AI: Coming to a Public Safety Network Near You

Years ago, public safety communications began a long journey toward greater interoperability and more dependable network performance using analog technologies. It has since embarked on a cycle of seemingly continuous iteration beginning with the adoption of digital public-safety standards: APCO P25, TETRA, or Digital Mobile Radio (DMR). More recently, public safety radios and networks began supporting and leveraging consumer technologies, such as Long Term Evolution (LTE) and wireless local area networking (WLAN), to achieve higher data rates. Despite these steady advances in both network and radio technologies, a technology lurking in the background stands to push public safety capabilities to a completely new level. That technology is artificial intelligence (AI).

For many, thinking of artificial intelligence (AI) in public safety brings to mind movies like “Minority Report.” That sci-fi film features “pre-cogs” that can foresee crimes and pinpoint details about them so that wrongdoers can be arrested before the event occurs. While predictive AI-based capabilities will undoubtedly emerge at some point, today’s early implementations are not nearly that sophisticated. However, much can be inferred and accomplished by using the power of AI to sort through hours of video footage or monitor a public appearance by an important figure like a government official, for example.

As a result, plans for many initial AI public-safety applications are to interpret video from closed-circuit television (CCTV) in cities, stadiums, and other public places. Immediate momentum can be gained by doing so in urgent situations, such as tracking a lost child. Descriptions of a vehicle or person—including hair color, clothing details, etc.—can be entered within a certain location range and quickly sorted by AI to gain helpful details.

Putting an End to Dark Data

These sources and others, such as body-worn police cameras, gather voluminous amounts of video data on an ongoing basis. Much of it may be important and usable. If it is not immediately related to a case or event, however, it is not recognized as such. With AI, video footage can quickly be searched for relevant information. Video files also can be stored and tagged so that an organized archive exists for future reference, helping people pull relevant files at a later time.

AI systems also can prevent events from being “missed” in real time via surveillance. Because AI can develop an understanding of how people behave in a certain scenario, anything unusual can be “flagged.” “Intelligent” new cameras, which already support AI, provide such monitoring in real time. Cognitive analytics is key here, as it can infer valuable information such as relationships between key individuals in footage.

As more of these systems come online, additional applications will emerge. The common thread running through all of them will be the use of AI to sort data, organize it, and recognize findings or patterns. The ongoing learning and development of computer vision, pattern recognition, and other elements will push the evolution of AI forward. Eventually, algorithms could be developed to predict criminal behavior, much as movies have predicted. For now, however, AI is improving public safety by gathering details and intelligence in a “needle in a haystack” maneuver that humans cannot match. From the adoption of digital communications technologies, public safety is moving closer to a complete digital transformation.

To see how Keysight supports the development, testing, and manufacturing of public-safety radios, click here.