TRAFFIC LIGHT PREDICTION USING DECENTRALIZED FEDERATED LEARNING

An example operation includes one or more of performing by at least one processor on a first vehicle, processing a first Vehicle Artificial Intelligence (VAI) model and a received off-vehicle Aggregation Artificial Intelligence (AAI) model configured to aggregate the first VAI model, transmitting the first VAI model to the off-vehicle AAI model, and responding to a traffic light prediction, by the first VAI model, by at least one of changing speed, changing direction, changing acceleration, continuing with no changes, or updating an infotainment system, of the first vehicle.