If you're a missile seeker, all you have as input is a two-dimensional forward view of where in your forward cone the target is - either deflection and azimuth,

If you're a missile seeker, all you have as input is a two-dimensional forward view of where in your forward cone the target is - either deflection and azimuth, or x and y offset - it's like a dot on a scope. Let's assume cartesian x and y input for simplicity. So as you track, you can measure where it is, and how fast it's apparently moving in those two degrees of freedom - so you know x, y, dx and dy (and I guess you know higher degree deltas too). To hit it (assuming it's in front of you) all you need to do is make control inputs to bring dx and dy to zero - you don't care what x and y are. If dx and dy are zero, if you're faster than the thing you're tracking, you're going to hit it - that's constant bearing, decreasing range. Making control inputs to bring x and y to zero as well as dx and dy is a harder problem! You need to control four degrees of freedom to do that. And making control inputs just to try to bring x and y to zero without controlling dx and dy is a good way to miss the target completely - obviously if you get x and y to 0 but dx and dy are nonzero, the target is going to move away from zero, requiring more control input, which might increase dx or dy, causing you to need even stronger control input... So the upshot is: you don't have to back out the target's absolute trajectory, figure out its true velocity and your own relative velocity, determine range to target and offset where you're pointing to compensate for time of flight - just get it to stay on the same bearing from your perspective, and you're done.