Choosing an IoT Application Processor

Thursday 2/3/22 08:12am
|
Posted By Rajan Mistry
  • Up1
  • Down0

Qualcomm products mentioned within this post are offered by
Qualcomm Technologies, Inc. and/or its subsidiaries.

As the world of IoT evolves, today’s uses cases and verticals demand that processing be done at or near the device edge. At the same time, device architects designing their next edge devices are compelled to balance a range of requirements including new form factors, various operating environments, power efficiency, and deployment budgets, all while satisfying a cost-effective bill of materials (BOM).

In some cases, this means selecting a powerful IoT processor with advanced capabilities, while in other cases, a simpler but more specialized processer fits the bill.

As you start designing your device, here are a few key questions to consider:

  • What sort of Internet backhaul/backbone (e.g., 5G, Ethernet, etc.) will the device require?
  • What local connectivity is needed (e.g., Cellular, Wi-Fi, Bluetooth, etc.)?
  • Is there a power source available to support hardwiring? Or will the device have to rely on a battery?
  • Is hardware-accelerated AI processing required or will the device depend on cloud services for heavy inference processing?
  • What video resolution, frame rate, and codec capabilities are needed?
  • Which embedded operating system is required (e.g., Linux or Android)?

Answering these questions can help you select an appropriate application processor. And to support you, we offer a range of IoT application processors to match your requirements.

Below are some of our IoT application processors built on our advanced mobile technology for you to consider:

Qualcomm QCS610 and Qualcomm QCS410
The QCS610 is a high-tier application processor and sibling to the mid-tier QCS410. Both are Linux-based systems designed for building solutions that require AI at the edge. This includes edge applications like industrial IoT, smart AI enterprise security, home IP cameras, dash cams, body cams, smart displays, and video conferencing.

Both application processors include our Qualcomm Hexagon DSP and Qualcomm Adreno 612 GPU, and multiple connectivity options including, Ethernet, Wi-Fi, and Bluetooth. The QCS610 features a more powerful Qualcomm Kryo CPU than the QCS410. The QCS610 also supports higher camera and video resolutions (4K versus HD) than its mid-tier sibling, and can simultaneously run multiple AI models for inference.

Since both application processors feature the Hexagon DSP, you can take advantage of the Qualcomm Neural Processing SDK to convert your models to our DLC format for execution on Hexagon and our AI Model Efficiency Toolkit (AIMET) to further optimize models. You can also use the Hexagon SDK for bare-metal DSP development.

Several developer kits and reference designs are available including, the eInfochips EIC-Q610-200 and Altek UPC610 Open Development Kit. See the Qualcomm Application Processors Selector Guide for additional kits and details.

For additional inspiration on what you can build with the QCS610, read about our Remote Kiosk Project and check out some of our other projects on QDN. We also have several learning resources that provide additional development details.

Qualcomm QCS2290
The QCS2290 is an entry-level, cost-effective SoC with LTE, Wi-Fi, and Bluetooth connectivity for industrial and retail IoT applications. Potential use cases suited for the QCS2290 include asset tracking, industrial handhelds, and smart cameras.

The QCS2290 shares pin-to-pin compatibility with the more-powerful QCS4290 application processor (described below), allowing you to take advantage of shared hardware designs across IoT devices.

If you are interested in using the QCS2290, check out the TurboX CM2290/C2290 Development Kit based around their CM2290/C2290 SoM.

Qualcomm QCS4290
Like the QCS2290, our mid-tier QCS4290 provides LTE, Wi-Fi, and Bluetooth connectivity but has more processing power thanks to its Kryo CPU and Hexagon DSP. This means you can take advantage of tools like the Qualcomm Neural Processing SDK to convert your models to our DLC format for execution on Hexagon, and our AI Model Efficiency Toolkit (AIMET) to further optimize your models. You can also use the Hexagon SDK for bare-metal Hexagon DSP development.

Potential use cases/verticals for the QCS4290 include industrial handhelds, security panels, cameras, logistics, and warehousing.

The QCS4290 shares pin-to-pin compatibility with the QCS2290 (described above), allowing you to take advantage of shared hardware designs across IoT devices.

Qualcomm QCS8250
The QCS8250 is our premium-tier, Android-based application processor for enterprise and commercial IoT applications. It specializes in edge AI and video analytics, making it suitable for video collaboration, smart cameras, connected healthcare, smart retail digital signage, and fleet management. You can also use it to build AI hubs to offload work from sensing devices such as multiple camera feeds – up to 24 in total.

Take advantage of the Qualcomm Neural Processing SDK to convert your models to our DLC format for execution on Hexagon and our AI Model Efficiency Toolkit (AIMET) to further optimize your models. You can also use the Hexagon SDK for bare-metal DSP development.

If you are interested in the QCS8250, check out the eInfochips EIC-QCS8250-210 development kit or the Thundercomm Turbox C865 Development Kit. See the Qualcomm Application Processors Selector Guide for additional kits and details.

You can find information on these application processors and others, along with supportive resources such as developer kits, learning resources, etc., on our Application Processor page on QDN.


Qualcomm QCS610, Qualcomm QCS410, Qualcomm Hexagon, Qualcomm Adreno, Qualcomm Neural Processing SDK, Qualcomm Kryo, Qualcomm QCS2290, Qualcomm QCS4290, and Qualcomm QCS8250 are products of Qualcomm Technologies, Inc. and/or its subsidiaries. AIMET is a product of Qualcomm Innovation Center, Inc.