aiMotive Services

simulation

Provides a simulation environment for verification/validation of AD/ADAS. Uses a proprietary rendering engine and is a tool that has been certified to ISO 26262 ASIL-D.

Data Pipeline Tools

This tool efficiently supports data-driven development by realizing data pipelines such as the collection, storage, and analysis of large amounts of data for AD/ADAS on a unique platform. Data can also be reused in conjunction with aiSim.

NPU License

This is a hardware license for an efficient NPU (Neural Processing Unit, an accelerator built into an SoC) for automobiles. Its unique algorithm and architecture enable it to increase the efficiency* of inference processing per second.

"aiSim" is a high-precision simulation tool ideal for autonomous driving and ADAS development

aiSim is a simulator that uses AI to create digital twins, high-speed sensor simulation, and a proprietary rendering engine to provide a highly reproducible environment. Its flexible architecture makes it easy to integrate with existing toolchains and contributes to reducing the cost of real-world testing. aiSim offers innovations that go beyond traditional automotive simulation, setting new standards for realism, adaptability, and industry reliability.

Physics-based sensors and environment simulation
Supports weather conditions (snowstorms, heavy rain, fog, etc.) and complex sensor setups with distributed rendering across multiple GPUs, all simulated based on the laws of physics.

3D Digital Twin Environment
Utilize digital twin 3D environments of real locations in all relevant ODDs (operational destination domains) such as highways, urban areas, and parking lots.

Automotive grade
aiSim is the world's first ISO 26262, ASIL-D certified automotive simulator. This certification is made possible by deterministic, physics-based sensor and environment simulation.

Flexible architecture and easy integration
It is designed to be modular with APIs for easy integration with any system under test, and an open SDK allows you to create your own test toolchain.

Extensive 3D asset library and mature content pipeline
It provides a comprehensive set of 3D assets and related tools needed to set up diverse, high-fidelity environments, including vehicles, vulnerable road users, maps, assets, and scenarios.

Smart Edge Case Detection
Our domain randomization feature (aiFab) allows you to create the large amounts of diverse synthetic data needed to replicate the real-world conditions required for training and testing your software.

Reasons to choose aiSim - The capabilities of development support tools required for autonomous driving and ADAS development

- A unique rendering engine enables deterministic, physics-based sensor simulation

By running pixel-level simulations of a wide variety of sensors and rapidly changing weather conditions, complex sensor setups can be validated, contributing to lower development and validation costs by reducing the number of people and vehicles required.

・Automotive-grade simulation tool that complies with ISO26262 ASIL-D

As the world's first automotive-grade simulator certified for ASIL-D, aiSim enables manufacturers to use it to develop and verify changes to their AD/ADAS algorithms and autonomous driving systems, reducing the cost and time required for safety certification.

- Mass generation of edge cases and rare scenarios through synthetic data generation

By creating large amounts of synthetic data scenarios by systematically or randomly varying the values of multiple parameters, it becomes easier to discover flaws in algorithms, software, and systems in edge cases, leading to improved product quality and safety.

・Convert real data into digital data for simulation

Real-world data can only be collected once, and even when driving in the same location, it is difficult to obtain data under the desired weather conditions or traffic conditions. However, by extracting scenarios from real-world data, it is possible to utilize existing data and eliminate the need to re-collect data, allowing for more efficient development.

・Open API/SDK enables integration with other simulation tools and HIL/SIL systems

The open platform concept based on standards enables linkage with various standard tools, which allows you to perform advanced simulations by compensating for the weaknesses of simulation tools and making the most of the tools you are currently using (simulators, HILS/SILS, etc.).

"aiData" data-driven development platform that accelerates autonomous driving and ADAS development

aiData is an automated data pipeline designed and streamlined to streamline the development of autonomous driving technology. It automates key stages of the MLOps workflow, from data collection and organization to annotation and validation, a key part of machine learning operations.

Accurate sensor calibration and time-synchronized data collection are essential to generate high-quality data for the development, testing, and validation of AD/ADAS (autonomous driving /Advanced Driver Assistance Systems). aiData Recorder is a highly adaptable smart data collection software that ensures recorded data is of the highest quality.

The development of AD/ADAS (autonomous driving /Advanced Driver Assistance Systems) requires a huge amount of data. After data acquisition, annotation is essential as a post-processing step, but traditionally this has required a lot of manual work and a large amount of human resources. aiData Auto-Annotator uses AI and GPUs to enable fast, automatic annotation after data acquisition.

Evaluating and validating algorithms and software developed for AD/ADAS (autonomous driving /advanced driver assistance systems) is often complex, and aiData Metrics helps validate datasets and compares them with performance requirements, enabling data gap analysis.

Data preparation is a time-consuming part of MLOps workflows in AD/ADAS (autonomous driving /advanced driver assistance systems). This is because preparing and organizing data takes time. The aiData Versioning System provides full transparency and traceability throughout the data flow, streamlining data preparation with the latest AI technologies such as text, image, and scenario-based search.

Why choose aiData? The power of development support tools that change the quality and efficiency of data utilization in autonomous driving development

aiData Recorder
・マルチモーダルセンサー設定の​オフラインキャリブレーション機能

・センサーの​正確な​時刻同期

・取得データの​自動アップロード

High-quality data can be acquired with inter-sensor calibration and precise time synchronization, and the acquired data can also be automatically uploaded. After uploading, automatic annotation dramatically improves the efficiency of annotation work.

aiData Auto Annotator
- Multi-sensor automatic annotation of dynamic/static objects using 4D (space and time) environment models
・Scenario extraction converts sensor data into virtual scenarios
・Quality control measures for automatic/manual annotation

By reducing the man-hours and time spent on annotation work, you can focus on core development and improve development speed. Scenario extraction also eliminates the need to utilize existing data or re-acquire data, improving development efficiency.

aiData Metrics
- Find gaps in your data and easily determine its usefulness
・Comprehensive evaluation of neural network algorithms and software

By benchmarking environment detection and object tracking, you can visualize the evaluation results for each version of your algorithm or software, and identify gaps. Repeated evaluation and improvement contributes to improving the quality of your algorithms and software.

aiData Versioning System
- Track the entire data flow for pre-processing/post-processing
- Manage data such as weather, map, and content parameter tagging
・Organize data using AI technology, such as text/image/scenario search

It tracks the entire data flow from acquisition to processing, making progress management easy.In addition to weather, location, and time information, advanced searches such as text/scene searches are also possible, reducing the effort required for data preparation.

What is aiMotive?

aiMotive (founded in 2015, headquartered in Budapest, Hungary) is an automotive technology company working on level-agnostic autonomous driving solutions. Combining in-house expertise with aiMotive's modular capabilities, the company offers an integrated portfolio of tools and embedded solutions to enable the rapid development and deployment of autonomous driving functions, while achieving significant reductions in development costs and time to market.