This Guidance demonstrates how customers can process and search high-accuracy, scenario-based data with the Autonomous Driving Data Framework (ADDF). Automotive teams who want to implement common tasks for autonomous vehicles (AV) and advanced driver-assistance systems (ADAS) can share, modify, or create fully customizable modules that reduce the amount of effort required to create and deploy this Guidance.
Initial data quality checks and data extraction with containers running on AWS Batch. Processed data is stored in Amazon S3.
Images are annotated with machine learning models to detect objects and road lanes. Low confidence predictions are set aside for manual annotation. Bounding boxes are used for blurring faces and license plates. Amazon SageMaker Ground Truth is used for labeling.
Use Amazon EMR in combination with Amazon S3 and AWS Batch to enrich sensor data with localized weather and map matching info. It also combines image annotations, and sensor data to detect various scenes like traffic intersections or people and objects in the street.
Host high-performance, on-demand visualization applications on Amazon Elastic Kubernetes Service (Amazon EKS) for engineers.
Scalable simulation and KPI calculation modules use Amazon EKS or AWS Batch. Amazon QuickSight is used to analyze KPIs and simulation results.
The AWS Well-Architected Framework helps you understand the pros and cons of the decisions you make when building systems in the cloud. The six pillars of the Framework allow you to learn architectural best practices for designing and operating reliable, secure, efficient, cost-effective, and sustainable systems. Using the AWS Well-Architected Tool, available at no charge in the AWS Management Console, you can review your workloads against these best practices by answering a set of questions for each pillar.
The architecture diagram above is an example of a Solution created with Well-Architected best practices in mind. To be fully Well-Architected, you should follow as many Well-Architected best practices as possible.
This Guidance offers a secure-by-default setup to allow users to safely operate and respond to incidents and events. If you decide to move into a production-like environment, the ADDF security and operations guide outlines best practices for securely deploying and operating ADDF in the AWS Cloud.
ADDF was built with security in mind. Before release to the public, AWS performed an initial, internal security review of ADDF and resolved any identified security issues. Both AWS and the open-source community contribute to ongoing security reviews of the framework.
Interfaces to the public internet are not exposed by the core modules. Services are only reachable as an authenticated user in the context of an AWS account.
Various built-in security features in ADDF are designed to help you set up a secure framework and help your organization meet common enterprise security requirements. AWS defined an ADDF shared responsibility model, as well as a secure setup and operation guide, to help you on your ADDF journey from a secure start through to production.
To implement a reliable architecture, each individual module is designed to cover module-specific throttling and limit-issues based on current experience. The default deployment options offer the end-user a sensible working baseline with common account limits. If the end-user decides to scale out, that user is responsible for considering any newly hit constraints or limits.
ADDF is an open-source project. The ADDF community constantly improves features based on customer or community input.
ADDF provides best-practice patterns that have been proven in challenging enterprise environments with customers. All selected services reflect the learnings from real-life customer use-cases. Amazon EKS hosts high-performance, on-demand visualization applications for engineers. For developer instances, Amazon EC2 and Nice DCV stage and share files using FSx for Lustre. Both patterns have proven to work at scale in enterprise environments.
The default deployment options offer the end-user a sensible working baseline. The user is free to change the default configuration of modules to scale up or down based on the use case.
This Guidance uses resources based on workload data and resource characteristics to keep up with demand.
In this Guidance, the ADDF modules describe patterns for running ADAS and AV workloads at an enterprise scale, containing common best-practices for scaling traffic and data access patterns.
Any compute intensive workload should have a default value that balances between a high baseline utilization and end-user usability. The ADDF modules provide a reference implementation, and all deployed resources are set to the minimum resources needed to support the ADDF models. This ensures a high baseline utilization.
The sample code; software libraries; command line tools; proofs of concept; templates; or other related technology (including any of the foregoing that are provided by our personnel) is provided to you as AWS Content under the AWS Customer Agreement, or the relevant written agreement between you and AWS (whichever applies). You should not use this AWS Content in your production accounts, or on production or other critical data. You are responsible for testing, securing, and optimizing the AWS Content, such as sample code, as appropriate for production grade use based on your specific quality control practices and standards. Deploying AWS Content may incur AWS charges for creating or using AWS chargeable resources, such as running Amazon EC2 instances or using Amazon S3 storage.