The most obvious solution to the problem is to use artificial hair cells. This would involve a “hair” mounted on to an analog stick. The hair would act like the kinocilium, with the potentiometer at the base of the analog stick acting as the synapse. The direction of the hair movement would be translated in to an electrical impulse, which could be interpreted to detect the direction of air flow around the UAV, much how the lateral lines of fish detect the flow of the water surrounding them.
There is one main problem with this methodology: up in the air, there is an incredible amount of vibration and crosswind. The control systems of UAVs already have to do a large amount of vibration filtering to avoid systems interpreting vibration as constant movement. These levels of filtering would also be required on the artificial lateral line system. Lateral lines are designed to detect small changes in surrounding flow, and it seems likely that the air displacement caused by a neighbouring swarm UAV would easily become lost in the general vibrations, and so the filtering. Even if the air changes were able to be discerned from the general vibrations, the cross winds would cause major problems. These would most certainly make it impossible to detect any meaningful movements within the surrounding area, except for perhaps the largest of projectiles passing nearby. Some shielding could potentially be placed around the lateral line to reduce cross winds, but this would hamper the detection abilities so much as to render the system useless.