As the complexity of the operating environment increases, the interface between the robotic system and the human operator becomes more critical. The domain of team-based human-robot interaction (HRI) stands to benefit significantly from existing research in the domain of human factors, human-machine interaction and the performance of human-human teams, where extensive literature exists on how human teams allocate tasks, convey information and trade-off between autonomy and command and control during the performance of collaborative tasks.
A key strength of human teams is the high operational capacity of each team member, adaptation and anticipation of unplanned events, and flexible and heterogeneous communication channels. Sources of human team failure include lack of communication, loss of focus, and individual team member overload.
We will draw on these existing findings in the human factors literature to develop the missing robot capabilities, particularly in three key areas: identifying essential information transfer for teleoperation, the anticipation of human behaviour and unplanned events through human motion and behaviour understanding, and skill and interaction improvement through learning and adaptation.
Within this research theme, we will focus primarily on three topics:
- Teleoperation/shared initiative, building on Professor Kulić's work on mapping human motion to humanoid robots;
- Human robot interaction, building on Professor Kulić's work on incremental learning of motion primitives and their sequencing during on-line observation of human motion; and
- Imitation learning, building on Professor Kulić's work on task abstraction and human motion analysis.