Loading...
Loading...
Loading...
AIJuly 14, 20265 MIN READ

Teaching Spaces to Think: Enabling Safer Public Places with Robotics and Spatial Intelligence

A busy airport terminal showing a split real-world and digital view. People are tracked with teal boxes. A security robot inspects an unattended orange bag with data overlays like 'UNATTENDED OBJECT DETECTED'. Futuristic UI overlays.

This is a Guest Post by Rob Watts, AI Architect, Intel

A family rushes toward their boarding plane, weaving through a busy terminal. In the shuffle, a carry-on is left behind. Minutes later, a robot rolls into view, stops at a safe distance, and begins a predefined safety check, while security teams see a clear incident on their screen: where the bag is, when it was dropped, and how the system recommends they respond. Earlier this year, this live “robot bag inspector” demo powered by production-ready technologies from OnLogic, Intel®, and Clearpath Robotics, showcased how spatial intelligence, powered by Intel SceneScape, enables public safety.

For CIOs, CTOs, and operations leaders responsible for airports, transit hubs, hospitals, retail environments, and critical infrastructure, the real story isn't about robots. It’s about implementing trustworthy systems that understand a physical space as continuously as a human security officer would: tracking people and objects across dozens of cameras, understanding context, and knowing what "normal" looks like well enough to flag what isn't.

What can spatial intelligence do to enable governance?

Most facilities already have cameras, sensors, and point solutions in play, but they lack a unified, real-time understanding of the space. The gaps between these elements expose business and safety risks such as discovering incidents too late or triggering false alarms that degrade trust. Just throwing more sensors and cameras at the problem increases complexity and cost but doesn’t necessarily yield more or better insights. 

To make physical spaces “aware” of what’s happening in real time, SceneScape (part of the Open Edge Platform) enables spatial intelligence, providing a live, 3D understanding of people and objects across multiple cameras and sensors versus a wall of disconnected video feeds. Its 3D tracking, scene hierarchy, pose estimation, sensor fusion, and re-identification capabilities align multiple cameras into one coordinate system, maintaining awareness of people and objects, even as they move, overlap, or briefly leave the frame. Industrial systems from OnLogic such as the Karbon 520 Series, built on Intel Core™ Ultra Series 2 processors, provide the reliable, accelerated edge compute that makes insights gathered from multiple cameras, sensors, and other inputs actionable, supporting both the AI analysis and the robotic response with the deterministic performance that safety-critical environments demand. Paired with SceneScape, these systems can make the difference to reduce false positives, and accelerate response, while simplifying the ability to scale around more cameras and use cases.

5 rugged computers showing varying front plates - the Karbon 520 Series

One platform, many mission-critical scenarios

The airport bag scenario is just one of many powerful capabilities that Intel and its partners are enabling across industries. These spatial intelligence use cases not only highlight what’s possible but also what’s becoming increasingly practical as innovation in edge systems powered by Intel silicon are now available.

  • Retail loss prevention: Re-identification and pose estimation make it possible to detect suspicious activities like concealment and loitering in crowded stores, across multiple cameras and scenes even when subjects are temporarily blocked from view.

  • Smart buildings and cities: Underpinning smart building use cases for commercial and public infrastructure, sensor fusion and camera data enable spatial monitoring, occupancy, and alerting across large facilities like transit hubs, stadiums, and campuses.

  • Healthcare: Pose estimation and spatial intelligence contribute to next-generation patient monitoring—from detecting infant position in the NICU, to patient falls, and caregiver presence - to deliver real-time alerts to clinical teams.

Re-identification—or the ability to continuously track a person or object across different cameras and scenes through crowds, temporary occlusion, and off-camera moments—adds context to 3D spatial intelligence, delivering real-time insights for appropriate response."

Under the hood of the “robot bag inspector” scenario

Using an OnLogic Karbon 520 Series system built on Intel® Core™ Series 2 processors as the “brain at the edge,” the demo runs SceneScape in a mock airport environment and coordinates a Clearpath Jackal unmanned ground vehicle (UGV) that serves as the inspection robot. Multiple cameras feed into SceneScape, to detect people and luggage, assigns them identities, and maintain their positions in a shared 3D model of the terminal space.

