Search
Free Trial

aiSim™

End-to-end automated driving simulation tool, the world's first ISO26262 ASIL-D certified simulator tool

Functions

Self-developed engine provides accurate environment and weather simulation.

Correctly simulating the environment sensed by the relevant sensors to create physical environments and weather effects with full certainty and reproducibility provides a solid foundation for neural network-based multi-sensor sensing systems.

  • Adopts light tracking and rasterization technology to simulate various weather conditions such as blizzard, rainstorm, foggy day, sunny day and so on.
  • Configure a variety of road parameters, including lane line wear road degradation.

Real-time simulation of physically based sensors

Comprehensive and accurate setup of simulation scenarios and interaction environments, with a library of sensor models including cameras, laser radar, radar, ultrasonic sensors, and more.

  • Based on a highly scalable framework, it realizes distributed light pursuit rendering on multiple GPUs to simulate highly complex multi-sensor configurations in real-time.
  • Test the effect of various camera degradation and lens distortion scenarios on the perceptual capabilities of the camera sensor.
  • The new sensor configuration and design enables rapid testing of its sensing performance.

Large-scale generation of synthetic data

Provides real-world scenes that are difficult to capture and edge cases during vehicle driving, enhancing the capabilities of ADAS and AD systems by expanding the dataset.

  • Large-scale generation of scenarios with randomized features
  • Maps and resources for multiple scenarios
  • Provide statistical information and visual feedback to help users understand the synthesized data

Cloud Native UI and Open SDKs

The GUI functionality provided by aiSim supports both local deployment and web-based forms to ensure a seamless user experience and deployment.

 

Equipped with a comprehensive SDK, the API and SDK enable developers to customize their requirements and achieve seamless and efficient integration of the user tool chain.

  • Sensor API allows custom development or integration of third-party sensor models
  • The Scene API allows developers and users to control static & dynamic objects in a scene.

Mature editing process for handling 3D resources

  • Rich and high quality vehicle models
  • Editable road user action model
  • Modularized map editing process that combines multiple operational design domains
  • Thousands of 3D road models
  • 1500+ templates for free creation and customization of scenes

Features and Benefits

Real-time, deterministic simulation engine

Designed to meet all the requirements of physically correct weather and environment simulations, aiSim's proprietary rendering engine provides advanced end-to-end testing capabilities from large-scale test pipelines to challenging HiL setups.

 

In aiSim 4, state-of-the-art rasterization and light-tracing rendering technologies can be used to simulate a wide range of weather and environmental conditions, including snowstorms, heavy rain, fog, and sunlight. In addition, road paint deterioration and road degradation can be easily configured to create more challenging environments for perceptual systems.

Physically based sensor modeling

Sensor simulation is a key part of the autonomous driving software stack testing process because information about the environment around the autonomous driving vehicle is obtained through sensors.

 

Modeling sensors requires a physically-based approach to accurately generate a comprehensive set of simulated conditions, environmental interactions, and sensor degradation. aiSim 4 provides an extensive library of physically-based and validated sensor models, including cameras, laser radar, radar, and ultrasonic sensors.

Large-scale synthetic data generation

aiSim, in conjunction with aiFab, supports simple generation of large-scale scenes with domain randomization to replicate large variations in real-world data.

 

It covers locations and assets for a variety of automated driving use cases, including highway, city, and parking scenarios. Once the data is generated, it provides statistics and visualized feedback so that users can understand their synthetic data in detail.

Configuration Recommendations

optimal configuration

Comprehensive Weather and Traffic Simulation Using the Widest Range of Sensor Models in Computationally Intensive Simulations

Computer Configuration

  • CPU: 7th Generation Inter Core i7 processor (or higher)
  • Memory: 32 GB
  • Hard disk space: 200 GB
  • Graphics: NVIDIA GeForce RTX 3080Ti (or better)
  • Optical tracking: For optical tracking sensor emulation (camera, laser radar, radar), a GPU with an optical tracking configuration is required, e.g. NVIDIA GeForce RTX 2080Ti
  • Titan RTX
  • GeForce RTX 3000 Series
  • GeForce RTX 4000 Series
  • GeForce RTX 3000 Series
  • Quadro RTX 3000
  • Quadro RTX 4000
  • NVIDIA RTX A5000

Software Configuration

  • Latest GPU drivers supporting Khronos Vulkan
  • For NVIDIA cards used on Windows 10, it is recommended to install the GeForce Game Ready driver.
  • For Ubuntu systems, it is recommended to get the latest Nvidia drivers through the Graphics Drivers PPA.
  • 7-Zip
  • Python 3
  • Git Bash for Windows
  • VK_EXT_SCALAR_BLOCK_LAYOUT
  • VK_KHR_SHADER_FLOAT_CONTROLS
  • VK_EXT_DESCRIPTOR_INDEXING
  • vk_khr_spirv_1_4
  • VK_KHR_ACCELERATION_STRUCTUR
  • vk_khr_ray_tracing_pipeline
  • VK_KHR_MAINTENANCE3
  • vk_khr_pipeline_library
  • VK_KHR_DEFERRED_HOST_OPERATIONS
  • vk_khr_buffer_device_address
  • CMake 3.26 or later
  • Python 3
  • Visual Studio 2019 16.10 and 16.11 → Microsoft Visual C++ Compiler 14.29
  • Visual Studio 2022 17.2 → Microsoft Visual C++ Compiler 14.32
  • Ubuntu 20.04 → GCC 9.4.0 and Clang 10.0 (with libstdc++)
  • Ubuntu 22.04 → GCC 11.4.0 and Clang 14.0 (with libstdc++)
  • CMake 3.26 or later

Supported Platforms

  • Windows 10 64 bit
  • Ubuntu 18.04 and 20.04


performances

  • Recommend the best configuration of aiSim to realize its full potential.

Base Configuration

CPU only simulation.and the use of object-level sensors.

  • CPU: 7th Generation Inter Core i7 processor (or higher)
  • Memory: 8 GB
  • Hard Disk: 200 GB

The aiSim Normal configuration is an example of deploying aiSim on a lower performance device. It may affect performance when simulating complex environments such as city scenes or scenes with many light sources at night; therefore, optimal performance cannot be guaranteed.


(The NVIDIA GTX 1050Ti does not support optical tracking sensor emulation.)

Upgraded Configurations

Combined with camera sensors, object-level emulation eliminates the need for more complex light tracking.

  • CPU: 7th Generation Inter Core i7 processor (or higher)
  • Memory: 32 GB
  • Graphics: NVIDIA GeForce GTX 1080Ti (or better)
  • Hard disk space: 200 GB
close menu icon

Contact Hongke to help you solve your problems.

Let's have a chat