Continuing the thread drift but there had also been an incident of multiple drones falling to the ground or into the water during a drone formation display in Melbourne, Australia in July 2023:

https://www.news.com.au/technology/h...aff58b3b6?amp=

I wonder how the formation control logic deals with a control failure in one or more drones? Does it perhaps have a cascading effect?

Now that we know who was responsible for this, it would be nice to know what they were doing (I'm standing by what I previously stated).

The initial report was released on the Florida drone show failure - apparently a couple of things were a problem. See NTSB Preliminary report DCA25LA065.

The main failure was mismanagement of the drone show bounding box, that is a volume which the drones are not expected to exit. The box is a tight/minimum box that some of the drones will be operating against. This gives the show operators information about how much room they require. Outside of this is the geofence that represents the volume the drones may not operate outside of. Typically there is a gap between the two to make up for small positioning errors. The zone is supposed to be 1 meters, but the show operators used 4 meters, possibly allowing drones to exceed the bounding box by an addition 4 meters, which may not have been reflected in audience setback.

The show and the bounding box are fixed at the time the show is made, except for it's location and orientation, which are changed depending on where the crowd will be.

The geofence, however, should be fixed with respect to the site and the crowd located outside of it.

The claim is that at some point a change in orientation was made, turning the show 7º degrees and the geofence with it, but the show organizers did not update the crowd location map. This meant that there were volumes where the show and the geofence were located to include the volume where the crowd was waiting. As far as the drones were concerned the majority of drones didn't leave the geofence.

The other operational error was some rebooting of the software that appears not to have been complete. This is what caused the initial problems with some drones leaving the ground late and running into drones above them. It was not clear in the report the nature of how problematic information was loaded onto the drones, if all the drones got the information they should have, etc.

I have not seen a reason for the one drone that cannonballed it's way towards the crowd and nearly killed the little kid. He is reported home following open heart surgery.

Very reassuring message from the White House that most of the drone sightings in New Jersey, but not all of them, were authorized for some reason by the FAA.

The claim is that some part of the government set up advanced drone tracking systems, which would not be required if they were authorized.