Adlink reveals two Linux-driven Coffee Lake systems for autonomous cars and C-V2...

Adlink reveals two Linux-driven Coffee Lake systems for autonomous cars and C-V2X

Nov 2, 2021 — by Eric Brown

34 views

Adlink unveiled a “AVA-3501” system used in the Indy Autonomous Challenge and announced a collaboration on 5G based C-V2X technology with Askey, III, and EOI using a similarly Ubuntu-on-Coffee Lake AVA-3510. Adlink also announced an edge visualization partnership with AUO.

Adlink has ramped up its involvement in autonomous vehicles in recent weeks with announcements of collaborations with Indy Autonomous Challenge, as well as several other companies working on C-V2X (Cellular Vehicle-to-Everything) technology. Today, it followed up by announcing a collaboration with optoelectronic firm AUO on edge visualization technologies for a list of applications led by automated vehicles and autonomous robots. An Adlink/AUO virtual forum on the topic started today and will run through Nov. 4 (see farther below).

AVA-3501 (left) and Dallara AV-21 racecar used in Indy Autonomous Challenge
(click images to enlarge)

Adlink also unveiled two new Ubuntu Linux-driven edge AI computers aimed primarily at autonomous vehicles. The Intel 9th Gen Coffee Lake powered AVA-3501 has been adopted by all the participants in the Indy Autonomous Challenge. The C-V2X collaboration is using Adlink’s Coffee Lake driven AVA-3510 computer. Both systems are listed on Adlink’s Autonomous Vehicles Solutions page along with a similarly “preliminary,” Jetson AGX Xavier based AVA-XV-V1 system, which we plan to cover soon.

Indy Autonomous Challenge

The AVA-3501 used in the Indy Autonomous Challenge (IAC) follows a similarly 9th Gen Coffee Lake based Adlink DLAP-8000 system that was used in last year’s IAC event. Nine teams from nine different countries representing 21 universities participated in the Oct. 23 IAC competition at the Indianapolis Motor Speedway (IMS). Sponsors included Adlink, Cisco, Ansys, Bridgestone, Intel, and others. TUM Autonomous Motorsport from Germany’s Technische Universität München won the $1 million Grand Prize.

Each participant raced a Dallara AV-21 racecar equipped with the AVA-3501. In partnership with the Autoware Foundation, Open Source Robotics Foundation, and the Eclipse Foundation, Adlink also supplied IAC teams with open source autonomous driving tools and expertise, including support with ROS 2, Eclipse Cyclone DDS and Zenoh, and Autoware.Auto.

As part of the event, Adlink sponsored a STEM Garage showcase of autonomous vehicle and advanced driver-assistance systems (ADAS) technologies in collaboration with the Autoware Foundation, Boston Dynamics, GAIA Platform, and Kvaser. Demonstrations included the AVA-3501 and Adlink’s ROScube robot controller, which similarly runs Ubuntu on Coffee Lake.

The Autoware Foundation and UC San Diego showed off an F1Tenth (F 1/10th) racetrack for small-scale autonomous vehicles, and GAIA demonstrated a racing simulator robot. Boston Dynamics demonstrated its dog-like Spot agile mobile robot and CAN specialists Kvaser, which provided USB-to-CAN interfaces for the race, showed its Kvaser DevKit racecar simulator.

AVA-3501

The AVA-3501 computer used for the IAC event appears to be a minor upgrade of the DLAP-8000. The system is built around a 9th Gen Xeon E-2278GE running Ubuntu 20.04. The octa-core 3.3GHz/4.7GHz CPU is accompanied by an Intel C246 chipset and 64GB DDR4. Dual PCIe x16 slots default to using an Nvidia RTX8000 graphics card. The system can also load up to 3TB storage via Samsung 970 EVO Plus NVME M.2 SSDs, as well as dual 512GB hot-swappable 2.5-inch SSD SATA drives.

AVA-3501 internal view
(click image to enlarge)

Other highlights include dual 40GbE QSFP+ connectivity and 6x CAN channels. The AVA-3501 also provides 3x GbE,6x USB, 2x DP++, DVI, DIO, and more. The 0 to 50°C tolerant system offers shock and vibration among other ruggedization features.

Adlink C-V2X collaboration

Last week, Adlink announced it was collaborating with Askey Computer Corp., the Institute for Information Industry (III), and Excellence Opto. Inc. (EOI), to develop an end-to-end C-V2X solution. The solution will incorporate Adlink autonomous vehicle computers and applications with 4G LTE/sub-6GHz 5G NR vehicle network connectivity, smart traffic lights, and signal controllers. The partners will participate in four C-V2X pilot projects underway across Taiwan, three of which involve autonomous vehicles.

