Neural-Box® industrial

Neural-Box® industrial

All-in-one Edge Computing Device with AI acceleration based on the NVIDIA Jetson platform, scalable in three performance levels. Stylish aluminium housing in IP20 for mounting on DIN rail or as desktop, connection via standard connectors and terminals. Communicative through numerous interfaces, both wired and wireless.

The operating system used is the Linux distribution Ubuntu, prepared by NVIDIA for AI acceleration. CODESYS 3.5, including web visualisation and ISOBUS, is implemented as the runtime system for PLC controls, into which real-time AI applications can also be easily embedded. For efficient data management, an Azure Edge database is prepared that can be automatically synchronised with a Cloud database. Our Edge Computing devices can be tracked, configured and applications updated via a web frontend.


Processor systems:
  • NVIDIA Jetson nano (Quad-Core CPU / 128-Core GPU) or
  • NVIDIA Jetson Xavier NX (Six-Core CPU / 384-Core GPU / 48 Tensor-Cores) or
  • NVIDIA Jetson Orin NX (Octo-Core CPU / 1024-Core GPU / 32 Tensor-Cores)
  • RAM 4 / 8 / 16 GB
  • eMMC 16 GB (nano + Xavier)
  • SSD 2 TB (optional for nano + Xavier)
  • NVRAM 64kB
  • microSD
  • 2 x Gigabit Ethernet 1000/100 BASE TX (RJ 45)
  • 2 x USB 3.2, 1 x CAN, 1 x RS485
Interfaces wireless:
  • 4G / LTE oder 5G (MiMo) with GNNS as module (optional)
  • WiFi (MiMo) / Bluetooth (optional)
Security: Infineon OPTIGA™ TPM2.0-Security-Chip
Voltage: 8–32 V
Housing: Aluminium housing IP20 with passive cooling, additional heat sink may be necessary


Powerful edge computing for the DIN rail, even with connection to 5G campus networks, distributed by our system partner RSConnect.

Neural-Box® visu

Powerful edge computing for direct use on machinery with bright and robust 21.5" multi-touch display.

ADVES Functional Framework

Implement entire control systems efficiently and safely.