/sci/ - Science & Math

What is the "easiest" modern approach for teaching a robot to do tasks in a real world environment using an image sensor as the input. A task like:

>Go from room A to B
>When in room B find C
>When find C approach to C
>When near C grab C
>When C grabbed go to room A
>When in room A drop C

What would be the list of algorithms you would use for State detection and action taking.