Founded in April 2017, Roadstar.ai has rich practical experiences in automatic driving and are the backbone of autonomous driving research and development. After more than a year of development, it has also made significant progress and is one of the stars of the L4 unmanned vehicle.
Previously, Roadstar.ai had conducted multiple road tests in. But compared to Silicon Valley, Shenzhen’s road conditions are more complex. There are more people and cars, and occasionally the take-out delivery electric vehicles, the retrograde of shared bicycles might be really dangerous. This is why Roadstar.ai chose the Shenzhen Xili road section to be more technically persuasive.
In January, Roadstar.ai completed its debut at the CES and Tokyo Motor Show. In February, it officially began preparations for a road test in Shenzhen; this time, the vehicle involved in the driving test was a red Lincoln MKZ. Five eye-catching laser radars and six cameras (three cameras on the front, one car camera is respectively on the left, right, and rear) and a GPS locator on the car roof.
The total road test length is 10 kilometers, which takes 25 minutes and the average speed is kept at the upper limit of the speed limit. Multiple traffic lights and zebra crossings and the sudden overtaking of the electric bike have also encountered on the road, Roadstar.ai’s test vehicle have all successfully avoided. During the test ride, at a distance of 150 meters from the traffic lights, the point cloud array of Roadstar.ai has marked where the traffic lights are located.
In addition, in the Shenzhen Xili section, the surrounding trees are very lush and the road conditions are extremely complicated. A single sensor does not play a significant role in this situation, and even GPS does not work at all. Rely on Roadstar.ai's pre-fusion-based multi-sensor sensing and location technology, the road test actually proves the feasibility and security of Roadstar.ai multi-sensor fusion. The Xili section is not a smooth road, and there are uphill and downhill sections in the middle. A fellow passenger, Han Zhibiao, a researcher at the University of Tokyo Matsuo Research Institute, said that the entire test ride feels better than the human driver. The whole process is following all the traffic rules and the vehicle responds very quickly.
