Systems and methods related to controlling autonomous vehicle(s)

A method for using one or more trained machine learning (ML) models by an autonomous vehicle, the method comprising:

• identifying, based on sensor data from one or more sensors of the autonomous vehicle, a plurality of actors in an environment of the autonomous vehicle;
• identifying, based on the plurality of actors in the environment of the autonomous vehicle, a plurality of streams, wherein each stream of the plurality of streams represents a candidate navigation path for the autonomous vehicle or for one or more of the plurality of actors in the environment of the autonomous vehicle;
• processing, using one or more ML layers of one or more of the trained ML models, the plurality of actors and the plurality of streams to generate trajectories for the autonomous vehicle and for each of the plurality of actors, each of the trajectories being forecast with respect to a stream of the plurality of streams;
• for a plurality of iterations:
• processing, using a stream connection function, the trajectories for the autonomous vehicle and for each of the plurality of actors to forecast each of the trajectories with respect to each stream of the plurality of streams;
• and
• processing, using one or more of the ML layers, the trajectories that are forecast with respect to each of the streams to update each of the trajectories with respect to the respective stream;
• and
• subsequent to processing the trajectories for the plurality of iterations:
• determining, based on the trajectories, one or more predicted outputs;
• and
• causing the autonomous vehicle to be controlled based on the one or more predicted outputs generated using one or more of the ML layers.