Edge AI Computing For Autonomous Vehicles

"We need a Edge AI Computing For Autonomous Vehicles"

Computing Type

Edge Computing

Product

X6-MTH-ORN

Brand

Timeline

March 2025

Industry

Transportation

A transport authority was rolling out a fleet of autonomous shuttle vehicles to connect underserved areas of a growing smart city. These vehicles needed to process video feeds, LiDAR data, and object recognition in real-time — all while navigating safely through unpredictable urban environments.

Their engineers had a clear checklist:

GPU support for AI models

Real-time inferencing with high throughput

Remote diagnostics and updates

Compatibility with CAN bus for in-vehicle communication

Compact form factor to fit tight on-vehicle spaces

They tried a few industrial PCs, but none ticked all the boxes. Either the CPU wasn’t powerful enough, or the AI accelerator lacked flexibility. Some systems needed manual BIOS updates (not ideal when your kit is already mounted inside a minibus), while others ran too hot or took up too much space.

Relec’s team worked closely with the transport authority to understand the operational challenges and performance targets. We recommended the DFI X6-MTH-ORN, a compact AI edge box that brings both brains and brawn to the road.

Dual Powerhouse Performance

Powered by Nvidia Jetson Orin NX (for up to 100 TOPS of AI compute) and Intel® Core™ Ultra Processor, it could run AI inference models while handling complex system tasks in parallel.

Remote Ready

The Out-of-Band (OOB) Management feature meant engineers could reboot systems, push firmware updates, or access diagnostics remotely — no need for onsite tinkering.

Automotive I/O

CAN bus support, multiple display outputs, USB 3.2, serial connections, and RJ45 ports made integration with existing vehicle systems a breeze.

Expandable and Future-Proof

With M.2 slots and an optional MXM GPU slot, it could evolve with their AI demands over time.

Relec helped them configure the solution, liaised with DFI for firmware-level tweaks, and provided pre-install support to streamline deployment.

With the X6-MTH-ORN fully integrated, the autonomous shuttle fleet went live on time — a rare thing in public infrastructure projects. Engineers were able to monitor and update systems remotely, which significantly reduced service downtime. The system handled real-time image recognition, pedestrian detection, and route optimisation without lag or thermal throttling.

The transport authority now has a scalable, future-ready computing for autonomous vehicles solution, and is exploring expanding the same tech into waste collection and emergency response fleets. And as for Relec? We’re still on hand, providing ongoing support if requested.