Embedded PCs for AI@Edge by Interworld Electronics

Description

AI@EDGE

Artificial intelligence (AI) processing today
is mostly done in a cloud-based data centre. The majority of AI processing is
dominated by training of deep learning models, which requires heavy compute
capacity and high costs and this has now forced a rethink of how to deploy some
of this AI inferencing locally and thereby reduce costs, prevent sensitive data
being sent to the cloud for processing and increase speed of decision making.

AI has now become the key driver for the
adoption of edge computing to deliver these benefits. When compared to the
enterprise data centre and public cloud infrastructure, edge computing has
limited resources and computing power. When deep learning models are deployed
at the edge, they don get the same horsepower as the public cloud which may
slow down inferencing, a process where trained models are used for
classification and prediction.

To bridge the
gap between the data centre and edge, chip manufacturers are building niche, purpose-built
accelerators that significantly speed up model inferencing. These modern
processors assist the CPU of the edge devices by taking over the complex
mathematical calculations needed for running deep learning models. While these
chips are not comparable to their counterparts – GPUs – running in the cloud,
they do accelerate the inferencing process. This results in faster prediction,
detection and classification of data ingested to the edge layer.

Interworld has a range of BOXER industrial
embedded computers using Intel Movidius™ Myriad™ and NVIDIA Jetson processors
to bring intelligent AI solutions to any application. With a wide range of
features and profiles, the BOXER edge computing systems are built to fit the
needs of any application, from robotics and drones to intelligent security
systems and unmanned stores.

Features of NVIDIA Jetson TX2 Boxer-81x0AI
series as follows:

8110AI 8120AI 8130AI 8150AI 8170AI CPU HMP
Dual Denver 2 + Quad ARM A57 HMP
Dual Denver 2 + Quad ARM A57 HMP
Dual Denver 2 + Quad ARM A57 HMP
Dual Denver 2 + Quad ARM A57 HMP
Dual Denver 2 + Quad ARM A57 AI
Solution
NVIDIA Jetson TX2 NVIDIA Jetson TX2 NVIDIA Jetson TX2 NVIDIA Jetson TX2 NVIDIA Jetson TX2 Memory 8GB 8GB 8GB 8GB 8GB GbE
LAN
1 4 1 1 1+4xPoE Serial 1 2 1 1 2 USB3.0 2 2 2 8 4 Other
I/O
1xCANBus

1xOTG

1xOTG 1xOTG 1xOTG 1xOTG Storage 32GB eMMC

MicroSD

32GB eMMC

MicroSD

32GB eMMC

MicroSD

32GB eMMC

MicroSD

32GB eMMC

MicroSD

Display HDMI HDMI HDMI 2xHDMI 2xHDMI O/S
Support
Linux Linux Linux Linux Linux Power 12VDC 10-24VDC 10-24VDC 10-24VDC 12-24VDC

At
the core of the BOXER-8300AI series is the innovative AI
Core X module from AAEON. Each AI Core X module features the Intel
Movidius™ Myriad™ X VPU. The Intel Movidius™ Myriad™ X provides high
performance processing, with speeds up to 105 fps (80 fps typical) and 1 TOPS
as a dedicated neural network. The BOXER-8300AI Series features improved
thermal design, allowing the Intel Movidius™ Myriad™ X to operate at higher
temperatures without loss of performance.

The
BOXER-8300AI series offers users several advantages over other embedded AI@Edge
systems. All BOXER-8300AI systems are compatible with the Intel Distribution
of OpenVINO toolkit, allowing users to run AI inferences on existing frameworks
(such as TensorFlow or Caffe) or to create their own. The BOXER-8300AI supports
expandable storage, supporting the full suite of Linux and Ubuntu. This also
allows the BOXER-8300AI series to be expanded with support for wireless cards
such as WiFi, 4G, or Bluetooth. Memory is also easy to replace and upgrade.

The Features of Boxer-83x0AI series using Intel Movidius™ Myriad™ as follows:

8310AI 8320AI 8330AI
CPU N4200/N3350 I3-6100U Core
i3/i5/i7

Xeon E3

AI
Solution
1x Intel
Movidius™ Myriad™ X
2x Intel
Movidius™ Myriad™ X
Upto 16x Intel
Movidius™ Myriad™ X
Memory Up
to 8GB
Up
to 16GB
Up
to 32GB
GbE
LAN
2 2 5
Serial 3 4 1
USB3.0 4 4 4
Audio Mic-In,
Line-Out
Mic-In,
Line-Out
Storage mSata 2.5 drive bay 2×2.5 drive bay
Display 1xVGA,
1xHDMI
1xVGA,
1xHDMI
O/S
Support
Linux Linux Linux
Power 9-24VDC 9-30VDC 12-24VDC