The collaboration, which includes participation by the Autoware Foundation and Open Source Robotics Foundation, combines Adlink’s new Xeon E based AVA-3510 system with Askey’s 5G RSU onboard/roadside unit (RSU/OBU). China’s III is contributing AI-enabled applications for image-based AI vehicle identification, tracking, and warning technology, and EOI is providing its smart traffic light and traffic signal controller.

Askey’s 5G RSU is built around the i.MX8 QuadXPlus, the quad-core version of the Cortex-A35/Cortex-M4 equipped i.MX8X, so it likely runs Linux. The IP67-protected, 245 x 201 x 74mm roadside unit device has a 5G NR Sub-6 NSA modem for C-V2X. Other features include 2GB LPDDR4, GbE with PoE, GNSS, and several 5G and C-V2X antennas.

V2X is an umbrella term that includes fast vehicle-to-vehicle (V2V), vehicle-to-person (V2P), and vehicle-to-infrastructure (V2I) connectivity. The technology is touted for improving traffic safety via continual communications between ADAS and autonomous cars with other vehicles and with infrastructure. The technology may also fill in the gaps in self-driving car technology that have slowed its adoption.

C-V2X is an emerging Vehicle-to-Everything standard backed by cellular chipmakers such as Qualcomm and the Chinese government. It challenges the dominant, WiFi-based DSRC standard based on 802.11p, which has been adopted by the FCC in the US.

NXP recently launched a headless i.MX8X Lite variant of the i.MX8X with a built in V2X accelerator designed to work with DSRC modems. The SoC showed up last week in SolidRun’s Linux-powered i.MX 8XLite SOM and dev kit.

DSRC has the advantage of being more decentralized, more mature, and less expensive than C-V2X, which is likely to offer a greater range, lower latency and higher data transfer. (For more details on V2X, see our i.MX8X Lite report.)

AVA-3510

The Adlink AVA-3510 used in the C-V2X project is dedicated to autonomous vehicle applications. The system runs Ubuntu on a Coffee Lake generation, octa-core Xeon E-2278GE and incorporates an Nvidia Turing Quadro RTX 5000 module via a PCIe-based MXM interface.

AVA-3510
(click image to enlarge)

The 335 x 225.1 x 95mm AVA-3510 is equipped with up to 64GB DDR4-2133 plus dual 2.5-inch SATA bays. An M.2 B+M 2242/2280 slot enables further storage expansion, including NVMe. There are also 2x full-sized mini-PCIe slots for wireless.

The AVA-3510 is equipped with 2x 10GbE, 5x GbE, 2x USB 3.0, 2x USB 2.0, and 2x RS-232/422/485 ports. There is a DP++ port in addition to the 2x DP available via the RTX 5000. Other features include 2x isolated CAN and 4-in/4-out isolated DIO. The -10 to 55°C tolerant system has a 9-36V input with ignition controls, and offers MIL-STD-810H compliant shock and vibration resistance.

Adlink and AUO collaborate on edge visualization

Today, Adlink announced a collaboration with Taiwanese optoelectronics company AUO, which makes micro-LED, tiled LED, and AmLED (Adaptive mini LED) display devices, as well as display panels and systems that incorporate them. The companies will work on developing AI-enabled “intelligent interactive human-machine interfaces” for “smart” applications in transportation, healthcare, enterprises, and manufacturing. Specific applications include HMIs for autonomous vehicles, traffic management, factory HMI, medical displays, retail PoS systems, and self-service kiosks.

Adlink and AUO plan to build a “private 5G factory of the future that leverages swarm intelligence among automated guided vehicles (AGVs) and autonomous mobile robots (AMRs).” The factory is expected to increase efficiency by over 33 percent within AUO’s Longtan manufacturing site.

The companies are hosting a virtual forum called Visualization at the Edge, running from today through Nov. 4 with keynotes from Adlink Chairman and CEO Jim Liu, AUO President and COO Frank Ko, and Taiwanese representatives of Arm and Intel. More information may be found here.

Further information

No pricing or availability information was provided for Adlink’s “preliminary” AVA-3501 and AVA-3510. More information may be found in its AVA-3501 and AVA-3510 product pages, as well as its Indy Autonomous Challenge, C-V2X, and AUO edge visualization announcements.