Gateworks Newport GW6903 SBC w/Google Coral TPU Machine Learning (ML) Accelerator for AI — What is it?

AI Graphic

This demo features a Gateworks Newport GW6903 Rugged Single Board Computer (SBC) running the Google Coral Edge Tensor Processing Unit (TPU) Mini-PCIe card.

The Coral TPU provides a means to perform advanced Machine Learning (ML) tasks in a low power, small form factor, Mini-PCIe card. The TPU Accelerator is based on a custom "Application Specific Integrated Circuit" (ASIC) that Google designed for hardware accelerated AI calculations. The TPU features the following:

The Gateworks Single Board Computer platform is an ideal fit for the Google Coral TPU Mini-PCIe card, especially because it supports half sized cards.

The above image shows the Google Coral TPU card in the more left Mini-PCIe slot on the bottom of the GW6903 SBC and a wireless radio in the more right upper slot

This TPU is appropriate for an application where identifying an object or pattern is required. This could be, but is not limited to:

  • Object detection
  • Pose or gesture estimation
  • Image segmentation
  • Key phrase detection

Some practical applications are:

  • Autonomous vehicles
  • Robots
  • Voice control/Language Processing
  • Monitoring devices

Nearly all industries can benefit from this technology. To name a few more specifically:

  • Health Care
  • Agriculture
  • Manufacturing
  • Oil and Gas
  • Security/Defense
  • Automated Kiosk

For more information on the Google Coral TPU Mini-PCIe card see the following links:

Getting started with the Coral Edge TPU on Gateworks SBCs

To begin you will need:

  • A workstation with Linux natively installed.
  • A Gateworks Newport SBC — Coral requires AARCH64.
  • A Coral EDGE TPU, our testing was done with the mPCI-e form factor model.
  • Network/Internet connection.
  • Optional: USB webcam

Compiling the kernel

The Gateworks kernel defconfig for Newport does not include support for video devices. To create an image with webcam support:

  • Acquire the Newport BSP, we will call the directory this repo has been sync'ed to the <BSP> directory.
  • Follow the steps here to modify the kernel and create and Ubuntu image.
    • In the menuconfig enable the module "USB_VIDEO_CLASS", this will allow you to use a USB webcam with v4l-utils.
    • Complete the procedure detailed in the aforementioned section, build your Bionic image and flash it to your SBC
    • Though it was not specifically created to address using the Newport BSP this video may provide some insight into the process of changing the kernel config.

Building and installing the Gasket and Apex modules

Gateworks Video detailing the process:

The source code for the modules can be downloaded here as a tar.gz file:

Build the source using the same method as you would for an out of tree module (out of tree = when the module source is not included in the kernel source).

  • cd to the <BSP> directory.
  • source setup-environment 
    • Doing this will configure the toolchain for building. Keep in mind some of the lines in this file use the argument $PWD thus it should not be sourced from any location other than the <BSP> folder.
  • You can verify the arguments have been exported by executing the command
    echo $ARCH

This will return "arm64"

  • cd to <the_module_directory>
  • cat the "Makefile", you'll see two variables which need to be set for the modules to build correctly.
  • Execute the following command
    make -C <BSP>/linux M=$PWD

This procedure will result in two modules being created, "apex.ko" and "gasket.ko". Copy the .ko files to your target board "/lib/modules/<kernel_version>/extra/" folder. Using SCP may be the simplest way to go about this.

With the modules copied to the board execute the following commands:

depmod -a
insmod gasket.ko

Remove power from the board and reboot. On reboot verify that "/dev/apex_0" device is present.

Installing and configuring Python

Python 3.7 is required to run the tenser flow examples. Other versions can be used, though at the time of writing this wiki 3.7 is the best option when using the Bionic Ubuntu for Newport BSP.

apt-get update
apt install python3.7 -y

Set Python 3.7 to have priority over 3.6.

update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.6 1
update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.7 2  
update-alternatives --config python3
#on this menu enter the number 2. 

You can verify you have been successful in changing the version with the following command:

python3 --version

Installing the TPU runtime

Install the util curl

apt-get install curl -y

Add Debian package repository to your system

echo "deb coral-edgetpu-stable main" | sudo tee /etc/apt/sources.list.d/coral-edgetpu.list
curl | sudo apt-key add -
sudo apt update

Install PIP and required libraries

apt-get install python3-pip libedgetpu1-std -y
pip3 install --upgrade pip setuptools wheelpi

Download the TFlite Runtime .whl


Note: If you have chosen not to use Python 3.7 you can find the .whl appropriate for your version here

Acquire necessary libraries to build the TFlite runtime

pip3 install cython
pip3 install numpy
apt-get install python3-pil -y

Install TFlite runtime

pip3 install tflite_runtime-2.1.0.post1-cp36-cp36m-linux_aarch64.whl  

Standard Demo: Downloading Classification example & Testing an Inference

Create a place for the Coral examples to reside.

mkdir coral && cd coral

Clone the examples, "cd" to examples directory.

git clone
cd tflite/python/examples/classification

Install prerequisite programs using supplied script.


Run the classification demo, passing in a image of a parrot, as shown with the last argument.

python3 \
--model models/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite \
--labels models/inat_bird_labels.txt \
--input images/parrot.jpg  

Example output:

root@bionic-newport:~/coral/tflite/python/examples/classification# python3 \
> --model models/mobilenet_v2_1.0_224_inat_bird_quant_edgetpu.tflite \
> --labels models/inat_bird_labels.txt \
> --input images/parrot.jpg  
Note: The first inference on Edge TPU is slow because it includes loading the model into Edge TPU memory.
Ara macao (Scarlet Macaw): 0.77734

Advanced Demo: Video Webcam Detection with GStreamer Example

This example will use a USB webcam and the TPU to identify objects presented to the webcam. The video output and overlay will be streamed to a location on the network for viewing.

Install Gstreamer, you will need this program on both the SBC and your workstation where you will be viewing the output.

apt-get install gstreamer1.0-x gstreamer1.0-tools gstreamer1.0-plugins-good gstreamer1.0-plugins-bad gstreamer1.0-alsa -y

Clone the Gstreamer example

mkdir google-coral && cd google-coral
git clone --depth 1

Download models. Models are the information that will be feed to the TPU for it to reference when identifying an object.

cd examples-camera

Configure the Gstreamer Python script

cd gstreamer

Adaptations to the existing will be required for this example to work. A modified script is available to download from here. Edit this file line 231 with the IP address of the desktop workstation you will be streaming to. This is what the edited line will look like.

  ! rsvgoverlay name=overlay ! videoconvert ! jpegenc ! tcpclientsink host= port=9001
  • The workstation being used in this example has an IP address of, the SBC is using an IP which is on the same subnet.

Launch Gstreamer on your workstation to receive the video from the Gateworks SBC:

gst-launch-1.0 tcpserversrc host= port=9001 ! jpegdec ! videoconvert ! autovideosink sync=false

On the Gateworks SBC run:


Here's a video demonstration of the output you can expect if everything is working:


Youtube screen

Going Further

Some additional links for models and creating your own models:

Other AI Modules

Note the below modules have not been verified compatible or tested on Gateworks SBCs:

Last modified 22 months ago Last modified on 09/02/2022 04:37:37 PM

Attachments (8)

Download all attachments as: .zip

Note: See TracWiki for help on using the wiki.