Baidu Inc (ADR) (NASDAQ:BIDU), sometimes called the Google of China, has recently announced that it is following in the footsteps of the American search giant in developing its own self-driving vehicle program. In a recent interview with TNW, Kai Yu, deputy director of Baidu’s Institute of Deep Learning, drew a distinction in saying that his company doesn’t view it’s project as a driverless car, as the driver actually remains in control at all times.
Highly autonomous car
“This is actually an intelligent assistant collecting data from road situations and then operating locally. We don’t call this a driverless car. I think a car should be helping people, not replacing people, so we call this a highly autonomous car,” Yu explained in the interview.
Yu went on to describe the differences between Baidu Inc (ADR) (NASDAQ:BIDU)’s and Google’s approach to self-driving cars. “Philosophically we have a fundamental difference to look at this type of things. I think in the future, a car should not totally replace the driver but should really give the driver freedom. Freedom means the car is intelligent enough to operate by itself, like a horse, and make decisions under different road situations. Whenever the driver wants to resume control, you can do that. It’s like riding on a horse, rather than just sitting in a car where you only have a button.”
Baidu: First prototypes in 2015
Yu also mentioned that the first prototypes will be produced some time in 2015, but would not give any hints as to the appearance of Baidu Inc (ADR) (NASDAQ:BIDU)’s new partial self-drive vehicle.
The Chinese internet giant is focused on improving human safety first, Yu said, given that pedestrians in China tend to behave recklessly and gargantuan traffic jams snarl the major cities such as Beijing and Shanghai.
However, as a Chinese search company, Baidu is in the ideal position to develop a self-drive car designed for Chinese drivers ad driving conditions. This is particularly true given that Baidu has also accumulated a great deal of data from its location-based system in the country, and all that data will be very useful in training cars to learn and respond to a wide variety of triggers.