PFLAB

Platform Lab (PFLAB) is dedicated to providing a comprehensive suite of common platforms designed to promote and accelerate research in autonomous driving, leveraging Autoware, the world’s first open-source software for autonomous driving.

PFLAB is a research laboratory formed within TIER IV, the deep-tech startup that pioneered Autoware, to facilitate longer-term research on autonomous systems. Our members include industrial research engineers from TIER IV as well as academic researchers from partner organizations. By combining industry and academia, we aim to expand our research potential through active collaboration, while also promoting the use of Autoware as a research platform.

Our mission

PFLAB is committed to expanding the community of Autoware users and contributors while advancing the technological capabilities of the open-source software by providing the necessary platforms for Autoware solutions to be built upon.

Recognizing that the success of open-source software for autonomous driving hinges on active global research collaboration, we aim to harness the Autoware project to foster research in the field. Our goal is to build the necessary platforms for information exchange, research execution and the continual expansion of autonomous driving technology.

The 3 platforms of PFLAB

Platform for collaborative research

Lowering the entry barrier to Autoware
PFLAB establishes a common platform encompassing hardware, software and data to foster collaborative research, the sharing of outcomes, common evaluations and datasets with Autoware as a vital educational resource.
Platform for advancing Autoware

Enhancing Autoware’s functionality
Unlocking the potential of Autoware through novel architectures, diverse information sources, varied operational design domains and new business models, PFLAB streamlines the contribution process to seamlessly integrate research output into the open-source autonomous driving project.
Platform for information sharing

Bridging the gap between R&D and autonomous driving technology
PFLAB facilitates connections between TIER IV’s R&D activities and academic and industrial research in autonomous driving technology through information-sharing activities such as research seminars and demonstrations.

Members

Simon Thompson

PFLAB Lead, TIER IV
Shinpei Kato

Founder, CEO and CTO, TIER IV
Yuuki Takano

Researcher, TIER IV
Manabu Tsukada

Associate Professor, The University of Tokyo
Eijiro Takeuchi

Researcher, TIER IV
Maxime Clement

Research Engineer, TIER IV
Daniel Shih

Professor, National Taiwan University
Ehsan Javanmardi

Project Assistant Professor, The university of Tokyo

Learn more

Research interests

PFLAB is dedicated to pushing the boundaries of autonomous driving technology across a wide spectrum of research areas. Our focus spans from fundamental algorithms such as perception and control to safety and verification of autonomous systems, to software-defined vehicles, high-performance automotive hardware platforms, connected infrastructure and fleet management, and the tools and education needed to drive research forward.

PFLAB supports cutting-edge research in autonomous driving through direct funding, collaborative research efforts and support for Autoware-based research projects. If you have innovative research ideas with the potential to advance autonomous driving and are passionate about our open-source approach, we encourage you to reach out and explore how PFLAB can support your endeavors!

Autonomous driving algorithms

Drive the evolution of fundamental algorithms in sensing, localization, perception, planning and control. Focus on cutting-edge technologies such as novel sensors, control at the limits of the vehicle, attention-based perception, and interaction-aware perception and planning.
Autonomous driving ecosystem

Investigate the comprehensive ecosystem supporting autonomous vehicles, including infrastructure, connected agents, traffic routing and information-sharing systems.
Research tools & education

Develop robust tools to facilitate research in Autoware. This includes creating evaluation metrics, simulators, benchmarking tools, open datasets, state-of-the-art algorithms for comparison. Additionally, design and deliver educational content and teaching methods for Autoware and autonomous driving tutorials to support ongoing learning and innovation.
Software-defined vehicles

Explore how software-defined vehicles can revolutionize autonomous driving development and deployment. Emphasize containerization, microservice architectures, heterogeneous computing and orchestration to enhance flexibility and scalability.
Accelerated AI & green compute

Focus on hardware acceleration and power efficiency for AI-based inference in autonomous driving. Optimize software and network architecture design to develop compact, efficient inference models on custom AI accelerators, promoting sustainability and performance.
Safety and verification

Advance hardware and OS development for creating verifiable systems that are real-time and memory safe. Implement rigorous software verification processes for autonomous driving algorithms, and establish comprehensive safety metrics and monitoring systems to ensure the reliability and safety of autonomous driving technology.