search

opendilab / InterFuser

[CoRL 2022] InterFuser: Safety-Enhanced Autonomous Driving Using Interpretable Sensor Fusion Transformer
Apache License 2.0
514 stars 42 forks source link

map issue #58

Open a1wj1 opened 11 months ago

a1wj1 commented 11 months ago

Hello, how do you consider the situation where the current frame of an object exists, but the next frame does not exist in the map?

deepcs233 commented 11 months ago

Hi! Our framework also predicts the velocity and the heading when detect the other objects. With these information, we can estimate their future positions. If they are far away from the ego-car, we will not consider them again.

a1wj1 commented 11 months ago

Hello, what I mean is during the training phase, how to make predictions when training data if an object in the future frame no longer exists

deepcs233 commented 11 months ago

In our current pipeline, the framework receives single-frame input and predict single-frame output. Or you can refer to ReasonNet to modify the framework to let it predict multiple frames in the future.