Four of Google's self-driving cars dinged in California

Google’s self-driving car is one of the most pioneering moonshots the search engine is testing, but in the public eye the car is still seen as a dangerous and destructive idea, to have a car run by itself with no human controls.

One argument for the self-driving car is safety, but new reports show Google’s cars aren’t without some flaws when it comes to driving on the road, with four of the 50 self-driving cars logging crashes since September 2014.

Google reported that two out of the four crashes happened while a human was in charge, and all of them were minor. It does not detail how long the human had been driving the car before the crash, or what lead to the crash in the first place. In addition, three of the crashes involed Lexus SUVs that Google had outfitted with sensors.

Both issues remain unsolved, due to Google not wanting to speak to the press about difficulties on its self-driving program. The pilot program is being tested in California and both the UK and Germany want to get involved this year.

Google has been testing the cars in urban environments, trying to get the cars to recognise everyday things on the road like children crossing the street, animals underneath cars and other common features on the road.

It wants to build a database capable of recognising all sorts of incoming objects before they are near, allowing the computer to make quick decisions on whether to stop, slow down or keep moving.

Google’s end goal is to offer a self-driving car without a steering wheel or pedals, although we expect the first model will have both as part of regulator agreements. This new information on the amount of crashes might push Google further back from achieving its end goal, although new moves by Tesla Motors, Uber, German car manufacturers and Apple might speed up the process considerably.