eastwest connect

Xpeng explains the advantages of G9 subverting the traditional Robotaxi | Will start autonomous driving passenger test for all users in Huangpu District, Guangzhou in the second half of 2023

Source: Xpeng

On May 1, Xpeng released an article explaining the advantages of G9 being used as Robotaxi.


In April 2023, Xpeng G9 officially obtained the Guangzhou Intelligent Connected Vehicle Passenger Test License, and the test scope can cover all general test roads in Guangzhou.

Since the autonomous driving road test was started in November last year, Xpeng G9 has obtained the Guangzhou ICV road test permit, becoming the industry’s first autopilot software upgrade and enhancement without changing the hardware, that is, to obtain the Chinese road test license. Zero-modification mass-produced vehicles that are qualified for self-driving road tests within the range.

In addition, we expect to start the autonomous driving passenger test for all users in Huangpu District, Guangzhou City in the second half of 2023. At that time, users can call the Xpeng G9 self-driving car with one click through the designated WeChat applet, enjoying a new experience of smart travel.


Xpeng G9 becomes the first mass-produced autonomous car in the industry without hardware modification

Xpeng G9 is the industry’s first mass-produced self-driving car with zero hardware modifications. It realizes the complete commonality of hardware such as perception, computing, and control of the autonomous driving system, and performs a number of optimizations for the Robotaxi scene to meet the user’s needs for Robotaxi. The new model of pre-installed mass production has greatly reduced the comprehensive cost of R&D, mass production and maintenance, subverting the traditional Robotaxi model.

Combined with Xpeng’ complete manufacturing, energy replenishment, and service network, Xpeng G9 makes the large-scale operation of Robotaxi possible by virtue of stronger software and hardware integrity and stability, as well as lower overall costs. It also allows more users to have the opportunity to personally experience the technological experience brought by autonomous driving and realize true technological inclusiveness.

At the same time, the driving performance of self-driving cars comes from massive data. Xpeng G9 Robotaxi can join the training system in parallel with mass-produced vehicles, greatly improving the source, volume and scene richness of training data.

At present, the training data of most Level 4 companies in the industry usually come from “tens to hundreds of self-owned Robotaxi fleets”, and usually operate within a “fixed road range” and “fixed time period”. Not only that, but the average cost per vehicle of modified Robotaxi often exceeds one million, and the cost of data is also very high.

Based on the same hardware architecture, Xpeng G9 Robotaxi not only has a lower overall cost, but also continuously improves the performance, stability and generalization capabilities of the autopilot system through data closed-loop.

In the second half of this year, we will carry out self-driving passenger-carrying tests for all users in Huangpu District, Guangzhou. At that time, users can make an appointment through the WeChat applet and call Xpeng Robotaxi with one click, opening a new smart travel experience covering high-frequency travel scenarios such as going to work, school, and shopping.

Equipped with the industry’s first and only mass-produced BEV sensor
With or without maps, intelligent travel is unimpeded

Xpeng G9 adopts China’s first and only mass-produced BEV perception – XNet deep visual neural network, through the multi-frame time-sequential fusion of data collected by multiple cameras, output 4D information based on dynamic objects under the BEV perspective (such as vehicle speed, distance, etc.), and 3D information based on static objects (such as lane lines, roadsides, etc.).

Among them, dynamic information can realize target trajectory prediction on the basis of visual recognition; Static information can provide real lane lines and drivable space recognition, realizing the ability to generate “high-precision maps” in real time.

As the application of XNet technology continues to mature, users in the future will not be limited by the area of high-precision maps, and can enjoy the safe and efficient travel experience brought by autonomous driving in more cities and scenes faster.

At the same time, with the large-scale operation and use of Robotaxi, the experience threshold of intelligent assisted driving will be further lowered, and more and more users will have the opportunity to experience the changes in travel methods brought about by technology.

In the future, Xpeng will continue to explore, through high-level assisted driving and automatic driving in parallel, make full use of system capabilities, realize the closed-loop feedback of data and technical capabilities, further promote the popularization of autonomous driving technology, and create a more convenient and beautiful smart travel life for users.