Indeed our ROAR is similar in principle to the RAG device. The main difference is the type of tags and the variation in stimuli that can be used. We added motion sensor activation recently also. I'd love to hear more about your preliminary results and agree, imitating a human is a darn good set of potentially aversive stimuli for many species. One of the sound files we are keen to trial on lions and feral cats in Australia is a recorded crowd, played at a loud volume.
Have you found any problems with habituation to the stimuli? That's a big part of the reason that we wanted activation contingent upon close proximity (also helps preserve battery life). We really aim to have animals associate their presence with the stimuli. However, having different random sounds that all share an aversive quality could help if habituation does start to creep in.
You may have heard of the "Scary Man" effigy, which Andelt et al. (1997 - full reference below) tested as a part of a battery of stimuli to help prevent bird predtion on fish at fish farms. The effigy randomly inflates and has lights and a wailing sound. I've attached a picture from Andelt et al. (1997) for you to get an idea of what it looked like. They had modest success, but suggested habituation over time.
It actually takes quite a bit of power (potentially) to keep reinflating an effigy I imagine, although I do like the idea. Our variation on that theme is to use something like a "cut out" of a human (e.g. using marine plywood) and have one or two motorised arms. Could also potentially motorise it to pivot in place. We aren't able to use audible sound or visible lights in our application as the herders sleep very near to where the effigies will need to be located, but you could definitely consider adding something like that. I thought about a bright, near infrared light, that would be very unlikely to disturb anyone, but would be very visible to most nocturnal predators. Ultrasound could be another option (although unless it's extremely loud, might itself not be overly aversive).
Anyway, I'd be very excited to discuss a collaboration on this sort of thing between us all if that's of interest? Looking forward to more discussion and thanks for reaching out.
Andelt, W. F., Woolley, T. P., & Hopper, S. N. (1997). Effectiveness of barriers, pyrotechnics, flashing lights, and Scarey Man® for deterring heron predation on fish. Wildlife Society Bulletin, 686-694.