A Mixed Initiative Approach to Human-Robot Interaction for Through-the-Door Operation

We propose an innovative, distributed multi-agent based mixed-initiative planning approach to facilitate true partnership among humans and robots useful in many tactical application environments including “through-the-door” and NIST’s search and rescue scenario. We leverage our distributed control framework (DCF) along with indoor positioning system like Cricket and OpenCV based vision for situational awareness, to coordinate shoulder-to-shoulder team building and provide intelligent support for agile execution. Algorithms wise, IAI will leverage hierarchical task network (HTN) representation to organize the team members and their task allocations efficiently in completely distributed and scalable manner. Additionally, we propose to build cognition models for humans and intent detection for overall optimal team performance. All capabilities required to demonstrate these tactical scenarios beginning with decomposition of tasks among the human and robots while coordinating partners’ actions in decision making will be developed and integrated seamlessly. Our approach is inherently robust and scalable as it leads to significant improvement in human situational awareness and human-robot camaraderie for better coordination and control of different activities required to complete missions in chaotic and disorganized scenarios. We will demonstrate the end to end scenario on distributed control platform (DCF) with multiple sensor and cognition modules readily available at IAI.