AImotive looks to Automotive General Intelligence with aiSim 5
AImotive, now part of car maker Stellantis, is looking to Automotive General Intelligence (autoGI) for its next generation of tools.
The latest simulation technology in aiSim 5 is aimed at autoGI for autonomous driving simulation saus AImotive in Hungary. This latest release combines simulation dynamics with a deterministic and reproducible environment using aiSim AIR for multi-sensor simulation.
AThe company has developed an intuitive user interface, aiFab, uses large language model-based techniques and is fully integrated into the aiSim workflow for smart scenario creation with features such as scenario variation and detailed statistics.
“aiSim 5 is a milestone, combining AGI with deterministic simulation,” said Daniel Tosoki, aiSim Product Director. “Its impact on the industry is profound, offering an unparalleled bridge between innovation and reliability. The integration of AutoGI not only propels the boundaries of what’s achievable in automotive simulation but also reinforces our commitment to shaping the future of mobility. In an era demanding cutting-edge technology and dependable solutions, aiSim 5 emerges as a catalyst for unprecedented advancements in the autonomous driving landscape.”
aiSim AR provides high-quality synthetic sensor data while optimizing resource usage with scalable parallelism, reduced memory footprint, balanced workloads and asynchronous data transfer while keeping ISO26262 compliance and determinism. This allows seamless transitioning between real-time Hardware-in-the-Loop (HiL) to Software-in-the-Loop (SiL) with rasterization, raytracing, and AI-based rendering techniques to apply technologies such as Neural Radiance Fields (NeRF) and 3D Gaussian splatting to bring the virtual worlds even closer to reality.
“Our version of AGI represents a pragmatic expansion of intelligence specifically tailored for the automotive domain,” said the company. “Since aiMotive’s inception, we’ve been at the forefront of harnessing AI for automotive applications. Neural networks have been the backbone of our systems, gathering, processing, and applying data for training, validation, and driving.”
“However, the road ahead demands more than understanding and respecting traffic rules; it requires a level of situational intelligence akin to human understanding.”
aiSim 5 has open APIs for sensors, scenario & map management, vehicle dynamics and a complete SDK with example codes on multiple platforms and integrates with industry-standard toolchains. It is also compatible with ROS2, Autoware, FMU, and ASAM’s OpenCRG, OpenSCENARIO, OpenDRIVE, and OpenMATERIAL.
The company has partnerships with tool vendors Foretellix, Mathworks, Macnica, HongKe, NI, dSpace and B-Plus and plans to use autoGI in its other products thorugh 2024.
“It’s essential to clarify that AutoGI, in our context, doesn’t imply superhuman intelligence but rather a form of generic intelligence. The goal is not to mimic human intelligence entirely but to broaden the capabilities of our AI to go beyond the limitations of conventional, automotive-centric intelligence,” said the company.
“As we embrace AutoGI, we acknowledge that it goes beyond the conventional definition of AI. Our version of AutoGI is about crafting technology that mirrors and anticipates the complexities of human intelligence, offering a level of adaptability and comprehension that sets a new standard for autonomous driving systems.”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
