AI has become a crucial component of autonomous vehicle technologies. The aROK 8210-AC4F, equipped with high-performance Intel® Core™ Processor (Series 2) and a powerful dedicated NVIDIA® GeForce RTX™ 5090 graphics card featuring 21,760 CUDA® cores, delivers up to hundreds of TOPS in AI inference performance. This makes it ideal for supporting AI aided transportation applications such as railway safety assurance which can be collaborating with existing ATP/ATO, Autonomous Mobile Robots (AMR), Machine Learning (ML), Intelligent Transportation Systems (ITS), and factory automation.
Thanks to NEXCOM’s advanced thermal design, the aROK 8120-AC4F can maintain high CPU computing performance even when paired with a graphics card like the NVIDIA® GeForce RTX™ 5090 with 600W thermal power, operating reliably under harsh conditions to enable consistent real-time AI inference.
The aROK 8120-AC4F is purpose-built for railway and transportation AI computing. It supports DC 24V non-isolated power input with IGN control and features four 1GbE M12 PoE connectors (IEEE 802.3at/at/bt). It can be optionally expanded to support up to twelve Ethernet ports via add-on cards, ideal for long-distance IP cameras and LiDAR sensors. It also offers a wide range of I/O interfaces, including six USB 3.2 ports, dual isolated CAN bus, dual COM ports (RS-232), dual COM ports (RS-232/422/485), digital I/Os (DI/DO), and dual HDMI® outputs. With optional 5G/NR and Wi-Fi 5/6 modules, the aROK 8120-AC4F can connect to Cyber-Physical Systems (CPS) for AI model retraining, enhancing inference accuracy in the field.
Built to withstand tough environments, the aROK 8120-AC4F supports an operating temperature range of -30°C to 70°C (without a graphics card installed) and complies with MIL-STD-810H standards for shock and vibration resistance (with a graphics card installed). It is certified to CE/FCC Class A, UKCA, EN 45545-2 and EN 50155 standards.