Avita 11 intelligent driving system does not rely on high-precision maps to achieve urban NCA

The intelligent driving system of Avita 11 adopts the urban NCA technology that does not rely on high-precision maps, solving the pain points of road changes. The system integrates the road topology reasoning network (RCR) and the general obstacle detection network (GOD), which can not only understand the road, but also understand the objects, and keenly identify various special-shaped obstacles. In addition, the system can also continuously optimize the algorithm through massive data training. It has accumulated 2.8E FLOPS learning and training examples, conducts 10 million kilometers of simulation training every day, and iterates the model every 5 days on average, making the car smarter and smarter.