I am a bit stuck trying to understand how a lane change is performed from an operational point of view.
Let's assume a self-driving car uses an occupancy grid map for local planning, this map may even have the detected lane boundaries. It's following a slow car and decides to overtake, but how does it know where the centre of the adjacent lane is? Does it use a separate map, or is there a separate data structure that is used to keep the lane information which informs the car where the centre of the adjacent lane is?
Alternatively, does the car just decide to start drifting off to the side until it picks up the lane boundaries and then centres itself?