Compact edge AI developer kit is based on Nvidia Jetson Xavier NX

The AIE110-XNX is the latest edge AI developer kit from Axiomtek. It is powered by the Nvidia Jetson Xavier NX which features six core Nvidia Carmel Arm v8.2 64-bit CPU delivering up to 21 Tera operations per second  (TOPS) of accelerated AI computing performance. It integrates an advanced 384-core Nvidia Volta GPU with 48 Tensor cores. Aimed to save effort on system integration and reduce complexity for embedded developers, the AIE110-XNX enables users to develop AI solutions on a friendly budget with no limit of minimum order. The palm-sized edge AI developer kit is ideal for makers, learners, and system integrators to build more AI innovation with little effort.

The AIE110-XNX can be quickly engaged to be the AIE100-903-FL-NX for smart city, the IP67-rated AIE800-904-FL for outdoor edge AI applications, the AIE900-XNX for robotics and AMR or self-service kiosk applications, advised Axiomtek. It can also be upgraded with AI performance such as Nvidia Jetson Orin NX, Nvidia Jetson Orin Nano. The AIE110-XNX has reserved mounting holes on the bracket, which helps system integrators to start development.

The edge AI developer kit AIE110-XNX has a rugged design for harsh environments, allowing it to operate under a wide temperature range from -10 to +60 degrees C and vibration of up to 3Grms. It has 8Gbyte of LPDDR4x memory and 16Gbyte eMMC onboard. One M.2 Key M 2280 SSD slot with a high-speed PCIe x4 NVMe interface and one Micro SD slot are available for extensive storage needs. One 15W GbE PoE port is available for both power and video transmission to easily deploy an IP camera. It also supports one PCI Express Mini Card slot for Wi-Fi, Bluetooth, LTE, and GPS. 

The AIE110-XNX supports the Nvidia JetPack, which provides a full development environment for end-to-end accelerated development on Nvidia Jetson modules, to foster the development of AI-assisted operations across industries.

The AIE110-XNX supports device monitoring and management services by Allxon, bringing comprehensive remote management onto edge AI devices. Axiomtek’s edge AI developer kit AIE110-XNX is available now.

http://www.axiomtek.com 

> Read More

Dual-band SoC supports long-range wireless protocols for robotics

A dual-band FG28 SoC, designed for long-range networks and proprietary protocols such as Amazon Sidewalk and Wi-SUN includes radios for sub-GHz and 2.4Ghz Bluetooth LE. Silicon Labs announced that the FG28 also has a built-in AI / ML accelerator for machine learning inference and its proprietary Secure Vault technology.

“The FG28 SoC addresses several key needs for our customers when it comes to developing and deploying low-power wide area networks,” said Daniel Cooley, CTO of Silicon Labs. “By including Bluetooth, it gives users an easy way to provision and deploy new devices onto the network, while the sub-Ghz band is designed to support device communications over one mile, allowing for new edge applications in areas like smart agriculture, smart cities, and neighbourhood networks like Amazon Sidewalk.”

Emerging applications require solutions that can provide a “Swiss army knife” of connectivity options, said Silicon Labs. By providing sub-Ghz connectivity using networks like Wi-SUN, the FG28 can serve as the battery-operated end node in smart cities, for example, as a tracker on refuse collection lorries to help locate them and check the last time they were emptied. Other uses are in irrigation systems across multi-acre commercial farming operations, or as livestock trackers and health monitors on sprawling ranches. Bluetooth connectivity not only allows for devices to easily be deployed on the network, but also allows for operators to connect to the device locally for diagnostics and data downloads.

Other features which open new applications for sub-Ghz devices are what is claimed to be the industry’s first integrated AI / ML hardware accelerator in a sub-Ghz SoC. This allows for machine learning inference at the edge to enable predictive maintenance warnings, monitor soil conditions for key conditions like moisture and pH levels.

There is also 1024kbytes of flash storage and 256kbytes of RAM, to meet the memory needs of a wide range of protocols and technology stacks, said Silicon Labs.

Its Secure Vault Mid and High support allow designers to choose the level of security needed for applications while the energy-efficient radio core with low active and sleep currents with fast wakeup times preserver battery life for end nodes.
There are also up to 49 general purpose I/O pins.

The FG28 is sampling now with general availability planned before the end of Q3 2023. Derivatives optimised for Amazon Sidewalk and Z-Wave Long-Range and a standalone MCU are also planned.

http://www.Silabs.com 

> Read More

u-blox bases two GNSS positioning modules on F9 

Two GNSS positioning modules, the NEO-F9P for industrial machinery and the ZED-F9P-15B for mobile robotics, have been introduced by u-blox.

