An ADS-B collision avoidance method and system are provided for aircraft having a speed profile that includes supersonic speeds. ADS-B messages are transmitted from an aircraft at a transmission rate and power predicated on the aircraft's speed. A trained neural network predicts a time-based trajectory of the aircraft using the position, altitude, and velocity of the aircraft when it is flying supersonic. Time-based zone boundaries are generated using the time-based trajectory. Each time-based zone boundary is disposed about the aircraft flying supersonic. Each time-based zone boundary is indicative of an amount of time for the aircraft to travel thereto along the time-based trajectory. Each time an aerial vehicle crosses one of the time-based zone boundaries when the aircraft is flying supersonic, at least one of an indication of the presence of the aerial vehicle, an indication of a potential collision, and a maneuver to avoid the potential collision are generated.