aiSim can simulate 24 virtual pinhole cameras or 4 fisheye cameras and 8 pinhole cameras real-time at 720p resolution each, at 25Hz refresh rate. This is a benefit of the task scheduling methods in the aiSim engine that utilize multi-CPU and multi-GPU platforms efficiently. This performance enables automotive OEMs to utilize completely automated real time hardware-in-the-loop and software-in-the-loop system testing. When real-time running is not required the simulator offers deterministic fixed-time step simulation for rapid software-in-the-loop testing. Utilizing the determinism and realism of developers can quickly identify corner cases in which their solutions fail and tackle these with further engineering efforts.
Relying on aiSim’s modular design suppliers of ADAS solutions and automotive hardware platforms can validate their algorithms or hardware in virtual environments. aiSim is capable of providing ground truth data for all modules not being tested, thus select functionalities can be tested without a complete software stack. Furthermore, when rendering is unneeded aiSim is a reliable tool for reinforcement learning while also enabling rapid augmented reality content creation. Hardware solutions can also be benchmarked, running any self-driving software. This allows technology providers and Tier1s to evaluate their systems safely while decreasing their reliance on costly road testing.
Able to virtualize several sensor types, aiSim is a powerful sensor modeling solution. As a result, sensor suppliers can rely on the simulator to evaluate the performance of their developing sensors. This enables them to gain almost instantaneous insight into the effects that changes to a sensor have on self-driving performance, thus accelerating their development efforts. Sensor providers can also elect to have their sensor featured in aiSim, making their models available to other users, including automated driving engineering teams. Due to the robust sensor models available, different sensor setups can be tested quickly in aiSim, to further support the creation of self-driving systems.