Nullmax pure vision technology opens up a new experience of urban navigation
2024-08-26
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
Autonomous driving technology is steadily moving towards a new milestone, and Nullmax has once again demonstrated its strength in this field with its leading pure vision autonomous driving technology. This technology, through the deep integration of high-precision cameras and advanced AI algorithms, enables autonomous vehicles to fully perceive and accurately understand the surrounding environment, setting a new benchmark for the autonomous driving industry.
Nullmax's pure vision autonomous driving system can capture every detail on the road, including curbs, fences, lane lines, pedestrians, vehicles, etc., through vision, just like a human driver, so as to achieve real-time perception of complex traffic environments. This capability is perfectly reflected in the 7V pure vision vehicle that Nullmax has recently demonstrated. The 7V pure vision vehicle can autonomously turn from one road to another, complete a series of complex operations such as changing lanes, merging into traffic, and bypassing parked vehicles on the roadside. The whole process is smooth and natural, as if an experienced driver is controlling it.
Nullmax's end-to-end autonomous driving model uses visual information as the main input. The vehicle only needs to be equipped with basic cameras. According to the instructions such as left and right turns provided by the navigation system, point-to-point intelligent driving in urban scenes can be achieved. And because the sensor configuration is simple, the overall computing power requirement is relatively lower. Nullmax's technical solution can achieve the navigation function of urban scenes within 100T sparse computing power. Even relatively cheap models, even if they are not equipped with expensive sensors and chips with super computing power, can enjoy a safe and intelligent experience.
In addition, Nullmax has also achieved fruitful results in the field of pure vision. The BEVSegFormer and CurveFormer algorithms developed by it have been included in two international academic conferences, WACV and ICRA. According to the review paper "Vision-Centric BEV Perception: A Survey" of IEEE TPAMI, the top journal in the field of computer vision, the BEVSegFormer proposed by Nullmax is the earliest representative research publicly published in China in the field of BEV perception. Based on the pure visual perception of environmental information, the upgraded algorithm CurveFormer++ proposed by Nullmax in February this year explored the sparse Query time series fusion method, allowing the static perception, dynamic perception and planning of autonomous driving to have "memory" like humans.
Nullmax firmly believes that pure vision is an essential and core technical capability for autonomous driving. It can not only independently realize a variety of tasks and functions, but also is the key to unmanned driving and embodied intelligence. With the continuous maturity of technology and the gradual expansion of the market, Nullmax's pure vision autonomous driving technology is expected to achieve large-scale commercial application in the next few years. By then, more consumers will be able to enjoy the safe, convenient and intelligent travel experience brought by this technology.
