Note - this answer is focusing on the question in comments 'how will that burn accurately take place in 3D space'.
In general spacecraft do not know where they are or where they are going. Instead ground control uses various sensors on the ground and on the craft to determine what the craft position and speed is, and send a plan to the craft to execute, observe the results and repeat as required to get the needed orbit/trajectory.
So our craft needs to work out how to execute those plans. Orientation generally comes from star trackers, so the craft can rotate to position engines facing in the required direction.
The tricky part is the actual thrust. If we are in orbit around earth and it is decaying we know we need to thrust in the direction of orbit to boost apogee back up by a certain amount.
The simplest and earliest option was to just use time, turn engine on for X seconds and stop, and worked even before true computers would fit into spacecraft. This is vulnerable to errors with engine thrust being higher or lower than expected, because fuel pressure and temperature will change during the burn changing performance and burnt will lower mass making on the ground computation complicated.
A potentially more accurate approach is to measure acceleration. If your craft is traveling a passive orbit with no forces applied, a scale onboard will measure zero. If we start the burn, a scale with a known mass sitting on it will start to read a weight, or more precisely a force. With a known force and a known mass we can calculate the acceleration. And if we know the acceleration and our target starting and ending speed we can start to use math to know when to stop, either assuming constant acceleration or using integration to handle variable thrust.
An example of this process in use, and going wrong is with the first flight of Ariane V