Simulation has long been an essential part of testing autonomous driving systems, but only recently has simulation been useful for building and training self-driving vehicles. These advances have set the stage for an influx of startups working on new simulation products.
Traditionally, simulation has been used to play back sensor data that cars collected in the real world, and verify that the vehicle’s self-driving software handles tricky situations appropriately. For example, a test vehicle’s cameras might record video of pedestrians crossing the street at night. Software developers can use that video every time they update their self-driving software, to verify that the software still detects the pedestrians correctly.