最新精品麻豆一区二区,久久国产精品99精品国产福利 ,亚洲国产精品乱码一区二区,亚洲精品欧洲久久婷婷99,亚洲精品一区二区在线,91孕妇精品一区二区三区,久久99精品九九九久久婷婷,最新精品伦理一区二区

Home> Updates

Pony.ai develops automotive grade computing unit

Updated: 2023-01-11

Over the past year, the automatic driving with "people in the front seat" of Pony.ai has been maturing, with its "unmanned in the front seat" automatic driving test currently in full swing in Beijing E-Town.

Meanwhile, the company continues to increase its interoperability and cooperation with upstream and downstream partners in the industry to promote the scale and mass production of autonomous driving.

11.png

Vehicles to be tested [Photo/beijingetown.com.cn]

Since the "unmanned front seat" test started, each car has accumulated a mileage of no less than 2,000 kilometers, said Pony.ai technicians. 

Since the early stage of "manned co-pilot", Pony.ai has stabilized the field of unmanned driving, and since November 16, after being approved to carry out the test of "no one in the front seat, people in the back seat", Pony.ai's disengagement rate and accident rate at 10,000 kilometers were both 0, and the test vehicle drove safely in the second phase of the test without incident.

In the vehicle maintenance room of Pony.ai, the Toyota Sienna hybrid electric vehicle equipped with Pony.ai's new generation of automatic driving system. 

The lidar on the top of the body and the multiple cameras embedded in the body are particularly eye-catching. 

In line with the design concept of mass production, the biggest feature of this car is the reduced lidar on the roof. It is also the first batch of models to be equipped with an exterior design of the autonomous driving software and hardware system, sensors and computing platform solutions. It is expected that this car will be used by Robotaxi in the first half of 2023, according to Pony.ai. 

In addition, this model is equipped with a self-developed automotive grade L4 autonomous driving computing unit, and a car-grade Nvidia DRIVE Orin system-level (SoC) chip. 

Compared with the previous generation computing unit, the computing power of the model is expected to increase by at least 30 percent, the weight reduced by at least 30 percent, and the cost reduced by at least 30 percent.