1. Definitely worth looking at, there are so many techniques out there, each one with some advantages and disadvantages plus it gets you thinking about how you could improve the algorithms (because of how painfully slow many GI methods are)
2. Trig, basic understanding of rays, ray propogation, matrices, vectors, sometimes the very basics of calculus and that should get you through most algorithms
3. theres ray tracing, path tracing, voxel cone tracing, metropolis light transport, ambient occlussion, photon mapping
Based on my experience you can add GI on your engine.....but two goals are difficult to achieve:
1. Even if global illumination stays into the low frequency domain....make it convincing and temporal coherent it's hard
2. Make it fast is even harder
just my 2 cents
p.s.
Is it worth since it will "open" your brain :-)
Can you do great researches without have good background about computer hardware ( Graphics card ) and some background on electrical engineering ?
I read papers on the topic almost daily. Usually you can get by just fine and if there isn't anything you don't know, it probably isn't too technical for you to not be able to figure out via google. Take it like this, I'm in high school, I get by fine with the normal math you learn at school ;)
