Google has been testing its self driving cars since 2009 and since then there were 16 minor accidents happened with 12 of them happened because the vehicles were rear-ended by a human driver. Since the beginning, the company has designed the autonomous cars to follow the rules of the road and now the company is teaching them to drive like people.
According to Jen-Hsun Huang, the Chief Executive of Nvidia Corp about the reason of why is it getting rear-ended, the answer is that it drives as computer. It is too save for its own good. For example, during the recent test drive involving two Wall Street Journal reporters, Google Lexus RX450h slowed down as it thought that a car which was travelling from behind was approaching so fast that the car would cut the Google’s car as it passed, however, in reality, the other car did not do that.
Another occasion happened when Google car was trying to make a left turn in a busy T-intersection with limited views of cross traffic and it waited for about 30 seconds before making the turn and stopped in the middle of intersection abruptly. It turned out that a woman from the other side of the street stood near the edge of the sidewalk which tricked Google car into thinking that she was trying to cross the road.
In regards of these situations, Google is currently building an algorithm to handle human flexibility and interpretation better.