Ventech China recently invested in Level 4 (completely human-free level in autonomous driving) self-driving solution provider Roadstar.ai. The company’s founding team are all from top tech companies in the Silicon Valley.

Professional team

The 3 founders all had top-tier educational backgrounds, plus core member experiences in leading tech companies. TONG Xianqiao the CEO, was a PhD from Virginia Polytechnic University. His experiences include Apple’s self-driving car developing team, NVIDA’s autonomous driving algorithm team, and tech lead in localization, mapping, and technology of Baidu in Silicon Valley. HENG Liang the CTO, was a bachelor from Tsinghua University and a PhD from Stanford University. He worked in Autopilot Team at Tesla & Street View Team at Google, and served as core member in Baidu’s ADU Technology Committee and tech lead on Sensing Team. ZHOU Guang, Chief Scientist in Robotics, graduated from Tsinghua University as bachelor and later acquired a PhD degree from University of Texas. In 2015, Zhou led a team and won the 1^st prize in DJI Developer Challenge. Zhou also worked in Baidu Silicon Valley in calibration and sensing.

Multiple sensor fusion solution

Roadstar.ai aims to provide a whole set of L4 self-driving solution, including working with car manufacturers to develop L4 modules, and provide multiple sensor-fusion technology, controlling software, and high definition maps. Roadstar.ai adopts multiple sensor fusion technology. While a single sensor cannot reliably guarantee the safety and stability of cars, the combination of visual sensors (cameras), LiDAR, etc. can take advantage of features from different sensors, improving the system’s accuracy and robustness.

Roadstar.ai thinks it seized the timing to enter the industry: on one hand, multiple companies from China and abroad are dedicated into the R&D of sensors, whose price may therefore become affordable over the next few years; on the other hand, deep learning, when still at its primary stage, has already contributed considerably to enhance performance of self-driving system, and there still remains huge potential. Roadstar.ai plans to complete its first prototype vehicle and start road test in early 2018.

Commercial scenarios to enter

Tong considers commercial vehicles as a proper entry angle for self-driving application. A notable example would be the logistics area. Operation expenses in the industry are high, while costs mainly arise from relatively simple driving scenarios like highways. The implementation of self-driving technology would largely cut the cost, improve efficiency and further prompt the industry to boom.

In addition, Roadstar.ai will generate centimeter-level HD maps through data on the surrounding environment collected during driving. The maps cannot only provide information for self-driving, but also become a complement for map products of traditional map suppliers.

Ventech China has been long paying attention to AI. With distinguished team background and deep understanding in self-driving, Roadstar.ai has the potential to become the leading runner in China.