crobato, as mentioned by ISRO official, normally the technique of avoiding collision with others is employed while launching satellites, i.e. the path of the launch rocket is not to intersect any other sat’s path– or if it does, the sat should be calculated to being absent there at that point of time.

However, in the ASAT test conducted by China, the rocket was set upon a ‘deliberate’ collision course. Like all rocket launches globally, it would have been tracked also, but Not command guided by powerful ground radars (like Arrow’s Greenpine or Aegis’s Raytheon radars). The rocket may have used terminal homing, in case the satellite was radioed to maneuver (by 20 kms as you said); however, in a disturbance free atmosphere, targeting a 20 km deviation is achievable given that BMDs achieve an interception against targets that cover a few kms per second.

The seeker is always needed because collisions themselves are very difficult to achieve especially when they are that small. The ASAT itself weighs no more than 40 to 50kg and that is something can only be as big as a chair, against a target that is only 3m2 or roughly 10 feet long. Considering the vast distances and speeds, even a slightest miscalculation or deviation caused by unknown factors would have meant a miss. Your INS have to be that precise, I can’t imagine this thing being done without a very precise atomic clock and gyroscope—this alone means major implications about our beliefs on the sheer accuracy of all Chinese missiles. The seeker should make up for any last moment variation and buffer for errors. For it to be that precise, the seeker must be at least on the infrared.

Simply said the ASAT never assumed an orbital path because it was never seen or tracked in the western hemisphere and never even when past the horizon. The entire ascent and collision was done completely in one sweep in within the eastern hemisphere.