Show details
Many robots need to predict the future. Consider path planning in order to safely navigate to a goal position, or the more fundamental problem of obstacle avoidance which implicitly anticipates the undesirable consequence of a collision. There are many ways we can build anticipation into robots, but the one I would like to discuss in this talk is robots with simulation-based internal models. Illustrated with results from recent experiments I will show how such internal models can improve safety or fault detection in robots swarms, or allow us to build robots with simple ethical behaviours. Perhaps this approach might open a path to robots with artificial theory of mind? |
Target audience Students plus general public |
York Doctoral Symposium
Alan Winfield Alan.Winfield@uwe.ac.uk +447775803034 |
University of York
Heslington East Campus YO105GE York United Kingdom |