I'm writing my MSE thesis report on an inertial navigation problem I've been working on. In my work, I use a complementary filter to track the orientation of a device. Now, since my fellow students are likely not familiar with complementary filters or Kalman filters, I wish to give a brief description of what a "complementary filter" is, in general.
A definition, if you will.
So, what defines a complementary filter? Is it any filter on the form
z = a * x + (1 - a) * y, where x and y are separate measurements of a single quantity?
It's fine to define it in terms of Kalman filters if that's appropriate.