HRC_neurobased_taskplanning

Human–robot collaborative task planning using anticipatory brain responses

This repository provides data accompanying the publication:

Ehrlich, S. K., Dean-Leon, E., Tacca, N., Armleder, S., Dimova-Edeleva, V., & Cheng, G. (2023).
Human-robot collaborative task planning using anticipatory brain responses.
PLOS ONE, 18(7), e0287958. https://doi.org/10.1371/journal.pone.0287958

The work explores the feasibility of using EEG-based neuro-cognitive measures as an implicit feedback signal to enable dynamic subtask assignment in human–robot collaboration (HRC). The repository contains data from an experimental HRI study (EEG recorded during takeover/role-switching situations)

---

Repository structure

HRC_neurobased_taskplanning/
├── s01/                          # EEG and metadata for subject 01
│   └── ...
├── s02/                          # EEG and metadata for subject 02
│   └── ...
...
├── s12/                          # EEG and metadata for subject 12
│   └── ...
├── README.md
└── ...

Study description (in brief)

Experimental HRI study (EEG during takeover anticipation)

Purpose: demonstrate EEG measures indicative of a human partner anticipating takeover situations (human→robot or robot→human) during collaboration, and assess whether these signals are decodable in single trials.

Design: A human participant collaborated with an industrial robot (UR10) in a trajectory-following task in a 7×7 grid world. Each movement from one grid tile to the next constitutes a trial. The task was designed to generate structured situations of takeover and non-takeover between human and robot.

Two collaboration scenarios were used:

• sequential collaboration (sC): human and robot are responsible for separate workspace areas; takeovers occur at the workspace boundary
• intermittent collaboration (iC): both operate in the full workspace, but each controls only two of the four movement directions; takeovers occur intermittently depending on required direction changes

Takeover situations were categorized as:

• HH: human continues
• HR: robot takeover from human
• RR: robot continues
• RH: human takeover from robot

EEG was recorded with a 32-channel setup (actiChamp, Brain Products; 1000 Hz sampling; average mastoid reference TP9/TP10; EOG channels included).

---

Citation

If you use this data, please cite:

@article{ehrlich2023hrc,
  title   = {Human-robot collaborative task planning using anticipatory brain responses},
  author  = {Ehrlich, Stefan K. and Dean-Leon, Emmanuel and Tacca, Nicholas and Armleder, Simon and Dimova-Edeleva, Viktorija and Cheng, Gordon},
  journal = {PLOS ONE},
  volume  = {18},
  number  = {7},
  pages   = {e0287958},
  year    = {2023},
  doi     = {10.1371/journal.pone.0287958}
}