The army is thinking of making the M1A2 Abrams tank a drone control platform:
The Army is preparing to configure Abrams tank prototypes able to control nearby “robotic” wing-man vehicles which fire weapons, carry ammunition and conduct reconnaissance missions for units on the move in combat, service officials said.
Although still in the early stages of discussion and conceptual development, the notion of manned-unmanned teaming for the Abrams continues to gain traction among Army and General Dynamics Land Systems developers.
Algorithms are progressing to the point wherein they will be able to allow an Abrams tank crew to operate multiple nearby “wing-man” robotic vehicles in a command and control capacity while on the move in combat.
As for how good an idea this is, it depends on the drone and depends on the mission. Reconnaissance seems like a natural, and some drones might work well at over-the-horizon top-down attack roles against other armored vehicles in rough terrain or beyond the range of the 120mm main gun (which is plenty far), assuming they can find a way to defeat active countermeasure systems like Drozd and Arena.
I can also imagine non-winged drones, like an updated, low-profile version of the old German Goliath tracked mine used for tasks like bunker clearing.
The M1A2s equipped with the System Enhancement Package (SEP) already have considerably upgraded electronics, and drone control is probably seen as a natural follow-on. The question is where do you store the drones and, if they’re not launched from the tank itself, what advantages does having local control over give you over remote control. I suspect that needing to launch the drones from the tank before rotating the turret is going to be a non-starter in most combat environments. Maybe that could be done automatically, assuming the kinks could be worked out in autonomous operation (a big if; all it takes is one friendly fire incident to scuttle that idea right quick).
(Hat tip: Stephen Green at Instapundit.)