When a new bag appears, SceneScape associates it with the person who placed it. If that person walks away and the bag remains, the system starts a timer and evaluates policies like “time unattended,” “location risk level,” and “crowd density nearby.” Once a configured threshold is crossed, the OnLogic system publishes a structured event into the security ecosystem, noting an unaccompanied bag, its specific location, confidence levels, and elapsed time. This enables a response ranging from a TSA officer dispatch to an autonomous robot inspection.

This same pattern scales to other industry scenarios like monitoring crowds, detecting unauthorized entry into restricted zones, or tracking suspicious behavior across large facilities like stadiums, transit hubs, and theme parks, without rebuilding the underlying perception stack for every new use case. This helps system builders spend more time shipping useful new features, instead of wrestling with infrastructure.

OnLogic: Making spatial intelligence deployable at the edge

Delivering real-time, safety-critical capability requires hardware built to run reliably where the action happens. OnLogic industrial systems, purpose-built for edge AI, provide the ruggedized, low-latency platform that keeps spatial intelligence workloads running continuously in demanding environments.

The OnLogic K520 Series is engineered for exactly this kind of always-on, mission-critical deployment. With shock and vibration resistance, wide operating temperature ranges, and flexible connectivity, the K520 was designed to operate in control rooms, ceilings, kiosks, vehicles, and outdoor enclosures without constant IT intervention, an important consideration for facility operators managing distributed infrastructure across dozens or hundreds of locations.

an engineer connecting a K523 rugged computer mounted in a rack in a warehouse

Intel Core Ultra Series 2 and Series 3 processors enable the heterogeneous compute spatial intelligence workloads demand: general-purpose CPU cores orchestrate real-time controls, integrated GPU runs AI inference, and NPU accelerates performance across multiple camera streams and multimodal pipelines. With detection, tracking, and sensor fusion running directly at the edge, facilities are less reliant on cloud, to meet strict response times. As deployments grow, the same platform scales to more cameras, additional robots, and new analytics workloads without a full redesign of the compute stack, a critical cost and risk consideration for any infrastructure leader planning a multi-year rollout.

A shared world model for people, robots, and systems

With SceneScape operating as a scene controller underpinned with a powerful system like the OnLogic K520 Series, a shared model of the physical environment becomes accessible to people, robots, and applications alike, bringing cameras and other sensors into a common coordinate frame. Structuring this with a "scene hierarchy" of zones, paths, and points of interest—such as gates, exits, restricted areas—provides an operational language for policies like 'flag bags unattended for more than five minutes in baggage claim' or 'alert when unauthorized people enter a restricted zone'.

For operators, this turns a wall of camera feeds into a live map of incidents, reducing cognitive load and enabling faster, more confident decisions. For integrators and ecosystem partners, it provides a consistent interface for adding new capabilities — AR overlays, additional robot types, specialized sensors — without rebuilding the perception and tracking stack each time.

From demo to deployment

What began as a "robot bag inspector" demonstration points to a broader shift: spatial intelligence becoming a core, reusable edge service across industries. Combined with the Open Edge Platform, OpenVINO™, and a growing partner ecosystem that includes OnLogic and system integrators such as WWT and Wipro, airports, hospitals, retailers, and critical infrastructure operators can move from pilots to scalable, repeatable edge AI deployments built on a shared, trusted understanding of the physical world.

For robots and automation to actively improve safety and operations, they need a reliable picture of the world around them and that starts with dependable spatial intelligence software and rugged, capable edge hardware. Scenescape delivers situational awareness; OnLogic Intel-powered systems deliver performance and reliability to run it where it matters. Together, they turn cameras, sensors, and robots into coordinated partners that help keep people moving safely, wherever the mission-critical work happens.

Learn more about Scenescape's spatial intelligence capabilities. 

Learn more about OnLogic Karbon 520 Series.

Headshot of Rob Watts, AI Architect in the Federal and Aerospace Division  of Intel's Edge Computing Group
Rob Watts
Rob Watts is AI Architect in the Federal and Aerospace Division of Intel's Edge Computing Group and lead architect and founder of Intel® Scenescape. Rob has extensive expertise across semiconductors, wireless, IoT, and end-to-end integrated solutions.

Stay in the know.

Get first access to the newest products and Industry 4.0 innovation, OnLogic events and more.