Tech Focus: Autonomous vehicle software
Autonomous vehicles, capable of sensing their environment and operating without human involvement, could help to prevent road accidents and save billions in damages across the world each year. Carmakers, engineering software providers, AI chipmakers and start-ups are all developing solutions to accelerate development and make autonomous vehicles a reality.
Advanced driving assistance systems (ADAS) are helping to lay the groundwork. Still, automotive software increasingly covers electronics, sensors and optics and widespread connectivity to share information across other vehicles on a network.The Society of Automotive Engineers (SAE) has clearly defined guidelines for automotive “autonomy levels” from Level 1 “Driver Only”, to Level 5 “Fully Automated”. Current ADAS systems are estimated to be able to prevent more than a third of all passenger vehicle crashes.
These systems generally follow SAE guidelines Level 2 “Assisted Driving.” The benefits of automated driving are clear, but real-world vehicles must overcome technical hurdles to reach Level 5 autonomy. The automotive industry needs to adopt advanced software and integrate that with the design of new vehicles and components. Automotive companies must design and verify tools that can help integrate sensors,connectivity, mapping and the design and validation of optical systems and computing frameworks into vehicle designs.
One thing that’s undeniable about GIGABYTE servers is their prodigious processing power, especially for their NVIDIA-based GPU Server line-up (the G-Series).
From 1GPU set-up for low-cost cloud use cases, to 2GPU and 3GPU options for mainstream HPC and containerised AI development, to 8GPU and 10GPU massive HPC and AI workloads, the G-Series portfolio is particularly remarkable for its wide range of model choices, the GPU computational density, and the capability of massive data processing.
A particular case in point is the new G493-ZB3-AAP1 GPU supercomputing solution. With its new PCIe Gen5 bandwidth for GPU devices, the new G493-ZB3-AAP1 will further augment computational power by adopting the new NVIDIA L40 GPU and the new NVIDIA H100 GPU.
The two GPU models feature the 4th-gen Tensor cores (with hardware support for structural
sparsity and optimised TF32 format), a Transformer Engine with FP8 precision, to achieve one petaflop of throughput for single- precision matrix-multiply operations, with zero code changes. For many emerging HPC/AI use cases, the new G493-ZB3-AAP1 is an all-in-one solution.
For more information: https://bit.ly/3G2djLj
Autonomous vehicle design and verification software on the market
Autonomous vehicles (AVs) and ADAS bring increased complexity and a need for more testing. Exploring all the required scenarios within product development timing requires advanced simulation.
Altair enables customers to deliver solutions that make cars and trucks safer today and on the road to driverless mobility. Altair offers solutions in a wide array of engineering for antennas, from design to placement to communication.
Altair Feko efficiently and accurately simulates radar antenna design as well as the integration aspects, including radome and bumper effects. Feko also provides a solution for ultrasonic sensors, and the Altair Partner Alliance (APA) offers access to TracePro for LiDAR modelling.
Altair Feko efficiently and accurately simulates radar antenna design as well as the integration aspects, including radome and bumper effects. Feko also provides a solution for ultrasonic sensors, and the Altair Partner Alliance (APA) offers access to TracePro for lidar modelling.
Altair PollEx provides integrated printed circuit board (PCB) design viewing, analysis and verification tools for electrical, electronics, and manufacturing engineers.PollEx transfers data between popular ECAD and simulation tools and lets engineers quickly visualise and review PCB designs. Its checking tools detect issues early in the design to avoid product failures and simplify manufacture and assembly.
Ansys Speos predicts systems’ illumination and optical performance to save on prototyping time and costs while improving a product’s efficiency.
Ansys HFSS is a 3D electromagnetic (EM) simulation software for designing and simulating high-frequency electronic products such as antennas, antenna arrays, RF or microwave components, high-speed interconnects, filters, connectors, IC packages and PCBs. Engineers worldwide use Ansys HFSS software to design high-frequency, high-speed electronics found in communications systems, ADAS, satellites, and internet- of-things (IoT) products.
