As industrial automation moves toward deeper integration of AI edge computing and real-time control, traditional PLCs are increasingly limited when handling AI vision inspection and real-time data analytics. On the other hand, general-purpose industrial computers may offer sufficient computing power but often struggle to meet the requirements for millisecond-level real-time control and industrial reliability.
This gap between IT and OT has become one of the key challenges in bringing Industrial AI into real-world applications.
The BLIIoT EdgePLC Industrial AI Edge Controller Series was designed to bridge this gap.
Built around a combination of high-performance ARMxy processors and distributed I/O expansion, EdgePLC integrates PLC real-time control, AI edge inference, and industrial communication into a single platform.
A single EdgePLC can simultaneously handle:
Real-time logic control
AI vision inference
Multi-protocol data conversion
Edge visualization
Remote device management
Instead of deploying a traditional combination of PLC + industrial PC + gateway, EdgePLC consolidates these functions into a single intelligent edge node, simplifying deployment, system architecture, and maintenance.
According to industry forecasts, the global Industrial Automation Edge AI market is expected to grow rapidly in the coming years, while the Industrial Edge Computing market continues to expand at a significant pace.
Behind these numbers is a clear shift:
Industrial AI is moving from “analyzing” to “taking action.”
AI is no longer limited to software tools that assist decision-making. It is moving directly onto the factory floor, where it can contribute to material handling, assembly, inspection, robotic grasping, and other production processes.
At the same time, industrial applications are creating a growing demand for integrated control and computing platforms.
AI inference and control logic need to move closer to the production line to reduce cloud transmission latency and overcome bandwidth limitations.
NPU and AI acceleration technologies are moving from centralized computing infrastructure toward industrial edge devices, making local AI inference increasingly practical.
Traditional PLC + industrial PC + gateway architectures can result in higher costs, greater system complexity, and more difficult maintenance.
The industry needs a platform that can compute, control, communicate, and act — at the edge.
The BL244 is the flagship model in the EdgePLC family, designed specifically for applications requiring heavy AI workloads combined with real-time industrial control.
Powered by the Rockchip RK3576J/RK3576 processor, the BL244 uses an 8nm process and features a heterogeneous octa-core architecture consisting of:
4 × Cortex-A72 high-performance cores
4 × Cortex-A53 high-efficiency cores
It also integrates 3 × independent Cortex-M0 real-time coprocessors running at 400MHz, dedicated to time-critical I/O responses and interrupt handling.
To help ensure long-term processor reliability, the default operating frequencies are 1.6GHz for the A72 cores and 1.4GHz for the A53 cores.
The BL244 integrates a dedicated 6 TOPS NPU supporting multiple precision formats, including:
INT4 / INT8 / INT16 / FP16 / BF16 / TF32
It is compatible with mainstream deep learning frameworks such as TensorFlow, PyTorch, Caffe, and MXNet.
According to test results, the 6 TOPS NPU can process AI analysis from up to 12 × 1080P camera streams simultaneously, with YOLOv5s inference reaching more than 45 FPS.
This makes the BL244 suitable for demanding multi-camera industrial vision applications.
The BL244 supports:
8K@30fps / 4K@120fps H.265 hardware decoding
4K@60fps H.265/H.264 hardware encoding
2 × USB 3.1, up to 5Gbps
Optional HDMI 2.1, up to 4K@120fps
NVMe M.2 SSD expansion
Mini PCIe for Wi-Fi / 4G / 5G connectivity
This provides the bandwidth and processing capability required for high-resolution industrial vision and multimedia applications.
The BL244 supports up to 32 N-Series distributed I/O modules, covering a wide range of industrial signal types:
DI
DO
AI
AO
RTD — PT100/PT1000
TC — Thermocouple
PWM
High-speed pulse counting
This modular architecture allows the controller to adapt to different machine and automation configurations without redesigning the entire control system.
Designed for demanding industrial environments, the BL244 supports:
-40°C to +85°C wide-temperature operation
Industrial EMC Level III protection
Independent hardware watchdog
DIN35 rail mounting
12–24VDC wide-voltage power input
It is built to operate where standard computing platforms may not be suitable.
The BL244 runs on Ubuntu 20.04 + Linux-RT 6.1.115 real-time kernel and supports a comprehensive industrial software stack, including:
CODESYS / OpenPLC / NexPLC — IEC 61131-3 control
IgH EtherCAT Master — hard real-time EtherCAT control
Node-RED — low-code workflow development
Docker / K3s — containerized application deployment
YOLOv5/YOLOv8 + OpenCV — AI vision development
From PLC logic to AI inference and industrial networking, the BL244 provides a ready-to-deploy platform for integrated edge applications.
Industries such as 3C electronics, new energy batteries, food, and pharmaceuticals increasingly require parallel visual inspection across multiple production stations.
With its 6 TOPS NPU, the BL244 can process AI analysis from up to 12 × 1080P camera streams simultaneously.
A single BL244 can reduce the need for multiple industrial PCs and dedicated AI boxes.
Industrial cameras can connect through USB 3.1, while YOLO models perform inference locally. Inspection results can then be displayed in real time through HDMI.
Warehouse AGVs and inspection robots often need to run several workloads simultaneously:
Visual SLAM + object detection + path planning + motor control
The BL244's 6 TOPS NPU handles AI inference locally, while the Cortex-M0 real-time coprocessors support time-critical motor responses.
With a compact 110 × 92 × 38mm form factor, the controller can also be integrated into space-constrained mobile robotic platforms.
Applications such as SCARA robots, Delta robots, and CNC machines require close integration between visual positioning and motion control.
The BL244 supports the IgH EtherCAT Master and can work with the CODESYS SoftMotion library to achieve synchronized multi-axis motion and vision-guided closed-loop control.
Energy storage and solar installations are often distributed across large areas and exposed to challenging environmental conditions.
They require a combination of:
Data acquisition + local control + AI-powered security
With its -40°C to +85°C operating temperature range, the BL244 is designed for demanding outdoor environments.
N-Series I/O modules can directly collect battery and inverter signals, while the NPU can analyze video streams for smoke, fire, and unauthorized personnel detection.
BLIoTLink can then transmit the data to cloud platforms through 4G/5G connectivity.
Modern logistics systems need to perform package identification, sorting control, and inventory updates simultaneously.
The BL244's NPU can perform real-time barcode recognition and OCR, while I/O modules control sorting mechanisms such as wheel diverters and cross-belt systems.
BLIoTLink can transmit the results to WMS systems in real time.
The core value of the BL244 is not simply its computing power.
It is what happens when 6 TOPS AI computing, 32-module I/O expansion, and hard real-time industrial control are brought together in one compact platform.
The BL244 combines:
Multi-camera AI inference
Hard real-time control
8K-class video processing
Flexible industrial I/O
Industrial-grade reliability
AI + PLC + Industrial Communication
A single BL244 can help replace the traditional stack of:
PLC + Industrial PC + AI Box + Gateway
It may not be the device with the highest AI computing performance on the market.
But at the 6 TOPS performance level, the BL244 delivers a highly practical combination of AI inference + 32-module I/O expansion + hard real-time control for demanding industrial applications.
AI doesn't just need to see what is happening.
In industrial environments, it needs to understand, decide, and act — in real time.
Shenzhen Beilai Technology Co., Ltd. | BLIIoT
www.BLIIoT.com
