In the last few decades, not only has computing power grown exponentially, the requirements of them have done so as well. And while manufacturers continue to push out devices with increased processing capabilities with reduced power consumption, the strain on these devices is becoming more apparent.
For example, it was not long ago that 480p video was considered acceptable by the masses; however, now with larger monitors and faster internet connections, users are not satisfied unless they have 4K full-resolution streaming.
What’s more, AI is fast becoming integrated into more and more products. And now with the need for low latency, AI is shifting away from cloud computing and into edge devices. Current single-board computers (SBCs) on the market that may have been an acceptable solution five years ago are no longer able to meet the latest demands from consumers.
This is where Arrow Electronics steps up to the task, with the announcement of the release of their two latest products: the AI-ML and Thor96 Single-Board Computers.
AI-ML Single-Board Computer
The AL-ML is a single-board computer powered by the i.MX 8X SoC, centered on Arm’s A35 quad-core 64-bit processor clocked at 1.2 GHz. The SoC integrates a Cortex-M4F clocked at 266 MHz, which is capable of running independently from the main processor and can perform real-time tasks such as GPIO access and digital signal processing. Worth noting is that the Cortex-M4F is capable of running when the main processor is powered down. This allows the system to monitor external GPIO and perform other needed routines (such as wake-ups and interrupts). Also on board is 2 GB of LPDDR RAM Industrial Temp by Micron. This combination of a quad-core processor with large RAM space ensures that the AI-ML can tackle nearly any and all customer requirements.
The AI-ML is capable of H.265 decoding up to 4K resolutions at 30 fps as well as encoding H.264 1080 video at 60 fps. This impressive graphical capability is thanks to the dedicated GC7000 lite GPU, which includes four shaders, 16 execution modules and a Vulkan High-Performance 2D Blit engine with OpenGL support. The ability to encode and decode video allows the AL-ML to perform a wide range of graphical applications, including 4K displays in billboards and advertising, machine learning and live-stream image capture for drones.
When it comes to connectivity, the AI-ML goes above and beyond with the integration of an Ethernet port as well as a Laird/LSR 450-0169C module, which is powered by Cypress’s industry-leading 802.11ac Wi-Fi and Dual-Mode Bluetooth chipset. Of course, no SBC is complete without I/O support, and this is something that the AL-ML excels at. Two dedicated connectors for high- and low-speed signals provide developers with two ports containing various busses including UART, SPI, I2C, GPIO, I2S, CAN, camera and USB.
Now, not only is there a vast range of different busses available, there are also six LED indicators, with two dedicated for wireless reception/transmission and four dedicated as generic LEDs that can be controlled by the designer.
Thor96 Single-Board Computer
At the heart of the Thor96 single-board computer is the i.MX 8M SoC powered by Arm’s A53 quad-core 64-bit processor running at 1.5 GHz. Like the i.MX 8X, the i.MX 8M also integrates a Cortex-M4F running at 266 MHz for real-time digital signal processing as well as operation when the main processor is powered down. While graphical capabilities of the Thor96 are also powered by the GC7000lite GPU, the Thor96 is capable of H.265 4K at 60 fps. This makes it very much ideal for media applications involving large 4K displays demonstrating 60-fps footage.
The GPU also includes OpenGL support and incorporates a Vulkan High-Performance 2D Blit engine. As with the AI-ML, the Thor96 includes 2 GB of LDDR RAM Industrial Temp by Micron, making it highly capable of handling today’s customer demands.
Networking on the Thor96 includes an Ethernet port and a Murata Sterling-LWB5 powered by Cypress’s industry-leading 802.11ac Wi-Fi and Dual-Mode Bluetooth chipset. However, Wi-Fi and Bluetooth are not the only connections available for the Thor96: A Silicon Labs MGM111 Thread/Zigbee Module further expands the connection options. Connectivity of the Thor96 is similar to the AI-ML with two connectors (one for high speed and the other for low speed) and includes UART, SPI, I2C, GPIO, I2S, CAN and USB. Unlike the AI-ML, however, the Thor96 also has dual-camera capabilities.
To complement the Thor96 video capabilities it also includes support for Analog Devices high performance, low latency A2B® technology providing connectivity to multiple microphone arrays and speakers. Running both power and multi-channel digital audio over a single unshielded twisted pair of cables reduces the installation’s weight, cost and complexity while enabling advanced voice recognition and active noise cancellation applications.
Single-Board Computer Applications
The applications that can be applied to the AI-ML and the Thor96 are wide and vast thanks to the powerful quad cores, co-processor, large RAM space, high connectivity and superior graphical capabilities.
Interfacing with Building & Home Automation
Home automation is a market that could heavily benefit from the AI-ML and the Thor96 SBCs. The GPIO capabilities and wireless connectivity enable these computers to interface with both wired and wireless hardware, while their graphical capabilities allow them to interface with even the largest screens in the modern home while retaining the all-important 4K resolution.
AI can readily be integrated into these computers thanks to the quad-core Arm processor. Coupled with camera input systems, a personalized home based on the owner’s preferences can be realized.
Drones
Drones and autonomous vehicles are areas that would benefit from the AI-ML and Thor96 computers thanks to their small physical size, visual capabilities and wireless connectivity. The ability to read camera information and encode on the fly could easily allow for real-time processing of the environment for obstacle avoidance, while the Cortex-MF4 can be used for reading sensor data such as orientation and acceleration.
Low-energy applications
While the AI-ML and Thor96 provide superior SBC processing capabilities, some applications may need to conserve as much power as possible. This is where the Cortex-MF4 becomes vital as it can operate even when the main processor is in low-power mode. This allows for the system to run real-time tasks in the background that can wake up the system when appropriate conditions are met. Such a system would be highly beneficial to low-power applications that may be in remote locations with the system spending most of its time monitoring its environment and then waking up to respond to an event that may require visual processing (such as a remote facial recognition entry system).
Yocto OS Compatible
The AI-ML and Thor96 boards are both fully compatible with the Yocto OS Project. Most OS options involve a pre-made OS, but the Yocto project allows designers to customize and build their own Linux distro best suited for their system. This allows for improved speed by removing unneeded software, provides a custom solution for each project and can even help to improve security by minimizing the number of possible entry points into the system. Being an open-source project, the cost to the developer is only the time that they wish to put into their OS with free community support and the ability to integrate their OS into the final product.
The parts powering the boards
The AI-ML and Thor96 are powered by the following manufacturers, whose parts enable the production of the next generation of SBCs!
- Analog Devices
- Silicon Labs
- Cypress Semiconductor
- NXP
- TE Connectivity
- AVX
- Littelfuse
- Micron
As performance expectations for single-board computers continue to increase, solutions like the AI-ML and Thor96 SBCs will help design engineers produce better-quality solutions for everything from IoT to displays and more.
Voir les produits connexes
Voir les produits connexes
