SoM is designed for advanced IoT vision systems
Inforce 68A1 is a system on module (SoM) designed to enable advanced IoT vision system devices by Smart Wireless Computing and based on Qualcomm’s QCS8250 SoC.
It is designed for high performance compute-intensive network edge IoT camera applications. The Inforce 68A1 SoM combines support for multiple high resolution 8K or 4K cameras with complex artificial intelligence (AI) inferencing at the edge to enable smart camera systems, conferencing or collaboration devices.
The high-resolution video capabilities include processing 4K60 Ultra HD content for multiple cameras and displays concurrently, as well as high bandwidth connectivity via 802.11ax DBS Wi-Fi 6E and Bluetooth 5 for connected camera networks. The module makes use of Qualcomm Technologies’ Hexagon DSP and Qualcomm AI Engine to run heterogeneous deep learning workloads and neural networks for high-performance, on-device AI inferencing in video analytics.
The SoC integrates the Qualcomm Kryo 585 CPU, optimised for multi-tasking, the Qualcomm Adreno 650 GPU, delivering twice the performance of the previous generation platform and 60 per cent greater power efficiency. There is also the Hexagon 690 DSP which runs power efficient machine learning with quad-threaded Scalar performance and twice the Vector eXtensions performance. A dedicated computer vision hardware block and new Hexagon Tensor Accelerator delivering 15 TOPS of AI performance are also included. The Qualcomm Spectra 480 ISP enables image capture at 64MPixel at 30 frames per second ZSL with no shutter lag and quick auto focus.
The Adreno DPU995 display engine enables dual concurrent display content at HEVC 10bit and VP9 decode.
Wired and wireless connectivity options include Wi-Fi 6, 5G, Bluetooth 5.1 RF front-end technologies and dual USB 3.1. There is also a secure processing unit which enables secure boot, DRM and content protection creating a secure environment for the camera system, says Smart Wireless Computing.
The Inforce 68A1 SoM also leverages PMIC regulators for large power requirements and on-board power management. The module also features on-board memory and storage with 8Gbyte DDR5 RAM and 64Gbyte universal flash storage (UFS) for fast multi-tasking.
The small form factor SoM measures just 35 x 53 x 7.5mm.
AI technology can simplify audio visual collaboration, especially as remote working becomes more widely adopted, says Smart Wireless Computing. It says that by integrating the Inforce 68A1 SoM into enterprise video conferencing solutions, businesses can deliver content, connect and interact effortlessly, from anywhere, for a flawless video experience.
It also has applications in the medical field. Applying computer vision AI to equipment such as portable ultrasound machines, it can help provide remote anatomical identification with fewer clicks and increased accuracy, says the company. Anatomical identification also helps automate measurements, make them more reproducible, and can give medical professionals tools to improve patient care, and enhance the imaging resources of hospitals and healthcare trusts.
In retail, businesses are using AI video content analysis for data processing which leads to a deeper understanding of consumer attitudes, preferences, and buying habits. By integrating the Inforce 68A1 SoM into retail camera systems and self-checkout technology, retailers can improve merchandising and enhance customer behaviour and engagement, says Smart Wireless Computing.
“With vast progress being made in AI and machine learning, applications that previously involved manual intervention can now be automated or enhanced to solve some of the most common challenges in our day-to-day lives,” said Todd Wynia, vice president, product marketing at Smart Wireless Computing. He said that the company’s application-ready embedded platforms are at the core of this trend, providing multiple options for IoT connectivity, multimedia capabilities and powerful edge computing power to change the way we communicate, shop and approach healthcare.
The Qualcomm QCS8250 SoC is optimised for maximum performance while balancing power efficiency with support for Wi-Fi 6 and 5G connectivity, explained Siddhartha Franco, director, business development, Qualcomm Technologies. He said the Inforce 68A1 SoM “brings technology at the connected intelligent edge to a broader ecosystem of developers, helping reduce the time and risks involved with integrating premium tier AI performance into their edge camera solutions.”
This module includes an Android 10 board support package with the Qualcomm Technologies’ software development kits, including Hexagon SDK and Qualcomm neural processing engine software development kit, enabled.
Samples are available to order from this month.
