Aetina Corporation has announced the DeviceEdge AIE-VX28/18-IP, a fanless, IP66-rated rugged in-vehicle edge AI system powered by NVIDIA Jetson AGX Orin, providing in-vehicle video analytics and supporting up to 12 camera inputs.
Engineered for construction, agricultural, mining and rail vehicles as well as roadside infrastructure, the AIEVX28/18-IP combines MIL-STD-810H durability with simultaneous support for four PoE and eight GMSL2 cameras to deliver real-time surround-view video analytics in harsh outdoor conditions.
Heavy-duty commercial and off-highway vehicles are rapidly adopting AI for obstacle avoidance and pedestrian detection; however, extreme operating environments expose conventional in-vehicle computers to relentless vibration, mud, rain and temperature swings. Simultaneously, system integrators face severe cost constraints and rigorous European automotive compliance standards. The AIE-VX28/18-IP directly overcomes these three critical bottlenecks: environmental survivability, high perception density and deployment readiness.
The AIE-VX28/18-IP features a sealed, fanless enclosure rated IP66 for dust and water protection, with waterproof M12 connectors and threaded locking that keep connections secure under sustained vibration.
Built to withstand heavy vibration and shock, the system operates from -25 degrees Celsius to 70 degrees Celsius and accepts 9–48 VDC input with ignition power control for surge protection and safe shutdown. By integrating IP66 ingress protection with MIL-STD-810H standards and E-Mark (E24) automotive certification, a single computing platform seamlessly serves on-road vehicles, off-highway machinery, and rail transit applications.
The AIE-VX28/18-IP integrates four M12 PoE ports supplying a total of 60W power for IP cameras, alongside eight GMSL2 cameras via mini-FAKRA connectors. Supporting up to 12 concurrent camera streams, the platform empowers 360-degree situational awareness and multi-sensor fusion. The PoE configuration simplifies cabling for high-definition IP cameras, while GMSL2 enables ultra-low-latency, long-distance video transmission across large vehicle chassis.
Powered by NVIDIA Jetson AGX Orin modules, the system delivers up to 275 TOPS of AI performance to run multiple concurrent deep neural networks and real-time vision pipelines on NVIDIA JetPack 6.2 or later.
With NVIDIA JetPack 7.2 introducing Super Mode, the NVIDIA Jetson AGX Orin 32GB module achieves an AI performance boost from 200 to 241 TOPS — unlocking substantial headroom for dense vision analytics without requiring an upgrade to higher-cost modules. Comprehensive I/O interfaces include M12 10GbE, isolated CAN FD, RS-232/422/485, and flexible M.2 expansion supporting NVMe high-speed storage, Wi-Fi/Bluetooth, and centimeter-level RTK positioning or LTE/5G cellular modules.
Engineered for demanding mobile fleets and harsh industrial environments, the AIE-VX28/18-IP powers mission-critical edge AI vision tasks across diverse sectors:
- Automated heavy machinery — 360-degree blind-spot obstacle avoidance and virtual geofence exclusion zone alerts for excavators and bulldozers.
- Smart agriculture — All-weather autonomous tractor guidance, precision variable-rate spraying, and crop health monitoring.
- Outdoor autonomous mobile robots (AMR) — High-density multi-camera perception for autonomous mobile robots operating across rugged industrial yards and construction sites.
- Intelligent transportation systems (ITS) — Ruggedized roadside AI units for real-time traffic flow analytics and incident detection.
