How can I make sure that my paper's work is unique?
The only thing you can do to ensure that your contribution is original, i.e., nobody has done the same thing before, is extensive literature research.
Use Google Scholar or Scopus to search for keywords related to your field and try to find something similar. You will notice that there is only so much literature that is closely related to your topic. If you are thorough, you should be able to say if you have done something original.
Your advisor doesn't seem to be keen on talking to you regularly. In the case that you want to continue working for him, you should probably do the literature research, write the article, send it to your advisor for corrections and hand it in. I doubt your advisor will take much time to correct your article, so you should have more experienced colleagues read your article before sending it to him. If he agrees, submit it to a journal and wait for the outcome.
Tactic #1: Keep up with the literature
I once wrote a blog post on this topic; here is the summary:
- Use search engines, especially Google Scholar
- Link forward through the literature
- Learn how to do effective keyword searches
- Use review articles
- Use RSS or email notifications for new journal issues
- Check preprint servers (like arXiv) and key author websites
The details are in the post linked above.
Tactic #2: Talk to people
Talk to everybody you can about your research topic. Ask them what they know about related research. Do this at your university, but especially at conferences or when you visit other institutions. When you have new results, email your close colleagues about them. Once you know a few people, this approach becomes even more effective (and much more efficient) than #1, since you are leveraging your colleagues' collective knowledge of the literature.
To address you concern that it is too late once it has been submitted to a conference.
Acceptance: Conference reviews are not the be all and end all. Getting accepted does not make it a good or correct paper. It does not mean that the paper makes a novel contribution. Acceptance reflects the views, knowledge and biases of the reviewers. This type of peer review is useful from the view that it is a a final last check on paper quality, novelty, etcetera. before being put out in the wild to see if your ideas are useful/deepen understanding/correct.
Therefore, you should only rely on the reviewers as a last line of defense against publishing something non-innovative, etcetera. and do a thorough literature search beforehand. You can never be sure, of course. Some authors put "to the best of our knowledge" before making claims of novelty.
Rejection: If you are rejected because it is not seen as novel is not as detrimental as you might think. Firstly, the next time you submit to a conference you might get ignorant but positive reviewers, or the reviewer who thought it was not novel may be someone who thinks everyone is copying them.
Therefore, if it is rejected for not being novel you should not worry too much. Most likely your paper might claim to be the first to do X and Y, but another paper already does X and Y. However, X and Y might be a very broad general contribution. Perhaps the existing work does X and Y using method M1 whilst yours uses M2. Or perhaps existing work makes a more specific contribution than X and Y, X1 and Y1, whilst yours makes X2 and Y2 and both pieces make contributions that could be described as X and Y. In other words, you would need to rephrase how your work is different by making comparisons on the specific details. Now, if the specific details are also the same (never seen this happen, personally), you can at least do a little bit more research to make it different.
To summarise, even if after doing an extensive literature search (which is necessary to make a contribution) you can always adjust how you sell your work (which is sufficient to make a contribution).