Advantech AOM 2721 SOM

The Advantech AOM-2721 is a System on Module (SOM), specifically a Computer-on-Module, that uses the OSM 1.1 form factor. It's designed to be a compact and integrated computing platform, particularly suited for embedded applications and edge AI.

Key features of the AOM-2721:

• Powered by Qualcomm QCS6490/5430 SoC

• 1 Kryo Gold plus up to 2.7 GHz

• 3 Kryo Gold at 2.4 GHz

• 4 Kryo Sliver at 1.9 GHz

• Andreo VPU 633 4K30 encode/Decode

• Andreo GPU 643, OpenGL ES3.2/OpenCL 2.0

• 12 TOPs NPU for AI applications

• Onboard LPDDR5 8GB, 8533MT/s

• 1x MIPI-DSI x4, 1x DP and 1x eDP1.4 for Displays

• 1x USB3.2 Gen1, 1x USB2.0, 2x PCIe Gen3.0 x1, 1x PCIe Gen3.0 x2, 2x I2S, 4x

• 4wire UART, 3x SPI,39x GPIO, 4x I2C, 2x MIPI-CSI x4

Setting Up Your Advantech AOM 2721 SOM

0. Preface

The AOM 2721 SOM is typically the primary system component to a complete system. In this documentation, we will use the Advantech EPC-R2860 device as the sample "AOM 2721 SOM".

For example, pictures below showing micro switches on the R2860 that are used during the flashing process should be common to any AOM 2721 compatible device.

In the following setup instructions, the EPC-R2860 will be the actual device shown in the pictures. Other devices, based upon the AOM 2721 SOM may look slightly different.

1. (OPTIONAL) Flashing your AOM 2721 SOM

Your actual device containing the AOM 2721 SOM may already ship with a running yocto-based image flashed into it. If so, you can optionally proceed to step 3). If you want to flash your device, you will need to follow the instructions located here

Official yocto images created by Advantech for the AOM 2721 SOM can be found here

Once flashed, proceed to step 3) below.

2. (OPTIONAL) Building your own Yocto image for your AOM 2721 SOM

Some may want to actually fully build their own Yocto image for their AOM 2721 SOM. In this case, please refer to the following instructions located here to setup a Yocto build host and build a compatible yocto image for your AOM 2721.

2. Starting up your AOM 2721 SOM and connecting to the Internet

1. Install the Edge Impulse CLI on your computer.

2. Connect power to the back of the AOM-2721 SOM.

3. If your device is equipped, connect the COM1 serial port to your host computer's USB port via a SERIAL-TO-USB converter similar to the following (gender changers may also be needed on some implementations):

1. Open a serial connection between your host computer and the board.

edge-impulse-run-impulse --raw

1. Press the power button and the device should begin to boot up.

2. After 30-60 seconds you should see a login prompt in your serial terminal. Log in with:

   • Username: root

   • Password oelinux123

3. Next, set up a network connection, either:

   1. Connect an Ethernet cable.

   2. Or, if you want to connect over WiFi:

      • Qualcomm Linux <1.3: edit the wpa_supplicant.conf.

      • Qualcomm Linux 1.3: use nmcli.

1. If you want to continue setting up over ssh (so you can unplug the device from your computer), find your IP address via:

   Copy

   $ ifconfig | grep "inet addr:" | grep -v "127.0.0.1"
   inet addr:192.168.1.38 Bcast:192.168.1.255 Mask:255.255.255.0

   Then log in via ssh (password: oelinux123):

   Copy

   $ ssh [email protected]

3. Installing the Edge Impulse Linux CLI

On the AOM-2721 SOM, install the Edge Impulse CLI and other dependencies via:

$ wget https://cdn.edgeimpulse.com/firmware/linux/setup-edge-impulse-qc-linux.sh
$ sh setup-edge-impulse-qc-linux.sh

4. Connecting to Edge Impulse

With all dependencies set up, run:

$ edge-impulse-linux

This will start a wizard which asks you to log in and choose an Edge Impulse project. If you want to switch projects, or use a different camera (e.g. a USB camera) run the command with the --clean argument.

5. Verifying that your device is connected

That's all! Your device is now connected to Edge Impulse. To verify this, go to your Edge Impulse project, and click Devices. The device will be listed here.

Next steps: building a machine learning model

With everything set up you can now build your first machine learning model with these tutorials:

• Responding to your voice

• Recognize sounds from audio

• Adding sight to your sensors

• Object detection

• Visual anomaly detection with FOMO-AD

Looking to connect different sensors? Our Linux SDK lets you easily send data from any sensor and any programming language (with examples in Node.js, Python, Go and C++) into Edge Impulse.

Profiling your models

To profile your models for the Advantech AOM-2721 SOM:

• Make sure to select the Qualcomm Dragonwing AOM-2721 SOM as your target device. You can change the target at the top of the page near your user's logo.

• Head to your Learning block page in Edge Impulse Studio.

• Click on the Calculate performance button.

To provide the on-device performance, we use Qualcomm AI Hub in the background (see the image below) which run the compiled model on a physical device to gather metrics such as the mapping of model layers to compute units, inference latency, and peak memory usage. See more on Qualcomm® AI Hub documentation page.

Deploying back to device

Using the Edge Impulse Linux CLI

To run your impulse locally on the RB3, open a terminal and run:

$ edge-impulse-linux-runner

This will automatically compile your model with full hardware acceleration, download the model to your AOM-2721 SOM, and then start classifying (use --clean to switch projects).

Alternatively, you can select the Linux (AARCH64 with Qualcomm QNN) option in the Deployment page.

This will download an .eim model that you can run on your board with the following command:

edge-impulse-linux-runner --model-file downloaded-model.eim

Using the Edge Impulse Linux Inferencing SDKs

Our Linux SDK has examples on how to integrate the .eim model with your favourite programming language.

Using the IM SDK GStreamer option

When selecting this option, you will obtain a .zip folder. We provide instructions in the README.md file included in the compressed folder.

See more information on Qualcomm IM SDK GStreamer pipeline.

Image model?

If you have an image model then you can get a peek of what your device sees by being on the same network as your device, and finding the 'Want to see a feed of the camera and live classification in your browser' message in the console. Open the URL in a browser and both the camera feed and the classification are shown: