Optical EEG for Man-Machine-Interfaces

Niels-Ole Rohweder, M.Sc.

State of the art / Motivation

• Measuring brain activity by electroencephalography (EEG) is the "gold standard" for determining various physical and psychological states such as fatigue, attention, distraction
• Up to now, current EEG procedures have relied on expensive and contact-based methods, which limit the possible applications, e.g. in robotics or semi-autonomous vehicles
• An alternative, non-contact method consists of measuring the pupil diameter with a camera, which correlates on medium and long time scales (> 30s) with brain activity in certain areas and frequency bands

Objectives

• Investigation of the correlation on short time scales
• Measurement of various, known EEG signals with the contactless method, and generation of indicators, which can be used within the HerMes project of Clausthal UT to determine human psychological parameters in cooperative man-machine teams
• Implementation of the sensor at the demonstrator for human-machine-collaboration for the joint project KEIKO - Kognitiv und Empathisch Intelligente Kollaborierende Roboter

Methods

• Development of a precise algorithm for determining the pupil diameter from camera images, based on ellipse detection
• Realization of an experimental setup, consisting of lighting, camera and EEG and stimulus generator
• Investigation of the influence of lighting
• Quantifying the sensitivity resp. the noise floor
• Conducting various correlation measurements with test persons

Results

• Construction of a high-resolution, sub-micrometer pupilometer
• Design of a Brain-Computer-Interface by using luminance stimuli (pupillary light reflex) at 2-7 Hz

Theses

Currently, the following Master projects are available:

• Linearity of the pupil at periodic luminance stimuli
• BCI: Optical flashing-light stimuli of various frequencies for controlling a cobot