Jim,
Two ideas. One, for the first stage, to replicate the "Spinnerons" as the Sidewinder missile uses. Although I've never done that and the gyroscopic physics/aerodynamic interaction is something i've not tried to fully understand, just take as a given. Biggest issue might be flutter, though the real thing does not have that problem (the gyroscopic effect may prevent flutter).
The other is..... add electronic roll control to the first stage booster, in between the fins. Would mean adding a separate battery/controller and at least two servos. Maybe three since the booster has three fins. Although, having two vanes near the very bottom, 180 degrees apart, ought to work in roll even if two fins are 30 degrees away (possible pitch-yaw interaction with those fins, maybe). I know, that's a lot to add.
As for the cost of testing, this is why it would be so good if you had a much smaller model to do testing with. So it was not only cheaper, but you could test it at a lot of other launches, even a local site when the weather was good. Several flights a day. But I know at this point, that would be a lot of extra work to do to make up a smaller model, where the vast majority of the work wold not be the rocket airframe but all the guidance related stuff.
That s not a complaint or Monday morning quarterbacking. You have done incredibly well with this project, and lot of impressive research, learning, input, and finding a guru to program the controller. I do get concerned at times about some others who might get into this and decide "yeah I'll do an K powered rocket for my first guidance project", not learn what they needed to, not ground test/simulate like they needed to, and make some mistake leading into a very serious out of control crash. The kind of mistake they could/should have learned in testing smaller models. OK, don't mean to excuse not learning enough to begin with, but part of learning unfortunately means the hard way (a good hard earned lesson - don't let your battery pack go dead before liftoff. Yep, did that once. Waited for a cloud to go by, in order to get a better flight on a sunguidance model. Did not realize the battery went dead during the delay, servo set at full pitch-up as it should have been. The model looped into the ground. And I did not do another pre-launch confirmation that the guidance was working (wave hand over the sensors), before liftoff, I'd done that 15 minutes before).
Fortunately, newbies at guidance having guidance realted accidents with big rockets, without learning on smaller ones first, has not become a problem, yet. I know of none so far. But it could be a problem eventually.
Perhaps that's more of a lesson for others who work up guidance for hobby rockets, as any who read this thread may be interested in working up their own. Start smaller so it's less expensive and very important - logistically you can fly a lot more often (I did my Sunguidance project using Staged D12 power in a BT-60 rocket. Others doing Arduino type guidance today use 2 to 2.6" with up to F power, maybe G, though many could build lighter than they chose to and keep the costs lower). And if the ultimate bad thing happens in testing, crash under power, it's not a big big rocket that it happened with (good feature about your design is there's no active steering of the flight path until after first stage burnout. So if it "went nuts" after booster sep it could not make the rocket veer into the ground. And IIRC you have a stage ignition lockout if it got way way off vertical before staging?).
Even vonBraun did not START with the V-2.....