Something like you describe really does sound like an ideal terminal stage defense system, with 20 or 35mm autocannon covering longer ranges.
Purely theory crafting, I think the final piece would be small, standalone AESA radar units perhaps the size of a man-carryable generator that can be distributed across a given area and networked to likewise distributed shooters, spreading out the targets, adding redundancy to the system, and reducing the odds of your dugout-with-air-defense-asset getting hit by an anti-radiation munition that homed in on a known emitter.
The reason that you don't need radar for this is that the small suicide drones use cameras (vision) to attack. If you can't see them, they can't see you! Contrast this to general aviation where planes can be kilometers up, behind cloud-cover and can attack you with radar-guided missiles.
Radar probably won't have high enough resolution and won't be able to differentiate between drones and birds. Not to mention that most drones are carbon fibre or plastic and won't have much of a radar return to begin with.
Conversely, vision systems could be trained to ID drones reliably, and could use data such as the high-frequency "beat" of the propellers to detect drones from just a few pixels of sensor data.
Purely theory crafting, I think the final piece would be small, standalone AESA radar units perhaps the size of a man-carryable generator that can be distributed across a given area and networked to likewise distributed shooters, spreading out the targets, adding redundancy to the system, and reducing the odds of your dugout-with-air-defense-asset getting hit by an anti-radiation munition that homed in on a known emitter.