Discovering Patterns in Human-Robot Interaction: New Tools for Complex Adaptive Social Systems - UQ Part
Abstract
The purpose of this project is to develop a set of complex systems analysis tools for analyzing human-robot interaction and social neuroscience data. Topological data analysis (TDA) is an exciting new set of techniques which allow for the quantification of topological features of data that is not easily captured by traditional geometric analyses. The AFOSR grant has ended, but the collaborative project with UCSD and AFRL/RH is ongoing, with further funding to this team provided by the Australian government. In the first year, the research team developed a novel TDA pipeline for analyzing and visualizing the simplicial structure of delay embedded time series data. This TDA pipeline was used to analyze both a social neuroscience and human-robot interaction dataset. Recurrence quantification analysis (RQA) isa time series analysis technique which allows for the visualization and quantification of the temporal patterning of when a dynamical system re-enters a similar part of phase space. The team generated RQA vectors for neural data and showed improved classification for some classes when compared with the classification of power spectral features. The Chiba lab (UCSD) and the Wiles lab (UQ) collaborated to develop a TDA and RQA pipeline and apply these methods to neuroscience, biomechanics, and human robot interaction datasets.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 28, 2018
- Accession Number
- AD1057268
Entities
People
- Andrea A Chiba
- Janet Wiles
Organizations
- University of Queensland