I know I'm just some dude and people on the ground likely have tried all of these ideas, but I guess then I would take the approach that you can use that sort of expensive sensor array on autonomous forward observer drones, which detect and map out jammer locations which can then be targeted with basic artillery. Basic terrain avoidance and autonomous (non-RF dependent) flight algorithms can be implemented on simple $500 drones with IP cameras and rpi onboard computers (I've built some of my own). I would imagine that within an order of magnitude of $500, you could get an autonomous drone with enough payload for any type of sensor array you could want for detecting jammers. The big problem that Ukraine has faced has been jammers operating within FPV ranges, which are typically about half the range of a 155mm artillery shell, so leave the targeting to them.
Well, on terrain avoidance, yes, that's fairly simple. What's not simple is avoiding terrain until you reach a certain point, after which you want to collide with the terrain. Identifying that point and autonomously targeting the correct bit of terrain is the problem.
There are systems that can detect and locate a signal like an RF jammer, but you have to 1) deploy that system into the area, 2) operate it, 3) maintain it, and 4) protect it. Then to make use of the intelligence it produces you need to have some sort of ordnance or PGM available. That is, you need to dedicate and risk plenty of materiel to take out a simple RF signal broadcaster, with the reward being the use of drones (including the enemy's drones).
I don't think your approach is ill-conceived or anything, but the reality for both Ukraine and Russia is that it's really difficult to properly implement that approach.
I think the calculation is pretty simple, really: the cost of developing and using a reliable jammer-seeking drone is much higher than the cost of a second jammer.
