Week In Review: Auto, Security, Pervasive Computing

Flex Logix on AI edge; DARPA SSIITH success; Emotet downed

popularity

Automotive/Mobility
Chip makers in Taiwan will “do their best” to “squeeze out more chips” said Taiwan’s Minister of Economic Affairs Wang Mei-hua after having lunch with representatives of TSMC, UMC, Vanguard International Semiconductor Corp, and Powerchip Semiconductor Manufacturing Co., according to the Taipei Times. After the auto industry initially cut automotive chip orders because COVID-19 lockdowns and economic slowdowns caused consumers to buy fewer vehicles, the auto industry saw demand reignite near the end of 2020. By then, chip production had shifted away from automotive chips.

Ansys added an optical simulation of lidar to its new Ansys 2021 R1 simulation tool offerings. Ansys 2021 R1 has tools for simulating real-time, physics-based radar sensor capability for automated vehicles (AVs). A closed-loop simulation validation enhances AV safety. Also, new scanning and rotating lidar models boost AV simulation reliability. Ansys says it improved modeling and code generation flexibility for AUTOSAR automotive software components and supports the FACE 3.0 Technical Standard for military avionics. Systems safety tools can add to EV and AV safety analyses by graphically pinpointing potential system failures, simplifying the software safety review process

General Motors says it will stop making gas- and diesel-powered light-duty vehicles (passenger cars and SUVs) by 2035. General Motors plans to have 30 electric vehicle models by 2025 and plans to be carbon neutral by 2040 — in its manufacturing and products’ tailpipe emissions — and will work on battery technology with its battery partner, according to GM CEO Mary Barra’s LinkedIn blog. If the U.S. federal government and Congress achieve their plans to invest in recharging infrastructure, and the state of California rearranges its tax breaks to make electric car ownership more affordable for lower income customers, the EV market has a brighter future in the U.S.

Learning from how humans have interacted with their products, rideshare scooters from Ford’s Spin will be able to reposition themselves autonomously or by remote control using onboard cameras and built-in machine learning software. Two benefits are avoiding sidewalk clutter or moving back to a charging station or designated area. The scooters have a third wheel, which makes them more stable and safer for the average rider and harder to push over. Also, the payment plan will be by-the-minute to discourage riders from leaving them in certain areas or throwing them into lakes.

Security
DARPA says processor hardware from its System Security Integration Through Hardware and Firmware (SSITH) program has weathered a planned bug bounty program. DARPA reviewed 13,000 hours of hacking conducted by over 580 cybersecurity researchers and affirmed that secure hardware architectures designed under SSITH did well. The Finding Exploits to Thwart Tampering (FETT) Bug Bounty also found some critical areas where defenses needed hardening. “Knowing that virtually no system is unhackable, we expected to discover bugs within the processors but FETT really showed us that the SSITH technologies are quite effective at protecting against classes of common software-based hardware exploits,” said Keith Rebello, the DARPA program manager leading SSITH and FETT said in a press release. “The majority of the bug reports did not come from exploitation of the vulnerable software applications that we provided to the researchers, but rather from our challenge to the researchers to develop any application with a vulnerability that could be exploited in contradiction with the SSITH processors’ security claims. We’re clearly developing hardware defenses that are raising the bar for attackers.”

The United States Department of Defense will start having third party entities review and certify the cybersecurity compliance of DoD contractors, according to a story in NextGov.

Emotet, one of the largest botnets responsible for data theft and ransomware, was taken down by a coalition of police agencies working in Europe and North America with Europol, according to a story in Wired. The command infrastructure — infected servers and equipment — were removed or taken down. Emotet, a for-hire network that started in 2014 and the go-to-solution for criminals according to a Europol press release, was the door opener into other computers. It was polymorphic and resilient. Spread by a Microsoft Word file on a few devices, the code was polymorphic, changing each time it was opened. The Emotet grew from an initial access on a few devices and spread resiliently.

Pervasive computing — IoT, edge, cloud, data center, and back
Flex Logix is pairing its InferX X1 AI inference accelerator with a high-bandwidth Winbond 4Gb LPDDR4X chip for AI on the edge. “Now for the first time, affordable edge computing systems can implement complex neural networking algorithms to achieve high accuracy in object detection and image recognition even when processing data-intensive high-definition video streams,” said Dana McCarty, VP of Sales & Marketing for Flex Logix’s AI inference products, in a press release.

A team of researchers around the world, led from Stanford University, used a set of low-power chips to create the illusion of a larger chip for putting AI on small, smart devices that run on low power. The chips avoids the wasteful memory wall that hands data to separate memory devices. Instead, the hybrid chips use RRAM on the chip and algorithms that “meld the eight, separate hybrid chips into one energy-efficient AI-processing engine,” according to a press release.  “If we could have built one massive, conventional chip with all the processing and memory needed, we’d have done so, but the amount of data it takes to solve AI problems makes that a dream,” said computer scientist Subhasish Mitra, senior author of a new study. “Instead, we trick the hybrids into thinking they’re one chip, which is why we call this the Illusion System.”

Aldec added VHDL-2019 (IEEE 1076-2019) to its Active-HDL IDE for FPGA development. Also added in the latest update are support for release 2020.08 of the open source VHDL verification methodology (OSVVM) and SystemVerilog enhancements that include initial support for multidimensional arrays of instances, preliminary support of unresolved user-defined nettypes, and preliminary support for unique constraints.

Job, Event and Webinar Boards: Find industry jobs and upcoming conferences and webinars all in one place on Semiconductor Engineering. Knowledge Center: Boost your semiconductor industry knowledge. Videos: See the latest Semiconductor Engineering videos.



Leave a Reply


(Note: This name will be displayed publicly)