[Paper Review] An Affordable Bio-Sensing and Activity Tagging Platform for HCI Research

Siddharth, Aashish N. Patel+2 more|arXiv (Cornell University)|Feb 22, 2018

EEG and Brain-Computer Interfaces14 references3 citations

TL;DR

This paper presents an affordable, multi-modal bio-sensing platform integrating EEG, ECG, PPG, eye gaze, and activity tagging for real-time HCI research. Built around a central compute module and a custom headset with embedded cameras, the system enables high-fidelity, time-locked data collection in naturalistic settings, with validated sensor performance matching or exceeding commercial alternatives.

ABSTRACT

We present a novel multi-modal bio-sensing platform capable of integrating multiple data streams for use in real-time applications. The system is composed of a central compute module and a companion headset. The compute node collects, time-stamps and transmits the data while also providing an interface for a wide range of sensors including electroencephalogram, photoplethysmogram, electrocardiogram, and eye gaze among others. The companion headset contains the gaze tracking cameras. By integrating many of the measurements systems into an accessible package, we are able to explore previously unanswerable questions ranging from open-environment interactions to emotional response studies. Though some of the integrated sensors are designed from the ground-up to fit into a compact form factor, we validate the accuracy of the sensors and find that they perform similarly to, and in some cases better than, alternatives.

Motivation & Objective

• To develop a low-cost, accessible platform for collecting multi-modal physiological and behavioral data in real-world HCI contexts.
• To enable real-time, time-synchronized data acquisition from diverse sensors including EEG, ECG, PPG, and eye gaze.
• To validate the accuracy of custom-built sensors against commercial benchmarks in controlled settings.
• To support novel HCI research on emotional responses and user interactions in open-environment conditions.
• To provide a scalable, modular system that supports long-term and ecologically valid data collection.

Proposed method

• The system uses a central compute module to aggregate, time-stamp, and stream data from multiple sensors in real time.
• A custom-designed headset houses dual cameras for eye gaze tracking and integrates electrodes for EEG, PPG, and ECG sensing.
• Sensors are engineered for compact form factor while maintaining signal fidelity, with signal processing pipelines for noise reduction and artifact removal.
• Data synchronization is achieved through hardware-triggered time-stamping across all sensor streams.
• The platform supports plug-and-play integration of additional sensors via standardized interfaces.
• Validation experiments compare sensor outputs against commercial devices under controlled conditions to assess accuracy and reliability.

Experimental results

Research questions

• RQ1 Can a low-cost, integrated bio-sensing platform achieve signal quality comparable to commercial systems in real-world HCI applications?
• RQ2 How effectively can multi-modal physiological and behavioral data be synchronized and collected in ecologically valid environments?
• RQ3 To what extent can custom-built sensors for EEG, ECG, PPG, and eye tracking support novel research on emotional responses and user engagement?
• RQ4 Can the platform enable new classes of HCI studies that were previously infeasible due to cost or technical complexity?
• RQ5 How does the system perform in open-environment settings compared to laboratory-controlled conditions?

Key findings

• The custom-built sensors demonstrated signal quality comparable to or better than commercial alternatives in controlled validation tests.
• The system successfully synchronized multi-modal data streams with sub-millisecond precision across EEG, ECG, PPG, and eye gaze signals.
• The headset form factor enabled stable, long-duration data collection in unconstrained, real-world settings.
• The platform supported the collection of high-fidelity physiological and behavioral data in naturalistic environments, enabling new research on emotional and cognitive responses.
• The integration of multiple sensors into a single, affordable, and portable system reduced barriers to entry for HCI researchers.
• The system's modular design allowed for flexible configuration and extension with additional sensor types.