KOTEI Launches Data-Driven ADAS Development Platform to Accelerate AI-Powered Autonomous Driving

At Auto China 2026, KOTEI officially unveiled its new data-driven ADAS development platform, designed to help automakers improve advanced driving algorithms with greater efficiency, lower development costs, and higher-quality data utilization.

As autonomous driving systems continue to evolve, data-driven engineering has become the foundation for large-scale model iteration. The quality of data — and how effectively it can be applied — increasingly defines the upper limit of intelligent driving performance.

KOTEI’s platform addresses several common industry challenges, including fragmented data utilization, low-value training datasets, and development workflows that rely heavily on manual operations. To solve these issues, the company has built a complete end-to-end data pipeline covering data collection, management, annotation, model training, simulation, and validation.

Powered by AI technologies, the platform establishes a closed-loop workflow centered around four core capabilities: scenario identification, value evaluation, scenario reconstruction, and scenario generalization. By extracting high-value scenarios from massive volumes of FOT (Field Operational Test) data and combining them with simulation technologies, the platform enables continuous conversion of road-test data into algorithm capabilities.

AI-Powered Scenario Identification

Traditional manual scenario recognition is often time-consuming and inconsistent. KOTEI applies AI technologies to automatically identify driving scenarios and generate semantic labels, significantly improving both efficiency and accuracy.

Based on KOTEI’s optimized VLM models, the system can accurately recognize complex driving environments while supporting customizable labeling systems tailored to different user requirements. Combined with expert review mechanisms, the platform continuously improves model performance through self-iterative learning.

AI-Based Scenario Value Evaluation

The quality of FOT data can vary significantly, making efficient data selection critical for algorithm training.

KOTEI’s platform uses AI to automatically evaluate scenario value and identify the most relevant high-quality datasets for different development objectives. Built around the ISO 34505 standard, the system provides quantifiable scenario evaluation while generating explainable causal analysis for scoring results, creating a more transparent and traceable evaluation process.

Through the combination of expert validation and model self-evolution, the platform helps developers efficiently identify truly valuable corner-case scenarios.

AI Scenario Reconstruction for High-Fidelity Simulation

Building realistic simulation environments traditionally requires significant manual effort and often lacks sufficient realism.

KOTEI leverages AI-powered reconstruction technologies to automatically generate virtual environments from real-world scenarios, greatly improving simulation efficiency and fidelity. Using generative 3D Gaussian Splatting (3DGS) technology, the platform increases modeling efficiency by up to ten times while accurately reproducing textures, lighting, and environmental details.

The result is highly interactive dynamic simulation scenarios that can be directly used for end-to-end closed-loop testing.

AI Scenario Generalization for Corner Case Expansion

High-value edge-case data remains one of the biggest bottlenecks in autonomous driving development.

To address this challenge, KOTEI applies AI-driven scenario generalization technologies that automatically generate expanded datasets from limited high-value samples. AI Agents analyze existing scenarios and create generalized variants, while AI-generated dynamic traffic flows enable broader testing coverage across different conditions.

The platform also supports parameter tuning and manual fine adjustment, allowing developers to customize extreme operating conditions as needed.

KOTEI and Ignite by FORVIA HELLA Demonstrate Joint Innovation in Compliance Validation

At the exhibition, KOTEI also partnered with Ignite by FORVIA HELLA to jointly showcase an autonomous driving compliance validation solution that combines high-fidelity simulation with “Regulation as Code” technologies.

Within the solution, KOTEI’s data-driven ADAS development platform provides interactive simulation environments and ADAS algorithm execution capabilities, while Ignite contributes TÜV-certified regulatory evaluation algorithms.

The entire validation workflow can be completed inside a simulation environment without on-road testing, reducing risk, manual workload, and certification costs. This enables OEMs to perform European regulatory compliance assessments during the development phase and accelerate global deployment of autonomous driving technologies.

KOTEI’s data-driven ADAS development platform represents a shift from traditional “data-driven training” toward a more advanced “scenario-driven development” approach. By helping automakers improve edge-case performance faster, more accurately, and at lower cost, the platform aims to accelerate the commercialization of next-generation intelligent driving technologies.

Looking ahead, KOTEI will continue collaborating with industry partners to explore new possibilities in autonomous driving development and help shape the future of intelligent mobility.