Shared Mobility

Baidu leads the self-driving race with an AI-packed ‘robocar’

Baidu leads the self-driving race with an AI-packed ‘robocar’

By Dashveenjit Kaur 

Back in May this year, a commercial driverless taxi service was launched in China by the country’s biggest search engine player, Baidu. It was the first mass-produced robotaxi in China and Baidu did not slow down ever since. The company that has been building AI programs to power self-driving cars just unveiled its first “robocar” and a second-generation artificial intelligence chip for driverless vehicles.

Announced at its annual Baidu World conference recently, the AI-packed announcement showcases the company’s advances in an eight-year effort to develop autonomous-driving technology. It also underscores how these new areas of technology are key to the company’s future growth.

As described by CEO of Baidu himself Robin Li, “We believe that cars of the future will be robocars.” In fact, Baidu’s latest ‘robocar’ prototype is capable of Level 5 autonomous driving capabilities. “They will drive autonomously, act as both an intelligent assistant and loyal companion, and be self-learning,” he told an online audience.Dashveenjit Kaur |

Li believes intelligent vehicles of the future will be more like intelligent robots. “They will move, they will communicate, and they will learn. The robocar will not only be your vehicle, but also your driver, secretary, and personal assistant. It will drive automatically, understand your words, take orders and continuously learn from you and upgrade to serve you a more personalized experience,” he explained.

Baidu’s ‘robocar’

The robocar adopts a boxy shape and features gullwing doors as well as a transparent glass roof and a display at the front to visually communicate with its surroundings. The body is integrated with a suite of external sensors to enable its autonomous driving capabilities, which at Level 5, means the vehicle is capable of driving on its own under all conditions

As of now, there are no Level 5 autonomous vehicles being sold in the market, and most robotaxis being tested by other companies operate at Level 4, requiring certain conditions to be met before they can engage in autonomous driving. The robocar’s interior reflects its self-driving capabilities, as it contains zero-gravity seats, a widescreen curved display, and a control pad, with the steering wheel and pedals notably missing. 

Other tech bits here include voice and facial recognition, along with an advanced AI system that analyses the internal and external surroundings to provide predictive suggestions to proactively serve the needs of passengers.

The robocar, however, is still a prototype and the company has yet to confirm if it will be mass-produced. Even before revealing the robocar, Baidu already runs its Apollo robotaxi service in some cities in China and as of the end of the second quarter of this year, the service has provided more than 400,000 rides and driven more than 14 million km.

Baidu began developing autonomous driving tech in 2013, ahead of its competitors, and in 2017, it launched the Apollo open-source self-driving development platform. The platform has attracted Chinese government support as well as powerful partners like Toyota Motor and Volkswagen, along with tech companies Intel and Nvidia.

AI chip for self-driving technology

In 2018, Baidu launched its first-generation Kunlun chip. On Wednesday, during a conference, the company said it has begun mass producing Kunlun 2, its second-generation AI chip. The semiconductor is designed to help devices process huge amounts of data and boost computing power.

Baidu says the chip can be used in areas such as autonomous driving as it races to become a key player in the chip industry which Beijing is trying to strengthen. The new generation of Kunlun AI chips, using 7 nm process technology, achieved a top computational capability two to three times the previous generation, Baidu said in a statement.

Kunlun chips have been used mostly by Baidu in smart electric vehicles and cloud computing.



Exit mobile version