heres my theory on that: in the desert, one occasionally enjoys a windless day where the air is so clear and calm, dust free, nearly moisture free, that distant objects seem to be close enough to touch. for those who have never been in the american western desert, it is a unique experience. go far out in the desert and enjoy something extremely rare: total silence. no cars, no people, no birds chirping or bugs buzzing. one can hear one's heartbeat. this is no exaggeration: you really can, given the right conditions. you might get days like that down there in texas.
one sighting was over the tonopah range. if it was a day like the one i describe, there would be no way to determine the distance to an aircraft flying overhead. now, suppose somebody developed a drone with a wingspan of about 2 meters, driven by a pulse det engine, made entirely of flat plates to scatter radar (like the f117). if he assumed the normal sized aircraft, he would overestimate the distance to the tiny aircraft by a factor of 4 or more, and would also goof the speed by the same factor.
nowthen, the mysterious part is that the observer did not report the angle between the velocity vector and the vector from observer to aircraft at the moment of hearing the sonic boom. this angle is the arcsin of the inverse of the mach number:
alpha = arcsin (1/m) where m is the mach number of the aircraft and alpha is the angle between the v vector and the aircraft direction at sonic boom arrival. or m = 1/sin (alpha) if you prefer.
this equation works regardless of the size of the aircraft or its distance. the pulse det engine would have made a hell of a racket at mach 3 (if it somehow got going that fast) but would have been silent until alpha was about 20 degrees regardless. an experienced aircraft watcher would know this, so i cant explain how it was reported that the "aurora" was travelling in excess of mach 3.
i can see why the government would intentionally create a small recon drone and fly it all over the desert, hoping many observers would greatly overestimate its speed capability, thereby providing a smokescreen for what it is *really* developing.