 nA to 10 μA.
• Configured in a current sink topology for accurate and stable current output.
• Multiple load options for testing and validation:
  • 8-bit 100 kΩ digital potentiometer
  • 3 fixed precision resistors: 100 kΩ, 470 kΩ, 1 MΩ
  • 3 variable rheostats: 100 kΩ, 500 kΩ, 2 MΩ
  • External electrode connectors for GSR measurements

Level Crossing ADC & ΣΔ ADC Integration

• Capable of simulating a Level Crossing ADC using two externally controlled discrete Sample-and-Hold (S/H) stages
• Integrated 12-bit ΣΔ ADC for high-resolution signal acquisition

Digital Communication & Compatibility

• Communicates via I2C, ensuring seamless integration with various microcontrollers and SoCs.
• Supports multiple logic voltage levels in the range 1.65 V to 5.5 V, enabling compatibility with virtually all microcontrollers and embedded platforms.
• Arduino-compatible software support included in the repository for easy interfacing and experimentation.

Repository Contents

• Firmware & Software: Arduino-compatible library for easy interfacing.
• Schematics & PCB Design: KiCad files for board design, production, and hand assembly.
• SPICE Simulations: TINA-TI simulations of the different analogue stages.

Dataset Source

This project utilizes the WEMAC dataset, an open multi-modal dataset designed for gender-based violence detection through physiological, speech, and self-reported emotional data.

Reference:

J. A. Miranda Calero, L. Gutiérrez-Martín, E. Rituerto-González, E. Romero-Perales, J. M. Lanza-Gutiérrez, C. Peláez-Moreno, and C. López-Ongil, “WEMAC: Women and Emotion Multi-modal Affective Computing dataset,” Scientific Data, vol. 11, no. 1, p. 1182, Oct. 2024, doi: 10.1038/s41597-024-04002-8.

Contact

Please contact the authors for additional information

• Alex López: alejandro.lopezrodriguez@epfl.ch
• Juan Sapriza: juan.sapriza@epfl.ch
• Matías Miguez: matias.miguez@epfl.ch
• José Miranda: jose.mirandacalero@epfl.ch

Developed at the Embedded Systems Laboratory (ESL), EPFL.