再看看occupancy flow
Predicting Occupancy and Occupancy Flow
Utilizing a combination of neural networks, the Occupancy Network predicts both occupancy and occupancy flow. Occupancy prediction involves determining whether a particular location within the environment is occupied or unoccupied. This prediction is crucial for path planning and obstacle avoidance. Occupancy flow, on the other HAND, focuses on tracking the movement of objects in relation to the vehicle or bot. By understanding occupancy flow, Tesla's autonomous systems can accurately anticipate the behavior and motion of objects, enabling dynamic navigation and real-time decision-making.
geniushanbiao 写了: 2024年 10月 31日 13:44
我仔细读了,无非就是cell grid,grip mapping,这种方法对物体距离是无效的,因为没有一个可靠的reference。