Rack-Mount AI Industrial Computer RK3588, Ubuntu OS | Scalable Al Acceleration | 1U Rack-Mount The USR-EG1000 is a high-performance ARM-based AI industrial computer powered by an 8-core, 64-bit RK3588 processor. It features an integrated 6 TOPS NPU and supports additional AI computing expansion, enabling deployment of large language models and AI video analytics algorithms.
Standard hardware includes four Gigabit Ethernet ports, up to 16 serial ports, and integrated 4G, Wi-Fi, USB, HDMI, and CAN interfaces for flexible deployment across a wide range of applications. Its built-in algorithm management platform supports connections to up to 16 cameras and offers more than 200 selectable algorithms, including face and license-plate recognition, enabling real-time local inference and decision-making.
Standard hardware includes four Gigabit Ethernet ports, up to 16 serial ports, and integrated 4G, Wi-Fi, USB, HDMI, and CAN interfaces for flexible deployment across a wide range of applications. Its built-in algorithm management platform supports connections to up to 16 cameras and offers more than 200 selectable algorithms, including face and license-plate recognition, enabling real-time local inference and decision-making.
Standard 19-Inch 1U Rack Design Housed in a standard 19-inch 1U rack chassis with high-strength aluminum handles on both sides, the EG1000 installs neatly in any cabinet. Reserved expansion slots let you add 4G, WiFi, storage drives, or AI accelerator module as your workloads grow.
Comprehensive Connectivity for Multiple Scenarios The EG1000 covers virtually every on-site scenario: 4 Gigabit Ethernet ports, up to 16 serial ports, 2 CAN buses, HDMI, and more - so cameras, controllers, and field devices connect directly, without extra converters.
AI Vision + Data Acquisition, All in One The EG1000 fuses AI vision with data acquisition in a single device. On-device AI models handle image recognition and data processing locally, then deliver results to on-site screens or cloud monitoring centers. Custom AI algorithms are available on request.
200+ Edge AI Vision Algorithms The USR-EG1000 ships with 200+ pre-trained edge AI vision algorithms, so you can go from unboxing to intelligent monitoring without training your own models. One rack-mount unit covers the scenarios that matter most on the factory floor and job site — smoke and fire detection, hard-hat and high-vis vest detection, license plate recognition, illegal parking, vehicle speeding, intrusion detection, people counting, and more. Algorithms run locally on-device, keeping video streams inside your facility while alerts reach your platform in real time.
Up to 32-Channel Video Input Connect up to 32 video channels to a single EG1000 and bring your entire camera estate into one AI system. The gateway works with mainstream camera brands — Hikvision ISC, EZVIZ, Dahua DSS, and Uniview — and speaks the protocols you already run, including ONVIF, GB28181, RTSP, RTMP, HLS, and V4L2, so existing cameras integrate without replacement. The current product version supports up to 16 channels; contact us about 32-channel compatibility when your deployment needs more.
Unlock Intelligent AI Video Applications Pair any standard camera with the EG1000's built-in AI algorithms, and ordinary video becomes actionable insight. The combination turns raw footage into computer-vision intelligence: smoke and fire detection, behavior analysis, intrusion detection, fall detection, absence detection, safety-helmet detection, and vehicle-violation detection. There's no need to replace working cameras or build AI models from scratch — simply connect the devices you already have, and the EG1000 handles detection, analysis, and alerting at the edge. Ideal for smart factories, warehouses, campuses, and traffic monitoring.
Mainstream AI Framework Support Deploy models trained in TensorFlow, Caffe, MxNet, TF Lite, ONNX, or Darknet as-is. INT4/INT8/INT16/FP16 inference, dual AI accelerator expansion (up to 46 TOPS), and 32-channel video input make the EG1000 ready for demanding vision workloads.
Built-In WukongEdge Suite The built-in WukongEdge suite brings edge computing, local HMI design, PLC programming, and network management into one platform. Build custom dashboards, wire up data flows with Node-RED, and get the network online in minutes.
WukongEdge Edge Engine WukongEdge, the built-in edge engine, unifies network management, data acquisition, computing, and reporting. Straightforward parameter configuration opens north-south data channels and supports a wide range of PLC and industrial protocols, so field devices - PLCs, sensors, cameras, scanners - connect quickly.
Ubuntu 24.04 Preinstalled, Custom OS on Request The EG1000 ships with Ubuntu 24.04, open APIs, and full development documentation for C/C++ and Python. Need something different? Android, Kylin OS, Debian, and HarmonyOS images are available as custom builds.
Industrial-Grade Design Dual Power Inputs: DC 12-48 V terminal power for cabinets and DC plants, or AC 110-250 V for standard wall outlets
Built-In Cooling Fan, Active Cooling: The AI Accelerator edition adds a built-in cooling fan with active heat dissipation, keeping the system cool and stable under sustained AI workloads.
Industrial-Grade Protection: Engineered for harsh sites: ESD protection at 4 kV (air and contact), surge and EFT immunity to 1 kV, and a wide -25 to +60 ℃ keep the EG1000 running in factories, substations, and outdoor cabinets.
Built-In Cooling Fan, Active Cooling: The AI Accelerator edition adds a built-in cooling fan with active heat dissipation, keeping the system cool and stable under sustained AI workloads.
Industrial-Grade Protection: Engineered for harsh sites: ESD protection at 4 kV (air and contact), surge and EFT immunity to 1 kV, and a wide -25 to +60 ℃ keep the EG1000 running in factories, substations, and outdoor cabinets.
Front Status LCD A built-in HD LCD shows network link states, CPU load, temperature, and uptime in real time, so field staff can check device health at a glance - without logging in.
Ordering Guide The EG1000 comes in two editions: Standard (6 TOPS NPU, 8 x RS485) and AI Accelerator (46 TOPS total with dual RK1820 cards, 16 x RS485). Both share the RK3588 CPU, 4GB memory, 64GB storage, Ubuntu 24.04, 4 Gigabit ports, and dual-band WiFi.
Industry Applications Typical deployments include smart-factory production line monitoring, smart O&M for distributed PV plants, cold-chain environmental monitoring, rail-transit onboard video analytics, C&I energy-storage EMS, distribution-grid automation, and green data-center energy management.
Packing List Standard shipment includes:
- 1 * USR-EG1000 unit
- 1 * Ethernet cable
- 3 * Terminal blocks
- 2 * Rack ears
- 2 * Serial adapter cables
- 1* WiFi antenna.
