ABSTRACT
We establish the context of our work as enabling mission-level cognitive autonomy across all five domains of warfare: air, land, maritime, space, and cyberspace [1]. Thus, we address both ‘autonomy at rest’ and ‘autonomy in motion’ as described by the Defense Science Board [2]. We focus on the ‘mission-level’ aspect rather than the ‘vehicular control’ aspect; that is, we address how autonomous systems work together in human/machine teams to achieve commander’s intent. By ‘autonomy,’ we mean the ability to achieve mission objectives without external control by humans or
machines while sensing and adapting to changes in the environment.1 While definitions of ‘cognitive’ abound, in this context, we mean the ability to learn from own experience and from others to detect unforeseen changes and to reason about unplanned events to achieve mission goals. Thus, the system should be able to deal with achieving mission objectives in uncertain environments. We have designed and implemented a reusable software framework automating the OODA loop2 for each individual autonomous system, enabling a seamless transition path from advanced automation3 to more cognitive autonomous systems. Figure 6.1 depicts the Cognitive Autonomy Engine (CAE™)4 framework in more detail. Communication and collaboration among autonomous systems and human teammates are also directly incorporated in the CAE framework.
