Somewhere, and I'm thinking it could be the RAF museum at Hendon, there's an excellent example of how ILS operates, but using old tech.
The first attempt at a localiser used good old radio braodcasts. Two lobes of radio energy were broadcast using directional antennae. Both simply broadcast a beep, with spaces between beeps of equal length to the beep itself. The beeps of each lobe were perfectly out of phase with one another but were broadcast on the same frequency, and here's the clever bit - the lobes had a small area of overlap. If you were in the overlap, you heard a solid sound with no gaps, and you were in the groove!
ILS is much the same in concept except with more of an overlap and the phase difference is measured by an instrument (that'd be the ILS receiver) in the cockpit. The sweet spot is still right in the middle, and the needle on the receiver's dial (or bug or whatever in a glass cockpit) shows you where you are in relation to that sweet spot. Left a bit, right a smidge, etc etc.
Going back to uk20's third post - you say the aircraft "goes through approach". Actually an aircraft of the airliner variety will usually spend every second of every flight, whether in the approach phase or not, under the control of ATC. There are more area/en-route (the latter name is used in North America) controllers than there are approach/terminal (same again) controllers, because aircraft spend a lot more time and travel a lot more distance in the cruise phase than the approach phase. The instructions we give aren't just climb/descend and turn left/turn right. Speed control is an essential element too. Then are tower controllers who usually have a radar screen or two nearby but actually control on the basis of what they can see out of the window. If you really want to know more, pay a visit to your nearest tower. If you can make it to one with an approach unit on site, so much the better.