I only just came across this story from May 2015 about an accident involving a Volvo car during a demonstration of its auto-braking features:
Although the article is not correct in talking about self-parking, the response from Volvo spokespeople is still incredible. It turns out that although these cars have auto-braking features, they are only designed to stop them crashing into other cars. To avoid crashing into pedestrians, there is an optional extra known as “pedestrian detection functionality” – yes, the buyer needs to pay extra to enable the car to detect and not knock over pedestrians in its way!
In an era where progress towards self-driving cars is seen as promising, stories like this should make us think hard about whether that’s a good thing or not. The idea that someone decided avoiding pedestrians could be a source of additional revenue in such an advanced car almost beggars belief.
Next time you are working through stories (or whatever means you use to decide what features make it into your product), take some time out to think what really makes sense to the user of your product and what their realistic expectations might be – this seems to have been lost at Volvo based on the separate treatment of pedestrians in the functionality of this auto-braking feature.
There are further comments from Volvo officials in this related article, remarkable reading:
http://fusion.net/story/139703/self-parking-car-accident-no-pedestrian-detection/
(This post was inspired by a tweet from Michael Bolton on the same subject.)
Rockin' and Testing All Over The World 
“As Designed”
“But…wait, what if…?”
“AS DESIGNED”
There is so much software out there that is “working as badly designed”