IMO:
R&D is all about specialization, so choose an area you'd like to work in. You don't need to choose immediately, you can play with various stuff to find the right kind of work for you (the factors that matter for me are: the amount of low-level/high-level coding, the heaviness of math, the size of projects).
As for PhD, I think that, as long as you're exceptional in your area of expertise, you can get R&D jobs without it. On the other hand, getting to that level can easily take as long or longer than doing a PhD. Also, I think you can probably gradually advance into more researchy positions through smart career progression, starting with more development-oriented jobs in R&D areas combined with doing plenty of research on your own.