Week In Review: Auto, Security, Pervasive Computing

AI on a Cortex-M; Hyundai EV platform skateboard; predicting in-chip failures.


Arm putting AI (artificial intelligence) and machine learning (ML) on the Cortex-M processor by offering IP for a microNPU for Cortex-M. The company says in a press release that it will deliver a 480x uplift in ML performance. The new Cortex-M IP is Arm Ethos-U55 NPU, which Arm says is the industry’s first microNPU (neural processing unit). Arm is hoping the new IP will start an explosion of AI on IoT devices, which are constrained by small sizes and power budgets.

For its AI processor design, Mythic has adopted the Analog FastSPICE and Symphony platforms from Mentor, a Siemens business. FastSPICE will be used for custom circuit verification and device noise analysis, and Symphony Mixed-Signal Platform will be used to verify that the analog and digital logic in Mythic’s Intelligence Processing Units (IPUs) functions correctly. “Mythic IPUs leverage analog computing to perform the calculations required for deep neural network (DNN) inference inside flash memory arrays. This requires us to simulate thousands of analog-to-digital converters (ADCs) at extremely demanding accuracy specifications,” said Ty Garibay, vice president of Engineering at Mythic, in a press release.

Xilinx added new machine learning capabilities to its Xilinx Pro AV and broadcast platforms including region-of-interest encoding, intelligent digital signage, automatic object tracking and window cropping, and speech recognition. The company also demoed its first HDMI 2.1 implementation on its 7nm Versal devices. These devices are built on an adaptive compute acceleration platform (ACAP), which Xilinx says is a new category of heterogeneous compute devices with capabilities that exceed conventional CPUs, GPUs and FPGAs.

Internet of Things
Synopsys launched a new DesignWare ARC-based IP subsystem for ultra-low bandwidth wireless IoT (NB-IoT) applications. The subsystem, DesignWare ARC IoT Communications IP Subsystem, has Synopsys’ low-power ARC EM11D processor, optimized for DSP and RISC, with single-cycle 16+16 MAC operations. Included are hardware and software accelerators, a communications library, an on-board power management unit. Target applications are IoT machine-to-machine learning and industrial automation, along with smart city and smart agriculture.

Broadcom announced a Wi-Fi 6E chip for mobile devices.

The GSMA cancelled Mobile World Congress 2020, the large annual conference about mobile industry held in Barcelona, due to the coronavirus.

Synopsys is releasing an update to its Polaris Software Integrity Platform that can help find and fix security weaknesses in chip designs. The updates, to be released on February 18, will extend static application security testing (SAST) and software composition analysis (SCA) via a Code Sight IDE plugin that can find the security issues in both open-source and proprietary code.

STMicroelectronics announced an ultra-low-power STM32L5x2 MCUs that emphasize security to assure better protected IoT-connected applications.

The U.S. Department of Justice is charging Huawei with 16 counts of racketeering in violation of the Racketeer Influenced and Corrupt Organizations Act (RICO) statute. Huawei is accused of taking trade secrets from U.S. companies.

DarkReading reports that the Fast Identity Online (FIDO) Alliance kicked off its first working group meeting on IoT authentication last week, according to Nick Steele, research and development technical leader for authentication provider Duo Security.

PDF Solutions and UltraSoC are working together to predict and prevent in-field chip failures, of interest especially to the automotive industry. Data from UltraSoC’s in-chip hardware-based behavioral monitors will be combined PDF Solutions’ semiconductor manufacturing data collected in its Exensio Software Platform. All this data will inform machine learning (ML) algorithms to create predictive analytics and proactive maintenance. The companies call this a fab-to-field analytics framework because it will historical data from semiconductor manufacturing, test, assembly, supply chain, and in field, according to a press release.

Hyundai Motor Group and Canoo are co-developing an all-electric vehicle (EV) platform, based on Canoo’s EV skateboard architecture, that may be the base for electric purpose-built vehicles, like a bus, a roach coach, a cargo carrier or camper. Canoo, based in Los Angeles, has been working on subscription-only electric vehicles, which are in beta stage now and have a waitlist. Canoo will provide the engineering to Hyundai to create a scalable all electric platform for the company, which has plenty of electric mobility ambition. Hyundai shared its vision for the future at CES 2020, breaking it into three types of mobility: urban air mobility (UAM), purpose-built vehicle (PBV), and Hub, a landing pad/bus stop/transit center. Hyundai and Kia recently announced that they are developing fully electric PBVs.

Hyundai Motor Group and Canoo are co-developing an all-electric vehicle (EV) platform, such as the one above conceived by Canoo. (Image: Canoo. Used with permission.)

Ford and Bosch are teaming up to train Ford’s mechanics using virtual reality headsets. “Technicians soon will be trained how to service and maintain the all-electric Mustang Mach-E without need to access a physical model – thanks to a new virtual reality (VR) training tool from Ford and Bosch,” according to a press release.

Leave a Reply

(Note: This name will be displayed publicly)