This ends up being a cheap jet engine with a quadcopter built around it, for dealing with aircraft.
For dealing with tanks, fpv suicide drones work well. A version of the drone guided by onboard AI, and upgrading the brick of HE that Ukraine seems to put on the front to a shaped charge, would deal with most armor. That is because the drone can just fly behind tanks and other armored vehicles to strike at the thinnest armor, usually the rear, roof, or from below.
You can see recent videos of this used in Ukraine, obviously they are pretty disturbing.
Current transformers networks seem to be more than capable of perceiving and analyzing the features of a complex target such as a Russian T-72 with addon armor. Biggest limitation now is the several H100s needed would require over a kilowatt of power, take several seconds per frame, and add $25k an H100 to the BOM of an under $1000 drone.
The one drone concept that seems to already be obsolete is the “missile” launching drone. Why launch a missile, just use the drone.
https://www.anduril.com/roadrunner/
This ends up being a cheap jet engine with a quadcopter built around it, for dealing with aircraft.
For dealing with tanks, fpv suicide drones work well. A version of the drone guided by onboard AI, and upgrading the brick of HE that Ukraine seems to put on the front to a shaped charge, would deal with most armor. That is because the drone can just fly behind tanks and other armored vehicles to strike at the thinnest armor, usually the rear, roof, or from below.
You can see recent videos of this used in Ukraine, obviously they are pretty disturbing.
Current transformers networks seem to be more than capable of perceiving and analyzing the features of a complex target such as a Russian T-72 with addon armor. Biggest limitation now is the several H100s needed would require over a kilowatt of power, take several seconds per frame, and add $25k an H100 to the BOM of an under $1000 drone.