Ansys AVxcelerate realistic sensor testing and validation enables users to test their autonomous vehicles, ADAS and sensors faster than with physical prototypes.
The Autoware Foundation is a non-profit organisation supporting open-source projects enabling self-driving mobility. The Autoware Foundation creates synergies between corporate development and academic research, enabling autonomous driving technology for everyone.
The Autoware defined Open AD Kit brings the best practices of software-defined vehicle (SDV) development to the Autoware ecosystem.
The Open AD Kit promises to accelerate the development of optimised hardware and software solutions for autonomous driving. The kit includes a complete AD reference platform for rapidly developing cloud-to-edge solutions and a cloud-native development and verification of AD solutions with system-level parity to edge platforms.
The kit includes a micro-services architecture orchestrating mixed critical workloads across heterogeneous compute platforms and an open ecosystem system of hardware/ software components and development tools.
The open Nvidia DRIVE SDK gives developers all the building blocks and algorithmic stacks needed for autonomous driving. It allows developers to build and deploy various AV applications more efficiently, including perception, localisation and mapping, planning and control, driver monitoring, and natural language processing.
The foundation of the DRIVE Software stack, DRIVE OS is a safe operating system for in- vehicle accelerated computing. It includes NvMedia for sensor input processing, Nvidia CUDA libraries for efficient parallel computing implementations, Nvidia TensorRT for real-time AI inference, and other developer tools and modules to access hardware engines.
Nvidia DriveWorks provides middleware functions on top of DRIVE OS that are fundamental to autonomous vehicle development. These consist of the sensor abstraction layer (SAL) and sensor plug-ins, data recorder, vehicle I/O support, and a deep neural network (DNN) framework. It’s modular, open, and designed to comply with automotive industry software standards.
The DRIVE AV software stack contains the perception, mapping, and planning layers and diverse DNNs trained on high-quality, real-world driving data. These rich perception outputs can be used for both autonomous driving and mapping. In the planning and control layer, the
Nvidia Safety Force Field computational module keeps a vehicle out of harm’s way to ensure it won’t contribute to or cause an unsafe situation.
Nvidia DRIVE Chauffeur is an AI-assisted driving platform based on the Nvidia DRIVE AV SDK that can handle both highway and urban traffic with increased safety. It can use the high-performance compute reference architecture and sensor set of Nvidia DRIVE Hyperion 8 to drive from address to address. For those who want to drive, the system also provides active safety features and the ability to intervene in dangerous scenarios.
Siemens Accelerated Product Development is a suite of tools and solutions designed to help automotive manufacturers to speed up the product development process. It includes various software tools, simulation and testing capabilities, and data management solutions. Siemens Accelerated Product Development includes software tools such as NX for product design and engineering, Simcenter for simulation and testing, and Teamcenter for data management and collaboration.
Siemens Accelerated Product Development can help automotive manufacturers reduce time-to-market, improve quality, increase efficiency, and reduce costs. It does this by enabling the digitalisation of the product development process, enhancing collaboration between teams, and providing advanced simulation and testing capabilities.
Siemens Teamcenterm enables engineering teams to use digital twins to connect and optimise designs, systems, software, simulation and visualisation processes.
Synopsys’ comprehensive Safety-Aware Solution helps engineers automate the design and verification of functional safety systems-on-chips (SoCs).
The Safety-Aware Solution helps to make automotive SoC designs highly predictable so that you can reach your chip’s target automotive safety ILevel (ASIL).
With the rapid shift to ADAS and autonomous driving, carmakers are transitioning to a more centralised domain architecture with central compute. Synopsys believes autonomous driving compute solutions will become more centralised within SoCs to rapidly fuse multimodal sensor data and safely choose and initiate self-driving actions. The SoC helps to ensure safe and secure operation.
To design SoCs that meet these requirements, system architects integrate automotive-grade IP and ISO 26262-certified safety-aware test, design implementation, and hardware verification solutions while consulting with automotive SoC experts. Selection of design flows and IP can simplify meeting functional safety objectives and achieving power-performance-area (PPA) targets.