ADAS Simulation on Cloud Continental AG

Format:

Book

Conference/Event

Author/Creator:

Ramapuram, Vinay Goud, author.

Contributor:

Dhar, Jayshri

Munaiahgari, Mallikarjuna Reddy

Conference Name:

11th SAEINDIA International Mobility Conference (SIIMC 2024) (2024-12-11 : New Delhi, India)

Language:

English

Physical Description:

1 online resource cm

Place of Publication:

Warrendale, PA SAE International 2024

Summary:

The rapid evolution of new technologies in the automotive sector is driving the demand for advanced simulation solutions, enabling faster software development cycles. Developers often encounter challenges in managing the vast amounts of data generated during testing. For example, a single Advanced Driver Assistance System (ADAS) test vehicle can produce several terabytes of data daily. Efficiently handling and distributing this data across multiple locations can introduce delays in the development process. Moreover, the large volume of test cases required for simulation and validation further exacerbates these delays. On-premises simulation setups, especially those dependent on High-Performance Computing (HPC) systems, pose several challenges, including limited computational resources, scalability issues, high capital and maintenance costs, resource management inefficiencies, and compatibility problems between GPU drivers and servers, all of which can impact both performance and costs.To address these issues, transitioning to cloud-based simulations using AWS servicessuch as S3, Batch, Auto Scaling Group (ASG), and Elastic Container Registry (ECR)allows for parallel execution of Software-In-the-Loop (SIL) tests. Multiple compute instances can be deployed simultaneously, significantly reducing testing times. Additionally, GPU-based simulations can be executed on cost-effective instances, potentially reducing simulation costs by up to threefold. The Docker container used in these simulations packages the necessary software algorithms, simulation tools, and test recordings. Key driving functions, including Blind Spot Detection (BSD), Lane Change Assist (LCA), and Occupant Safety Exit (OSE), are tested using real-world driving scenario data to ensure their effectiveness. The workbench offers an innovative environment for developing software solutions using virtual ECUs, even before the physical hardware becomes available. Furthermore, the automotive operating system utilized for software development provides robust middleware, ensuring secure and standardized communication between computers and the cloud

Notes:

Vendor supplied data

Publisher Number:

2024-28-0166

Access Restriction:

Restricted for use by site license