Comparison between Xpeng G6 and Jiyue 01 smart driving: Which one is better, lidar or pure vision?

New energy vehicles are not only the core and foundation of global energy conservation and emission reduction, but also the expansion trend of our "double carbon" goal of green travel; as we continue to make breakthroughs in various technologies, intelligence has also reached a higher level, from From human driving to assisted driving, it seems that great results have been achieved in just a few years. And with the gradual relaxation of L3 and L4 levels, in the near future, it may be claimed that it is a "driving vehicle". In fact, it means entering the car and planning the route. When it is time to watch videos, when it is time to sing karaoke, when it is time to sleep. It’s as simple as sleeping and reminding you to get off when you arrive at the address. This is really not just a casual comment. If you look at the previous lidar detection and now on to the pure visual solution, everything is accelerating. So what are the differences between lidar intelligent driving and pure visual intelligent driving? Whose plan is closer to reality? Today we will compare the Xpeng G6 which are representative of each solution .

Xpeng G6 currently has 6 models on sale, positioned as a medium-sized SUV, with a guide price range of 199,900-276,900 yuan. It currently has a discount of 20,000-35,000 yuan, and offers three cruising range options of 580km/755km/700km. Equipped with 800V fast charging function, the time is 0.33 hours.

There are currently 3 models of Jiyue 01 on sale, with a guide price range of 219,900-309,900 yuan. The price is almost the same as the Xpeng G6, but Jiyue 01 is positioned as a medium and large SUV; it offers three ranges of 580km/780km/700km. Mileage selection, fast charging requires 0 .5 hours; current car purchases include lifetime vehicle and three-electronic warranty, lifetime free roadside assistance, free home charging piles/55,000 Jiyue points/choose 1 from 2, and 3-month free subscription to ASD Apollo high-end smart driving/limited time buyout discount The price is RMB 4,999 (consult local dealers for details).

Xpeng G6's configuration distribution in the smart driving field is uneven. The entry-level model uses BEV visual perception technology and XNet deep visual neural network, matched with a single NVIDIA DRIVE Orin X chip, with a computing power of only 254TOPS; only the high-end Max version is equipped with it. It uses 2 Sagitar Jutron lidars and matches 2 NVIDIA DRIVE Orin X chips to achieve multi-sensor fusion perception and improve the perception accuracy of complex environments.

In the smart driving field, Jiyue 01 adopts the strategy of standard equipment for all series, which increases the cost-effectiveness of the new car. Jiyue 01 adopts a pure visual solution and is equipped with the ASD Apollo high-end intelligent driving system based on Baidu Apollo L4 autonomous driving. The entire series is equipped with 34 perception hardware as standard, matched with 2 NVIDIA DRIVE Orin X chips, and has a computing power of 508TOPS; and OCC occupation network technology has also been added to upgrade the vehicle's field of view from "two-dimensional" to "three-dimensional". The model iteration speed is fast, which reduces the car purchase cost for users to experience high-end intelligent driving.

The intelligent driving performance of both models on highways and urban expressways is very mature and stable. In terms of high-speed smart driving performance, Xpeng G6's straight-line driving is stable and reliable. After entering the ramp, it will intelligently choose fast and slow lanes according to different vehicle speeds. When merging or diverging, it will change lanes earlier, which performs well.

Jiyue 01 can intelligently complete overtaking, merging, up and down ramps and other actions on the highway. It can also automatically avoid when encountering a large truck. The following distance can be adjusted in three gears according to the actual speed, making the adjustment more abundant. In comparison, Jiyue 01 responds faster when exiting the ramp and can smoothly merge into traffic; it is also more flexible in handling traffic jams.

When driving intelligently on urban sections, Xpeng G6 can better complete lane keeping, following and changing lanes; however, when entering complex sections, facing congested road conditions, safety and stability may be more focused, and it is obvious that the efficiency of continuous lane changes is not good. High, slower. Jiyue 01 can recognize zebra crossings and traffic lights on urban road sections, and can smoothly complete pedestrians, unprotected left turns and avoid some sudden non-motorized vehicles; in complex road conditions, the pure visual solution can better identify various obstacles. and traffic scenes, but the faster response is sometimes accompanied by sudden braking, which may cause some shaking discomfort to the people in the car.

As for the smart parking function, the memory parking function of Xpeng G6 includes parking space recognition and tracking reversing. It is generally more practical, but it is somewhat limited in the types and options of parking space recognition. Jiyue 01 also supports valet parking and off-car voice parking functions on its basis. It has richer parking space recognition and selectivity, and even reaches a range of indoor and outdoor parking options at a distance of 2km.

In general, the two models are equally matched in the field of smart driving, and they also represent the different characteristics of lidar solutions and pure visual solutions. Looking at the two models alone, the smart driving configuration of the Xpeng G6 is uneven. The low-end models are subject to many restrictions. Although the lidar solution is accurate, the response speed is not good. Jiyue 01's smart driving system has completed the implementation of standard configurations, both in terms of computing power and system. The pure visual solution is more flexible when dealing with complex road conditions, but it does not work well when encountering insufficient light or bad weather. What about performance when it's not good? Which smart driving solution do you think will be the leader in the future?