Embedded Computer Systems: COM and AI Edge Computing Technology

Ultra-Miniature Computer-on-Module for On-Device AI Applications

Advantech's ROM-2860 provides 12 TOPS of AI capability and supports dual-camera input By Mike Ball / 25 Jun 2024
Follow UST

Advantech has launched the ROM-2860, a new Size-L Open Standard Module (OSM) designed to meet the demanding requirements of autonomous mobile robotics (AMR) and Artificial Intelligence of Things (AIoT) applications.

Ultra-Miniature Computer-on-Module for On-Device AI Applications

Powered by the Qualcomm Octa-core QCS6490 processor with a 6.9-watt TDP, the new module achieves nearly 20% better performance than mid-tier x86 processors, coupled with a 43-75% reduction in power consumption and an ultra-compact 45 x 45 mm form factor.

By supporting Windows 11 on ARM, the ROM-2860 maximizes its potential with Arm architecture, enhancing battery life and ensuring consistent performance. Arm Systems on Chip (SoC) often come with essential features including CPU, GPU, Wi-Fi, mobile data networks, and Neural Processor Units (NPUs) to handle AI workloads. This provides users seamless compatibility with existing apps and tools on new Arm-powered devices.

The ROM-2860 accommodates diverse embedded machine vision needs, featuring 12 TOPS of AI capability, dual-camera input support, and a native video codec. These capabilities make it ideal for applications requiring seamless video capture, low-latency signals, limited installation space, and resilience to vibrations. Compared to traditional devices, the ROM-2860’s on-device AI improves operational efficiency in use cases such as drones and robotics, allowing them to use cameras, radar, and LiDAR for initial AI modeling and then to gain inference capabilities for tasks like object detection, route optimization, localization, and collision avoidance.

The ROM-2860 also supports various development tools like Microsoft Azure AI Service, Qualcomm Snapdragon Neural Processing Engine (SNPE) SDK, and Qualcomm AI Hub, streamlining machine learning deployment and converting models from Aware, Azure, and other cloud service providers. With the Qualcomm QCS6490 CPU, it offers 64-bit program support. Developers can update Visual Studio to version 15.9 or above for ARM64 compatibility, simplifying binary compilation for existing projects.

The ROM-2860, with its compact OSM form factor, supports a variety of interfaces including PCIe Gen3 for 5G / Wi-Fi 6, LPDDR5, UFS, USB 3.2, URAT, GPIO, and more. Additionally, it offers dual-display capability with support for up to 2K resolution, including DSI, DP, and eDP. These features provide diverse I/O options that facilitate seamless connections within AIoT applications.

To learn more, contact Advantech: Visit Website Send Message View Supplier Profile
Posted by Mike Ball Mike Ball is our resident technical editor here at Unmanned Systems Technology. Combining his passion for teaching, advanced engineering and all things unmanned, Mike keeps a watchful eye over everything related to the unmanned technical sector. With over 10 years’ experience in the unmanned field and a degree in engineering, Mike’s been heading up our technical team here for the last 8 years. Connect & Contact