Added more examples and cleaned up explanations on modifying R and Q. In particular, the filter setting for R was not the same as the simulated noise for no obvious reason. Also, names were bad R instead of R_var, for example.