Embedded World 2018
Embedded World broke its own records—again—this year, with over 1,000 exhibitor companies and more than 32,000 international visitors from the embedded community.
Nuremberg, Germany was a winter wonderland for this year’s Embedded World, as freezing temperatures gripped the Nuremberg Messe showground. Inside the six halls of the world’s biggest embedded industry event, however, was a hotbed of IoT, automation, automotive, communications, and networking innovation.
All types and sizes of modules were on display, illustrating the diversity of choice available in the market today. One of the most interesting was the System on Module (SoM) for industrial-grade Linux designs by Microchip. The ATSAMA5D27-SOM1 (Figure 2) is designed to remove the complexity of developing an industrial system based on a microprocessor, running Linux® OS. Lucio di Jasio, Business Development Manager, Europe, at Microchip, explains that the 40 x 40mm board will help engineers with PCB layout in these applications. It has the company’s ATSAMA5D27C-D1G-CU System in Package (SiP) and uses the SAMA5D2 MPU. The small form factor manages to integrate power management, non-volatile boot memory, Ethernet PHY and high speed DDR2 memory to develop a Linux-based system or it can be used as a reference design. Schematics, design, and Gerber files are available online and free of charge.
For security, the SAMA5D2 family has integrated Arm TrustZone® and capabilities for tamper detection, secure data and program storage, hardware encryption engine and secure boot. The SoM also contains Microchip’s QSPI NOR Flash memory, a Power Management Integrated Circuit (PMIC), an Ethernet PHY and serial EEPROM with a Media Access Control (MAC) address.
There is a choice of three DDR2 memory sizes for the SiP, 128 and 512Mbit and 1Gbit, all optimized for bare metal, Real-Time Operating Systems (RTOS) and Linux OS. All of Microchip’s Linux development code for the SiP and SOM are in the Linux communities.
Customers can transition from the SOM to the SiP or the MPU, depending on the needs of the design, adds di Jasio.
The company also announced two new microcontroller families, one for the PIC range and one for megaAVR series. The PIC16F18446 microcontrollers are suitable for use in sensor nodes, while the ATmega4809 is the first megaAVR device to include Core Independent Peripherals (CIPs), to execute tasks in hardware instead of through software, decreasing the amount of code required to speed time to market.
A COM Express Type 6 module from congatec shows how the company is wasting no time in exploiting the prowess of AMD’s latest Ryzen™ processor. The conga-TR4 (Figure 3) is based on neighboring exhibitor AMD’s Ryzen Embedded V1000 processors.
The company has identified embedded computing systems that need the graphics performance of the Ryzen processor for medical imaging, broadcast, infotainment and gambling, digital signage, surveillance systems, optical quality control in automation and 3D simulators.
The Ryzen Embedded V1000 was launched days before Embedded World, together with the EPYC™ Embedded 3000 processor, and made up the ‘Zen’ zone in the company’s booth, as the two processors enter a new age for high-performance embedded processors, explains Alisha Perkins, Embedded Marketing Manager at AMD.
Focusing on the Embedded V1000, Perkins explained that it targets medical imaging, industrial systems, digital gaming, and thin clients to the edge of the network. It increases performance by a factor of two, compared with earlier generations, has up to three times more GPU performance than any processor currently available, has nearly half as much again (up to 46%) more multi-threaded performance than competing alternatives and, crucially for mobile or portable applications, is 36% smaller than its competitors.
AMD couples its Accelerated Processing Unit (APU) with Zen Central Processing Units (CPUs) and Vega Graphics Processing Units (GPUs) on a single die, offering up to four CPU cores/eight threads and up to 11 GPU compute units for a 3.6TFLOPS processing throughput. On the stand were examples of medical imaging stations that could be wheeled between wards, a dashboard for remote monitoring of utilities and an automated beer bottle checking visual inspection station, using the high-performance graphics and computing powers of the processor.
At congatec, the COM Express Type 6 module was also cited as being suitable for smart robotics and autonomous vehicles, where its Thermal Design Power (TDP) is scalable from 12 to 54W to optimize size, weight, power and costs (SWaP-C) at high graphics performance, says Christian Eder, Director of Marketing, congatec.
For the smart, connected factory, Texas Instruments introduced its latest SimpleLink™ microcontrollers (MCUs) devices, with concurrent multi-standard and multi-band connectivity for Thread, Zigbee®, Bluetooth® 5 and Sub-1 GHz. Designers can reuse code across the company’s Arm® Cortex®-M4-based MCUs in sensor networks to the cloud.
The additions announced in Nuremberg expand the SimpleLink MCU platform to support connectivity protocols and standards for 2.4 GHz and Sub-1 GHz bands, including the latest Thread and Zigbee standards, Bluetooth low energy, IEEE 802.15.4g and Wireless M-Bus. The multi-band CC1352P wireless MCU, for example, has an integrated power amplifier to extend the range for metering and building automation applications, while maintaining a low transmit current of 60mA.
