How Japanese startup Turing inc. used MICHELIN SIMIX chassis simulation datasets to accelerate its autonomous driving validation: faster, safer, and without costly real-world testing.
WHAT CHANGED FOR Turing inc
WITH MICHELIN SIMIX
Accelerated simulation:
Instant access to validated tire models saved months of setup.
Safer AI training:
Teams could simulate extreme conditions without risking real drivers.
Proven value:
Reliable manufacturer data without the heavy testing budget.
THE CHALLENGE: MISSING TIRE DATA
AND HIGH TESTING COSTS
Developing a realistic vehicle dynamics model requires precise tire data. But for Turing, a small startup working on a non-standard vehicle, accessing validated tire models was nearly impossible.
“We’re a startup, so at that time we didn’t have much data to work with,” explains Bon Aizawa, E2E Advanced Development Team of Turing Inc.
“The Toyota Alphard is not a vehicle famous for autonomous driving, and of course, the tire data is not available since we are not the car manufacturer.”
Generating this data manually would have meant expensive test campaigns, time-consuming measurements, and safety risks.
“We thought about obtaining the data by ourselves, but testing services were very expensive, and we don’t have the experience to conduct proper tests.”
Without simulation-ready data, engineers couldn’t model tire grip or predict behaviour in real-world conditions.
THE SOLUTION: MICHELIN SIMIX CHASSIS SIMULATION DATASETS
During their research, the Turing team found MICHELIN SIMIX, a public platform giving access to high-quality tire datasets for simulation.
“As far as I know, only Michelin provides an online tire library,” says Bon. “We don’t have any close relationship with other tire makers, so it was naturally the only option for us.”
“THE SERVICE OBVIOUSLY SAVES US A LOT OF TIME AND MONEY.”
MICHELIN SIMIX provided instant access to validated tire models compatible with Turing’s in-house simulation tools: no lab work or data crafting required.
“The price is reasonable: not cheap for a small startup, but still cheaper than doing the whole tests by ourselves. The service obviously saves us a lot of time and money.”
INTEGRATING SIMIX INTO AN AI SIMULATION WORKFLOW
Turing’s AI platform is an end-to-end vehicle simulation system that learns directly from real-world data. By integrating MICHELIN SIMIX simulation-ready datasets, the team could model tire grip and vehicle-road interaction under extreme conditions.
“Our model is end-to-end. We don’t rely on rule-based fine-tuning; we just understand how much this tire is loaded. With MICHELIN SIMIX data, we can check if a speed or curvature is acceptable, then our AI learns to drive safely in those conditions.”
This integration allowed Turing to replace dangerous physical tests with safe virtual simulations, reducing costs while improving learning precision.
RESULTS: RELIABLE AUTONOMOUS DRIVING IN WET AND CURVY ROADS
With MICHELIN SIMIX tire datasets, Turing achieved realistic autonomous driving simulation in wet and curvy conditions, a scenario that’s notoriously difficult to reproduce safely.
“We don’t have to push the real vehicle to the limit in the real world. With the dataset, we can already know when the tire would start to slip at certain speeds, so we can plan our wet tests with a much safer boundary. That’s a huge benefit for us.”
The dataset helped simulate tire dynamics and grip limits with accuracy, enabling safer and faster AI validation. Their test video, showing the Alphard performing on wet curves, demonstrates what’s possible when simulation data has a real-world correlated and ready to use.
IMPACT – FASTER TIMELINES, LOWER COSTS, SAFER R&D
With MICHELIN SIMIX, Turing inc :
- Shortened testing cycles by eliminating track dependency
- Increased AI reliability through using data from the manufacturer
- Improved safety by avoiding high-risk real-world limit testing
“The timeline is definitely shorter. We don’t have to book testing
services or repeat failed tests. Michelin’s service is from the
manufacturer, so it’s 100% guaranteed : which is really valuable for us.”
A NEW STANDARD FOR DATA OPENNESS
Beyond technical benefits, Bon Aizawa highlights something deeper: the openness of Michelin’s approach.
“I appreciate your openness about data, because so many tire manufacturers try to keep their data secret. This openness can enable so many companies to accelerate the automotive industry. It’s just great.”
For Turing, this openness symbolizes a new era of collaboration between OEMs and simulation innovators.
WHAT’S NEXT
Building on this success, Turing plans to continue using MICHELIN SIMIX datasets for new vehicles and advanced AI training projects.
“We plan to move on to our next vehicle and test new tires with MICHELIN SIMIX. It automatically lets us choose from Michelin tires, which is very good for us.”
ABOUT
Turing inc
NAME: TURING CO., LTD.
(チューリング株式会社)
FOUNDED: AUGUST 20, 2021
HEADQUARTERS: TOKYO, JAPAN
EMPLOYEES ~95:
WEBSITE: チューリング株式会社
Explore datasets used by Turing inc
