Paul Duckworth

Paul Duckworth

Profile

My research is in the area of unsupervised human-activity analysis performed by a mobile robot. It combines information retrieval techniques, computer vision, robotics and qualitative spatial representations.

In particularly I am  interested in devising techniques for representing a visual scene using qualitative calculi to learn and recognise repeated patterns in sensor data available on a mobile robot. We have developed ways for building probabilistic models of human behaviour from observing human populated ewnvironments over long periods of time. These models have been used to understand what areas are useful for certain behaviours, and for detecting atypical events.

Research groups and institutes

  • Artificial Intelligence
  • Robotics at Leeds

Qualifications

  • BSc in Mathematics and Statistics. Lancaster University, 2008
  • MSc in Mathematics and Computational Science, University of Manchester 2010
  • PhD candidate, Leeds, 2013-present

Research groups and institutes

  • Artificial Intelligence