Welcome to the Human-Robot Interaction 2013 Workshop on Collaborative Manipulation!
Challenges of Collaboration (Discussion Summary)
- Intent Transparency: One of the main challenges that a collaborative robot faces is making its intent transparent to the human collaborator --- what is the robot about to do?
- Trust: Connected to intent transparency is the idea of trust --- how can the robot instill the right level of trust, so that the collaborator neither over-trusts the robot (which can lead to dangerous situations if the robot makes a mistake), nor under-trusts it (which can lead to under-use of the robot's assistance)? Trust can perhaps also be viewed as mental model transparency.
- Level of Engagement: How passive/timid vs. active/aggressive should the robot be in the task (e.g. how much should it assist the collaborator, how much should it keep hold of the floor, etc.)? This depends on the context, e.g. in certain domains, such as rehabilitation, it is not the task that matters, but how much progress the user is making in actuating his or her arm.
- Complimentary Capabilities: How can the collaboration be structured such that the human and the robot make best use of their individual strengths? In some cases, robots are better at directing the task, while in others they best serve as end effectors.