SimpleLink MSP432P4 MCUs have an integrated 16-bit precision ADC and can host multiple wireless connectivity stacks and a 320-segment Liquid Crystal Display with extended temperature range for industrial applications.
Security is addressed with new hardware accelerators in the CC13x2 and CC26x2 wireless MCUs, for AES-128/256, SHA2-512, Elliptic Curve Cryptography (ECC), RSA-2048 and true random number generator (TRNG) encryption protocols.
Code compatibility: These new products are all supported by the SimpleLink software development kit (SDK) and provide a unified framework for platform extension through 100 percent application code reuse.
Still with automation, ADLINK’s Jim Liu, CEO, has his sights set on Artificial Intelligence (AI). “We have gone from being a pure embedded CPU vendor to an AI engine vendor,” he says, introducing autonomous mobile robotics and ‘AI at the Edge’ solutions using NVIDIA technology.
Its embedded systems and connectivity couple with NVIDIA’s AI and deep learning technologies to target compute-intensive applications, such as robotics, autonomous vehicles and healthcare. Demonstrations included an autonomous mobile robot platform using ROS 2, an open source software stack specifically designed for factory-of-the-future connected solutions. There was a smart camera technology that can scan barcodes on irregularly shaped objects and differentiate between them (Figure 4). Another demonstration calculated vehicle flow to improve traffic management in a smart city.
Arm was also moving to the edge of computing, with machine learning and a display that fascinated many—a robotic Rubik’s Cube solver (Figure 5). John Ronco, Vice President and General Manager, Embedded and Auto Line of Business at Arm, sounds a cautious note: “Inference at the edge of the cloud has network, power and latency issues; there are also privacy issues,” he says. Ahead of Embedded World, the company announced its Project Trillium, promoting its Machine Learning (ML) technology, using an ML processor that is capable of over 4.6 Trillion Operations per Second (TOPs) with a power conserving efficiency of 3TOPs/W, and an object detection processor.
Swedish embedded software tools and services company, IAR Systems shared news of its many recent partnerships. The first, with Data I/O, is to bridge the development -manufacturing gap by integrating the software with the latter’s data programming and secure provisioning to transition the microcontroller firmware development to manufacture. The two share many customers within the automotive, IoT, medical, wireless, consumer electronics, and industrial controls markets, although at separate stages of the design and manufacturing process. To address the growing complexity of designs and the security concerns in the embedded market, explains Tora Fridholm, Product Marketing Manager at IAR Systems, the two companies have established a roadmap based on customer requirements for a workflow where resources such as images, configuration files, and documents can be securely shared. Customers thus enjoy an efficient design to manufacturing workflow that reduces time to market.
If adding device-specific security credentials, such as keys and certificates, both companies are committed to integrate the appropriate processes and tools.
Another announcement was with Renesas Electronics, whereby its Synergy™ can use the Advanced IAR C/C++ Compiler™ in e² Studio Integrated Development Environment (IDE) to reduce application code, allowing more features to be added to Synergy microcontrollers. There is also the benefit of the compiler’s execution speed, which allows the microcontroller to remain in low power mode to conserve battery life.
Synergy microcontrollers are used in IoT devices to monitor the environment in buildings and industrial automation, energy management, and healthcare equipment.
An essential part of embedded design is board technology, and this year’s show did not disappoint. WinSystems was highlighting two of its latest single board computers, the PX1-C415 (Figure 6) which manages IoT nodes, and the SBC35-C427, based on the Intel Atom® E3900 processor series.
The first uses Microsoft® Windows® 10 IoT Core OS to support IoT development, the second is designed for industrial IoT, with an onboard ADC input, General Purpose Input/Output (GPIO), dual Ethernet, two USB3.0 and four USB 2.0 channels. It can be used in transportation, energy, industrial control, digital signage and industrial IoT applications.
The SBC supports up to three video displays via DisplayPort and LVDS interfaces. It can be expanded using the Mini-PCie socket, an M.2 connector (E-Key) and the company’s own Modular I/O 80 interface.
A COM Express Type 7 module by ConnectTech was among the company’s booth highlights. The Com Express Type 7 + GPU Embedded System (Figure 7), can be used across four, independent or as a headless processing system. It has Intel Xeon® D x86 processors, NVIDIA Quadro® and Tesla® GPUs in a 216 x 164mm form factor. It anticipates the needs of high-performance applications that require 10GbE and Gigabit Ethernet, USB 3.0 and USB 2.0, HDMI, SATA II, I2C, M.2 and miniPCIe for encode/decode video, GPGPU CUDA® processing, deep learning and AI applications.
Caroline Hayes has been a journalist covering the electronics sector for more than 20 years. She has worked on several European titles, reporting on a variety of industries, including communications, broadcast and automotive.