Autonomous Virtual Assistant (AVA)
Navy SBIR FY2018.1

Sol No.: Navy SBIR FY2018.1
Topic No.: N181-033
Topic Title: Autonomous Virtual Assistant (AVA)
Proposal No.: N181-033-0245
Firm: Soar Technology, Inc.
3600 Green Court
Suite 600
Ann Arbor, Michigan 48105
Contact: Glenn Taylor
Phone: (734) 887-7620
Web Site:
Abstract: AEGIS watchstanders monitor tracks in the battlespace around the ship, tracking, identifying, and engaging threats to ensure the safety of the ship. This track management task can be overwhelming today, and will become more difficult with new adversary approaches such as unmanned air systems and swarming surface tactics including saturation raids. While AEGIS provides some capabilities for automatically identifying potential threats, making recommendations, and even engaging threats, its decisions are based on fairly rigid rules that do not take into account the current context in which the ship is operating. Instead, watchstanders must build up and provide their own context about the battlespace, the geo-political environment, and the normal background activities that occur in the environment. By not taking this information into account, current AEGIS tools leave the operators with too many tracks to manage on their own. SoarTech and team propose to design, develop and assess a ship-board Virtual Assistant to significantly reduce the workload of AEGIS operators. This Virtual Assistant will help watchstanders make decisions by learning relevant information from the environment and the watchstanders, providing additional context for decision-making, and making recommendations about tracks and potential responses.
Benefits: Providing virtual assistants to Navy watchstanders will help reduce their workload by augmenting the recommendations that AEGIS (and other combat systems) provides, providing addition context to help them make decisions. SoarTechƒ?Ts Automated Virtual Assistant (ƒ?oAVAƒ??) will provide not only tailored personal assistants for particular jobs in the CIC and on the bridge, but it will also provide a reusable framework for building virtual assistants for other roles in the Navy and beyond. Virtual assistants have the potential to decrease workload and help watchstanders and officers make better decisions in ship defensive and offensive operations. The same toolset could also be adapted to other defense-oriented domains not afloat, including land-based AEGIS installations, airfield defense, or ground-based missile defense applications. Beyond Navy applications, the entire commercial shipping industry could use AVA-like applications to avoid collisions (using AIS data), help make decisions or receive recommendations and reminders about which rules of the road apply in the current location. While the initial application would be for ship defense, many DoD operations require complex decision making that could benefit from a virtual assistant that helps automate tasks and marshal information to make better decisions. For example, Army and Marine Corps officers managing a battle from inside a Tactical Operations Center (TOC) must make time-critical decisions about how to employ forces in real time, based on data that is coming in from the field. An AVA tailored to TOCs, implemented in tools such as the Command Post of the Future (CPOF) or, at smaller scales on mobile tools like JPC-P running in Stryker vehicles, would be beneficial. Even smaller civilian operations, such as search and rescue or disaster relieve operations that require a number of organizations across police, fire, rescue, and NGOs, could benefit. Additionally, there is a clear use for virtual assistants in other high-data types of environments, such as business analytics and planning, even if the pressures are different than military operations. Each of these different applications could leverage the core AVA architecture, and then would have domain-specific interfaces and services built for them.