Both are based on the company’s F9 platform. The low-power NEO-F9P supports navigation and automation of moving industrial machinery and the ZED-F9P-15B provides customers in the mobile robotics market with an L1/L5 option in addition to the L1/L2 bands.

The u-blox NEO-F9P GNSS module features include concurrent reception of GPS, Galileo and BeiDou, multi-band L1/L5 RTK and short convergence times. It delivers centimetre-level accuracy in seconds and comes in the smallest form factor for a high-precision module, said u-blox. It is 50 per cent smaller than the regular u-blox ZED. 

The small size, coupled with very low power consumption and ANN-MB1 antenna compatibility, makes the u-blox NEO-F9P suitable for navigation and automation in smart antennas, UAVs (unmanned aerial vehicles) and mobile robotics. It supports open as well as standards-based correction services for enhanced performance, such as the u-blox PointPerfect GNSS augmentation service.

The modules are the company’s response to the increased requirement for precise navigation and automation in industrial applications, particularly in the UAV and mobile robotics sectors, commented Alex Ngi, product manager, industrial navigation and robotics at u-blox. “By delivering centimeter-level accuracy in a compact and energy-efficient module, the NEO-F9P, u-blox proven RTK engine, is ideal for space-constrained designs,” he added.  

The NEO-F9P offers customers an effective upgrade path from the u-blox NEO-M8P model and the ZED-F9P-15B provides existing ZED customers with an L1/L5 alternative to L1/L2.

u-blox specialises in positioning and wireless communication in automotive, industrial and consumer markets, providing services and products which let people, vehicles and machines determine their precise position and communicate wirelessly over cellular and short range networks. The company’s portfolio include chips, modules and secure data services and connectivity to empower customers to develop innovative and reliable solutions for the IoT.

The company is headquartered in Thalwil, Switzerland and has offices in Europe, Asia, and the USA. 

http://www.u-blox.com

> Read More

Green Hills Software supports Imagination’s RISC-V CPUs 

At AutoTech Detroit this week, Imagination Technologies will be demonstrating Green Hills’ µ-velOSity running on its RISC-V CPUs’ RISC-V Catapult CPUs. The two companies have announced a partnership in which Green Hills’ µ-velOSity RTOS will support Imagination Technologies’ CPUs to accelerate real-time safety and security in automotive and industrial designs.

This partnership will extend support to Imagination’s Catapult CPU family roadmap in the future, confirmed the two companies. 

The µ-velOSity RTOS is the smallest of Green Hills Software’s family of real time operating systems. It has been updated and optimised to support the RISC-V architecture and has been certified to meet industry standards for functional safety and security. It offers seamless integration with the safety-certified Green Hills MULTI integrated development environment (IDE) and C/C++ compilers, in order to make it easy for developers to learn and use to create high performance, small footprint designs for automotive, industrial, and IoT applications.

“We have a long-standing relationship with Imagination and are thrilled to extend our support to the company’s RISC-V CPUs,” said Dan Mender, vice president of business development at Green Hills Software. “The adoption of the RISC-V instruction set architecture is increasing rapidly in countless markets, particularly in IoT, industrial, and embedded applications. By using Imagination’s Catapult family and our µ-velOSity RTOS, advanced debugger and optimising C/C++ compilers, customers can efficiently create and confidently deploy safety- and security-critical RISC-V applications,” he added.

Catapult is Imagination’s RISC-V product line designed from the ground up for deployment and is configurable for use in a variety of markets, said the company.

Imagination’s RISC-V real-time CPUs can be integrated into complex SoCs for applications as wide ranging as networking, packet management, storage controllers and sensor management for AI cameras and smart metering. The company also provides GPU and AI accelerator IP and confirmed it will be expanding its RISC-V CPU product line with more applications-focused solutions in the future. All Imagination GPUs are compatible with RISC-V SoCs.

Imagination will be demonstrating Green Hills’ µ-velOSity running on Imagination’s RISC-V CPUs at AutoTech Detroit (7 to 8 June) at Booth 827. 

https://www.imaginationtech.com

> Read More

About Smart Cities

This news story is brought to you by smartcitieselectronics.com, the specialist site dedicated to delivering information about what’s new in the Smart City Electronics industry, with daily news updates, new products and industry news. To stay up-to-date, register to receive our weekly newsletters and keep yourself informed on the latest technology news and new products from around the globe. Simply click this link to register here: Smart Cities Registration