Ji Yue CEO Xia Yiping delineated the technical route: divided into two stages of research and development, "Pure Vision" has been popular in downtown Shanghai.
According to the noon news on October 17, on the eve of the launch of the car robot Ji Yue 01, Sina Science and Technology learned from insiders that the first "BEV+Transformer" pure visual intelligent driving scheme has now achieved pilot-assisted driving on "downtown" roads such as the Bund and Lujiazui in Shanghai. At the same time, the occupied network technology jointly developed by Ji Yue and Baidu will also be released soon, or carried in Ji Yue 01.
At present, there are two main routes in the global self-driving industry at the level of environmental awareness:
The first is the pure visual environment perception route represented by Tesla.
The second is other types of self-driving enterprises that are mainly based on multi-lidar sensing.
In these two routes, the pure visual environment perception route is difficult, but the technical cost is low after successful research and development. although the lidar is achieved quickly in the early stage, the overall cost is high, and the short-term compression cost is difficult.
Xia Yiping, CEO of Ji Yue, revealed when talking about the technical route of Ji Yue, "Ji Yue is the only Chinese car brand that adopts a pure visual route to develop smart cars." This means that with the successful test drive of a series of pure visual core technologies such as BEV+Transformer and OCC in the core routes of domestic first-tier cities, Jiyue will become the only enterprise in China with pure visual autopilot perception and lidar perception at the same time.
According to the insider, at present, Jiyue is promoting the research and development and landing application of pure vision technology in stages.
In the 1.0 stage of intelligent driving research and development: "pure vision + lidar" integrated intelligent driving scheme, two sets of dual independent systems backup and complement each other.
To the intelligent driving research and development stage 2.0: based on the mature "pure vision" perception architecture, as well as the complete data closed-loop system, the growth efficiency of the high-level intelligent driving technology system becomes efficient.
