http://www.sfgate.com/business/article/What-s-holding-back-Tesla-s-self-driving-car-6391158.php
What should be noted is that humans struggle with the same thing the car is struggling with. On the opposite coast of the 405, NYCDOT often does not maintain pavement markings very well (to be expected, considering they often do not maintain the pavement itself very well), and you have a lot of cases where a road is clearly supposed to have two lanes going the same way but there is little to no white paint to be found. Sometimes you will get cars driving down the middle because they don't realize there are supposed to be two lanes.
What's holding it back is the lack of human ability to assess new visual information based on a system of experience and judgement.
A self-driving car can sit for hours in a traffic jam just as one with a driver can.
Quote from: Pete from Boston on July 17, 2015, 07:52:54 PM
What's holding it back is the lack of human ability to assess new visual information based on a system of experience and judgement.
Assessing new visual information based on a system of experience and judgement is exactly how self driving cars work. That's basically the definition of a machine learning algorithm.
I found a good article that explains at a high level how this works. It's a fascinating area of research. http://www.theatlantic.com/technology/archive/2014/05/all-the-world-a-track-the-trick-that-makes-googles-self-driving-cars-work/370871/