Cars must react in real-time to unmapped items like fallen trees, debris, and people (not to mention other cars), so what additional value do high-res maps provide?
Do they reduce processing time because now cars simply scan for deltas instead of the entire environment?
Another reason is that autonomous vehicles take data from a wide variety of sensors, including cameras, RADAR, LIDAR and others.
None of these modalities is perfectly, and there is always ambiguity and drift. High fidelity, high resolution maps provide a strong prior that helps to resolve these issues.
Besides the sensing issues, high fidelity maps also provide priors with respect to planning and prediction of the behaviors of other road users. A busy intersection should be approaches differently to a small backroad.
In the end, autonomy needs both real-time processing and high fidelity maps to perform effectively.
In order to know where the car is located at any given time, the vehicle uses data from the gps and a combination of any of the other sensors to help localize itself in the world. Relying on just GPS is risky and error prone, so fuzing that position with matching landmarks from the field of view, or lidar is a way to improve positioning.
This is especially important for determining which lane the vehicle is in, where it needs to be in order to take a proper exit or turn, and what potential obstacles or traffic scenarios it might encounter depending on where it is located on a road.
They go into details on use cases for HD maps