Chinese internet search company Baidu has unveiled its plans to make its entry in the emerging field of artificial intelligence, including a partnership with Nvidia to develop robot taxis.
Baidu's Vision
According to the San Francisco Gate, Baidu also announced that the California Department of Motor Vehicles granted the company a permit to test driverless vehicles. Baidu executive Jing Wang said in a statement that the company will soon start testing its self-driving technologies on California's public roads.
Mr Wang added that Baidu has the company will be able to make rapid progress by being able to test its driverless technologies on public roads. Up to date, Baidu already built a strong research and development team in Silicon Valley to work on autonomous driving technologies.
According to Business Insider, by 2018, Baidu is planning to use its self-driving technology for a robot taxis service. Jing Wang, a senior vice president who heads Baidu's business unit developing self-driving cars, said that the company could also produce autonomous cars for private use after the year 2018.
Most of the other companies are working to get their autonomous cars on the road by 2020. Only Tesla has also committed to a 2018 timeline, according to TechInsider.
Baidu's Two Key Technologies
In order to advance its self-driving car efforts, Baidu is focusing its efforts into two key technologies. The first is deep learning artificial intelligence that can provide autonomous cars optimal image sensing and recognition.
Self-driving cars can learn on their own using deep learning algorithms. This way, they can evaluate the environment around them and make faster decisions.
Baidu now uses in its driverless cars a high-performance, image recognition server created by Chinese tech company Inspur in partnership with Nvidia. Baidu is also partners with Nvidia in order to develop a cloud platform that will provide Level 3 autonomous vehicle control.
The second key technology Baidu is focusing in order to advance its autonomous car efforts is its HD maps. Cars compare what is outlined on the mapping routes to what they're "seeing" with radar, LiDAR and sensors in order to detect and avoid obstacles.
