Factories, warehouses, and industrial sites are getting smarter and more automated, but that also means safety and security must evolve at the same speed. To tackle this challenge, Fujisoft, a global IT solutions provider, has built a next‑generation AI‑powered physical security system on the AMD Embedded+ platform, blending real‑time object detection, flexible detection zones, and instant alerts into a single, deployable unit.

For a broader, official‑style look at how AI and embedded hardware are transforming industrial safety, you can refer to the Artificial intelligence – Wikipedia page Artificial intelligence – Wikipedia.

Fujisoft’s AI Physical Security System

Detail	Info
Solution name	AI‑based physical security / site safety system developed by Fujisoft.
Underlying platform	AMD Embedded+, which combines AMD Ryzen Embedded processors and Versal adaptive SoCs on a single board.
Key capability	Uses AI models to identify people and objects in real time, not just “motion,” reducing false alarms.
Deployment stage	A demonstration unit was completed in 2025; Fujisoft is now refining it for wider rollout.
Target environments	Factories, warehouses, logistics hubs, and other high‑activity operational sites.

What Makes This Security System “Smart”?

Traditional security setups often suffer from two big problems: too many false alarms and rigid, hard‑to‑adjust detection zones. Fujisoft’s new AI system fixes both with a more “intelligent” approach.

AI‑based person and object detection: Instead of simply reacting to movement, the system uses models running on the AMD Embedded+ platform to classify humans, vehicles, and key objects, which makes intrusion detection far more accurate and context‑aware.
Automatic detection area configuration: One standout feature is that users can define the detection area just by placing four stop signs at the corners of the target zone; the system then auto‑configures the area, cutting the need for manual camera‑by‑camera setup.
Real‑time monitoring and alerts: The system provides continuous live‑feed analysis, with instant visual and audible alerts whenever a hazard, unauthorized entry, or safety violation is detected, letting site managers respond quickly.

Inside the box, AMD Ryzen AI Software and Vitis AI tools help Fujisoft tune and deploy models efficiently, while the integration of x86 processing and FPGA‑style adaptability lets the hardware adapt to different sensor types and workloads without a full redesign.

Why AMD Embedded+ Is The Right Fit?

The choice of AMD Embedded+ as the backbone of this solution plays a big role in its compactness and efficiency.

Single‑board, hybrid architecture: By putting Ryzen Embedded processors and Versal adaptive SoCs on one PCB, AMD reduces design complexity and BOM cost, helping Fujisoft get to market faster.
Flexible, field‑programmable I/O: The FPGA‑style side of the Versal SoC lets the system connect to a wide range of cameras, sensors, and industrial I/Os without needing extra interface adapters.
AI‑ready software ecosystem: Sample applications, development tools, and early‑access support from AMD mean Fujisoft’s team can focus on AI logic instead of low‑level integration, accelerating the whole development cycle.

A Fujisoft manager, Naoya Yanagitsubo, noted that combining adaptive hardware and x86 processing in one platform perfectly aligns with Fujisoft’s strengths in both hardware and software, letting them build a more accurate, flexible, and long‑term deployable solution.

System‑Features Table

Feature	How It Helps
AI‑based person/object detection	Reduces false positives by understanding what’s in the camera feed, not just whether something moved.
Auto‑configured detection areas	Lets users set up zones with physical stop signs instead of complex software configuration, cutting setup time and errors.
Real‑time alerts	Sends instant signals when a hazard or unauthorized activity is spotted, improving safety response times.
AMD Embedded+ hardware	Combines x86 compute with FPGA‑like flexibility and AI‑ready tools, giving Fujisoft a compact, future‑proof engine under the hood.

FAQs

Q: What kind of security problem does Fujisoft’s AI system solve differently?

A: It moves beyond simple motion detection to AI‑based identification of people and objects, which cuts false alarms caused by shadows, weather, or small animals, while still catching real intrusions and safety issues

Q: How does the AMD Embedded+ platform help Fujisoft build this solution faster?

A: By integrating Ryzen Embedded processors and Versal adaptive SoCs on one board, plus providing AI development tools and samples, AMD reduces design complexity and speeds up both hardware and AI‑model integration.