|Human-Autonomy Teaming: Supporting Dynamically Adjustable Collaboration
|Human Factors and Medicine|
Decision Support, Dynamics, Flexible levels of automation, Human / Vehicle Ratio, Human-Autonomy Teaming, Interface Design, Interoperability, Multi-Modal Interfaces, Supervisory Control, Supporting Dynamically Adjustable Collaboration, Uninhabited Vehicles
Based on progress in Robotics, Artificial Intelligence and Human Performance Modeling, Human-Agent-Robot Teamwork (HART) systems are being developed and tested in which humans and autonomy dynamically adjust and cooperate to accomplish a joint objective, often in shared spaces. In these systems, team members’ responsibilities and commitments are managed such that the human and automation jointly enhance performance and manage contingencies. Dynamically adjustable autonomous behaviors and supporting control architectures are critical to meet the many different contexts envisioned for these systems. Future autonomous systems, whether they be unmanned vehicles (UVs) or autonomous software agents, may be considered partners or even team members where tasks are adaptively re-allocated between operator and autonomy. The policies (e.g., authorities and obligations) for the situated work organization are tailored to the operational context by the dynamic setting of working agreements (e.g., the parameters and thresholds for dynamic task re-allocation) to maximize mission effectiveness. Adaptive work distribution and coordination are facilitated by sharing knowledge on task progress and the states of human and automation actors (e.g., intentions, abilities, momentary resources). User modeling methods support continuous monitoring, interpretation and learning of team members’ states (e.g., workload), skills and performance.
Given the evolving roles of human and autonomy, we need to strive toward flexible/adjustable, trustworthy automation being operated under user sovereignty. This RTO TG will explore the rapidly developing area of human-autonomy teaming. It directly leverages HFM Task Group HFM-RTG-170 and 217, which developed a framework for supervisory control of human robotic systems and leveraged the development of 15 emerging technologies across 8 NATO nations. Building off this acquired knowledge, this TG will identify and demonstrate successful teaming methodologies and interface design practices that allow for shared SA of task and environment, bi-directional understanding of intent, dynamic work distribution, and effective human-autonomy mission collaboration. Relevant issues and challenges are listed below.
- Adjustable & adaptive autonomous systems
- Authority sharing architectures and interface concepts
- Manned-unmanned team situation awareness and performance
- Bi-directional conveyance of intent
- Goal-based control
- Human-autonomy problem solving/cooperative dialog
- Decision Support Interfaces
- Situation assessment aids, feedback of action impact
- Predictive/look ahead tools, anticipatory support
- Intelligent aiding for time-critical team decision making.
- Multi-modal interfaces, intuitive interfaces, natural language interfaces.
- Networked telepresence