Using context and robot-human communication to resolve unexpected situational conflicts

摘要

While efforts to develop cognitive abilities for robots have made progress from the perspective of goal-directed task performance, research has shown that additional cognitive capabilities are needed to enable robots to interact, cooperate, and act as teammates with humans. In particular, robots need additional teamwork and coordination knowledge and an ability to apply this knowledge to a model of context that is at least homologous to the context models that people use in reasoning about environmental interactions. The Context-Augmented Robotic Interface Layer (CARIL) provides a robot with a cognitively-motivated computational capability for situation assessment and situational adaptation. CARIL is used to analyze and develop context-based reasoning strategies that allow a robot to coordinate its behavior and spatial movements with humans when they are working on shared tasks and/or in shared space. Both communication-free and communications approaches are addressed and tested in a simulated environment.