Model-based Opportunistic Robot Planning for Human-Robot Collaboration (MORPH)

• Applicant: Professorin Dr. Alexandra Kirsch
• Subject Area: Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
• Term: from 2012 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 222640427

The goal of the MORPH project is to explore the concept of opportunities to make robot behaviour more efficient and more understandable for humans. We explore situations in everyday household scenarios, in which a human and a robot cooperate on a joint task like setting the table.The proposed approach is based on reactive planning with plan transformation and uses prediction models to suggest and evaluate favourable opportunities. We will tackle three main challenges: 1) The representation of general plans to allow for the exploitation of opportunities, developing specific language constructs in a reactive plan language. 2) The efficient detection of opportunities, so that the plan execution process is not slowed down by the opportunity detection process, introducing the concept of opportunity models as a modular, declarative way to evaluate alternative plans. 3) Trading-off between commitment to the current plan and opportunistic behaviour, using a specific cost function to consider the efficiency gain of an opportunity in comparison to the commitment to the current plan.The approach will be evaluated in several simulated household environments and on a real robot, taking into account the efficiency of goal achievement as well as the legibility and acceptance of the robot behaviour by users.

• DFG Programme: Research Grants