Using Queryable Networks to Optimize the Search Space

We started with creating trajectories using classical optimization approaches where the constraints lie I described would be added incrementally. We run two sets of lightweight queryable networks, both really augmenting each other. One is trained from interventions from the FST beta fleet, which gives us score on how likely is a given maneuver to result in interventions over the next few seconds. The scoring helps us to tune the search space, keep branching further on the interactions and focus the compute on the most promising outcomes. Now if you model the spawn regions and the state transitions of this ghost objects correctly, if you tune your control response as a function of their existence likelihood, you can extract some really nice human-like behaviors