CHS: Small: Wearable Interfaces to Direct Agent Teams with Adaptive Autonomy

无人驾驶机器人系统将彻底改变许多重要的人类活动，包括灾难响应、公共安全、公民科学和农业，但这样的系统很复杂，需要多名飞行员。随着算法的接管和控制的简化，工人们从指挥而不是控制这些系统中受益。这种简化可以使工人使用他们的手，并将他们的感知集中在物理世界中，依靠可穿戴界面（例如，和弦键盘，手势输入）来管理无人驾驶车辆团队。自适应自主，即无人系统根据环境的复杂性改变对人类注意力的需求，提供了一种解决方案，使工作人员可以使用最少的投入来实施改变。目前的研究将可穿戴接口与自适应自治相结合，以指导模拟无人机器人系统的软件代理团队。该成果将支持下一代无人机器人系统接口。该项目的目标是为一组软件代理开发可穿戴接口，这些软件代理利用自适应自主性，并确定接口设计对指导代理的有效性。本研究开发了一个可穿戴网络人系统设计的测试平台，该平台使用软件代理作为无人机器人系统仿真，并使用自适应自治算法来驱动代理。该研究开发了一个框架，将可穿戴接口模式与它们最支持的活动联系起来。开发的系统将通过混合现实环境进行验证，参与者将在物理世界中指挥软件代理。主要假设是，可以开发一组相互连接的接口，通过适当的控制算法，最大限度地提高操作员对一组代理的控制范围，并优化操作员的身体工作量、精神工作量和情况感知。

Unmanned robotic systems are set to revolutionize a number of vital human activities, including disaster response, public safety, citizen science, and agriculture, yet such systems are complex and require multiple pilots. As algorithms take over, and controls are simplified, workers benefit from directing, rather than controlling, these systems. Such simplifications could enable workers to use their hands and focus their perception in the physical world, relying on wearable interfaces (e.g., chording keyboards, gesture inputs) to manage teams of unmanned vehicles. Adaptive autonomy, in which unmanned systems alter their need for human attention in response to complexities in the environment, offers a solution in which workers can use minimal input to enact change. The present research combines wearable interfaces with adaptive autonomy to direct teams of software agents, which simulate unmanned robotic systems. The outcomes will support next-generation unmanned robotic system interfaces. The objective of this project is to develop wearable interfaces for the direction of a team of software agents that make use of adaptive autonomy and ascertain the effectiveness of interface designs to direct agents. This research develops a testbed for wearable cyber-human system designs that uses software agents as unmanned robotic system simulations and uses adaptive-autonomy algorithms to drive the agents. The research develops a framework connecting wearable interface modalities to the activities they best support. Developed systems will be validated through mixed reality environments in which participants will direct software agents while acting in the physical world. The principal hypothesis is that a set of interconnected interfaces can be developed that, through appropriate control algorithms, maximizes an operator's control span over a team of agents and optimizes the operator's physical workload, mental workload, and situation awareness.