Astemo And TIER IV To Develop End-To-End Autonomous Driving AI

TIER IV and Astemo logos marking their joint development platform for end-to-end autonomous driving AI

Summary: TIER IV has signed a memorandum of understanding with Astemo to build a joint software development platform for end-to-end autonomous driving AI. Astemo will licence TIER IV’s Co-MLOps architecture, with commercialisation of the platform targeted around 2030.

Key engineering takeaway: The platform integrates the collection, processing, training, evaluation and continuous improvement of end-to-end autonomous driving AI models, the kind of AI development infrastructure TIER IV says is needed to bring such models to production readiness.

Why it matters: Astemo brings functional safety compliance and vehicle integration across braking, suspension, powertrain and ADAS. Pairing that with TIER IV’s data platform is an attempt to make end-to-end autonomous driving AI meet mass-production quality requirements, with Level 4+ deployment in passenger cars targeted for the early 2030s.

TIER IV, the pioneering force behind open-source software for autonomous driving, has signed a memorandum of understanding (MoU) with Astemo, to jointly develop a next-generation software development platform for autonomous driving systems.

Through this collaboration, the two companies will leverage TIER IV’s Co-MLOps solution – collaborative data-sharing technology – to build Astemo’s next-generation development platform, enabling the continuous development and improvement of production-quality end-to-end (E2E) autonomous driving AI models, with commercialization targeted around 2030. Astemo aims to deploy E2E autonomous driving AI models in passenger vehicles in the early 2030s using this development platform.

E2E autonomous driving AI models – which integrate perception, planning and control into a single learning process – are emerging as a foundational technology for next-generation advanced driver-assistance systems (ADAS) and autonomous driving systems.

Production Readiness And AI Development Infrastructure

As autonomous driving technology advances, TIER IV is working toward Level 4+ autonomy to gradually expand fully autonomous driving into more complex environments. E2E AI models are a key technology for achieving this goal. Bringing these models to production readiness requires compliance with automotive safety and manufacturing standards. It also demands an AI development infrastructure capable of continuously collecting, processing, training and evaluating large-scale driving data. Such infrastructure forms the foundation for the continuous evolution of E2E AI models and will be a key source of competitiveness for future production autonomous driving systems.

Through their involvement in the Autoware Foundation (AWF) and SOAFEE, TIER IV and Astemo previously collaborated on the reference implementation of Open AD Kit, a cloud-native framework for autonomous driving development based on Autoware. Astemo is also involved in TIER IV’s Co-MLOps project, launched in January 2024. This latest initiative deepens their partnership and expands TIER IV’s data-sharing platform for autonomous driving AI development.

Licensing The Co-MLOps Architecture

TIER IV leads the development of Autoware and advances AI-based Level 4 autonomous driving worldwide while building AI development platforms through its Co-MLOps solution. This MoU enables the companies to combine TIER IV’s expertise in AI development platforms with Astemo’s industry expertise as a Tier 1 supplier. This expertise includes functional safety compliance and vehicle integration across braking, suspension, powertrain and ADAS technologies. Together, the companies will build a development platform that enables the continuous evolution of production-ready E2E autonomous driving AI models, accelerating the development of next-generation autonomous driving systems.

This collaboration will advance the following initiatives:

  1. Creation of a next-generation development platform by harnessing Co-MLOps solution
    Astemo will license TIER IV’s Co-MLOps architecture – the core technology underpinning Co-MLOps solution – along with related technologies. Leveraging these assets, TIER IV and Astemo will build a next-generation development platform that integrates the collection, processing, training, evaluation and continuous improvement of AI models, with commercialization targeted around 2030. In parallel, Astemo will establish an internal framework for the continuous development and improvement of E2E autonomous driving AI models that meet the quality and safety requirements for mass-production vehicles.
  2. Joint study on E2E autonomous driving AI models for Level 4+ autonomy
    By sharing the insights and technological assets gained from the development of the planned platform, TIER IV and Astemo will define and align the technical requirements for E2E autonomous driving AI models, while developing a medium- to long-term roadmap.

Using this platform, Astemo aims to continuously advance its E2E autonomous driving AI models and deploy Level 4+ E2E AI models in passenger vehicles in the early 2030s.

Bringing E2E autonomous driving AI models into real-world deployment requires advancing both mass-production readiness and AI development infrastructure in parallel. We are delighted to deepen our collaboration with Astemo by providing licenses for TIER IV’s Co-MLOps solution and related technologies. Astemo’s track record as a global Tier 1 supplier makes it an ideal partner for this initiative. Together, we will further expand the Autoware ecosystem and accelerate the deployment of safe, scalable autonomous driving technologies.

Shinpei Kato, founder and CEO of TIER IV

We believe that next-generation autonomous driving systems will be driven by E2E AI models and data-centric AI development platforms that enable their continuous evolution. Through our collaboration with TIER IV, whose expertise spans autonomous driving and data platform technologies, we aim to further strengthen the internet-of-vehicles platform that Astemo is developing. By leveraging the combined ecosystems of our two companies, we will expand these technologies beyond the automotive industry into broader mobility applications, helping shape the future of continuously evolving mobility.

Hiroshi Tamagawa, Executive Vice President, CTO and CISO of Astemo

End-To-End Autonomous Driving AI: Frequently Asked Questions

What is end-to-end autonomous driving AI?

End-to-end autonomous driving AI integrates perception, planning and control into a single learning process, rather than running them as separate, individually engineered modules that pass results down a pipeline.

How does the Co-MLOps solution work?

Co-MLOps is TIER IV’s collaborative data-sharing technology. It provides the infrastructure to continuously collect, process, train and evaluate large-scale driving data, so AI models can be improved on a repeating loop rather than retrained as one-off exercises.

What is the difference between E2E AI and modular autonomous driving software?

Modular software splits the task into discrete perception, prediction, planning and control components with defined interfaces between them. An E2E model integrates perception, planning and control into a single learning process, learning the mapping from sensor input to driving action rather than passing results between separately engineered modules.

When will Astemo deploy end-to-end autonomous driving AI in production cars?

Astemo is targeting deployment of Level 4+ E2E AI models in passenger vehicles in the early 2030s. The development platform itself is targeted for commercialisation around 2030.

What is Level 4+ autonomy?

Level 4 describes a system that can perform the full driving task without human intervention within a defined operational design domain. TIER IV uses Level 4+ to describe the progressive expansion of that domain into more complex environments.

Further Reading

For more autonomous driving news, click here.

Source

The latest technology and engineering news direct to your inbox.

Discover more from Auto Tech News

Subscribe now to keep reading and get access to the full archive.

Continue reading