SBIR-STTR Award

Transitioning responsibilities presents a state of risk to organizations and the success of their missions. Effective handoffs depend on adequately transferring the most mission-relevant critical knowledge, or context. Determining and prioritizing information can be difficult because it is subject to variables including work domain, current system state, tendencies of the parties involved, and scale and timeline of the handoff. To mitigate risk involved in handoffs, Aptima will develop SPIDERSENSE: Situationally Pertinent Information DEtection and Relevance SENSing Engine. In Phase I SPIDERSENSE will be a theoretical framework and application methodology that permits the identification and operationalization of the critical knowledge necessary to enable successful handoffs, thereby increasing the safety of US Warfighters and the success of their missions. Aptima will leverage a threefold approach to this framework: (1) generalized task ontology to frame the work and the transition while accounting for time and work scales; (2) structured collection and analysis of passive data streams predominantly related to the task domain to provide raw information, both documented and implied; and (3) contextualized representation of user interactions proactively collected from the system and the parties. Together these will allow for the defining, sensing, reasoning over, and delivering of human-centered context to users.

Complex military missions have failed in the past due to poor planning, wherein critical contextual elements of the mission environment, team, equipment, and/or contingencies were missed. Furthermore, important information is often lost in transition from planning to action stages in teamwork due to memory leaks and communication errors. The ultimate success of a mission relies on gathering and capturing critical contextual elements to build a shared situational awareness, a transactive memory system (TMS) to capture the gathered knowledge throughout the cycle of teamwork, and the ability to quickly re-instantiate a team when a loss or disruption has occurred. Although optimizing team transitions and formations are time-consuming and resource intensive, artificial social intelligence (ASI) agents can optimize the work by capturing, reasoning, and predicting over team and mission data to support complex transitions. To enable application of ASI to planning problems, SPIDERSENSE II will create interfaces that support collaborative team planning, generation of machine-readable plans, and flexible team formation approaches that incorporate ASI reasoning. These will be created as components connected to an existing human-agent team (HAT) testbed developed in the DARPA ASIST program, resulting in a testbed aligned to support research on critical operational problems.