If you feel that USB is too boring for you for 1 whole year, at least for the current one, you could pick any other topic, even graphics file formats and graphics algorithms (bitmaps, compression, vector, video...). Pick a specific file format, compression or graphics algorithm for that year, and you will have something.
For example, for GIF, I've seen that it could otherwise take 2 years of wondering how it works, but I've implemented an HTML5 program that can render almost any GIF (only some old ones with bugs/features I don't understand yet), but can help you master GIFs in a shorter time, also download GIF87a/GIF89a standards, Dr. Dobb's articles, etc., and the newest GIFV from imgur that I haven't studied, but GIF87/89a can be unified in the handling code:
http://www.archefire.org/GIF_Explainer/ (main project)
http://devel.archefire.org/forum/viewforum.php?f=22http://devel.archefire.org/forum/viewforum.php?f=9I'm not sure you wouldn't finish it, just make a numbered list of the hardest topics in an OS, and if you dedicate 1 year for completing each in a logical order, as you feel and see that you are advancing and can/want do each because you feel it's entertaining for that year, you will be able to have something. Only by reaching the point where you could use official device drivers for any OS from your OS in 1 year, and 1 year for running any applications, third party and own, you would be on your way of having a capable OS, at least in 1 or 2 decades, unlike doing things without structure nor sufficient concentration/enthusiasm and then achieving 10 times less or so in the same time.
There will always be small tasks to be done and smaller goals in a year, so you will always have something to contribute.
If you explain the things you do in detail in a year, there will always be people willing to help, public domain or open-source in general, starting with enthusiast people who really don't care about licensing, just doing what they can for learning to do interesting stuff. At least I help and leave some contribution whenever I fully understand something so that everyone benefits and accelerate their efforts.