I am trying to understand the technology/procedure related to one of the trivial functions of a modern self-driving cars.

Let us consider the example of Tesla's Model S (or any).

The car is parked at the garage and needs to get get its batteries charged at the local charging station/gas station. It needs to self-drive to this station (say Monty's) and then, on reaching there, find where the charging dock is, and finally, align itself such that the charging port on the car's body, is precisely in front of the charging hose (or cable).

For this, the Model-S needs to do the following:

  1. Locate Monty's and find a way to self-drive there.
  2. Stop at the station and align its charging port to the charging dock's hose.

What is the method(s) used by the robots/cars to achieve these two functions ?


1 Answer 1

  1. High level path planning is present in any navigation system. You can select a goal and a start position (or your current position) and a (drivable) path will be planned by any navigation system, incl. google maps.

  2. Tesla cars use a neural network based control system. Video and radar feeds are combined and a semantic segmented map of the environment is available. This will identify a charger for example. Aligning to that charger, technology-wise could be done by a neural network (being trained by a vast amount of labeled charging videos, or by some classical trajectory planing application for non-holomonic systems (a car is a non-holomonic system). I do not know which option is used by Tesla.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.