Having reviewed GRFP applications, and gone through the NSF orientation sessions, I'd say it depends very much on the level of the applicant, but generally, literature reviews are not anywhere near as important as you thought they were.
This process only vaguely resembles a research grant review, primarily because there is ZERO effort to match reviewers' expertise to those of the applicants. Broadly, electrical engineers will review applications of electrical engineers, and biologists will review applications from biologists, but that's as fine-grained as the process will guarantee. This means that you can't display fundamental gaps in knowledge that any practitioner in your broad field should know, but in all likelihood, nobody is going to go check your sources, or know much about the very special second messenger system that impacts a new subtype of cation channel that 2 people in the world study.
As to where the research plan comes from, that's going to vary incredibly by the level of the applicant. A student about to start a program, who hasn't even set foot on their campus yet, is very different from a student that's been placed in a PI's lab for a year, and both can get funded! For the young student, reviewers are looking for your ability to put together a cogent hypothesis-driven study and communicate it well (and not many reviewers would believe this is the study you will actually do your graduate work in) along with a somewhat viable path to get yourself established with the type of resources you would need to actually do similar research, and somewhat more than that for an experienced student. My recollection (though I could be wrong) is that there is some earmarking of awards to the four levels of application status, and that young applicants will get a competitive amount of awards, even though the experienced students clearly have more polished packages.
Reviewers are trained to assess you in two areas; "Intellectual Merit" and "Broader Impact". Both are rolled together to make a holistic assessment of your application package. Reviewers are looking for evidence that 1) you've got the intellectual goods to thrive in science, 2) you've made opportunities for yourself and have taken advantage of them, and 2) that you see areas where you can bring your skills to the larger community. I can't say that you'll be successful if you can demonstrate success in these areas, but I can say that you won't if you can't. Using your literature would be a part of this, but I'd say that if you don't make any mistakes in your use of the literature, you won't be hurting your application.
As to whether you should apply this year or wait until next year, that's a tough call. The risk with the "make believe" type of applications, where you aren't even affiliated with a lab yet, is that one reviewer might not buy that the path you've outlined is viable, and considers this to important to recommend funding (I try hard not to do this, myself, though naivete can certainly be a strike against intellectual merit). This will kill your application. I don't believe you are excluded from applying a second time, though, so the risk of application this round seems minimal. You should verify eligibility for a second submission on your own, though (i.e., don't take my word for it).