Vanguard Magazine

fully demonstrated, especially in a network- contested environment. Other challenges that must also be addressed, include the insufficiency of C4ISR infrastructure on- board RCN platforms to meet the compu- tational demands of AI. Nowhere are these deficiencies more apparent, or relevant, than in support of real-time decision-mak- ing at the edge, on a ship, at sea, on an op- eration, in a contested environment. Given these challenges, the RCN will need to collaborate closely with private in- dustry to operationalize AI. Trust, whether human or machine, is earned through a robust force development, generation and employment process. To realize this, indus- try experts must be able to work together with the operators and analysts living these challenges daily. To ensure AI technologies integrate into the existing operational sys- tems architecture, an agile engineering and design process that delivers iterative, incre- mental improvements to capabilities will be paramount. If these conditions can be met, AI will have its best chance of evolv- ing beyond a lab environment to its pro- curement, integration and adoption by the RCN organizations. Working with Xtract AI through the IDEaS program offers just this opportunity. Reach out to Cornell P. Pich VP Business Development cornell@xtract.ai 604-653-8628. Xtract AI is an approved vendor for key procure- ment groups, including Government of Canada Approved Source List Tier 1, 2, & 3 AI Vendor. www.vanguardcanada.com AUGUST/SEPTEMBER 2020 37 for the Royal Canadian Navy aRtifiCial intElligEnCE, SURVEillanCE and REConnaiSSanCE PERSPECtiVE R oyal Canadian Navy (RCN) capabilities are rapidly evolv- ing through its aggressive defence-policy mandated modernization program. The modernization of two capabilities in par- ticular will significantly impact maritime operations: Intelligence, Surveillance, Tar- get Acquisition, and Reconnaissance (IS- TAR); and Electronic Warfare (EW). With new and more powerful sensor technolo- gies integrated into airborne, surface and subsurface sensor systems, the RCN will have more data than ever to contribute to the comprehensive Recognized Maritime Picture (RMP). The problem, however, is that the value of this massive amount of raw, single and multi-source data is limited due to the complex task of transforming it into decision-quality information and intelligence. At Xtract AI, we believe Artificial Intel- ligence (AI) technologies offer the most significant opportunity to address this problem. As a component 1B company for the Department of National Defence and Canadian Armed Forces' Innovation for Defence Excellence and Security (IDEaS) program, Xtract AI is currently working on the Detection and Classification of Objects of Interest, offering an exciting opportu- nity for the RCN to work with Xtract AI to solve this problem and shape the solution. Xtract AI is now seeking stakeholders from the RCN, as well as business partners, to collaboratively pursue solutions that lever- age sensor data in a manner that shortens the decision-action cycle. The opportunities for AI to augment sensor operators in the collection, pro- cessing, exploitation, and dissemination of data, information and intelligence, are vast. For example, AI can be built into a manned or unmanned ISTAR or into EW platforms to help reduce bandwidth re- quirements when transmitting sensor data; into the terminals at which operators in- terpret sensor data; or into a suite of intel- ligence analysis tools used by analysts on Her Majesty's Canadian Ship (HMCS) Trinity. In all three cases, AI-enabled detection, classification and tracking of objects of interest—such as an airborne, surface, or subsurface target contact by electro-optical, infra-red, radar, or sonar sensors—offer opportunities to reduce the cognitive load of sensor operators and el- evate them to higher-level tasks. This will also help achieve economies of effort and scale, as well as significantly improve sen- sor utilization rates supporting situational awareness and decision-making. There are several obstacles to realizing the opportunity AI provides, however. While the technology has evolved rapidly in support of private industry, military ap- plications have lagged due to the obstacles presented by the unique conditions in which the RCN operates. These obstacles include, but are not lim- ited to, insufficient volumes of quality data to confidently train models, and labour in- tensive data labelling and complex system integration requirements. The security classification of sensor data also challenges the RCN's ability to work with tier-1 com- puter science and engineering talent found in the private sector. Significant resistance to change is expected given that the real- world operational value of AI has yet to be Sponsored Content Skeldar. Photo: Saab AB Puma™ 3 AE. Photo: AeroVironment, Inc.

• Cover