Anticipation in robotics

Winfield, Alan F.; Hafner, Verena V.

Abstract

In this chapter, we introduce anticipatory robotic systems. We show how intelligent robots can anticipate the future, by outlining two broad approaches: the first shows how robots can use anticipation to learn how to control their own bodies and the second shows how robots can use anticipation to predict the behavior of themselves interacting with others, and hence demonstrate improved safety, or simple 'ethical' behaviors. Both approaches are illustrated with experimental results from recent work. We show that, with practical realizable embedded artificial intelligence, robots can indeed predict the future and that this is a technology with significant potential for improved safety and human-robot interaction.

• Publication Date: Jan 1, 2018
• Deposit Date: Dec 15, 2017
• Peer Reviewed: Peer Reviewed
• Book Title: Handbook of Anticipation: Theoretical and Applied Aspects of the Use of Future in Decision Making
• ISBN: 9783319317373
• DOI: https://doi.org/10.1007/978-3-319-31737-3_73-1
• Keywords: anticipation, robotics, internal simulations, predictive models, sensorimotor learning, multi-agent systems, developmental robotics, theory of mind, robot safety, robot ethics, expectations, forward models, exploration, tool use, sensory attenuation, human-robot interaction
• Public URL: https://uwe-repository.worktribe.com/output/876535
• Publisher URL: https://doi.org/10.1007/978-3-319-31737-3_73-1
• Contract Date: Dec 15, 2017