Advantech Releases New Industrial Edge AI Acceleration Module

Advantech showcased its new cutting-edge AI module VEGA-300 series: powered by Intel® Movidius™ Myriad™ X VPU, supported by Intel® OpenVINO Development Kit and Advantech Edge AI Suite, to enable rapid development and deployment of edge AI, which can be used for versatile, vision-based AI applications in fields such as robotics, drones, retail, and transportation.

The VEGA-300 series offers two compact, low-power modules: VEGA-320 M.2 A+E Key and VGEA-330 miniPCIe interface, which helps accelerate computer vision and edge deep learning reasoning. To speed up the development of artificial intelligence, Advantech also provides an edge artificial intelligence kit that integrates the OpenVino™ toolkit distributed by Intel®, pre-trained models, and edge management functions, so that customers can quickly run reasoning on edge devices for real-time decision-making without being affected by latency, cost, bandwidth, and power consumption.

Accelerate edge computing with optimal performance efficiency

The VEGA-300 edge AI module is powered by the Intel® Movidius™ Myriad™ X VPU, providing the best performance/power across the entire AI workload with maximum flexibility on edge devices, while being 1.5 times faster than a single CPU computing platform. The VEGA-300 series comes in two forms: one is the VEGA-320 series M.2 module, which has an Intel® Movidius™ Myriad™ X VPU onboard; the other is the VEGA-300 series Minipcie module, which has two Intel® Movidius™ Myriad™ X VPUs onboard. The VEGA-300 series is compact enough to be plugged into edge devices to perform deep learning inference, transforming data into valuable information with higher accuracy and faster response speeds, and a single VPU consumes less than 3.8W of power.

Accelerate AI application development

To help customers quickly implement deep learning inference on their edge devices and use the VEGA-300 module, Advantech also provides the Edge AI Kit. Through the integration of the OpenVino™ toolkit and pre-trained models released by Intel, the Edge AI Kit provides a deep learning model optimizer, inference engine, pre-trained models, and a friendly GUI toolkit. Customers can use python-based tools to quickly import existing Caffe, TensorFlow, and Apache MXNet trained models and deploy them on edge devices. The Edge AI Kit also integrates VPU/CPU/memory for real-time status monitoring, as well as model update functions for managing distributed Edge devices.

Currently, VEGA-320 is already available, and VEGA-330 will be available in June. For more information, please contact your local Advantech branch through the WeChat public platform "Advantech Embedded Community" or call Advantech Embedded Service Hotline 400-001-9088.