Rugged AI platform enhances inference computing in data-driven embedded systems

05-11-2021 | Elma Electronics | Subs & Systems

The new Jetsys-5320 from Elma Electronic Inc satisfies the growing data processing requirements of very rugged and mobile embedded computing applications by effortlessly handling data-intensive computation tasks. It offers complex GPGPU inference computing at the edge for DL and ML operations in AI applications.

Used for high-performance intelligent video analytics, VR, AR and AI at the edge, and in robots and unmanned or autonomous vehicles, the new system merges the power of NVIDIA-based computation with unmatched ruggedisation in a compact, low-power platform.

The new device enhances visual intelligence throughout transportation and defence applications, including high-resolution sensor systems, automatic target recognition, movement tracking security systems, threat location detection and prediction, as well as machine condition-based monitoring and predictive maintenance semi-autonomous driving and driver advisory systems.

Ram Rajan, vice president of engineering at Elma Electronic Inc., noted: “AI-based computing is enabling multiple levels of intelligence and safety advancements throughout the embedded computing industry. We’re seeing a huge increase in object detection and tracking, video surveillance, target recognition and condition-based monitoring, so we built an AI platform that could handle the processing requirements for these kinds of applications used in challenging environments.”

At the heart of the device is the NVIDIA Jetson TX2i SoM, allowing real-time AI inferencing and DL/ML capabilities. Paired with a dual-core Denver 2 64-bit CPU and quad-core ARM A57 Complex and 256 CUDA cores, NVIDIA’s Pascal architecture provides 1.3TFLOP of performance.

The rugged system also offers HD-SDI, Gigabit Ethernet with PoE, USB 3.0 interfaces for video capture and miniPCIe expansion slots. It operates in temperatures from -40C to 71C, provides IP67 rated ingress protection and meets MIL-STD-810G for harsh environments.

By Natasha Shek