A simulation engineer’s journey from the race track to virtual roads

TIER IV People shines a light on the employees and teams whose unique experiences, backgrounds, and stories bring our mission to life. In this installment, simulation engineer Takatoki Makino shares his journey from racing circuits and simulation technology through to the development of AWSIM, an open-source autonomous driving simulator.

– What sparked your interest in car simulations?

My parents were car enthusiasts, and they took me to race tracks from an early age. I bought a [Mazda RX-7 FD] sports car when I was a university student and started driving it on circuits. The more I raced, the more repairs it needed. With funds limited, I tried my best to fix it myself, but sometimes I ended up making it worse and the repairs took a long time. While the car was out of action, I couldn't stand not driving so started creating simulators.


– How did you arrive at TIER IV?

I’ve always loved cars. My experience with vehicle simulations started in my student days. My first job involved working with the technology, and I also worked on projects in my spare time. I always knew that one day I would work on simulations for next-generation vehicles, which ultimately led to me getting a job at TIER IV, a company I’d known about for some time. 

I learned about vehicle dynamics and tire models while developing simulators and have experience with physics engines and vehicle simulations, which I thought would all come in handy at TIER IV. I also had experience with sensor simulation in game development, so I felt I could make use of my graphics and game engine-related skills, too.

AWSIM

– What have you been working on since joining the company?

I’m on the simulation team, working on the development of AWSIM, an open-source autonomous driving simulator. It’s designed for end-to-end simulations of Autoware, open-source software for autonomous driving pioneered by TIER IV. Vehicles, sensors, driving environments, and traffic flow can be replicated in AWSIM. It’s possible to conduct full-stack tests of each Autoware component, excluding hardware.

I was involved in setting up the repository and creating the first commit, and now I’m working on the overall development of the software, creating roadmaps, writing documentation, and reviewing licenses, as well as handling design and implementation. I also gather feature requests internally and externally, handle issues concerning open-source software releases, and do PR activities such as speaking at events.

At first, I worked on the foundations of AWSIM mostly on my own. More recently, I’ve been collaborating with overseas members of the Autoware Foundation (AWF). We’re currently developing AWSIM as open-source software with the involvement of Autoware users both inside and outside the company.

– What challenges have you experienced so far?

There’ve been so many! First of all, we had to understand the various technologies involved in autonomous driving. For example, vehicle control, sensors, communications, and maps. I reached out to various departments within the company to gather information on areas I didn't understand, and was able to incorporate it into the implementation. I’m extremely grateful for the many things they taught me.

Sensors can be simulated using Robotec GPU LiDAR.

Next was performance tuning unique to autonomous driving simulators. It's necessary to constantly sense the entire surroundings of the vehicle using multiple cameras and LiDAR sensors. Therefore, techniques like occlusion culling and frustum culling, which discard unnecessary rendering in certain areas, are not as effective. Additionally, considerations need to be made for the impact of Level of Detail (LOD) and tessellation on 3D models, as they can affect point cloud data acquired by LiDAR. To maximize performance, we relied on GPUs as much as possible. For example, NVIDIA RTX ray tracing helped speed up processing in LiDAR simulations.

A lot of research and development is ongoing in the field of autonomous driving. The same is true for simulators. For well-established technologies, we made it a point to implement them quickly so that we could spend more time on areas that involved research-level tech. We aimed to operate as efficiently as possible, sharing roles among team members. During the development of AWSIM, we were able to extract certain features as libraries, such as RGL and ros2-for-unity, and released them as open-source resources.

– Do you have any future goals for AWSIM?

The release of AWSIM was a significant starting point. There are two main goals we hope to achieve. The first is to enhance Autoware's feature development and testing. By virtually replicating vehicle dynamics, sensors, and driving environments, we aim to reduce the costs and risks involved in the development of autonomous driving systems. The second goal is to promote the widespread adoption of Autoware, which users can easily try in the simulator without the need for hardware. Both the autonomous driving operating system (Autoware) and the simulator (AWSIM) are open source, meaning anyone can participate in the development of the technology.

We aim to push the technology forward, focusing on feature enhancements and performance improvements, while ensuring AWSIM retains its status as the reference end-to-end simulator for Autoware. We also look forward to collaborating with our game engine development partner Unity Technologies.

– What kind of candidates would help TIER IV achieve those goals?

People with a high level of expertise in their respective fields. Although it’s sometimes necessary to learn about fields you’re unfamiliar with here, rest assured, TIER IV is a place where team members collaborate and support each other.

– What’s the best thing about TIER IV?

The company brings together highly specialized members from a variety of fields. There are people here who worked on incredible products, such as [Nissan’s] Skyline GT-R and the K computer. Also, many of the executives have backgrounds as engineers, computer science researchers and university professors. Their experience means discussions on technical matters progress seamlessly, which is particularly appealing from an engineer’s standpoint.

The fact that TIER IV founder Shinpei Kato has served on the faculty at the University of Tokyo makes it easier for us to approach academia. In fact, we’re currently conducting a joint research project involving AWSIM with Meijo University and Nagoya University.

There’ve been times when I’ve thought to myself, "I'd like to learn about the product of a particular company," and it turned out that someone at TIER IV had connections there, so we were able to make contact swiftly.

Also, more than 60 companies and universities are involved in the AWF. I believe such communities are symbolic of the power of open-source software.

– Any last words?

If you want to make next-generation vehicles, TIER IV is the place for you!

・・・

TIER IV is always on the lookout for passionate individuals to join our journey. If you share our vision of making autonomous driving accessible to all, get in touch.

We’re currently hiring for the following related positions:

Visit our careers page to view all job openings.

If you’re uncertain about which roles align best with your experience, or if the current job openings don’t quite match your preferences, register your interest here. We’ll get in touch if a role that matches your experience becomes available, and schedule an informal interview.

・・・


Media contact

pr@tier4.jp

Business inquiries

sales@tier4.jp

Social Media

X (Japan/Global) | LinkedIn | Facebook | Instagram | YouTube


More

Previous
Previous

System Software Team x CTO 座談会 【後編)

Next
Next

Co-MLOps initiative and Edge.Auto take center stage