The companies are planning to make a series of hardware demonstrators that will combine image sensors, radar sensors and processor chipsets from On Semi with machine learning algorithms, machine learning hardware acceleration and simulation capabilities from AImotive.
The prototypes will utilize AImotive’s aiWare neural network acceleration IP in FPGA-based prototypes as well as modules of software from its aiDrive software portfolio and its aiSim simulation environment.
A move to AI-enabled subsystems within an automobile is expected to provide superior accuracy and robustness and lower power and lower latency due to reduced amounts of data being transferred between nodes inside the vehicle.
One of the areas where sensor fusion is being considered is in the fusion of images and data from lidar, optical cameras and radar. Combining high resolution image sensors with radar sensors is expected to improve perception performance in all weather conditions.